Page 1
UNIT 1 1 ROLE OF BUSINESS RESEARCH
Unit Structure
1.1 Overview
1.2 Introduction
1.3 The Nature of Business Research
1.3.1 Business Research Defined
1.4 Managerial Value of Business Research
1.5 When is Business Research Needed?
1.6 Business Research in the Twenty -First Century
1.7 Information Systems and Knowledge Management
1.8 Summary
1.9 Review Question
1.10 References
1.1 Overview
After studying this chapter, the learner should be able to:
➢ Understand how research contributes to business success.
➢ Know how to define business research.
➢ Understand the difference between basic and applied business research.
➢ Understand how research activities can be used to address business decisions.
➢ Know whe n business research should be conducted.
1.2 Introduction
Research is perhaps as old as mankind. If necessity was the mother of invention, it
was also the mother of discovery. The primitive man’s need must have sent him in
search not only of food, but also of knowledge. The process was basically the
1munotes.in
Page 2
2 RESEARCH IN COMPUTING
acquisition o f knowledge, the quest for truth, the exploration of the unexplored.
Since the area unexplored was at that time vast, every discovery must have been a
grand thrill.
As per Grinell, “ The word research is composed of two syllables, re and search .
The dictionary defines the former as a prefix meaning again, a new or over again
and the latter as a verb meaning to examine closely and carefully, to test and try,
or to probe. Together they f orm a noun describing a careful, systematic, patient
study and investigation in some field of knowledge, undertaken to establish facts or
principles .”
According to Robert Ross, “ Research is essentially an investigation, a recording
and an analysis of evidence for the purpose of gaining knowledge .” It can generally
be defined as a systematic method of finding solutions to problems. In the opinion
of Redman and Mory, “ It is a systematized effort to gain new Knowledge .”
A research may need not only lead t o ideal solution but also give rise to new
problems which may require further research. In other words , research may not be
an end to a problem since every research could have the capability of pointing to a
new question. It is carried on both for discover ing new facts and verification of old
ones. Further, being knowledgeable about research and research methods helps professional managers to:
1. Identify and effectively solve minor problems in the work setting.
2. Know how to discriminate good from bad research.
3. Appreciate and be constantly aware of the multiple influences and multiple
effects of factors impinging on a situation.
4. Take calculated risks in decision making, knowing full well the probabilities
associated with the different possible outcomes.
5. Prevent possible vested interests from exercising their influence in a situation.
6. Relate to hired researchers and consultants more effectively.
7. Combine experience with scientific knowledge while making decisions.
munotes.in
Page 3
3Chapter 1: Role of Business Research
1.3 The Nature of Business Research
Business research covers a wide range of phenomena. For managers, the purpose
of research is to provide knowledge regarding the organization, the market, the
economy, or another area of uncertainty. A financial manager may ask, “Will the
environment for lo ng-term financing be better two years from now?” A personnel
manager may ask, “What kind of training is necessary for production employees?”
or “What is the reason for the company’s high employee turnover?” A marketing
manager may ask, “How can I monitor m y retail sales and retail trade activities?” Each of these questions requires information about how the environment, employees, customers, or the economy will respond to executives’ decisions.
Research is one of the principal tools for answering these prac tical questions.
Within an organization, a business researcher may be referred to as a marketing
researcher, an organizational researcher, a director of financial and economic
research, or one of many other titles. Although business researchers are often
specialized, the term business research encompasses all these functional specialties. While researchers in different functional areas may investigate different phenomena, they are like one another because they share similar research methods.
It has been said that “every business issue ultimately boils down to an information
problem.” Can the right information be delivered? The goal of research is to supply
accurate information that reduces the uncertainty in managerial decision making.
Very often, decisions are made with little information for various reasons, including cost considerations, insufficient time to conduct research, or management’s belief that enough is already known. Relying on seat -of-the pants
decision making —decision making without r esearch —is like betting on a long shot
at the racetrack because the horse’s name is appealing. Occasionally there are
successes, but in the long run, intuition without research leads to losses. Business
research helps decision makers shift from intuitive information gathering to systematic and objective investigation.
1.3.1 Business Research Defined
Business research is the application of the scientific method in searching for the truth about business phenomena. These activities include defining business opportunities and problems, generating, and evaluating alternative courses of
action, and monitoring employee and organizational performance. Business research is more than conducting surveys. This process includes idea and theory development, problem definition, searching for and collecting information, analyzing data, and communicating the findings a nd their implications. munotes.in
Page 4
4 RESEARCH IN COMPUTING
This definition suggests that business research information is not intuitive or
haphazardly gathered. Literally, research (re -search) means “to search again.” The
term connotes patient study and scientific investigation wherein the researcher
takes another, more careful look at the data to discover all that is known about the
subject. Ultimately, all findings are tied back to the underlying theory.
The definition also emphasizes, through reference to the scientific method, that any
information generated should be accurate and objective. The nineteenth -century
American humorist Artemus Ward claimed, “It ain’t the things we don’t know that
gets us in trouble. It’s the things we know that ain’t so.” In other words, research is
not perfo rmed to support preconceived ideas but to test them. The researcher must
be personally detached and free of bias in attempting to find truth. If bias enters the
research process, the value of the research is considerably reduced. We will discuss
this furth er in a subsequent chapter.
Our definition makes it clear that business research is designed to facilitate the
managerial decision -making process for all aspects of the business: finance,
marketing, human resources, and so on. Business research is an esse ntial tool for
management in virtually all problem -solving and decision -making activities. By
providing the necessary information on which to base business decisions, research
can decrease the risk of making a wrong decision in each area. However, it is
important to note that research is an aid to managerial decision making, never a
substitute for it.
Finally, this definition of business research is limited by one’s definition of business. Certainly, research regarding production, finance, marketing, and management in for -profit corporations like DuPont is business research. However,
business research also profits, and governmental agencies can use research in much
the same way as managers at Starbucks, Jelly Belly, or DuPont. While the focus is
on for -profit organizations, this book explores business research as it applies to all
institutions.
1.4 Managerial Value of Business Research
The prime managerial value of business research is that it reduces uncertainty by
providing information that facilitates decision making about strategies and the
tactics used to achieve an organization’s strategic goals. The decision -making
process involves three interrelated stages.
The prime managerial value of business research is that it reduces uncertai nty by
providing information that facilitates decision making about strategies and the
tactics used to achieve an organization’s strategic goals.
The decision -making process involves three interrelated stages. munotes.in
Page 5
5Chapter 1: Role of Business Research
The prime managerial value of business resear ch is that it reduces uncertainty by
providing information that facilitates decision making about strategies and the
tactics used to achieve an organization’s strategic goals. The decision -making
process involves four interrelated stages.
A. Identifying the existence of problems and opportunities
Before any strategy can be developed, an organization must determine where it
wants to go and how it will get there. Business research can help managers plan
strategies by determining the nature of si tuations by identifying the existence of
problems or opportunities present in the organization.
B. Diagnosis and Assessment
After an organization recognizes a problem or identifies a potential opportunity,
an important aspect of business research is often the provision of diagnostic
information that clarifies the situation. Managers need to gain insight about the
underlying factors causing the situation. If there is a problem , they need to
specify what happened and why. If an opportunity exists , they may need to
explore, clarify, and refine the nature of the opportunity.
C. Selecting and implementing a course of action
Business research is often conducted to obtain specific information to help
evaluate the various alternatives, and to select the best cou rse of action based on
certain performance criteria.
D. Evaluation of the course of action
Evaluation research is conducted to inform managers whether planned activities were properly executed and whether they accomplished what they were expected to do. It se rves an evaluation and control function. Evaluation research
is a formal, objective appraisal that provides information about objectives and whether the planned activities accomplished what they were expected to accomplish. This can be done through perform ance-monitoring research, which
is a form of research that regularly provides feedback for evaluation and control
of business activity. If this research indicates things are not going as planned,
further research may be required to explain why something “w ent wrong.”
Some organizations have their own consulting or research department, which might
be called the Management Services Department, the Organization and Methods
Department, R & D (research and development department), or some other name.
This depar tment serves as the internal consultant to subunits of the organization munotes.in
Page 6
6 RESEARCH IN COMPUTING
that face certain problems and seek help. Such a unit within the organization, if it
exists, is useful in several ways, and enlisting its help might be advantageous under
some circumst ances, but not others.
The manager often must decide whether to use internal or external researchers. To
reach a decision, the manager should be aware of the strengths and weaknesses of
both, and weigh the advantages and disadvantages of using either, bas ed on the
needs of the situation. Some of the advantages and disadvantages of both internal
and external teams are discussed below:
Advantages of internal consultants/researchers
There are at least four advantages in engaging an internal team to do the res earch
project:
1. The internal team stands a better chance of being readily accepted by the
employees in the sub -unit of the organization where research needs to be done.
2. The team requires much less time to understand the structure, the philosophy
and climate, and the functioning and work systems of the organization.
3. They are available to implement their recommendations after the research
findings have been accepted. This is very important because any “bugs” in the
implementation of the recommendations may be removed with their help. They
are also available to evaluate the effectiveness of the changes, and to consider
further changes when necessary.
4. The internal team might cost considerably less than an external team for the
department enlisting help in problem solving, because they will need less time
to understand the system due to their continuous involvement with various
units of the organization. For problems of low complexity, the internal team
would be ideal.
Disadvantages of internal consultant s/researchers
There are also certain disadvantages to engaging internal research teams for the
purposes of problem solving. The four most critical ones are:
1. In view of their long tenure as internal consultants, the internal team may quite
possibly fall int o a stereotyped way of looking at the organization and its
problems. This inhibits any fresh ideas and perspectives that might be needed
to correct the problem. This is a handicap for situations in which weighty issues
and complex problems are to be invest igated. munotes.in
Page 7
7Chapter 1: Role of Business Research
2. There is scope for certain powerful coalitions in the organization to influence
the internal team to conceal, distort, or misrepresent certain facts. In other
words, certain vested interests could dominate, especially in securing a sizable
portion of the available scant resources.
3. There is also a possibility that even the most highly qualified internal research
teams are not perceived as “experts” by the staff and management, and hence
their recommendation may not get the consideration and attention they deserve.
4. Certain organizational biases of the internal research team might, in some
instances, make the findings less objective and consequently less scientific.
Advantages of external consultants/researchers
1. For decades, large consulting firms have maintained their reputation for having
the best strategy consulting practices. However, along with the expertise comes
a hefty price tag that many client companies are not so willing to cough up. Yet
their reputatio n precedes them for good reason, after working with multiple
large, influential corporations. Despite the cost, external consultants have
advantages which internal consultants cannot necessarily replicate due to their
long-term reputation for good work and for hiring the best graduates from best
schools.
2. Coming from an outside perspective allows consultants to have a more objective, bird’s eye view of the company and the industry . Instead of becoming too engrossed within a specific company, external consultants
should be on top of the industry . Not only do they have a broad perspective,
but an experienced consultant will have had multiple experiences working with
other companies in the same industry and that faced similar challenges.
Therefore, they can apply experience from the past into their current projects
and engagements.
3. Another advantage of not being as integrated into the work project environment
is the ability to be regarded a s an expert and not a peer. Due to the lack of
concrete understanding of the role, internal consultants can be viewed just another pair of hands to make changes within the organization. Instead, external consultants are hired for the sole purpose of their expertise and ability
to create change for a specific business problem. This brings more clarity and
focus to the role, and helps concentrate efforts on the project at hand, and often,
helps ensure client buy -in.
munotes.in
Page 8
8 RESEARCH IN COMPUTING
Disadvantages of external consultants/rese archers
1. While internal consultants battle company politics from inside the company,
employees often regard external consultants with suspicion. This reputation,
unfortunately, is often deserved. Many consultants come into an organization
without an underst anding of the company or a willingness to hear opinions.
Instead, they often try to implement one -size-fits-all strategies either taught to
them by their consulting firms or from past consulting experiences. While
knowledge from other consultants can often be helpful, it does not necessarily
apply in all similar situations.
2. External consultants also face the bad rap of coming in, presenting solutions,
and leaving. This conduct leaves many firms without a solid game plan and
causes them to flounder in the im plementation process. External consultants do
not tend to stay aboard after proposing their various strategies, and most clients
do not want to pay them afterward either. Unfortunately, this leaves clients
spending far too much money for too little change.
1.5 When Is Business Research Needed? The need to make intelligent, informed decisions ultimately motivates an organization to engage in business research. Not every decision requires research.
Thus, when confronting a key decision, a manager must initially decide whether to
conduct business research. The determination of the need for research centers on:
(1) time constraints
(2) the availability of data
(3) the nature of the decision to be made.
(4) the value of the research information in relation to costs
Most work in business organizations, in whatever sector or ownership, will require
research activities. When choosing an area for research, we usually start either with
a broad area of management, which particularly interests us, say, marketing or
operatio ns management, or we start with practical questions, which need answers
to help with managerial decision -making. Refining from this point to a researchable
question or objective is not easy.
We need to do several things:
• Narrow down the study topic to one which we are both interested in and have
the time to investigate thoroughly. munotes.in
Page 9
9Chapter 1: Role of Business Research• Choose a topic context where we can find some access to practitioners if possible; either a direct connection with an organization or professional body,
or a context which is wel l documented either on the web or in the literature. • Identify relevant theory or domains of knowledge around the question for
reading and background understanding. • Write and re -write the question or working title, checking thoroughly the
implications of ea ch phrase or word to check assumptions and ensure we really
mean what we write. This is often best done with other people to help us check
assumptions and see the topic more clearly. • Use the published literature and discussion with others to help us narrow down
firmly to an angle or gap in the business literature, which will be worthwhile
to explore. • Identify the possible outcomes from this research topic, both theoretical and
practical. If they are not clear, can we refine the topic so that they become
clear? 1.6 Business Research in The Twenty -First Century We can now define business research as an organized, systematic, data -based, critical, objective, scientific inquiry, or investigation into a specific problem, undertaken with the purpose of finding answers or solutions to it. Some commonly researched topical areas in business are specified below: • Employee behaviors such as performance, absenteeism, and turnover. • Employee attitudes such as job satisfaction, loyalty, and organizational commitment. • Supervisory performance, managerial leadership style, and performance appraisal systems. • Employee selection, recruitment, training, and retention. • Validation of performance appraisal systems. • Human resource management choices and organizational strategy. • Evaluation of assessment centers. • The dynamics of rating and rating errors in the judgment of human performance. • Strategy formulation and implementation. munotes.in
Page 10
10 RESEARCH IN COMPUTING
• Just-in-time systems, continuous -improvement strategies, and production efficiencies.
• Updating policies and procedures in keeping with latest government regulations and organizational changes.
• Organizational outcomes such as increased sales, market share, profits, growth,
and effectiveness.
• Consumer decision making.
• Customer relationship management.
• Consumer satisfaction, complaints, customer loyalty, and word -of-mouth
communication.
• Complaint handling.
• Delivering and performing service.
• Product life cycle, and product innovation.
• Impression management, logos, and image building.
• Product positioning, product modification, and new product development.
• Cost of capital, val uation of firms, dividend policies, and investment decisions.
• Risk assessment, exchange rate fluctuations, and foreign investment.
• Tax implications of reorganization of firms or acquisition of companies.
• Collection of accounts receivable.
• Development of ef fective cost accounting procedures.
• Qualified pension plans and cafeteria -type benefits for employees.
• Deferred compensation plans.
• Installation of effective management information systems.
• Advanced manufacturing technologies and information systems.
• Desig n of career paths for spouses in dual -career families.
• Creative management of a diverse workforce.
• Cultural differences and the dynamics of managing a multinational firm.
• Alternative work patterns: job sharing, flexitime, flexiplace, and part -time
work. munotes.in
Page 11
11Chapter 1: Role of Business Research
• Downsizing.
• Participative management and performance effectiveness.
• Differences in leadership positions, salaries, and leadership styles.
• Instrument development for assessing “true” gender differences.
• Installation, adaptation, and updating of computer netwo rks and software
suitable for creating effective information systems for organizations.
• Installation of an effective data warehouse and data mining system for the
organization.
• Keeping ahead of the competition
Business research, like all business activity, continues to change. Changes in
communication technologies and the trend toward an ever more global marketplace
have played a large role in many of these changes. Like all business activities,
business research has become increasingly global as m ore and more firms operate with few, if any, geographic boundaries. Some companies have extensive international research operations.
With the constantly evolving data inputs, data types and changing customer behavior, it is important for any business resea rch to remain managerial relevant
and timely. With the emergence of digital data, researchers have shifted toward
collecting objective, secondary data. Archival or proprietary information, such as
historical data, is readily available to researchers (Verma , Agarwal, Kachroo &
Krishen 2017). Academicians and practitioners are increasingly using sophisticated
software to execute predictive analytics.
The era of big data has provided new and varied sources of data and has placed
additional requirements on the analytical techniques that must handle these data
sources. There are, therefore, unique challenges facing today’s analyst with the
many issues in big data and analytics.
With an expanded focus on data challenges and data analytics, data scientists will
be able to create awareness and motivate management to invest in these promising
developments. If this happens, business research and business decisions in general
will be based on better data and more effective analytical techniques, and thus be
more know ledge based.
An article of economic times dated Dec 03, 2020 stated : “Firms have accelerated
the adoption of automation, digital business models: Infosys -HFS Research study”. munotes.in
Page 12
12 RESEARCH IN COMPUTING
Another article dated Jan 07, 2021 read: “Digital Identity Research Initiative
collaborates with Transerve for its India Pulse Program”.
1.7 Information Systems and Knowledge Management
Terms like information and data are often used interchangeably. Researchers use
these terms in specific ways that emphasize how useful each can be. Data are
simply facts or recorded measures of certain phenomena (things or events).
Information is data formatted (structured) to support decision making or define the
relationship between two facts. Business intelligence is the subset of data and
information that has some explanatory power enabling effective managerial decisions to be made. So, there is more data tha n information, and more information
than intelligence.
The question of defining knowledge has occupied the minds of philosophers since the classical Greek era and has led to many epistemological debates. It is unnecessary for the purposes of this paper to engage in a debate to probe, question, or reframe the term knowledge, or discover the “universal truth” from the perspective of ancient or modern philosophy. This is because such an understanding of knowledge was neither a determinant factor in building th e
knowledge -based theory of the firm nor in triggering researcher and practitioner
interest in managing organizational knowledge. It is, however, useful to consider
the manifold views of knowledge as discussed in the information technology (IT),
strategic management, and organizational theory literature. This will enable us to uncover some assumptions about knowledge that underlie organizational knowledge management processes and KMS.
A great deal of emphasis is given to understanding the difference among d ata, information, and knowledge and drawing implications from the difference. Because knowledge is personalized, for an individual’s or a group’s knowledge to
be useful for others, it must be expressed in such a manner as to be interpretable by
the receive rs. Hordes of information are of little value; only that information which
is actively processed in the mind of an individual through a process of reflection,
enlightenment, or learning can be useful.
The recent interest in organizational knowledge has prompted the issue of managing the knowledge to the organization’s benefit. Knowledge management
refers to identifying and leveraging the collective knowledge in an organization to
help the organization compete (von Krogh 1998). Knowledge management is
purpor ted to increase innovativeness and responsiveness (Hackbarth 1998). munotes.in
Page 13
13Chapter 1: Role of Business Research
A survey of European firms by KPMG Peat Marwick (1998b) found that almost
half of the companies reported having suffered a significant setback from losing
key staff with 43% experiencing impaired client or supplier relations and 13%
facing a loss of income because of the departure of a single employee. In another
survey, most organizations believed that much of the knowledge they needed
existed inside the organization, but that identifying that it existed, finding it, and
leveraging it remained problematic (Cranfield University 1998). Such problems
maintaining, locating, and applying knowledge have led to systematic attempts to
manage knowledge.
The usefulness of data to management can be described based on four characteristics: relevance, quality, timeliness, and completeness. Relevant data
have the characteristic of pertinence to the situation at hand. The information is
useful. The quality of information is the degree to which data repres ent the true
situation. High -quality data are accurate, valid, and reliable. High -quality data
represent reality faithfully and present a good picture of reality. Timely information
is obtained at the right time.
Computerized information systems can recor d events and present information as a
transaction takes place, improving timeliness. Complete information is the right
quantity of information. Managers must have sufficient information to relate all
aspects of their decisions together. A computer -based de cision support system
helps decision makers confront problems through direct interactions with databases
and analytical models. A DSS stores data and transforms them into organized
information that is easily accessible to managers.
A database is a collecti on of raw data arranged logically and organized in a form
that can be stored and processed by a computer. Business data come from four major sources: internal records, proprietary business research, business intelligence, and outside vendors and external d istributors. Each source can provide valuable input. Because most companies compile and store many different databases, they often develop data warehousing systems. Data warehousing is the
process allowing important day -to-day operational data to be stored and organized
for simplified access. More specifically, a data warehouse is the multitiered
computer storehouse of current and historical data. Data warehouse management
requires that the detailed data from operational systems be extracted, transformed,
and stored (warehoused) so that the various database tables from both inside and
outside the company are consistent. All this feeds into the decision support system
that automates or assists business decision making. munotes.in
Page 14
14 RESEARCH IN COMPUTING
Numerous database search and retrieval systems are available by subscription or in
libraries. Computer -assisted database searching has made the collection of external
data faster and easier. Managers refer to many different types of databases.
Although personal computers work independently, the y can connect to other
computers in networks to share data and software. Electronic data interchange (EDI) allows one company’s computer system to join directly to another company’s system.
Knowledge management systems (KMS) refer to a class of information systems
applied to managing organizational knowledge. Reviewing the literature discussing
applications of IT to organizational knowledge management initiatives reveals
three common applications:
(1) the coding and sharing of best practices,
(2) the creation of corporate knowledge directories, and
(3) the creation of knowledge networks. One of the most common applications is
internal benchmarking with the aim of transferring internal best practices
(KPMG 199 8a; O’Dell and Grayson 1998).
Information systems designed to support and augment organizational knowledge management need to complement and enhance the knowledge management activities of individuals and the collectivity. To achieve this, the design of information systems should be rooted in and guided by an understanding of the
nature and types of organizational knowledge.
An understanding is needed for formulating a knowledge management strategy and
in analyzing the role of information technology in facilitating knowledge management. In the information systems (IS) field, it has been common to design systems primarily focused on the codified knowledge (that is, explicit organizational knowledge). Management reporting systems, decision support systems, and executive support systems have all focused on the collection and
dissemination of this knowledge type.
IT as applied to KM need not be constrained to certain types of knowledge, because
the advances in communication and information technologies enable gr eater
possibilities than existed with previous classes of information systems. While the preponderance of knowledge management theory stems from strategy and organizational theory research, most knowledge management initiatives involve at
least in part, if not to a significant degree, information technology. Yet little IT munotes.in
Page 15
15Chapter 1: Role of Business Research
research exists on the design, use, or success of systems to support knowledge
management.
According to a framework, grounded in the sociology of knowledge (Berger and
Luckman 1967; Gurvitch 1971; Holzner and Marx 1979, organizations as knowledge systems consist of four sets of socially enacted “knowledge processes”:
(1) creation (also referred to as construction)
(2) storage/retrieval
(3) transfer
(4) application (Holzner and Marx 1979; Pentland 1995)
This view of organizations as knowledge systems represents both the cognitive and
social nature of organizational knowledge and its embodiment in the individual ’s
cognition and practices as well as the collective (i.e., organizational) practices and
culture. These processes do not represent a monolithic set of activities, but an
interconnected and intertwined set of activities, as explained later in this section.
Knowledge Creation
Organizational knowledge creation involves developing new content or replacing
existing content within the organization’s tacit and explicit knowledge (Pentland
1995). Through social and collaborative processes as well as an individual’s
cognitive processes (e.g., reflection), knowledge is created, shared, amplified,
enlarg ed, and justified in organizational settings (Nonaka 1994). This model views
organizational knowledge creation as involving a continual interplay between the
tacit and explicit dimensions of knowledge and a growing spiral flow as knowledge
moves through in dividual, group, and organizational levels. Four modes of knowledge creation have been identified: socialization, externalization, internalization, and combination (Nonaka 1994).
The socialization mode refers to conversion of tacit knowledge to new tacit
knowledge through social interactions and shared experience among organizational
members (e.g., apprenticeship).
The combination mode refers to the creation of new explicit knowledge by
merging, categorizing, reclassifying, and synthesizing existing explicit knowledge
(e.g., literature survey reports). munotes.in
Page 16
16 RESEARCH IN COMPUTING
The other two modes involve interactions and conversion between tacit and explicit
knowledge. Externalization refers to converting tacit knowledge to new explicit
knowledge (e.g., articulation of best practices or lessons learned).
Internalization refers to creation of new tacit knowledge from explicit knowledge
(e.g., the learning and understanding that results from reading or discussion).
The four knowledge creation modes are not pure, but highly interdependent and
intertwined. That is, each mode relies on, contributes to, and benefits from other
modes. For example, the socialization mode can result in creation of new knowledge when an individual obtains a new insight triggered by interaction with
another. On the other hand, the socialization mode may involve transferring
existing tacit knowledge from one member to a nother through discussion of ideas.
New organizational knowledge per se may not be created, but only knowledge that is new to the recipient. The combination mode in most cases involves an intermediate step - that of an individual drawing insight from explicit sources (i.e., internalization) and then coding the new knowledge into an explicit form (externalization).
Finally, internalization may consist of the simple conversion of existing explicit
knowledge to an individual ’s tacit knowledge as well as creation of new organizational knowledge when the explicit source triggers a new insight.
Knowledge Storage/Retrieval
Empirical studies have shown that while organizations create knowledge and learn,
they also fo rget (i.e., do not remember or lose track of the acquired knowledge)
(Argote et al. 1990; Darr et al. 1995). Thus, the storage, organization, and retrieval
of organizational knowledge, also referred to as organizational memory (Stein and
Zwass 1995; Walsh and Ungson 1991), constitute an important aspect of effective
organizational knowledge management.
Organizational memory includes knowledge residing in various component forms, including written documentation, structured information stored in electronic databases, codified human knowledge stored in expert systems, documented organizational procedures and processes and tacit knowledge acquired by individuals and networks of individuals (Tan et al. 1999).
Like the knowledge creation process, a distinction between individual and organizational memory has been made in the literature. Individual memory is
developed based on a person ’s observations, experiences, and actions (Argyris and munotes.in
Page 17
17Chapter 1: Role of Business ResearchSchön 1978; Nystrom and Starbuck 1981; Sanderlands and Stablein 1987). Collective or organizational memory is defined as “the means by which knowledge
from the past, experience, and events influence present organizational activities”
(Stein and Zwass 1995, p. 85).
Organizational memory extends beyond the individual ’s memory to include other
components such as organizational culture, transformations (production processes
and work procedures), structure (formal organizational roles), ecology (physical
work setting) and information archives (both internal and external to the organization) (Walsh and Ungson 1991).
Organizational memory is classified as semantic or episodic (El Sawy et al. 1996;
Stein and Zwass 1995). Semantic memory refers to general, explicit, and articulated
knowledge (e.g., organizational archives of annual reports), whereas episodic
memory refers to context -specific and situated knowledge (e.g., specific circumstances of organizational decisions and their outcomes, place, and time).
Memory may have both positive and negative potential influences on behavior and
performance. On the positive side, basing and relating organizational change in
experience facilitates implementation of the change (Wilkins and Bristow 1987).
Memory also helps in st oring and reapplying workable solutions in the form of
standards and procedures, which in turn avoid the waste of organizational resources
in replicating previous work. On the other hand, memory has a potential negative
influence on individual and organiza tional performance.
At the individual level, memory can result in decision -making bias (Starbuck and
Hedberg 1977). At the organizational level, memory may lead to maintaining the
status quo by reinforcing single loop learning (defined as a process of det ecting and
correcting errors) (Argyris and Schön 1978). This could in turn lead to stable,
consistent organizational cultures that are resistant to change (Denison and Mishra
1995).
Knowledge Transfer Considering the distributed nature of organizational cognition, an important process of knowledge management in organizational settings is the transfer of
knowledge to locations where it is needed and can be used. However, this is not a
simple process in t hat organizations often do not know what they know and have
weak systems for locating and retrieving knowledge that resides in them (Huber
1991). munotes.in
Page 18
18 RESEARCH IN COMPUTING
Communication processes and information flows drive knowledge transfer in
organizations. Gupta and Govindaraj an (2000) have conceptualized knowledge
transfer (knowledge flows in their terminology) in terms of five elements:
(1) perceived value of the source unit’s knowledge
(2) motivational disposition of the source (i.e., their willingness to share knowledge)
(3) existence and richness of transmission channels
(4) motivational disposition of the receiving unit (i.e., their willingness to acquire
knowledge from the source)
(5) the absorptive capacity of the receiving unit, defined as the ability not only to
acquire and assimilate bu t also to use knowledge (Cohen and Levinthal 1990)
The least controllable element is the fifth: knowledge must go through a recreation
process in the mind of the receiver (El Sawy et al. 1998). This recreation depends
on the recipient’s cognitive capacity to process the incoming stimuli (Vance and
Eynon 19 98).
Knowledge Application
An important aspect of the knowledge -based theory of the firm is that the source
of competitive advantage resides in the application of the knowledge rather than in
the knowledge itself. Grant (1996b) identifies three primary mec hanisms for the integration of knowledge to create organizational capability: directives, organizational routines, and self -contained task teams.
Directives refer to the specific set of rules, standards, procedures, and instructions
developed through the conversion of specialists’ tacit knowledge to explicit and
integrated knowledge for efficient communication to non -specialists (Demsetz
1991). Examples include directives for hazardous waste disposal or airplane safety
checks and maintenance.
Organizational routines refer to the development of task performance and coordination patterns, interaction protocols, and process specifications that allow
individuals to apply and integrate their specialized knowledge without the need to
articulate and communicate w hat they know to others. Routines may be relatively
simple (e.g., organizing activities based on time patterned sequences such as an
assembly line), or highly complex (e.g., a cockpit crew flying a large passenger
airplane). munotes.in
Page 19
19Chapter 1: Role of Business Research
The third knowledge integratio n mechanism is the creation of self -contained task teams. In situations in which task uncertainty and complexity prevent the specification of directives and organizational routines, teams of individuals with
prerequisite knowledge and specialty are formed for problem solving.
Technology can support knowledge application by embedding knowledge into
organizational routines. Procedures that are culture -bound can be embedded into
IT so that the systems themselves become examples of organizational norms.
Knowledge Management may be viewed in terms of:
• People – how do you increase the ability of an individual in the organisation to
influence others with their knowledge.
• Processes – Its approach varies from organization to organization. There is no
limit on the number of processes.
• Technology – It needs to be chosen only after all the requirements of a
knowledge management initiative have been established.
At present knowl edge and its proper management became an essential issue for
every organization. In the modern globalized world, organizations cannot survive in a sustainable way without efficient knowledge management. Knowledge management cycle (KMC) is a process of tran sforming information into knowledge
within an organization, which explains how knowledge is captured, processed, and
distributed in an organization. For the better performance organizations require a
practical and coherent strategy and comprehensive KMC.
Wiig (1999a) discussed about two Knowledge Management C ycles :
a) Institutional Knowledge Evolution Cycle
b) Personal Knowledge Evolution Cycle.
They can help organizations to structure their activities and priorities.
The Institutional Knowledge Evolution Cycle considers five stages as follows
(Wiig, 1999b):
• Knowledge development: Knowledge is developed through learning, innovation, creativity, and importation from outside.
• Knowledge acquisition : Knowledge is captured and retained for use and
further treatment. munotes.in
Page 20
20 RESEARCH IN COMPUTING
• Knowledge refinement : Knowledge is organized, transformed, or includ ed
in written material, knowledge bases, and so on to make it available to be
useful.
• Knowledge distribution and deployment : Knowledge is distributed to
Points -of-Action through education, training programs, automated knowledge -based systems, expert networks, to name a few; to people, practices, embedded in technology and procedures, etc.
• Knowledge leveraging : Knowledge is applied or otherwise leveraged. By
using knowledge, it becomes the basis for further learning and innovation as
explained by other mec hanisms.
The Personal Knowledge Evolution Cycle also has five stages that depict how
knowledge, as it becomes better established in an individual’s mind, migrates from
barely perceived notions to be better understood and useful.
The five stages of this c ycle are as follows (Wiig, 1999b):
• Tacit subliminal knowledge : This knowledge is mostly non -conscious and
is not well understood. It is often the first glimpse we have of a new concept.
• Idealistic vision and paradigm knowledge : Part of this knowledge is well
known to us and explicit and we work consciously with it. Much of it such
as, our visions and mental models is not well known, it is tacit and only
accessible by non -consciously.
• Systematic schema and reference methodology knowledge: Our knowledge of underlying systems, general principles, and problem -solving
strategies is, to a large extent, explicit and mostly well known to us.
• Pragmatic decision -making and factual knowledge : Decision -making
knowledge is practical and mostly explicit. It supports ev eryday work and
decisions, is well known, and is used consciously.
• Automatic routine working knowledge : We know this knowledge so well
that we have automated it. Most has become tacit and we use it to perform
tasks automatically, without conscious reasoni ng.
A major advantage of the Wiig approach to the KMC is the clear and detailed
description of how organizational memory is put into use to generate value for
individuals, groups, and the organization itself. The ways in which knowledge can munotes.in
Page 21
21Chapter 1: Role of Business Research
be applied and used are linked to decision making sequences and individual
characteristics.
Wiig also emphasizes the role of knowledge and skill, the bus iness use of that
knowledge, constraints that may prevent that knowledge from being fully used,
opportunities and alternatives to manage that knowledge, and the expected value
added to the organization (Dalkir, 2005). In brief, the strength of the Wiig KMC is
that it has a clear description of how organizational memory is put into use to
generate value for individuals, groups, and the organization.
Summary
• Business research is the application of the scientific method in searching for
truth about business phenomena.
• Applied business research seeks to facilitate managerial decision making. It
is directed toward a specific managerial decision in a particular o rganization.
• Basic or pure research seeks to increase knowledge of theories and concepts.
• Businesses can make more accurate decisions about dealing with problems
and/or the opportunities to pursue and how to best pursue them.
• Evaluation research is conducted to inform managers whether planned activities were properly executed and whether they accomplished what they
were expected to do.
• Managers determine whether research should be conducted based on time
constraints, availability of data, the nature of t he decision to be made, and the
benefit of the research information versus its cost .
• Business research is the application of the scientific method in searching for
the truth about business phenomena.
• The need to make intelligent, informed decisions ultimat ely motivates an
organization to engage in business research.
• A database is a collection of raw data arranged logically and organized in a
form that can be stored and processed by a computer.
• Knowledge management systems (KMS) refer to a class of information
systems applied to managing organizational knowledge. munotes.in
Page 22
22 RESEARCH IN COMPUTING
• Four modes of knowledge creation have been identified: socialization, externalization, internalization, and combination (Nonaka 1994).
• Wiig (1999a) discussed about two Knowledge Management Cycles: a) Institutional Knowledge Evolution Cycle and b) Personal Knowledge Evolution Cycle.
Review Question
1. Briefly describe decision -making process.
2. State some advantages of internal consultants/researchers.
3. State some disadvantages of external consultants/researchers.
4. When is a business research needed?
5. Write a note on Knowledge Management System.
6. What are the five knowledge transfer elements?
7. Explain the Institutional Knowledge Evolution Cycle.
8. How many stages does the Personal Knowledge Evolution Cycle have?
Briefly explain them.
9. How is Organizational memory classified?
References
• Business Research Methods , Eighth Edition by William G.Zikmund, B.J
Babin, J.C. Carr , Atanu Adhikari, M.Griffin Publisher : Cenage Learning
• Romance of Research, 1933 edition by A.V.H. Redman, L.V and Mory.
Publisher: The Williams and Wilkins Company (1933)
• Legal Research; Some thoughts 78 AIR 1991; J 130 Cited from Book at
Supra Note 11
• Commonly researched topical areas list mentioned in – Research Methods
for Business: A skill building approach , Fifth Edition by Uma Sekaran and
Roger Bougie. Copyright © John Wiley & Sons Ltd
• Research paper titled “Marketing research in the 21st century: opportunities
and challenges”, by Joe F. Hair Jr., Dana E. Harrison and Jeffrey J. Risher.
Available at Research Gate portal : munotes.in
Page 23
23Chapter 1: Role of Business Research
https://www.researchgate.net/publication/328361726_Marketing_Researc
h_in_the_21st_Century_Opportunities_and_Challenges
• Alavi, Maryam & Leidner, Dorothy. (1999). Knowledge Management
Systems: Issues, Challenges, and Benefits. Communication of the Association for Information Systems. 1. 1 -28. 10.17705/1CAIS.00107.
• Paper published by Mohajan, Haradhan, dated 5 October 2016 , on “A
Comprehensive Analysis of Knowledge Management Cycles” available at
https://mpra.ub.uni -muenchen.de/83088/1/MPRA_paper_83088.pdf
• An online article on Internal versus Exter nal Consulting – Advantages and
Disadvantages, by Reagan Cerisano available at
https://www.9lenses.com/internal -versus -external -consulting/?s
munotes.in
Page 24
24 RESEARCH IN COMPUTING
UNIT 2
2 ROLE OF BUSINESS RESEARCH
Unit Structure
2.1 Overview
2.2 Introduction
2.3 Theory Building
2.4 Organization Ethics and Issues
2.5 Summary
2.6 Review Question
2.7 References
2.1 Overview
After studying this chapter, the learner should be able to:
➢ Know general method of theory building research in applied disciplines .
➢ Understand multiple purposes of theory building .
➢ Have in -depth knowledge on research processes.
➢ Understand the perspective of Organizational Ethics .
➢ Gain insights about research problems.
2.2 Introduction
Before considering the generic methodological components of theory building, it
might be helpful to highlight and discuss considerations general to theory -building
research. The first is the notion of the multiple purposes of theory -building r esearch
methods. Second is a brief presentation and description of two commonly used
strategies in theory building. And finally, consideration is given to the requirement
of expertise in both knowledge of and experience with the phenomenon that is the
focus of the theory -building endeavor. munotes.in
Page 25
25Chapter 2: Role of Business Research
Like all abstractions, the word “theory” has been used in many ways , in many different contexts, at times so broadly as to include almost all descriptive statements about a class of phenomena, and at other times so narrow ly as to exclude
everything but a series of terms and their relationships that satisfies certain logical
requirements.
2.3 Theory Building
Figure 2.1: The General Method of Theory Building Research
in Applied Disciplines
Following concepts provide a brief description of each of the five phases of the
general method of applied theory -building research as portrayed in Figure 2.1 ,
referenced from the research paper by Lynham, Susan (2002) – “The General
Method of Theory -Building Research in Ap plied Disciplines.” .
Conceptual Development
Concepts serve critical functions in science, through their descriptive powers and
as the building -blocks of theory. When concepts are immature, therefore, science
suffers. Consequently, concept development ought to be considered a fundamental
scientific activity. Knowledge of the different approaches to concept development,
however, is relatively limited in the management discipline. Concepts abstract
munotes.in
Page 26
26 RESEARCH IN COMPUTING
reality. That is, concepts express in words various events or objects. Concepts,
however, may vary in degree of abstraction. For example, the concept of an asset
is an abstract term that may, in the concrete world of reality, refer to a wide variety
of things, including a specific punch press machine in a production shop.
The different approaches to concept development are known variously as:
• concept analysis (Bohman, 1992; Klausmeier & Goodwin, 1975; Meleis,
1991; Messias, 1996; Rodgers, 1989; Sartori, 1984; Walker & Avant, 1983;
Wilson, 1963; Wuest, 1994)
• concept c larification (Berthold, 1964; Meleis, 1991; Morse, 1995; Norris,
1982)
• concept comparison (Morse, 1995)
• concept correction (Morse, 1995)
• concept delineation (Morse, 1995)
• concept derivation (Walker & Avant, 1983)
• concept development (Chinn & Jacobs, 1983; Meleis, 1991; Schwartz -
Barcott & Kim, 1986)
• concept exploration (Messias, 1996)
• concept identification (Morse, 1995)
• concept integration (Meleis, 1996)
• concept synthesis (Walker & Avant, 1983)
In the end, researchers are concerned with the observable world, or what we shall
loosely term reality. Theorists translate their conceptualization of reality into
abstract ideas. Thus, theory deals with abstraction. Things are not the essence of
theory; ideas are. Concepts in isolation are not theorie s.
Operationalization
Operationalization works by identifying specific indicators that will be taken to
represent the ideas we are interested in studying. It basically involves spelling out
precisely how a concept will be measured. Operational definitions must include the
variable, the measure, and how you plan to interpret the measure. Indexes, scales,
and typologies are used to measure multi -dimensional concepts. It can be helpful
to look at how researchers have measured the concept in previous studies. The
purpose of the operationalization phase of theory -building research is essentially
an explicit connection between the conceptualization phase and practice. munotes.in
Page 27
27Chapter 2: Role of Business ResearchOperationalization reaches towards an overlap between the theorizing and practice components of the theory-building research process. A primary output of the theorizing component of theory-building research in applied disciplines is therefore an operationalized theoretical framework, that is, an informed theoretical framework that has been converted into components or elements that can be further inquired into and confirmed through rigorous research and relevant application. Figure 2.2: Research Process: Conceptualization & Operationalization Researchers need to be aware of the various dimensions of the concepts that are in vogue and clarify which ones they are interested in the context of a research problem. Dimensions are usually ‘concepts’ themselves. In practice often the terms: concept and dimension are used interchangeably. In empirical research we are more interested in dimensions rather than in concepts, per se. Concepts having more than one dimension are known as constructs. Research Problem Concepts
Reliability and Validity
Review of Literature
Nominal Definition
Dimension(s)
Indicators
Levels of Measurement
Indices STAGE I (CONCEPTUALIZATION)
STAGE II STAGE III OPERATIONALIZATION munotes.in
Page 28
28 RESEARCH IN COMPUTING
Once the researcher finalizes the relevant concepts (constructs), their working
definitions and the theoretical framework which links them, the researcher must
determine the appropriate scales of measuring the variability in them. That is the
researcher now must m ove to the operationalization or measurement stage of
concepts of interest. Here the researcher must specify the operations that will
indicate the value of cases on a concept (variable).
Confirmation or Disconfirmation
The confirmation or disconfirmation phase falls within the practice component of
applied theory building. This theory -building phase involves the planning, design,
implementation, and evaluation of an appropriate research agenda and studies to
purposefully inform and intentionally confirm or disconfirm the theoretical framework central to the theory. When adequately addressed, this third phase
results in a confirmed and trustworthy theory that can then be used with some
confidence to inform better action and practice . This is the process of choosing
cases that either serve as supplementary examples that lend further support,
richness and depth to patterns emerging from data analysis (confirming cases) or
serve as examples that do not fit emergent patterns and allow the research team to
evaluate rival explanations (disconfirming cases). This can help the research team
understand and define the limitations of research findings.
Application
A theory that has been confirmed in the contextual world to which it applies (i.e.,
operatio nalized) and has, at least to some extent, gone through inquiry in the
practical world is not enough. A theory must also be threaded through the
application phase. The application of the theory to the problem, phenomenon, or
issue in the world of practice is in the practice component of the general theory -
building research method. Application of the theory enables further study, inquiry,
and understanding of the theory in action. An important outcome of this application
phase of theory building is therefore that it enables the theorist to use the experience
and learning from the real -world application of the theory to further inform,
develop, and refine the theory. It is in the application of a theory that practice gets
to judge and inform the usefulness and relevance of the theory for improved action
and problem solving (Lynham, 2000b). And it is through this application that the
practical world becomes an essential source of knowledge and experience for
ongoing development of applied theory (Ruona & Lynham, 1999; Swanson, 1997) .
munotes.in
Page 29
29Chapter 2: Role of Business Research
Continuous Refinement and Development
Because a theory is never “complete,” it is necessary that the theory be continually
refined and developed (Cohen, 1991; Root, 1993). This recursive nature of applied
theory -building research r equires the ongoing study, adaptation, development, and
improvement of the theory in action and ensures that the relevance and rigor of the
theory are continuously attended to and improved on by theorists through further
inquiry and application in the real world. This continuing phase marks a further
overlap between the practice and theorizing components of applied theory -building
research. This phase addresses the responsibility of continuous attention to the trustworthiness and substantive quality of the theory that is the burden of the
theorist (Dubin, 1978; Van de Ven, 1989). The intentional outcome of this phase is
thus to ensure that the theory is kept current and relevant and that it continues to
work and have utility in the practical world. It al so ensures that when the theory is
no longer useful, or is found to be “false,” that it is shown to be as such and adapted
or discarded accordingly.
At the abstract, conceptual level, a theory may be developed with deductive
reasoning by going from a gener al statement to a specific assertion. Deductive
reasoning is the logical process of deriving a conclusion about a specific instance
based on a known general premise or something known to be true. At the empirical
level, a theory may be developed with induc tive reasoning. Inductive reasoning is
the logical process of establishing a general proposition based on observation of
particular facts. Over the course of time, theory construction is often the result of a
combination of deductive and inductive reasonin g. Our experiences lead us to draw
conclusions that we then try to verify empirically by using the scientific method.
The scientific method is a set of prescribed procedures for establishing and
connecting theoretical statements about events, for analyzing empirical evidence,
and for predicting events yet unknown. It is useful to look at the analytic process
of scientific theory building as a series of stages .
2.4 Organizational Ethics and Issues
Ethics are of interest to business scholars because they influence decisions, behaviors, and outcomes. While scholars have increasingly shown interest in
business ethics as a research topic, there are a mounting number of studies that
examine ethical issues at the organizational level of analysis. munotes.in
Page 30
30 RESEARCH IN COMPUTING
When we talk about organizational ethics, we are referring to the set of values that
identify an organization, from within (or, to put another way, how those working
in the organization understand it) as well as from without (the perception of the
organization by those who have dealings with it). Such a set of values can be
considered in a broad sense (that is, the set of values structuring the organization
and its practices, be they instrumental or final values, positive or negative) or in a
stricter sense (where we shall refer only to those values that express the vision, the
raison d’ être and the commitments of the organization, and that are linked to their
corporate and moral identity).
Generalizing, we could say that in the first case we would find those organiza tions
that ask themselves how to make progress “in search of excellence”; in the second,
those organizations that ask themselves “what is necessary for corporate moral
excellence?”
This means that when speaking of O rganizational Ethics , one can speak from
various perspectives:
One can focus on the practices: from this perspective what is relevant is to
identify the values which in fact structure organizational practices. That is,
basically to become aware.
One can focus on formal statements: from this perspective what is relevant is
to elaborate the discourse which is proposed as a value reference of the
organization. That basically involves formal declarations or statements.
One can focus on the processes: relevant to Organizational Ethics from this
perspective are organizational learning processes which permit continual re -
elaboration and reinterpretation of the relationship between statements and
practices. That is, basically to narrate and institutionalize.
One can focus on the project, stressing wha t, from this perspective, is relevant
to innovation and the creation of corporate identity. Both should be an
expression of the contribution that an organization makes to society in so far
as it is, simultaneously, economic actor and social actor. That is, basically to
build and to institutionalize.
We should not understand these four perspectives as being mutually exclusive –
quite the opposite. However, we should consider their different ‘accents’ and that
they can form an evolutionary sequence. At any r ate, these ‘accents’ should make
us aware of the fact that when we speak of O rganizational Ethics, we are not
referring to a standardized concept but to an option concerning our very
understanding of Organizational Ethics . munotes.in
Page 31
31Chapter 2: Role of Business Research
In our case, we adopt a perspect ive that conceives of Organizational Ethics as an
opportunity for learning and innovation. This also means that we position ourselves
within what we termed earlier a stricter view ( which is one that is not merely
descriptive or instrumental) and which will lead us on to speak of reflective
Organizational Ethics . In other words, it will go beyond Organizational Ethics as a
process of awareness to one in which Organizational Ethics is understood as a
project.
Business ethics is the application of morals t o behavior related to the business
environment or context. Generally, good ethics conforms to the notion of “right,”
and a lack of ethics conforms to the notion of “wrong.” Highly ethical behavior can
be characterized as being fair, just, and acceptable. E thical values can be highly
influenced by one’s moral standards. Moral standards are principles that reflect
beliefs about what is ethical and what is unethical. More simply, they can be
thought of as rules distinguishing right from wrong. The Golden Rule, “Do unto
others as you would have them do unto you,” is one such ethical principle.
An ethical dilemma simply refers to a situation in which one chooses from alternative courses of actions, each with different ethical implications. Each individual develop s a philosophy or way of thinking that is applied to resolve the
dilemmas they face. Many people use moral standards to guide their actions when
confronted with an ethical dilemma. Others adapt an ethical orientation that rejects
absolute principles. Their ethics are based more on the social or cultural acceptability of behavior. If it conforms to social or cultural norms, then it is ethical.
From a moral theory standpoint, idealism is a term that reflects the degree to which
one accepts moral standards as a guide for behavior. Relativism is a term that
reflects the degree to which one rejects moral standards in favor of the acceptability
of some action.
This way of thinking rejects absolute principles in favor of situation -based
evaluations. Thus, an action that is judged ethical in one situation can be deemed
unethical in another. In contrast, idealism is a term that reflects the degree to which
one bases one’s morality on moral standards. Someone who is an ethical idealist
will try to apply ethical princip les like the golden rule in all ethical dilemmas.
Research Problem
A code of ethics is a formal statement of the organization’s ethics and values that
is designed to guide the employees conduct in a variety of business situations.
Business ethics relate to corporate credos like the popular Johnson & Johnson
Credo. A corporate credo indicates a company’s responsibility to its stakeholders, munotes.in
Page 32
32 RESEARCH IN COMPUTING
such as individuals and groups who have an interest in the performance of the
enterprise and how it uses its resources. Gomez -Mejia and Balkin (2002) posit that
stakeholders include employees, customers, and shareholders. They state also that
a corporate credo focuses on principles and beliefs that can provide direction in a
variety of ethically challenging situations. Good corporate credos often emphasize corporate social responsibility (CSR) ethical corporate social responsibility (ECSR) good corporate governance (GCG) as well as the need for business
profitability and sustainability. Sustainability is often confused with CSR, but the
two are not the same.
The return of interest in business ethics that began in the 1970s was in realization
that businesses could be tempted to act immorally and unethically whenever
necessary in pursuit of profit. This interest grew rapidly in later years and almost
reached a crescendo in the 2000s when it became clearer that many heavy global
businesses like Enron collapsed for the most part, due to breaches in GCG and
business ethics. It is now believed, more than ever before, that business ethics are
also instrumental to the pursuit of long -term profit for the business, as well as
prosperity and sustainability for the organization and society.
Organizational sustainability thrives on integrity of the board of dir ectors (BODs).
Integrity is an ethical issue and for sound corporate performance, the Organization
for Economic Co -operation and Development ( OECD ) principles of corporate
governance state that the BODs should exercise leadership and judgment, with
enterpr ise and integrity, to achieve continuing prosperity for the corporation. The
BODs should also act in the best interests of the business enterprise in a manner based on transparency, accountability to shareholders and responsibility to stakeholders. Emphasi s is put on integrity as an important ethical factor in
enterprise prosperity and continuity (Ezeh, 2019).
Organizations will not adequately meet its goals and ensure sustainability where
there are breeches in business ethics and standards. For example, t he accounting
and auditing scandals that led to the, collapse of Enron, WorldCom and many banks in the 1990s/2000s, and the misfortune of Cadbury Nigeria Plc border on management ineffectiveness, indiscipline and failure to observe the principles of
busine ss ethics. Discipline relates to the theory of ethics which Kant (1724 – 1804)
thought as what is morally right or wrong in social conduct. Business ethics
therefore demands a high dose of discipline among members of the BODs of a
company or any other orga nization to be able to run the organization professionally
along ethical lines. munotes.in
Page 33
33Chapter 2: Role of Business ResearchA major challenge of the application of business ethics in many global organizations is the quest for profitability instead of sustainability. Profitability
often has a short -term dimension as against sustainability dimension that is future -
oriented. The inability of BODs to clearly interpret and understand the Friedmanian
theory that while CSR of any business is to increase its profits ethical values drive
wealth creation and rooted in the organizational cultures of the wealth -creating
opportunities. Although this is frequently forgotten because of the prominence
usually given to the value -empty economic theory of profitability. The issue is that
the BODs should apply business ethics by responding to public opinion as expressed by customers, by pressure groups or trade unions or by rules, regulations
and laws. This will reflect an organization suitably structured to effect GCG, as
well as reporting systems structured to provide ethical values, transparency and
accountability. This will support the Porterian theory of creating shared value and
recognizing that the interest of different stakeholders receives responsible weight
(Trevino, 1986; Gellerman, 1989; Van Marrewijk, 2003; A deyemi and Olamide,
2011; Porter and Kramer, 2011; Okaro and Okafor, 2013). In view of the various
organizational problems linked to breaches in business ethics, this researcher
believes that the solution lays in continuous search for the right answer. Eve n
though business ethicists propagate that every business should take the ethical path
to ensure sustainability the extent to which this can be achieved is yet to be
determined.
The research findings by Pflugrath, Martinov -Bennie and Chen state that the co de
of ethics (namely International Standard on Quality Control 1: ISQC1) has a
positive impact on the quality of auditors’ judgments. These findings confirm that the code of ethics benefits the organization in terms of ethical performance improvement and corporate ethical reputation. The code of ethics not only encourages employees to perform ethically and professionally, but it also shows a
signal to the public that a code of ethics exists within a company. Using the code
of ethics is helpful for a company to promote its positive image and reputation. However, the accomplishment of such sustainable development programs as mentioned earlier, requires the support of ethical culture.
A study by Park and Blenkinsopp argues that employees’ compliance to the cod e
of ethics is increased when it is supported by ethical culture. Some evidence
supports that a strong ethical culture in the organization may provide it with
financial benefits. Their reputation for ethical culture is associated with a high level
of emplo yee and customer loyalty. Hence, it can be said that doing business ethically can improve financial performance. Many companies realize the significance of ethical culture, but they need to know how to promote ethical munotes.in
Page 34
34 RESEARCH IN COMPUTING
culture in the workplace. The corporat e ethical virtues (CEV) model developed by
Kaptein is one approach that could be applied to promote the ethical culture of an
organization (Figure 2.3).
Figure 2.3 highlights the link between the eight organizational virtues (clarity, congruency of supervisors, congruency of senior management, feasibility, supportability, transparency, discussability, and sanctionability) and ethical culture. The presence of these organizational virtues could be helpful to develop organizational ethical culture. These organizational virtues are discussed as follows.
Figure 2.3 The CEV model
• First, an organization with the clarity of normative expectations regarding
ethic al conduct assists its employees to distinguish between ethical and
unethical conduct.
• Second, congruency of supervisors and third, congruency of senior management, with these normative expectations, encourage employees to
behave ethically. If supervisors and senior management show an example of
ethical behavior, employees will be motivated to do the same.
• Fourth, feasibility refers to the extent to which an organization sets the
conditions which enable employees to act in accordance with the normative
munotes.in
Page 35
35Chapter 2: Role of Business Research
expectations. Unethical behavior of employees might occur when they lack
the conditions or resources (such as time, budgets, information, etc.) provided
by an organization to fulfill their responsibilities.
• Fifth, supportability refers to the extent to wh ich an organization supports its
employees to meet normative expectations. The organization should be able
to motivate its members to comply with ethical standards.
• Sixth, transparency or visibility in the organization is important to the
encouragement of ethical decision -making and behavior among employees.
It is defined as “the degree to which employee conduct and its consequences
are perceptible to those who can act upon it”.
• Seventh, discussability is the opportunities to discuss ethical issues in a
company. A high level of discussability helps employees to realize their
ethical/unethical behavior to make them feel they are ready to improve their
conduct.
• Eighth, sanctionability refers to the extent to which the organization reinforces ethical behavior in the workplace by rewarding ethical behavior and punishing unethical behavior. The organizational virtue of sanctionability reflects that the organization values ethical behavior .
The CEV model provides an understanding of the determinants of ethical culture.
Organizations with the capability to create ethical virtues can gain success in
developing an ethical culture in the workplace. However, the CEV model should
be investigated further, particularly in different areas or regions where people hold
different cultural values. The CEV model may need to be adapted to a specific area
to effectively produce an ethical culture in various types of organizations as well
as in various coun tries.
Summary
• Concepts serve critical functions in science, through their descriptive powers
and as the building -blocks of theory
• Operationalization reaches towards an overlap between the theorizing and
practice components of the theory -building research process
• The confirmation or disconfirmation phase falls within the practice component of applied theory building.
• At the abstract, conceptual level, a theory may be developed with deductive
reasoning by going from a general statement to a specific asserti on. munotes.in
Page 36
36 RESEARCH IN COMPUTING• One can focus on the practices: from this perspective what is relevant is to identify the values which in fact structure organizational practices. That is,
basically to become aware.
• One can focus on formal statements: from this perspective what is re levant is
to elaborate the discourse which is proposed as a value reference of the
organization. That basically involves formal declarations or statements.
• One can focus on the processes: relevant to Organizational Ethics from this
perspective are organiza tional learning processes which permit continual re -
elaboration and reinterpretation of the relationship between statements and
practices. That is, basically to narrate and institutionalize.
• One can focus on the project, stressing what, from this perspecti ve, is relevant
to innovation and the creation of corporate identity. Both should be an
expression of the contribution that an organization makes to society in so far
as it is, simultaneously, economic actor and social actor. That is, basically to
build an d to institutionalize.
• Business ethics is the application of morals to behavior related to the business
environment or context.
• A code of ethics is a formal statement of the organization’s ethics and values
that is designed to guide the employees conduct in a variety of business
situations.
• As per a research, there are eight organizational virtues.
• First, an organization with the clarity of normative expectations regarding
ethical conduct assists its employees to distinguish between ethical and
unethical conduct.
• Second, congruency of supervisors and third, congruency of senior management, with thes e normative expectations, encourage employees to
behave ethically. If supervisors and senior management show an example of
ethical behavior, employees will be motivated to do the same.
• Fourth, feasibility refers to the extent to which an organization sets the
conditions which enable employees to act in accordance with the normative
expectations. Unethical behavior of employees might occur when they lack
the conditions or resources (such as time, budgets, information, etc.) provided
by an organization to fu lfill their responsibilities. munotes.in
Page 37
37Chapter 2: Role of Business Research
• Fifth, supportability refers to the extent to which an organization supports its
employees to meet normative expectations. The organization should be able
to motivate its members to comply with ethical standards.
• Sixth, tran sparency or visibility in the organization is important to the
encouragement of ethical decision -making and behavior among employees.
It is defined as “the degree to which employee conduct and its consequences
are perceptible to those who can act upon it”.
• Seventh, discussability is the opportunities to discuss ethical issues in a
company. A high level of discussability helps employees to realize their
ethical/unethical behavior to make them feel they are ready to improve their
conduct.
• Eighth, sanctionability refers to the extent to which the organization reinforces ethical behavior in the workplace by rewarding ethical behavior and punishing unethical behavior. The organizational virtue of sanctionability reflects that the organization values ethical beh avior.
Review Question
1. Explain the general method of theory building research.
2. What is operationalization?
3. What is the role of Continuous Refinement and Development in Business
Research?
4. Write a note on Organizational Ethics.
5. Briefly explain corporate ethical virtues model.
References
• Business Research Methods , Eighth Edition by William G.Zikmund, B.J
Babin, J.C. Carr, Atanu Adhikari, M.Griffin Publisher : Cenage Learning
• Lynham, Susan. (2002). The General Method of Theory -Building Research
in Applied Disciplines. Advances in Developing Human Resources. 4. 221 -
241. 10.1177/15234223 02043002. munotes.in
Page 38
38 RESEARCH IN COMPUTING
• Branch, John & Rocchi, Francesco. (2015). Concept Development: A
Primer. Philosophy of Management. 14. 111 -133. 10.1007/s40926 -015-
0011 -9.
• Operationalization concept available online at https://scientificinquiryinsocialwork.pressbooks.com/chapter/9 -3-
operationalization/
• Rao, Dr & R eddy Asi, Dr Vasudeva. (2013). An Examination of The Role
of Conceptualization And Operationalization In Empirical Social Research .
3. 108 -114. (Figure 2.2 has been referenced from the same source)
• McLeod, Michael & Payne, G. Tyge & Evert, Robert. (2014).
Organizational Ethics Research: A Systematic Review of Methods and
Analytical Techniques. Journal of Business Ethics. 134. 1 -15.
10.1007/s10551 -014-2436 -9.
• Lozano, Josep. (2005). An Approach to Organizational Ethics. Ethical
Perspectives. 10. 10.21 43/EP.10.1.503870.
• Ugoani, John. (2019). Business Ethics and its Effect on Organizational
Sustainability. Global Journal of Social Sciences Studies. 5. 119 -131.
10.20448/807.5.2.119.131.
• Wesarat, Phathara -On & Sharif, Mohmad & Majid, A.. (2017). Role of
Organizational Ethics in Sustainable Development: A Conceptual Framework. International Journal Sustainable Future for Human Security.
5. 67 -76. 10.24910/jsustain/5.1/6776. (Figure 2. 3 has been referenced from
the same source)
munotes.in
Page 39
39Chapter 3: Problem Definition: The Foundation of Business Research
UNIT 2
3 PROBLEM DEFINITION: THE FOUNDATION
OF BUSINESS RESEARCH
Unit Structure
3.0 Objective
3.1 Importance of Starting with a Good Problem Definition
3.2 Problem Complexity
3.3 The Problem -Definition Process
3.4 The Problem -Definition Process Steps
3.5 Clarity in Research Questions and Hypotheses
3.6 The Research Proposal
3.7 Anticipating Outcomes
3.0 Objective
• Importance of research
• Understanding the different phases of research
• Beginning with research process
3.1 Importance of Starting with a Good Problem Definition
When the client fails to understand their situation or insists on studying an
irrelevant problem, the research is very likely to fail, even if it is done perfectly.
Translating a business situation into something that can be researched is somewhat
like translating one language int o another. It begins by coming to a consensus on a
decision statement or question. A decision statement is a written expression of the
key question(s) that a research user wishes to answer. It is the reason that research
is being considered. It must be well stated and relevant. The researcher translates
this into research terms by rephrasing the decision statement into one or more
research objectives. These are expressed as deliverables in the research proposal.
The researcher then further expresses these in precise and scientific research
terminology by creating research hypotheses from the research objectives. munotes.in
Page 40
40 RESEARCH IN COMPUTING
For simplicity, the term problem definition is adapted here to refer to the process
of defining and developing a decision statement and the steps involved in
translating it into more precise research terminology, including a set of research
objectives. If this process breaks down at any point, the research will almost
certainly be useless or even harmful. It will be useless if it presents results that
simply are deemed irrelevant and do not assist in decision making. It can be harmful
both because of the wasted resources and because it may misdirect the company in
a poor direction. Ultimately, it is difficult to say that any one step in the research
process is most important. However, formally defining the problem to be attacked
by developing decision statements and translating them into actionable research
objectives must be done well or the rest of the research process is misdirected. Even
a good road map is useless unless you know just where you are going. All of the
roads can be correctly drawn, but they still don’t get you where you want to be.
Similarly, even the best research procedures will not overcome poor problem
definition.
3.2 Problem Complexity
Ultimately, the quality of business research in improving business decisions is
limited by the quality of the problem definition stage. This is far from the easiest
stage of the research process. In deed, it can be the most complex. Figure below
helps to illustrate factors that influence how complex the process can be.
1. Situation Frequency
Many business situations are cyclical. Cyclical business situations lead to
recurring business problems. These pr oblems can even become routine. In
these cases, it is easy to define problems and identify the types of research
that are needed. In some cases, problems are so routine that they can be solved
without any additional research. Recurring problems can even be automated
through a company’s DSS.
For example, pricing problems often occur routinely. Just think about how
the price of gas fluctuates when several stations are located within sight of
each other. One station’s prices definitely affect the sales of the other stations
as well as of the station itself. Similarly, automobile companies, airline
companies, and computer companies, to name just a few, face recurring
pricing issues. Because these situations recur so frequently, addressing them
becomes routin e. Decision makers know how to communicate them to
researchers and researchers know what data are needed. munotes.in
Page 41
41Chapter 3: Problem Definition: The Foundation of Business Research
Figure a: how to define research objectives?
2. Dramatic Changes
When a sudden change in the business situation takes place, it can be easier
to define the problem. For example, if Deland’s business had increased
sharply at the beginning of the year, the key factors to study could be isolated
by identifying other factors that have changed in that same time period. It
could be that a very large truc king contract had been obtained, or that a current customer dramatically increased their distribution needs, which Deland is benefiting from. In contrast, when changes are very subtle and take
effect over a long period of time, it can be more difficult to define the actual
decision and research problems. Detecting trends that would permanently
affect the recruitment challenges that Deland faces can be difficult. It may be
difficult to detect the beginning of such a trend and even more difficult to
know whet her such a trend is relatively permanent or simply a temporary
occurrence.
3. How Widespread are The Symptoms?
The more scattered any symptoms are, the more difficult it is to put them together into some coherent problem statement. In contrast, firms may sometimes face situations in which multiple symptoms exist, but they are all
munotes.in
Page 42
42 RESEARCH IN COMPUTINGpointing to some specific business area. For instance, an automobile manufacturing company may exhibit symptoms such as increased complaints
about a car’s handling, increased warrant y costs due to repairs, higher labor
costs due to inefficiency, and lower performance ratings by consumer advocates such as Consumer Reports. All of these symptoms point to
production as a likely problem area. This may lead to research questions that
deal with supplier -manufacturer relationships, job performance, job satisfaction, supervisory support, and performance. Although having a lot of
problems in one area may not sound very positive, it can be very helpful in
pointing out the direction that is most in need of attention and improvement.
In contrast, when the problems are more widespread, it can be very difficult
to develop useful research questions. If consumer complaints dealt with the
handling and the appearance of the car, and these were accompanie d by
symptoms including consumer beliefs that gas mileage could be better and
that dealerships did not have a pleasant environment, it may be more difficult
to put these scattered symptoms together into one or a few related research
questions. Later in the chapter, we’ll discuss some tools for trying to analyze
symptoms in an effort to find some potential common cause.
4. Symptom Ambiguity
Ambiguity is almost always unpleasant. People simply are uncomfortable
with the uncertainty that comes with ambiguity. Si milarly, an environmental
scan of a business situation may lead to many symptoms, none of which seem
to point in a clear and logical direction. In this case, the problem area remains
vague and the alternative directions are difficult to ascertain. A retail store
may face a situation in which sales and traffic are up, but margins are down.
They may have decreased employee turnover, but lower job satisfaction. In
addition, there may be several issues that arise with their suppliers, none of
which is clearly p ositive or negative. In this case, it may be very difficult to
sort through the evidence and reach a definitive decision statement or list of
research objectives.
3.3 The Problem -Definition Process
3.3.1 Problems Mean Gaps
A problem occurs when there is a difference between the current conditions and a
more preferable set of conditions. In other words, a gap exists between the way
things are now and a way that things could be better. The gap can come about in a
number of ways: munotes.in
Page 43
43Chapter 3: Problem Definition: The Foundation of Business Research
1. Business performance is worse than expected business performance. For
instance, sales, profits, and margins could be below targets set by management. This is a very typical type of problem analysis. Think of all the
new products that fail to meet their targeted goals. Trend analysis would also
be included in this type of problem. Management is constantly monitoring key
performance variables. Previous performance usually provides a benchmark
forming expectations. Sales, for example, are general ly expected to increase
a certain percentage each year. When sales fall below this expectation, or
particularly when they fall below the previous year’s sales, management
usually recognizes that they have a potential problem on their hands. The
Research Sn apshot on the next page illustrates this point.
2. Actual business performance is less than possible business performance.
Realization of this gap first requires that management have some idea of what
is possible. This may form a research problem in and of it self. Opportunity -
seeking often falls into this type of problem -definition process. Many American and European Union companies have redefined what possible sales
levels are based upon the expansion of free markets around the world.
China’s Civil Aviation A dministration has relaxed requirements opening the
Chinese air travel market to private airlines. Suddenly, the possible market
size for air travel has increased significantly, creating opportunities for
growth.
3. Expected business performance is greater than possible business performance. Sometimes, management has unrealistic views of possible
performance levels —either too high or too low. One key problem with new
product introductions involves identifying realistic possibilities for sales.
While you may have heard the old adage that 90 percent of all new products
fail, how many of the failures had a realistic sales ceiling? In other words, did
the company know the possible size of the market? In this case, the problem
is not with the product but with the plan. Some product “failures” may
actually have been successful if management had a more accurate idea of the
total market potential. Managemen t can close this gap through decision
making. Researchers help managers make decisions by providing relevant
input.
munotes.in
Page 44
44 RESEARCH IN COMPUTING
3.4 The Problem -Definition Process Steps
The problem -definition process involves several interrelated steps . Sometimes, the
boundaries be tween each step aren’t exactly clear. But generally, completing one
step leads to the other and by the time the problem is defined, each of these steps
has been addressed in some way. The steps are
1. Understand the business situation —identify key symptoms
2. Identify key problem(s) from symptoms
3. Write managerial decision statement and corresponding research objectives
4. Determine the unit of analysis
5. Determine the relevant variables
6. Write research questions and /or research hypotheses
Fig b: The Problem -Definition Process
1-Understand the Business Decision
A situation analysis involves the gathering of background information to familiarize researchers and managers with the decision -making environment. The
situation analysis can be written up as a way of documenting the problem -definition
process. Gaining an awareness of marketplace conditions and an appreciation of
the situation often requires exploratory research. Researchers sometimes apply
qualitative research with the objective of better p roblem definition. The situation
analysis begins with an interview between the researcher and management.
munotes.in
Page 45
45Chapter 3: Problem Definition: The Foundation of Business Research
Interview Process
The researcher must enter a dialogue with the key decision makers in an effort to
fully understand the situation that has motivated a research effort. This process is
critical and the researcher should be granted access to all individuals who have
specific knowledge of or insight into this situation. Researchers working wit h
managers who want the information “yesterday” often get little assistance when
they ask, “What are your objectives for this study?” Nevertheless, even decision
makers who have only a gut feeling that the research might be a good idea benefit
greatly if t hey work with the researcher to articulate precise research objectives.
Researchers may often be tempted to accept the first plausible problem statement
offered by management. For instance, in the opening vignette, it is clear that David
believes there is a recruitment problem. However, it is very important that the
researcher not blindly accept a convenient problem definition for expediency’s
sake. In fact, research demonstrates that people who are better problem solvers
generally reject problem definitio ns as given to them. Rather, they take information
provided by others and re -associate it with other information in a creative way. This
allows them to develop more innovative and more effective decision statements.
There are many ways to discover problems and spot opportunities.
There is certainly much art involved in translating scattered pieces of evidence
about some business situation into relevant problem statements and then relevant
research objectives. While there are other sources that address crea tive thinking in
detail, some helpful hints that can be useful in the interview process include :
i. Develop many alternative problem statements. These can emerge from the
interview material or from simply rephrasing decision statements and problem statements.
ii. Think about potential solutions to the problem. Ultimately, for the research
to be actionable, some plausible solution must exist. After pairing decision
statements with research objectives, think about the solutions that might
result. This can help make sure any research that results is useful.
iii. Make lists. Use free -association techniques to generate lists of ideas. The
more ideas, the better. Use interrogative techniques to generate lists of
potential questions that can be used in the interview process. Interrogative
techniques simply involve asking multiple what, where, who, when, why,
and how questions. They can also be used to provoke introspection, which
can assist with problem definition. munotes.in
Page 46
46 RESEARCH IN COMPUTING
iv. Be open -minded. It is very important to consider all ideas as plausible in the
beginning stages of problem solving. One sure way to stifle progress is to
think only like those intimately i nvolved in the business situation or only like
those in other industries. Analogies can be useful in thinking more creatively.
Identifying Symptoms
Interviews with key decision makers also can be one of the best ways to identify
key problem symptoms. Recall that all problems have symptoms just as human
disease is diagnosed through symptoms. Once symptoms are identified, then the
researcher must probe to identify possible causes of these changes. Probing is an
interview technique that tries to draw deeper and more elaborate explanations from
the discussion. This discussion may involve potential problem causes. This probing
process will likely be very helpful in identifying key variables that are prim e
candidates for study.
One of the most important questions the researcher can ask during these interviews
is, “what has changed?” Then, the researcher should probe to identify potential
causes of the change. At the risk of seeming repetitive, it is impor tant that the
researcher repeat this process to make sure that some important change has not been
left out. In addition, the researcher should look for changes in company documents,
including financial statements and operating reports. Changes may also be
identified by tracking down news about competitors and customers.
2-Identifying the Relevant Issues from the Symptoms
Anticipating the many influences and dimensions of a problem is impossible for
any researcher or executive. The preceding interview is extremely useful in
translating the decision situation into a working problem definition by focusing on
symptoms. However, the researcher needs to be doubly certain that the research
attacks real problems and not superficial symptoms. For instance, when a firm has
a problem with advertising effectiveness, the possible causes of this problem may
be low brand awareness, the wrong brand image, use of the wrong media, or perhaps
too small a budget. Certain occurrences that appear to be the problem may be only
symptoms of a deeper problem.
3-Writing Managerial Decision Statements and Corresponding Research Objectives
The situation analysis ends once researchers have a clear idea of the managerial
objectives from the research effort. Decision statements capture these objectives in
a way that invites multiple solutions. Multiple solutions are encouraged by using munotes.in
Page 47
47Chapter 3: Problem Definition: The Foundation of Business Research
plural nouns to describe solutions. In other words, a decision statement that says in
what “ways” a problem can be solved is better than one that says in what “way” a
problem can be solved. Ultimately, research may provide evidence showing results
of several ways a problem can be attacked.
Decision statements must be translated into research objectives. At this point, the
researcher is starting to visualize what will need to be measured and what type of
study will be needed.
What information or data will be needed to help answer this question? Obviously,
we’ll need to study the drive r census and the number of hires needed to fill open
positions. James needs to find out what might cause employee dissatisfaction and
cause turnover to increase. Thinking back to the interview, James knows that there
have been several changes in the compan y itself, many related to saving costs.
Saving costs sounds like a good idea; however, if it harms driver loyalty Even
slightly, it probably isn’t worthwhile. Thus, the corresponding research objectives
are stated as follows:
Determine what key variables r elate to driver loyalty within the company,
meaning (1) how does the lower level of pay impact driver retention and (2)
what does the increase in long-haul trucking do to Deland Trucking’s ability to
increase retention?
Assess the impact of different inter vention strategies on driver satisfaction These research objectives are the deliverables of the research project. A research study will be conducted that (1) shows how much each of several key
variables relates to loyalty and retention and (2) provides a description of
likelihood of different intervention strategies on driver satisfaction.
The researcher should reach a consensus agreement with the decision maker regarding
the overall decision statement(s) and research objectives. If the decision maker agrees
that the statement captures the situation well and understands how the research
objectives, if accomplished, will help address the situation, then the researcher can
proceed. The researcher should make every effort to ensure that the decision maker
understands what a research project can deliver. If there is no agreement on the
decision statement or research objectives, more dialogue between decision makers and
researchers is needed.
4-Determine the Unit of Analysis
The unit of analysis for a study indicates what or who should provide the data and
at what level of aggregation. Researchers specify whether an investigation will collect data about individuals (such as customers, employees, and owners), munotes.in
Page 48
48 RESEARCH IN COMPUTING
households (families, extended families, and so forth), organizations (businesses
and business units), departments (sales, finance, and so forth), geographical areas,
or objects (products, advertisements, and so forth). In studies of home buying, for
example, the husband/wife d yad typically is the unit of analysis rather than the
individual because many purchase decisions are made jointly by husband and wife.
Researchers who think carefully and creatively about situations often discover that
a problem can be investigated at more than one level of analysis. For example, a
lack of worker productivity could be due to problems that face individual
employees or it could reflect problems that are present in entire business units.
Determining the unit of analysis should not be overlooked during the problem -
definition stage of the research.
5-Determine Relevant Variables
What is a Variable?
What things should be studied to address a decision statement? Researchers answer
this question by identifying key variables. A variable is anything that varies or
changes from one instance to another. Variables can exhibit differences in value,
usually in magnitude or strength, or in direction. In research, a variable is either
observed or manipulated, in which case it is an experimental variable.
The converse of a variable is a constant . A constant is something that does not
change. Constants are not use ful in addressing research questions. Since constants
don’t change, management isn’t very interested in hearing the key to the problem
is something that won’t or can’t be changed. In causal research, it can be important
to make sure that some potential var iable is actually held constant while studying the cause and effect between two other variables. In this way, a spurious relationship can be ruled out. At this point however, the notion of a constant is more
important in helping to understand how it differ s from a variable.
Types of Variables
There are several key terms that help describe types of variables. The variance in
variables is captured either with numerical differences or by an identified category
membership. In addition, different terms describe whether a variable is a potential
cause or an effect.
A continuous variable is one that can take on a range of values that correspond to
some quantitative amount. Consumer attitude toward different airlines is a variable
that would generally be captured by numbers, with higher numbers indicating a
more positive attitude than lower numbers. Each attribute of airlines’ services, such munotes.in
Page 49
49Chapter 3: Problem Definition: The Foundation of Business Research
as safety, seat comfort, and baggage handling can be numerically scored in this
way. Sales volume, profits, and margin are common business metrics that represent
continuous variables.
A categorical variable is one that indicates membership in some group. The term
classificatory variable is sometimes also used and is generally interchangeable
with categorical variable . Categorical variables sometimes represent quantities
that take on only a small number of values (one, two, or three). However,
categorical variables more often simply identify membership.
For example, people can be categorized as either male or female. A variable
representing biological sex describes this important difference. The variable values
can be an “M” for membership in the male category and an “F” for membership in
the female category. Alternatively, the res earcher could assign a “0” for men and a
“1” for women. In either case, the same information is represented.
A common categorical variable in consumer research is adoption, meaning the
consumer either did or did not purchase a new product. Thus, the two g roups,
purchase or not purchase, comprise the variable. Similarly, turnover, or whether an
employee has quit or not, is a common organizational variable.
In descriptive and causal research, the terms dependent variable and independent
variable describe di fferent variable types. This distinction becomes very important
in understanding how business processes can be modeled by a researcher. The
distinction must be clear before one can correctly apply certain statistical procedures like multiple regression ana lysis. In some cases, however, such as when
only one variable is involved in a hypothesis, the researcher need not make this
distinction.
A dependent variable is a process outcome or a variable that is predicted and/or
explained by other variables. An independent variable is a variable that is
expected to influence the dependent variable in some way. Such variables are
independent in the sense that they are determined outside of the process being
studied. That is another way of saying that dependent variables do not change
independent variables.
For example, average customer loyalty may be a dependent variable that is influenced or predicted by an independent variable such as perceptions of restaurant food quality, service quality, and customer satisfaction. Thus, a process
is described by which several variables together help create and explain how much
customer loyalt y exists. In other words, if we know how a customer rates the food munotes.in
Page 50
50 RESEARCH IN COMPUTING
quality, service quality, and satisfaction with a restaurant, then we can predict that
customer’s loyalty toward that restaurant. Note that this does not mean that we can
predict food quali ty or service quality with customer loyalty.
Dependent variables are conventionally represented by the letter Y. Independent
variables are conventionally represented by the letter X. If research involves two
dependent variables and two or more independent variables, subscripts may also be
used to indicate Y1, Y2 and X1, X2, and so on.
Ultimately, theory is critical in building processes that include both independent
and dependent variables. Managers and researchers must be careful to identify
relevant and actionable variables. Relevant means that a change in the variable
matters and actionable means that a variable can be controlled by managerial
action. Superfluous variables are those that are neither relevant nor actionable and
should not be included in a study. Theory should help distinguish relevant from
superfluous variables. The process of identifying the relevant variables overlaps
with the process of determining the research objectives. Typically, each research
objective will mention a variable or v ariables to be measured or analyzed. As the
translation process proceeds through research objectives, research questions, and
research hypotheses, it is usually possible to emphasize the variables that should be
included in a study .
6-Write Research Objectives and Questions
Both managers and researchers expect problem -definition efforts to result in
statements of research questions and research objectives. At the end of the problem -
definition stage, the researcher should prepare a written sta tement that clarifies any
ambiguity about what the research hopes to accomplish. This completes the
translation process.
Research questions express the research objectives in terms of questions that can
be addressed by research. For example, one of the ke y research questions involved
in the opening vignette is “Are wages and long -haul distance related to driver
loyalty and retention?” Hypotheses are more specific than research questions. One
key distinction between research questions and hypotheses is that hypotheses can
generally specify the direction of a relationship. In other words, when an
independent variable goes up, we have sufficient knowledge to predict that the
dependent variable should also go up (or down as the case may be). One key
research hy pothesis for Deland Trucking is:
Higher cents per mile are related positively to driver loyalty. munotes.in
Page 51
51Chapter 3: Problem Definition: The Foundation of Business Research
At times, a researcher may suspect that two variables are related but have
insufficient theoretical rationale to support the relationship as positive or negative.
In this case, hypotheses cannot be offered. At times in research, particularly in
exploratory research, a proposal can only offer research questions. Research
hypotheses are much more specific and therefore require considerably more theoretical support. In addition, research questions are interrogative, whereas
research hypotheses are declarative.
3.5 Clarity in Research Questions and Hypotheses
Research questions make it easier to understand what is perplexing managers and
to indicate what issues have to be resolved. A research question is the researcher’s
translation of the marketing problem into a specific inquiry .
A research question can be too vague and general, such a s “Is advertising copy 1
better than advertising copy 2?” Advertising effectiveness can be variously
measured by sales, recall of sales message, brand awareness, intention to buy,
recognition, or knowledge, to name a few possibilities. Asking a more specif ic
research question (such as, “Which advertisement has a higher day after recall
score?”) helps the researcher design a study that will produce useful results, as seen
in the Research Snapshot above. Research question answers should provide input
that can be used as a standard for selecting from among alternative solutions.
Problem definition seeks to state research questions clearly and to develop well -
formulated, specific hypotheses.
A sales manager may hypothesize that salespeople who show the highest j ob
satisfaction will be the most productive. An advertising manager may believe that
if consumers’ attitudes toward a product are changed in a positive direction,
consumption of the product also will increase.
Hypotheses are statements that can be empirica lly tested. A formal hypothesis has considerable practical value in planning and designing research. It forces researchers to be clear about what they expect to find through the study, and it
raises crucial questions about data required. When evaluating a hypothesis,
researchers should ensure that the information collected will be useful in decision
making. Notice how the following hypotheses express expected relationships
between variables:
• There is a positive relationship between buying on the Internet and the
presence of younger children in the home. munotes.in
Page 52
52 RESEARCH IN COMPUTING
• Sales are lower for salespeople in regions that receive less advertising
support .
• Consumers will experience cognitive dissonance after the decision to adopt
a TiVo personal video recorder.
• Opinion leaders are more affected by mass media communication sources
than are non-leaders.
• Among non -exporters, the degree of perceived importance of overcoming
barriers to exporting is related positively to general interest in exporting
(export intentions).
Management is often faced with a “go/no go” decision. In such cases, a
research questio n or hypothesis may be expressed in terms of a meaningful
barrier that represents the turning point in such a decision. In this case, the
research involves a managerial action standard that specifies a specific
performance criterion upon which a decision can be based.
Figure c : Influence of Decision Statement of Marketing Problem on
Research Objectives and Research Designs
3.5.1 How Much Time Should Be Spent on Problem Definition?
Budget constraints usually influence how much effort is spent on problem definition. Business situations can be complex and numerous variables may be
relevant. Searching for every conceivable cause and minor influence is impractical.
munotes.in
Page 53
53Chapter 3: Problem Definition: The Foundation of Business Research
The more i mportant the decision faced by management, the more resources should
be allocated toward problem definition. While not a guarantee, allowing more time
and spending more money will help make sure the research objectives that result
are relevant and can demonstrate which influences management should focus on.
Managers, being re sponsible for decision making, may wish the problem -definition
process to proceed quickly. Researchers who take a long time to produce a set of
research objectives can frustrate managers. However, the time taken to identify the
correct problem is usually t ime well spent.
3.6 The Research Proposal
The research proposal is a written statement of the research design. It always
includes a statement explaining the purpose of the study (in the form of research
objectives or deliverables) and a definition of the p roblem, often in the form of a
decision statement. A good proposal systematically outlines the particular research
methodology and details procedures that will be used during each stage of the
research process. Normally a schedule of costs and deadlines is included in the
research proposal. The research proposal becomes the primary communication
document between the researcher and the research user.
The Proposal as a Planning Tool
Preparation of a research proposal forces the researcher to think critically about
each stage of the research process. Vague plans, abstract ideas, and sweeping
generalizations about problems or procedures must become concrete and precise
statements about specific events. Data requirements and research procedures must
be specified clearly so others may understand their exact implications. All ambiguities about why and how the research will be conducted must be clarified
before the proposal is complete.
The researcher submits the proposal to management for acceptance, modification,
or rejection. Research clients (management) evaluate the proposed study with
particular emphasis on whether or not it will provide useful information, and
whether it will do so wi thin a reasonable resource budget. Initial proposals are
almost always revised after the first review.
The proposal helps managers decide if the proper information will be obtained and
if the proposed research will accomplish what is desired. If the probl em has not
been adequately translated into a set of specific research objectives and a research munotes.in
Page 54
54 RESEARCH IN COMPUTING
design, the client’s assessment of the proposal will help ensure that the researchers
revise it to meet the client’s information needs.
An effective proposal co mmunicates exactly what information will be obtained,
where it will be obtained, and how it will be obtained. For this reason, it must be
explicit about sample selection, measurement, fieldwork, and data analysis. For
instance, most proposals involving descriptive research include a proposed questionnaire (or at least some sample questions).
The Proposal as a Contract
When the research will be conducted by a consultant or an outside research
supplier, the written proposal serves as th at person’s bid to offer a specific service.
Typically, a client solicits several competitive proposals, and these written offers
help management judge the relative quality of alternative research suppliers.
A wise researcher will not agree to do a resear ch job for which no written proposal
exists. The proposal also serves as a contract that describes the product the research
user will buy. In fact, the proposal is in many ways the same as the final research
report without the actual results. Misstatements and faulty communication may
occur if the parties rely only on each individual’s memory of what occurred at a
planning meeting. The proposal creates a record, which greatly reduces conflicts
that might arise after the research has been conducted. Both the researcher and the
research client should sign the proposal indicating agreement on what will be done.
The proposal then functions as a formal, written statement of agreement between
marketing executives and researchers. As such, it protects the researche r from
criticisms such as, “Shouldn’t we have had a larger sample?” or “Why didn’t you
use a focus group approach?” As a record of the researcher’s obligation, the
proposal also provides a standard for determining whether the actual research was
conducted as originally planned. Suppose in our Deland Trucking case, following
the research, David is unhappy with the nature of the results because they indicate
that higher cents per mile do, in fact, impact driver loyalty.
Funded business research generally ref ers to basic research usually performed
by academic researchers and supported by some public or private institution. Most commonly, researchers pursue federal government grants. A very detailed proposal is usually needed for federal grants, and the agreeme nt for funding is
predicated on the research actually delivering the results described in the proposal.
One important comment needs to be made about the nature of research proposals.
Not all proposals follow the same format. A researcher can adapt his or her munotes.in
Page 55
55Chapter 3: Problem Definition: The Foundation of Business Research
Figure d: Basic Points Addressed by Research Proposals
proposal to the target audience or situation. An extremely brief proposal submitted
by an organization’s internal research department to its own executives bears little
resemblance to a complex proposal submitted by a university professor to a federal
government ag ency to research a basic consumer issue.
munotes.in
Page 56
56 RESEARCH IN COMPUTING
3.7 Anticipating Outcomes
As mentioned above, the proposal and the final research report will contain much
of the same information. The proposal describes the data collection, measurement,
data analysis, and so f orth, in future tense. In the report, the actual results are
presented. In this sense, the proposal anticipates the research outcome. Experienced
researchers know that research fails more often because the problem -definition
process breaks down or because the research client never truly understood what a
research project could or couldn’t do. While it probably seems as though the
proposal should make this clear, any shortcoming in the proposal can contribute to
a communication failure. Thus, any tool that helps communication become as clear
as can be is valued very highly.
Dummy Tables
One such tool that is perhaps the best way to let management know exactly what
kind of results wil l be produced by research is the dummy table. Dummy tables are
placed in research proposals and are exact representations of the actual tables
that will show results in the final report with one exception: The results are
hypothetical. They get the name because the researcher fills in, or “dummies up,”
the tables with likely but fictitious data. Dummy tables include the tables that will
present hypothesis test results. In this way, they are linked directly to research
objectives
A research analyst can present dummy tables to the decision maker and ask, “Given
findings like these, will you be able to make a decision?” If the decision maker says
yes, the proposal may be accepted. However, if the decision maker cannot see how
results like those in the dummy tables will help make the needed decision(s), it may
be back to the drawing board. In other words, the client and researcher need to
rethink what research results are n ecessary to solve the problem. Sometimes,
examining the dummy tables may reveal that a key variable is missing or that some
dependent variable is really not relevant. In other words, the problem is clarified by
deciding on action standards or performance c riteria and recognizing the types of
research findings necessary to make specific decisions.
Summary
The chapter explains the importance of research along the process of research. It
also discuss about Cyclical business situations lead to recurring busines s
problems. The problem definition process identifies the problem gaps and also
includes various steps such as writing the decision statement, determining the unit munotes.in
Page 57
57Chapter 3: Problem Definition: The Foundation of Business Research
of analysis and relevant variables. A good proposal systematically outlines the
particular r esearch methodology and details procedures that will be used during
each stage of the research process. A research analyst can present dummy tables
to the decision maker and ask, “Given findings like these, will you be able to make
a decision?” If the deci sion maker says yes, the proposa l may be accepted research
findings necessary to make specific decisions.
Questions
1. Explain the process of research with sui table example.
2. Explain how research objectives are defined.
3. List the activities use for defining and identifying the research problem.
4. What is variable? Explain the types of variables used.
5. Explain the importance of survey and interview with respect to research
problem.
6. What is the influence of decision statement of marketing problem on resea rch
objectives and research designs?
7. Discuss the relation of hypothesis with respect to research.
8. Write a note on Dummy table .
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• 2. Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable,
• Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing
Corporation, ISBN: 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016,
• Business Research Methods, Edition 8, Cengage Publication
munotes.in
Page 58
58 RESEARCH IN COMPUTING
UNIT 2
4 QUALITATIVE RESEARCH TOOLS
4.0 Objectives
4.1 Introduction
4.2 What Is Qualitative Research?
4.3 Uses of Qualitative Research
4.4 Qualitative “versus” Quantitative Research
4.5 Contrasting Qualitative and Quantitative Methods
4.6 Contrasting Exploratory and Confirmatory Research
4.7 Orientations to Qualitative Research
4.8 Phenomenology
4.9 What Is Hermeneutics?
4.10 Ethnography
4.11 Observation in Ethnography
4.12 Grounded Theory
4.13 How Is Grounded Theory Used?
4.14 Case Studies
4.15 How are Case Studies Used?
4.16 Common Techniques Used in Qualitative Research
4.17 Focus Group Illustration
4.18 Environmental Conditions
4.19 The Focus Group Moderator
4.20 Planning The Focus Group Outline
4.21 Focus Groups As Diagnostic Tools
4.22 Video Conferencing And Focus Groups
4.23 Interactive Media and Online Focus Groups
4.24 Online versus Face -To-Face Focus Group Techniques
4.25 Projective Research Techniques
4.26 Exploratory Research in Science and in Practice
4.27 Scientific Decision Processes
58munotes.in
Page 59
59Chapter 4: Qualitative Research Tools
4.0 Objectives
• Understanding types of research
• Process of qualitative research
• Process of quantitative research
• Exploratory research
4.1 Introduction:
Chemists sometimes use the term qualitative analysis to mean research that
determines what some compound is made of. In other words, the focus is on the
inner meaning of the chemical — its qualities . As the word implies, qualitati ve
research is interested more in qualities than quantities. Therefore, qualitative
research is not about applying specific numbers to measure variables or using
statistical procedures to numerically specify a relationship’s strength.
4.2 What Is Qualitati ve Research?
Qualitative business research is research that addresses business objectives
through techniques that allow the researcher to provide elaborate interpretations of
market phenomena without depending on numerical measurement. Its focus is on
discovering true inner meanings and new insights. Qualitative research is very
widely applied in practice. There are many research firms that specialize in
qualitative research.
Qualitative research is less structured than most quantitative approaches. It doe s not
rely on self - response questionnaires containing structured response formats. Instead, it is more researcher -dependent in that the researcher must extract
meaning from unstructured responses, such as text from a recorded interview or a
collage representing the meaning of some experience, such as skateboarding. The
researcher interprets the data to extract its meaning and converts it to information.
4.3 Uses of Qualitative Research
The researcher has many tools available and the research design s hould try to match
the best tool to the research objective. Not every researcher has expertise with tools
that would comprise qualitative research. Generally, the less specific the research
objective, the more likely that qualitative research tools will be appropriate. Also, munotes.in
Page 60
60 RESEARCH IN COMPUTING
when the emphasis is on a deeper understanding of motivations or on developing
novel concepts, qualitative research is very appropriate. The following list
represents common situations that often call for qualitative research: 1. When it is difficult to develop specific and actionable problem statements or
research objectives. For instance, if after several interviews with the research client
the researcher still can’t determine exactly what needs to be measured, then
qualitative research approaches may help with problem definition. Qualitative
research is often useful to gain further insight and crystallize the research problem. 2. When the research objective is to develop an understanding of some phenomena in
great detail and in much depth. Qualitative research tools are aimed at discovering
the primary themes indicating human motivations and the documentation of
activities is usually very complete. Often qualitative research provides richer
information than quantitative approaches. 3. When the research objective is to learn how a phenomena occurs in its natural
setting or to learn how to express some concept in colloquial terms. For example,
how do consumers actually use a product? Or, exactly how does the accounting
department process invoices? While a survey can probably ask many useful
questions, observing a product in use or watching the invoice process will usually
be more insightful. Qualitative research produces many product and process
improvement ideas. 4. When some behavior the re searcher is studying is particularly context dependent —
meaning the reasons something is liked or some behavior is performed depend very
much on the particular situation surrounding the event. Understanding why Vans
are liked is probably difficult to dete rmine correctly outside the skating
environment. 5. When a fresh approach to studying some problem is needed. This is particularly the
case when quantitative research has yielded less than satisfying results. Qualitative
tools can yield unique insights, many of which may lead the organization in new
directions.
4.4 Qualitative “versus” Quantitative Research
In social science, one can find many debates about the superiority of qualitati ve
research over quantitative research or vice versa. We’ll begin by saying that this is
largely a superfluous argument in either direction. The truth is that qualitative
research can accomplish research objectives that quantitative research cannot.
Simila rly, truthful, but no more so, quantitative research can accomplish objectives
that qualitative research cannot. The key to successfully using either is to match the munotes.in
Page 61
61Chapter 4: Qualitative Research Tools
right approach to the right research context.
Many good research projects combine both qualitative and quantitative research.
For instance, developing valid survey measures requires first a deep understanding
of the concept to be measured and a descriptio n of the way these ideas are expressed
in everyday language. Both of these are tasks best suited for qualitative research.
However, validating the measure formally to make sure it can reliably capture the
intended concept will likely require quantitative r esearch. Also, qualitative research
may be needed to separate symptoms from problems and then quantitative research
can follow up to test relationships among relevant variables.
Quantitative business research can be defined as business research that addre sses research objectives through empirical assessments that involve numerical measurement and analysis approaches. Qualitative research is more apt to stand on
its own in the sense that it requires less interpretation. For example, quantitative
research is quite appropriate when a research objective involves a managerial
action standard. For example, a salad dressing company considered changing its
recipe. The new recipe was tested with a sample of consumers. Each consumer
rated the product using numeric sc ales. Management established a rule that a
majority of consumers rating the new product higher than the old product would
have to be established with 90 percent confidence before replacing the old formula.
A project like this can involve both quantitative measurement in the form of
numeric rating scales and quantitative analysis in the form of applied statistical
procedures.
4.5 Contrasting Qualitative and Quantitative Methods
Quantitative researchers direct a considerable amount of activity toward measurin g
concepts with scales that either directly or indirectly provide numeric values. The
numeric values can then be used in statistical computations and hypothesis testing.
In contrast, qualitative researchers are more interested in observing, listening, and
interpreting. As such, the researcher is intimately involved in the research process
and in constructing the results. For these reasons, qualitative research is said to be
more subjective , meaning that the results are researcher -dependent. Different
resear chers may reach different conclusions based on the same interview. In that
respect, qualitative research lacks intersubjective certifiability , the ability of
different individuals following the same procedures to produce the same results or
come to the sam e conclusion. This should not necessarily be considered a weakness
of qualitative research; rather it is simply a characteristic that yields differing
insights. In contrast, when a survey respondent provides a commitment score on a munotes.in
Page 62
62 RESEARCH IN COMPUTING
quantitative scale, it i s thought to be more objective because the number will be the
same no matter what researcher is involved in the analysis.
Exhibit below on the next page illustrates some differences between qualitative and
quantitative research. Certainly, these are generalities and exceptions may apply.
However, it covers some of the key distinctions. Qualitative research seldom involves samples with hundreds of respondents. Instead, a ha ndful of people are usually the source of qualitative data. This is
perfectly acceptable in discovery -oriented research. All ideas would still have to be
tested before adopted. Does a smaller sample mean that qualitative research is
cheaper than qualitative? Perhaps not. Although fewer respondents must be interviewed, the greater researcher involvement in both the data collection and
analysis can drive up the costs of qualitative research.
Given the close relationship between qualitative research and explor atory designs,
it should not be surprising that qualitative research is most often used in exploratory
designs. Small samples, interpretive procedures that require subjective judgments,
and the unstructured interview format all make traditional hypotheses testing
difficult with qualitative research. Thus, these procedures are not best suited for drawing definitive conclusions, as would be expected from causal designs involving experiments. These disadvantages for drawing inferences, however,
become advantag es when the goal is to draw out potential explanations because the
researcher spends more time with each respondent and is able to explore much more
ground due to the flexibility of the procedures.
Figure a: Comparing Qualitative and Quantitative Research
munotes.in
Page 63
63Chapter 4: Qualitative Research Tools
4.6 Contrasting Exploratory and Confirmatory Research
Philosophically, research can be considered as either exploratory or confirmatory.
Most exploratory research designs produce qualitative data . Exploratory designs
do not usually produce quantitative data, which represent phenomena by assigning numbers in an ordered and meaningful way. Rat her than numbers, the
focus of qualitative research is on stories, visual portrayals, meaningful characterizations, interpretations, and other expressive descriptions. Often, exploratory research may be needed to develop the ideas that lead to research
hypotheses. In other words, in some situations the outcome of exploratory research
is a testable research hypothesis. Confirmatory research then tests these hypotheses
with quantitative data. The results of these tests help decision making by suggesting
a spe cific course of action. For example, an exploratory researcher is more likely
to adopt a qualitative approach that might involve trying to develop a deeper
understanding of how families are impacted by changing economic conditions,
investigating how people suffering economically spend scarce resources. This may
lead to the development of a hypothesis that during challenging economic times
consumers seek low -cost entertainment such as movie rentals, but would not test
this hypothesis. In contrast, a quantita tive researcher may search for numbers that
indicate economic trends. This may lead to hypothesis tests concerning how much
the economy influences rental movie consumption.
Some types of qualitative studies can be conducted very quickly. Others take a ver y
long time. For example, a single focus group analysis involving a large bottling
company’s sales force can likely be conducted and interpreted in a matter of days.
This would provide faster results than most descriptive or causal designs. However,
other types of qualitative research, such as a participant - observer study aimed at understanding skateboarding, could take months to complete. A qualitative approach can, but does not necessarily, save time.
In summary, when researchers have limited experienc e or knowledge about a
research issue, exploratory research is a useful step. Exploratory research, which
often involves qualitative methods, can be an essential first step to a more conclusive, confirmatory study by reducing the chance of beginning with an
inadequate, incorrect, or misleading set of research objectives.
4.7 Orientations to Qualitative Research
Qualitative research can be performed in many ways using many techniques.
Orientations to qualitative research are very much influen ced by the different fields munotes.in
Page 64
64 RESEARCH IN COMPUTING
of study involved in research. These orientations are each associated with a
category of qualitative research. The major categories of qualitative research
include
1. Phenomenology —originating in philosophy and psychology
2. Ethnograp hy—originating in anthropology
3. Grounded theory —originating in sociology
4. Case studies —originating in psychology and in business research
Precise lines between these approaches are difficult to draw and there are clearly
links among these orientations. In addition, a particular qualitative research study
may involve elements of two or more approaches. However, each category does
reflect a somewhat unique approach to human inquiry and approaches to discovering knowledge. Each will be described briefly below.
4.8 Phenomenology
What Is A Phenomenological Approach To Research?
Phenomenology represents a philosophical approach to studying human experiences based on the idea that human experience itself is inherently subjective
and determined by the context in which people live.10 The phenomenological
researcher focuses on how a person’s behavi or is shaped by the relationship he or
she has with the physical environment, objects, people, and situations. Phenomenological inquiry seeks to describe, reflect upon, and interpret
experiences.
Researchers with a phenomenological orientation rely largely on conversational
interview tools. When conversational interviews are face to face, they are recorded either with video or audiotape and then interpreted by the researcher. The phenomenological i nterviewer is careful to avoid asking direct questions when at
all possible. Instead, the research respondent is asked to tell a story about some
experience. In addition, the researcher must do everything possible to make sure a
respondent is comfortable telling his or her story. One way to accomplish this is to
become a member of the group . Another way may be to avoid having the person
use his or her real name. This might be particularly necessary in studying
potentially sensitive topics such as smoking, drug usage, shoplifting, or employee
theft. munotes.in
Page 65
65Chapter 4: Qualitative Research Tools
Therefore, a phenomenological approach to studying the meaning of Vans may
require considerable time. The researcher may first spend weeks or months fitting
in with the person or group of interest to establish a comfort level. During this time,
careful notes of conversations are made. If an interview is sought, the researcher
would likely not begin by asking a skateboarder to describe his or her shoes. Rather,
asking for favorite skateboard incidents or talking about what makes a skateboarder
unique may generate productive conversation. Generally, the approach is very
unstructured as a way of avoiding leading questions and to provide every opportunity for new insights.
4.9 What Is Hermeneutics?
The term hermeneutics is important in phenomenology. Hermeneutics is an
approach to understanding phenomenology that relies on analysis of texts in which
a person tells a story about him or herself.12 Meaning is then drawn by connecting
text passages to one another or to themes expressed outside the story. These
connections are usually facilitated by coding the key meanings expressed in the
story. While a full understanding of hermeneutics is beyond the scope of this text,
some of the terminology is used wh en applying qualitative tools. For instance, a
hermeneutic unit refers to a text passage from a respondent’s story that is linked
with a key theme from within this story or provided by the researcher. These
passages are an important way in which data are i nterpreted.
Computerized software exists to assist in coding and interpreting texts and images.
ATLAS.ti is one such software package that adopts the term hermeneutic unit in
referring to groups of phrases that are linked with meaning. Hermeneutic units a nd
computerized software are also very appropriate in grounded theory approaches.
One useful component of computerized approaches is a word counter. The word
counter will return counts of how many times words were used in a story. Often,
frequently occurri ng words suggest a key theme. The Research Snapshot above
demonstrates the use of hermeneutics in interpreting a story about a consumer
shopping for a car
4.10 Ethnography
What Is Ethnography?
Ethnography represents ways of studying cultures through methods that involve
becoming highly active within that culture. Participant -observation typifies an munotes.in
Page 66
66 RESEARCH IN COMPUTINGethnographic research approach. Participantobservation means the researcher becomes immersed within the cultur e that he or she is studying and draws data from
his or her observations. A culture can be either a broad culture, like American
culture, or a narrow culture, like urban gangs, Harley -Davidson owners, or
skateboarding enthusiasts.
Organizational culture w ould also be relevant for ethnographic study.15 At times,
researchers have actually become employees of an organization for an extended
period of time. In doing so, they become part of the culture and over time other
employees come to act quite naturally a round the researcher. The researcher may
observe behaviors that the employee would never reveal otherwise. For instance, a
researcher investigating the ethical behavior of salespeople may have difficulty
getting a car salesperson to reveal any potentially deceptive sales tactics in a
traditional interview. However, ethnographic techniques may result in the
salesperson letting down his or her guard, resulting in more valid discoveries about
the car selling culture.
4.11 Observation in Ethnography
Observation plays a key role in ethnography. Researchers today sometimes ask
households for permission to place video cameras in their home. In doing so, the
ethnographer can study the consumer in a “natural habitat” and use the observations
to test new pr oducts, develop new product ideas, and develop strategies in general.
Ethnographic study can be particularly useful when a certain culture is comprised
of individuals who cannot or will not verbalize their thoughts and feelings. For
instance, ethnography h as advantages for discovering insights among children
since it does not rely largely on their answers to questions. Instead, the researcher
can simply become part of the environment, allow the children to do what they do
naturally, and record their behavior.17 The opening vignette describing a participant -observer approach to learning about skateboarding culture represents an ethnographic approach. Here, the researcher would draw insight from observations and personal experiences with the culture.
4.12 Grou nded Theory
What Is Grounded Theory?
Grounded theory is probably applied less often in business research than is either
phenomenology or ethnography. munotes.in
Page 67
67Chapter 4: Qualitative Research Tools
Grounded theory represents an inductive investigation in which the researcher
poses questions about info rmation provided by respondents or taken from historical records. The researcher asks the questions to him or herself and repeatedly questions the responses to derive deeper explanations. Grounded theory is
particularly applicable in highly dynamic situati ons involving rapid and significant
change. Two key questions asked by the grounded theory researcher are “What is
happening here?” and “How is it different?”19 The distinguishing characteristic of
grounded theory is that it does not begin with a theory bu t instead extracts one from
whatever emerges from an area of inquiry .
4.13 How Is Grounded Theory Used ?
Consider a company that approaches a researcher to study whether or not its sales
force is as effective as it has been over the past five years. The researcher uses
grounded theory to discover a potential explanation. A theory is inductively
developed bas ed on text analysis of dozens of sales meetings that had been recorded
over the previous five years. By questioning the events discussed in the sales
interviews and analyzing differences in the situations that may have led to the
discussion, the researcher is able to develop a theory. The theory suggests that with
an increasing reliance on e -mail and other technological devices for communication, the salespeople do not communicate with each other informally as
much as they did five years previously. As a re sult, the salespeople had failed to
bond into a close -knit “community.”
Computerized software also can be useful in developing grounded theory. In our
Vans example, the researcher may interpret skateboarders’ stories of good and bad
skating experiences by questioning the events and changes described. These may
yield theories about the role that certain brands play in shaping a good or bad
experience. Alternatively, grounded theorists often rely on visual representations.
Thus, the skateboarder could develo p collages representing good and bad
experiences. Just as with the text, questions can be applied to the visuals in an effort
to develop theory.
4.14 Case Studies
What Are Case Studies?
Case studies simply refer to the documented history of a particular person, group,
organization, or event. Typically, a case study may describe the events of a specific
company as it faces an important decision or situation, such as introducing a new munotes.in
Page 68
68 RESEARCH IN COMPUTING
product or dealing with some management crisis. Textbook cases typify this kind
of case study. Clinical interviews of managers, employees, or customers can
represent a case study.
The case studies can then be analyzed for important themes. Themes are identified
by the frequency with which the same term (or a synonym ) arises in the narrative
description. The themes may be useful in discovering variables that are relevant to
potential explanations.
4.15 HOW ARE CASE STUDIES USED?
Case studies are commonly applied in business. For instance, case studies of brands
that s ell “luxury” products helped provide insight into what makes up a prestigious
brand. A business researcher carefully conducted case (no pun intended) studies of
higher end wine labels (such as Penfold’s Grange) including the methods of
production and distr ibution. This analysis suggested that a key ingredient to a
prestige brand may well be authenticity. When consumers know something is
authentic, they attach more esteem to that product or brand.
Case studies often overlap with one of the other categories o f qualitative research.
The Research Snapshot on the next page illustrates how observation was useful in
discovering insights leading to important business changes.
A primary advantage of the case study is that an entire organization or entity can
be inve stigated in depth with meticulous attention to detail. This highly focused
attention enables the researcher to carefully study the order of events as they occur
or to concentrate on identifying the relationships among functions, individuals, or
entities. C onducting a case study often requires the cooperation of the party whose
history is being studied. This freedom to search for whatever data an investigator
deems important makes the success of any case study highly dependent on the
alertness, creativity, i ntelligence, and motivation of the individual performing the
case analysis.
4.16 Common Techniques Used in Qualitative Research Qualitative researchers apply a nearly endless number of techniques. These techniques overlap more than one of the orientations previously discussed, although each category may display a preference for certain techniques.
1- Focus Group Interview
What Is a Focus Group Interview?
The focus group interview is so widely used that many advertising and research
agencies do nothing but focus group interviews. In that sense, it is wrongly munotes.in
Page 69
69Chapter 4: Qualitative Research Tools
synonymous with qualitative research. Nonetheless, focus groups are a very
important qualitative research technique and deserve considerable discussion.
A focus group interview is an unstructured, free -flowing interview with a small
group of people, usually between six and ten. Focus groups are led by a trained
moderator who follows a flexible format encouraging dialogue among respondents.
Common focus group topics include employee programs, employee satisfaction,
brand meanings, problems with products, advertising themes, or new -product
concepts.
The group meets at a central location at a designated time. Participants may range
from consumers talking about hair coloring, petroleum engineers talking about
problems in the “oil patch,” children talking about toys, or employees talking about
their jobs. A moderator begins by providing some opening statement to broadly
steer discussion in the intended direction. Ideally, discussion topic s emerge at the group’s initiative, not the moderator’s. Consistent with phenomenological approaches, moderators should avoid direct questioning unless absolutely
necessary.
2. Advantages Of Focus Group Interviews
Focus groups allow people to discuss their true feelings, anxieties, and frustrations, as well as the depth of their convictions, in their own words. While other approaches may also do much the same, focus groups offer several advantages:
1. Relatively fast
2. Easy to execute
3. Allow respondents to piggyback off each other’s ideas
4. Provide multiple perspectives
5. Flexibility to allow more detailed descriptions
6. High degree of scrutiny D Speed and Ease
In an emergency situation, three or four group sessions can be conducted, analyzed,
and reported in a week or so. The large number of research firms that conduct focus
group interviews makes it easy to find someone to host and conduct the research.
Practic ally every state in the United States contains multiple research firms that
have their own focus group facilities. Companies with large research departments
likely have at least one qualified focus group moderator so that they need not
outsource the focus group. munotes.in
Page 70
70 RESEARCH IN COMPUTINGE Piggybacking and Multiple Perspectives
Furthermore, the group approach may produce thoughts that would not be produced
otherwise. The interplay between respondents allows them to piggyback off of
each other’s ideas. In other words, one respondent stimulates thought among the
others and, as this process continues, increasingly creative insights are possible. A
comment by one individual often triggers a chain of responses from the other
participants. The social nature of the focus group also helps bring out multiple
views as each person shares a p articular perspective. F Flexibility The flexibility of focus group interviews is advantageous, especially when compared with the more structured and rigid survey format. Numerous topics can
be discussed and many insights can be gained, particularly with reg ard to the
variations in consumer behavior in different situations. G Scrutiny
A focus group interview allows closer scrutiny in several ways. First, the session
can be observed by several people, as it is usually conducted in a room containing
a two-way mirror. The respondents and moderator are on one side, and an invited
audience that may include both researchers and decision makers is on the other. If
the decision makers are located in another city or country, the session may be shown
via a live video hookup. Either through live video or a two -way mirror, some check
on the eventual interpretations is provided through the ability to actually watch the
research being conducted. If the obse rvers have questions that are not being asked
or want the moderator to probe on an issue, they can send a quick text message with
instructions to the moderator.
4.17 Focus Group Illustration
Focus groups often are used for concept screening and concept ref inement. The
concept may be continually modified, refined, and retested until management
believes it is acceptable. While RJR’s initial attempts at smokeless cigarettes failed
in the United States, Philip Morris is developing a smokeless cigarette for the U.K.
market. Focus groups are being used to help understand how the product will be
received and how it might be improved. The voluntary focus group respondents are
presented with samples of the product and then they discuss it among themselves.
The interv iew results suggest that the key product features that must be conveyed
are the fact that it produces no ashes, no side smoke, and very little odor. These
beliefs are expected to lead to a positive attitude. Focus group respondents show munotes.in
Page 71
71Chapter 4: Qualitative Research Tools
little concern about how the cigarette actually functioned. Smokers believe they
will use the product if nonsmokers are not irritated by being near someone using
the “electronic cigarette.” Thus, the focus groups are useful in refining the product
and developing a theory of how it should be marketed.
Group Composition
The ideal size of the focus group is six to ten people. If the group is too small, one
or two members may intimidate the others. Groups that a re too large may not allow
for adequate participation by each group member. Homogeneous groups seem to
work best because they allow researchers to concentrate on consumers with similar
lifestyles, experiences, and communication skills. The session does not become rife with too many arguments and different viewpoints stemming from diverse backgrounds. Also, from an ethnographic perspective, the respondents should all
be members of a unique and identifiable culture. Vans may benefit from a focus
group intervi ew comprised only of skateboard enthusiasts. Perhaps participants can
be recruited from a local skate park. However, additional group(s) of participants
that are not boarders might be useful in gaining a different perspective.
4.18 Environmental Conditions
A focus group session may typically take place at the research agency in a room specifically designed for this purpose. Research suppliers that specialize in conducting focus groups operate from commercial facilities that have vide otape
cameras in observation rooms behind two -way mirrors and microphone systems
connected to tape recorders and speakers to allow greater scrutiny as discussed
above. Refreshments are provided to help create a more relaxed atmosphere
conducive to a free e xchange of ideas. More open and intimate reports of personal
experiences and sentiments can be obtained under these conditions.
4.19 The Focus Group Moderator
The moderator essentially runs the focus group and plays a critical role in its
success. There ar e several qualities that a good moderator must possess:
1. The moderator must be able to develop rapport with the group to promote
interaction among all participants. The moderator should be someone who is
really interested in people, who listens carefully to what others have to say, and
who can readily establish rapport, gain people’s confidence, and make them
feel relaxed and eager to talk. munotes.in
Page 72
72 RESEARCH IN COMPUTING
2. The moderator must be a good listener. Careful listening is especially important
because the group interview’s purpose is to stimulate spontaneous responses.
Without good listening skills, the moderator may direct the group in an
unproductive direction.
3. The moderator must try not to interject his or her own opinions. Good
moderators usually say less rather than more. They can stimulate productive
discussion with generalized follow -ups such as, “Tell us more about that
incident,” or “How are your experiences similar or different from the one you
just heard?” The moderator must be particularly careful not to ask leading
questions such as “You are happy to work at Acme, aren’t you?”
4. The moderator must be able to control discussion without being overbearing.
The moderator’s role is also to focus the discussion on the areas of concern.
When a topic is no longer generating fresh ideas, the effective moderator
changes the flow of discussion. The moderator does not give the group total
control of the discussion, but he or she normally has prepared questions on
topics that concern management. However, the timing of these questions in the
discussion and the manner in which they are raised are left to the moderator’s
discretion. The term focus group thus stems from the moderator’s task. He or
she starts out by asking for a general discussion but usually focuses in on
specific topics during the session.
4.20 Planning the Focus Group Outline
Focus group researchers use a discussion guide to help control the interview and
guide the discussion into product areas. A discussion guide includes written
introductory comments informing the group about the focus group purpose and
rules and then outline s topics or questions to be addressed in the group session.
Thus, the discussion guide serves as the focus group outline. Some discussion
guides will have only a few phrases in the entire document. Others may be more
detailed. The amount of content depends on the nature and experience of the
researcher and the complexity of the topic. A cancer center that wanted to warn the
public about the effects of the sun used the discussion guide in Exhibit below. The
business researchers had several objectives for thi s question guide: The first question was very general, asking that respondents describe their feelings
about being out in the sun. This opening question aimed to elicit the full range of
views within the group. Some individuals might view being out in the sun as a
healthful practice, whereas others view the sun as deadly. The hope is that by munotes.in
Page 73
73Chapter 4: Qualitative Research Tools
exposing the full range of opinions, respondents would be motivated to fully
explain their own position. This was the only question asked specifically of every
respondent. Each respondent had to give an answer before free discussion began.
In this way, individuals experience a nonthreatening environment encouraging their
free and full opinion. A general question seeking a reaction serves as an effective
icebreaker. The second question asks whether participants could think of any reason they
should be warned about sunlight exposure. This question was simply designed to
introduce the idea of a warning label. Subsequent questions were asked and became increasingly specif ic. They were first
asked about possible warning formats that might be effective. Respondents are
allowed to react to any formats suggested by any other respondent. After this
discussion, the moderator will introduce some specific formats the cancer cente r
personnel have in mind. Finally, the “bottom -line” question is asked: “What format would be most likely to
induce people to take protective measures?” There would be probing follow -ups of
each opinion so that a respondent couldn’t simply say something li ke “The second
one.” All focus groups finish up with a catchall question asking for any comments
including any thoughts they wanted passed along to the sponsor (which in this case
was only then revealed as the Houston -based cancer center).
In general, the following steps should be used to conduct an e ffective focus group
discussion guide: 1. Welcome and introductions should take place first. 2. Begin the interview with a broad iceb reaker that does not reveal too many specifics
about the inter view. Sometimes, this may even involve respondents providing some
written story or their reaction to some stimulus like a photograph, film, product, or
advertisement. 3. Questions become increasingly more specific as the inte rview proceeds. However,
the Moderator will notice that a good interview will cover the specific question
topics before they Have to be asked. This is preferable as respondents are clearly not fo rced to react
to the specific issue; it just emerges naturally. munotes.in
Page 74
74 RESEARCH IN COMPUTING4. If there is a very specific objective to be accomplished, such as explaining why a
respondent would either buy or not buy a product, that question should probably be
saved for last. 5. A debriefing statement should p rovide respondents with the actual focus group
objectives and answering any questions they may have. This is also a final sho t to
gain some insight from the group.
4.21 Focus Groups as Diagnostic Tools
Focus groups are perhaps the predominant means by whic h business researchers
implement exploratory research designs. Focus groups also can be helpful in later
stages of a research project, particularly when the findings from surveys or other
quantitative techniques raise more questions than they answer. Managers who are
puzzled about the meaning of survey research results may use focus groups to better
understand what survey results indicate. In such a situation, the focus group
supplies diagnostic help after quantitative research has been conducted.
Focus groups are also excellent diagnostic tools for spotting problems with ideas.
For instance, idea scree ning is often done with focus groups. An initial concept is
presented to the group and then they are allowed to comment on it in detail. This
usually leads to lengthy lists of potential product problems and some ideas for
overcoming them. Mature products c an also be “focusgrouped” in this manner.
4.22 Video Conferencing and Focus Groups
With the widespread utilization of videoconferencing, the number of companies
using these systems to conduct focus groups has increased. With videoconference
focus groups, managers can stay home and watch on television rather than having
to take a trip to a focus group facility.
FocusVision ( http://www.focusvision.com/ ) is a business research company that
provides videoconfere ncing equipment and services. The FocusVision system is
modular, allowing for easy movement and an ability to capture each group member
close up. The system operates via a remote keypad that allows observers in a far -
off location to pan the focus group roo m or zoom in on a particular participant.
Managers viewing at remote locations can send the moderator messages during the
interview.
munotes.in
Page 75
75Chapter 4: Qualitative Research Tools
4.23 Interactive Media and Online Focus Groups
Internet applications of qualitative exploratory research are growing rapidly and
involve both formal and informal applications. Formally, the term online focus
group refers to a qualitative research effort in which a group of individuals provides
unstruct ured comments by entering their remarks into an electronic Internet display
board of some type, such as a chat -room session or in the form of a blog. Because respondents enter their comments into the computer, transcripts of verbatim responses are availabl e immediately after the group session. Online groups can be
quick and cost efficient. However, because there is less personal interaction between participants, group synergy and snowballing of ideas may be diminished.
Several companies have established a form of informal, “continuous” focus group
by establishing an Internet blog for that purpose. We might call this technique a
focus blog when the intention is to mine the site for business research purposes.
4.24 Online versus Face -To-Face Focus Group Techniques
A research company can facilitate a formal online focus group by setting up a
private chat room for that purpose. Participants in formal and informal online focus
groups feel that their anonymity is very secure. Often respondents will say things
in this environment that they would never say otherwise. For example, a lingerie
company was able to get insights into how it could design sexy products for larger
women. Online, these women freely discussed what it would take “to feel better
about being naked.”26 One can hardly imagine how difficult such a discussion
might be face to face. Increased anonymity can be a major advantage for a company
investigating sensitive or embarrassing issues.
Disadvantages Of Focus Groups
Focus groups offer many advanta ges as a form of qualitative research. Like
practically every other research technique, the focus group has some limitations
and disadvantages as well. Problems with focus groups include those discussed
below.
First, focus groups require objective, sensit ive, and effective moderators. It is very
difficult for a moderator to remain completely objective about most topics. In large
research firms, the moderator may be provided only enough information to
effectively conduct the interview, no more. The focus gr oup interview obviously
shouldn’t reduce to, or even be influenced by, the moderator’s opinion. Also,
without a good moderator, one or two participants may dominate a session, yielding munotes.in
Page 76
76 RESEARCH IN COMPUTING
results that are really the opinion of one or two people, not the group. The
moderator has to try very hard to make sure that all respondents feel comfortable
giving their opinions and even a timid respondent’s opinion is given due consideration. While many people, even some with little or no background to do so,
conduct focus groups, good moderators become effective through a combination of
naturally good people skills, training (in qualitative research), and experience.
Second, some unique sampling problems arise with focus groups. Researchers often select focus group participants because they have similar backgrounds and experiences or because screening indicates that the participants are more articulate
or gregarious than the typica l consumer. Such participants may not be
representative of the entire target market. Thus, focus group results are not intended
to be representative of a larger population.
Third, although not so much an issue with online formats where respondents can
rema in anonymous, traditional face -to-face focus groups may not be useful for
discussing sensitive topics. A focus group is a social setting and usually involves
people with little to no familiarity with each other. Therefore, issues that people
normally do no t like to discuss in public may also prove difficult to discuss in a
focus group. Fourth, focus groups do cost a considerable amount of money,
particularly when they are not conducted by someone employed by the company
desiring the focus group. As research projects go, there are many more expensive
approaches, including a full -blown mail survey using a national random sample.
This may cost thousands of dollars to conduct and thousands of dollars to analyze
and disseminate .
2-Depth Interviews
An alternative to a focus group is a depth interview. A depth interview is a one -
on-one interview between a professional researcher and a research respondent.
Depth interviews are much the same as a psychological, clinical interview, but with
a different p urpose. The researcher asks many questions and follows up each
answer with probes for additional elaboration.
Like focus group moderators, the interviewer’s role is critical in a depth interview.
He or she must be a highly skilled individual who can encour age the respondent to
talk freely without influencing the direction of the conversation. Probing questions
are critical. Laddering is a term used for a particular approach to probing, asking
respondents to compare differences between brands at different levels. What
usually results is that the first distinctions are attribute -level distinctions, the second
are benefit -level distinctions, and the third are at the value or motivation level. munotes.in
Page 77
77Chapter 4: Qualitative Research Tools
Laddering can then distinguish two brands of skateboarding shoes based on a) the
materials they are made of, b) the comfort they provide, and c) the excitement they
create.
Each depth interview may last more than an hour. Thus, it is a time -consuming
process if multiple interviews are conducted. Not only does the interview have to
be conducted, but each interview prod uces about the same amount of text as does a
focus group interview. This has to be analyzed and interpreted by the researcher. A
third major issue stems from the necessity of recording both surface reactions and
subconscious motivations of the respondent. Analysis and interpretation of such
data are highly subjective, and it is difficult to settle on a true interpretation.
Depth interviews provide more insight into a particular individual than do focus
groups. In addition, since the setting isn’t really soc ial, respondents are more likely
to discuss sensitive topics than are those in a focus group. Depth interviews are
particularly advantageous when some unique or unusual behavior is being studied.
For instance, depth interviews have been usefully applied to reveal characteristics
of adolescent behavior, ranging from the ways they get what they want from their
parents to shopping, smoking, and shoplifting.
Depth interviews are similar to focus groups in many ways. The costs are similar if
only a few interview s are conducted. However, if a dozen or more interviews are
included in a report, the costs are higher than focus group interviews due to the
increased interviewing and analysis time.
3-Conversations
Holding conversations in qualitative research is an informal data -gathering
approach in which the researcher engages a respondent in a discussion of the
relevant subject matter. This approach is almost completely unstructured and the
researcher enters the conversation with few exp ectations. The goal is to have the
respondent produce a dialogue about his or her lived experiences. Meaning will be
extracted from the resulting dialogue. A conversational approach to qualitative
research is particularly appropriate in phenomenological research and for developing grounded theory. In our Vans experience, the researcher may simply
tape-record a conversation about becoming a “skater.” The resulting dialogue can
then be analyzed for themes and plots. The result may be some interesting and nov el
insight into the consumption patterns of skaters, for example, if the respondent said,
“I knew I was a real skater when I just had to have Vans, not just for boarding, but
for wearing.” munotes.in
Page 78
78 RESEARCH IN COMPUTING
This theme may connect to a right -of-passage plot and show how Van s play a role
in this process. Technology is also influencing conversational research. Online communications such as the reviews posted about book purchases at http://www.barnesandnoble.com can be treated as a conversation. Companies
may discover product problems and ideas for overcoming them by analyzing these
computer -based consumer dialogues. A conversational app roach is advantageous
because each interview is usually inexpensive to conduct. Respondents often need
not be paid. They are relatively effective at getting at sensitive issues once the
researcher establishes a rapport with them. Conversational approaches, however,
are prone to produce little relevant information since little effort is made to steer
the conversation. Additionally, the data analysis is very much researcher -
dependent.
4-Semi -Structured Interviews
Semi -structured interviews usually come in written form and ask respondents for
short essay responses to specific open -ended questions. Respondents are free to
write as much or as little as they want. The questions would be divided into
sections, typically, and within each section, the opening questio n would be
followed by some probing questions. When these are performed face to face, there
is room for less structured follow -ups. The advantages to this approach include an
ability to address more specific issues. Responses are usually easier to interpre t
than other qualitative approaches. Since the researcher can simply prepare the
questions in writing ahead of time, and if in writing, the questions are administered
without the presence of an interviewer, semi -structured interviews can be relatively
costeffective. Some researchers interested in studying car salesperson stereotypes
used qualitative semistructured interviews to map consumers’ cognitions (memory). The semi -structured interview began with a free -association task:
List the first five things that come into your mind when you think of a “car
salesman.”
This was followed up with a probing question:
Describe the way a typical “car salesman” looks.
This was followed with questions about how the car salesperson acts and how the
respond ent feels in the presence of a car salesperson. The results led to research
showing how the information that consumers process differs in the presence of a
typical car salesperson, as opposed to a less typical car salesperson
munotes.in
Page 79
79Chapter 4: Qualitative Research Tools
5-Social Networking
Social networking is one of the most impactful trends in recent times. For many consumers, particularly younger generations, social networking sites like MySpace, Second Life, Zebo, and others have become the primary tool for communicating with friends bot h far and near and known and unknown. Social
networking has replaced large volumes of e -mail and, many would say, face – to-face communications as well. While the impact that social networking will eventually have on society is an interesting question, what is most relevant to
marketing research is the large portion of this information that discusses marketing
and consumer related information. Companies can assign research assistants to
monitor these sites for information related to their particular brands. The
information can be coded as either positive or negative. When too much negative
information is being spread, the company can try to react to change the opinions.
In addition, many companies like P&G and Ford maintain their own social
networking sites f or the purpose of gathering research data. In a way, these social networking sites are a way that companies can eavesdrop on consumer conversations and discover key information about their products. The textual data
that consumers willingly put up becomes like a conversation. When researchers get
the opportunity to react with consumers or employees through a social network site,
they can function much like an online focus group or interview.
6-Free-Association/Sentence Completion Method
Free -association techniques simply record a respondent’s first cognitive reactions
(top-of- mind) to some stimulus. The Rorschach or inkblot test typifies the free -
association method. Respondents view an ambiguous figure and are asked to say
the first thing that comes to their mind. Free -association techniques allow researchers to map a respondent’s thoughts or memory.
The sentence completion method is based on free -association principles. Respondents simply are required to complete a few partial sentences with the first
word or phrase that comes to mind. For example:
People who drink beer are .
A man who drinks a dark beer is . Imported beer is most liked by .
The woman drinking beer in the commercial .
Answers to sentence -completion questions tend to be more extensive than responses to word association tests. Although the responses lack the ability to probe
for meaning as in other qualitative techniques, they are very effective in finding out
what is on a responden t’s mind . They can also do so in a quick and very cost -
effective manner. Free -association and sentence - completion tasks are sometimes munotes.in
Page 80
80 RESEARCH IN COMPUTING
used in conjunction with other approaches. For instance, they can sometimes be
used as effective icebreakers in focus gro up interviews.
7-Observation
Observation can be a very important qualitative tool. The participant -observer
approach typifies how observation can be used to explore various issues. Meaning
is extracted from field notes. Field notes are the researchers’ des criptions of what
actually happens in the field. These notes then become the text from which meaning
is extracted.
Observation may also take place in visual form. Researchers may observe
employees in their workplace, consumers in their home, or try to gai n knowledge
from photographic records of one type or another. Observation can either be very
inexpensive, such as when a research associate sits and simply observes behavior,
or it can be very expensive, as in most participant -observer studies.
8-Collages
Business researchers sometimes have respondents prepare a collage to represent
their experiences. The collages are then analyzed for meaning much in the same
manner as text dialogues are analyzed. Computer software can even be applied to
help develop poten tial grounded theories from the visual representations.
Like sentence completion and word association, collages are often used within
some other approach, such as a focus group or a depth interview. Collages offer the
advantage of flexibility but are also very much subject to the researcher’s interpretations.
4.25 Projective Research Techniques
A projective technique is an indirect means of questioning enabling respondents
to project beliefs and feelings onto a third party, an inanimate object, or a task situation. Projective techniques usually encourage respondents to describe a situation in their own words with little prompting by the interviewer. Individuals
are expected to interpret the situation within the context of their own experiences,
attitu des, and personalities and to express opinions and emotions that may be
hidden from others and possibly themselves. Projective techniques are particularly
useful in studying sensitive issues.
10-Thematic Apperception Test (TAT )
A thematic apperception te st (TAT) , sometimes called the picture interpretation munotes.in
Page 81
81Chapter 4: Qualitative Research Tools
technique, presents subjects with an ambiguous picture(s) and asks the subject to
tell what is happening in the picture(s) now and what might happen next. Hence,
themes ( thematic ) are elicited on the basis of the perceptual -interpretive
(apperception ) use of the pictures. The researcher then analyzes the contents of the
stories that the subjects relate. A TAT represents a projective research technique.
Frequently, the TAT consists of a series of pictures with some continuity so that
stories may be constructed in a variety of settings. The first picture might portray a
person working at their desk; in the second picture, a person that could be a
supervisor is talking to the worker; the final picture might show the original
employee and another having a discussion at the water cooler. A Vans TAT might
include several ambiguous pictures of a skateboarder and then show him or her
heading to the store. This might reveal ideas about the b rands and products that fit
the role of skateboarder.
The picture or cartoon stimulus must be sufficiently interesting to encourage
discussion but ambiguous enough not to disclose the nature of the research project.
Clues should not be given to the charact er’s positive or negative predisposition. A
pretest of a TAT investigating why men might purchase chainsaws used a picture
of a man looking at a very large tree. The research respondents were homeowners
and weekend woodcutters. They almost unanimously said that they would get
professional help from a tree surgeon to deal with this situation. Thus, early in pretesting, the researchers found out that the picture was not sufficiently ambiguous. The tree was too large and did not allow respondents to identify w ith
the tree -cutting task. If subjects are to project their own views into the situation, the
environmental setting should be a well -defined, familiar problem, but the solution
should be ambiguous.
4.26 Exploratory Research in Science and in Practice
Misuses of Exploratory and Qualitative Research
Any research tool can be misapplied. Exploratory research cannot take the place of
conclusive, confirmatory research. Thus, since many qualitative tools are best
applied in exploratory design, they a re likewise limited in the ability to draw
conclusive inferences —test hypotheses. One of the biggest drawbacks is the
subjectivity that comes along with “interpretation.” In fact, sometimes the term
interpretive research is used synonymously with qualitati ve research. When only
one researcher interprets the meaning of what a single person said in a depth
interview or similar technique, one should be very cautious before major business
decisions are made based on these results. Is the result replicable? Repl ication
means that the same results and conclusions will be drawn if the study is repeated
by different researchers with different respondents following the same methods. In munotes.in
Page 82
82 RESEARCH IN COMPUTING
other words, would the same conclusion be reached based on another researcher’s
interpretation?
Indeed, some qualitative research methodologies were generally frown ed upon for
years based on a few early and public misapplications during what became known
as the “motivational research” era. While many of the ideas produced during this
time had some merit, as can sometimes be the case, too few researchers did too
much interpretation of too few respondents . Compounding this, managers were
quick to act on the results, believing that the results peaked inside one’s subliminal
consciousness and therefore held some type of extra power. Thus, often the
research was flawed bas ed on poor interpretation, and the decision process was
flawed because the deciders acted prematurely.
4.27 Scientific Decision Processes
Objectivity and replicability are two characteristics of scientific inquiry. Are focus
groups objective and replicabl e? Would three different researchers all interpret
focus group data identically? How should a facial expression or nod of the head be
interpreted? Have subjects fully grasped the idea or concept behind a nonexistent
product? Have respondents overstated the ir satisfaction because they think their
supervisor will read the report and recognize them from their comments? Many of
these questions are reduced to a matter of opinion that may vary from researcher to
researcher and from one respondent group to another . Therefore, a focus group, or
a depth interview, or TAT alone does not best represent a complete scientific
inquiry.
However, if the thoughts discovered through these techniques survive preliminary
evaluations and are developed into research hypotheses, they can be further tested.
These tests may involve survey research or an experiment testing an idea very
specifically (for example, if a certain advertising slogan is more effective than
another). Thus, exploratory research approaches using qualitative research tools are
very much a part of scientific inquiry. However, before making a scientific
decision, a research project should include a confirmatory study using objective
tools and an adequate sample in terms of both size and how well it represents a
population.
But is a scientific decision approach always used or needed? In practice, many
business decisions are based solely on the results of focus group interviews or som e
other exploratory result. The primary reasons for this are
(1) Time,
(2) Money, munotes.in
Page 83
83Chapter 4: Qualitative Research Tools
(3) Emotion.
1) Time
Sometimes, res earchers simply are not given enough time to follow up on exploratory research results. Companies feel an increasingly urgent need to get new
products to the market faster. Thus, a seemingly good idea generated in a focus
group (like Clear, Vanilla, or Che rry Dr Pepper) is simply not tested with a more
conclusive study. The risk of delaying a decision may be seen as greater than the
risk of proceeding without completing the scientific process. Thus, although the
researcher may warn against it, there may be logical reasons for such action. The
decision makers should be aware, though, that the conclusions drawn from
exploratory research designs are just that — exploratory. Thus, there is less
likelihood of good results from the decision than if the research process had
involved further testing.
2) Money
Similarly, researchers sometimes do not follow up on explorator y research results
because they believe the cost is too high. Realize that tens of thousands of dollars
may have already been spent on qualitative research. Managers who are unfamiliar
with research will be very tempted to wonder, “Why do I need yet anothe r study?”
and “What did I spend all that money for?” Thus, they choose to proceed based
only on exploratory results. Again, the researcher has fulfilled the professional
obligation as long as the tentative nature of any ideas derived from exploratory
resea rch has been relayed through the research report.
3) Emotion
Time, money, and emotion are all related. Decision makers sometimes become so
anxious to have something resolved, or they get so excited about some novel
discovery resulting from a focus group inter view, that they may act rashly. Perhaps
some of the ideas produced during the motivational research era sounded so
enticing that decision makers got caught up in the emotion of the moment and
proceeded without the proper amount of testing. Thus, as in life, when we fall in
love with something, we are prone to act irrationally. The chances of emotion
interfering in this way are lessened, but not reduced, by making sure multiple
decision makers are involved in the decision process.
Summary Qualitative business research is research that addresses business objectives through techniques that allow the researcher to provide elaborate interpretations
of market phenomena without depending on numerical measurement. Quantitative munotes.in
Page 84
84 RESEARCH IN COMPUTING
researchers direct a considerable amount of activity toward measuring concepts
with scales that either directly or indirectly provide numeric values. Qualitative
research seldom involves samples with hundreds of respondents. Instead, a
handful of people are usually the source of qualitative data. Ethnography represents ways of studying cultures through methods that involve becoming
highly active within that culture. Grounded theory is particularly applicable in
highly dynamic situations involving rapid and significant change. Focus groups
are led by a trained moderator who follows a flexible format encouraging dialogue
among respondents. exploratory research approaches using qualitative research
tools are very much a part of scientific inquiry.
Questions
1. Wha t is quantitative research?
2. What is qualitative research?
3. Give comparison between quantitative and qualitative research.
4. What is a phenomenological approach to research?
5. What are hermeneutics?
6. What is ethnography?
7. Explain in brief about exploratory researc h.
8. Explain the scientific decision process with suitable example.
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• 2. Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Con cepts: Variable,
• Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN: 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016,
• Business Research Methods, Edition 8, Cengage Publication
munotes.in
Page 85
85Chapter 5: Secondary Data Research in a Digital Age
UNIT 2
5 SECONDARY DATA RESEARCH IN A
DIGITAL AGE
Unit Structure
5.0 Objectives
5.1 Introduction
5.2 Secondary Data
5.3 Sources of Secondary Data
5.4 Typical Objectives for Secondary -Data Research Designs
5.5 Identification of Consumer Behavior for a Product Category
5.6 Trend Analysis
5.7 Analysis of Trade Areas And Sites
5.8 External Data: The Distribution System
5.9 Information as a Product and Its Distribution Channels
5.10 Single -Source Data -Integrated Information
5.0 Objectives
• Understanding about the use of secondary data
• Single -Source Data -Integrated Information
• Use of social networking and other resources in the research
5.1 Introduction
Research projects often begin with secondary data , which are gathered and
recorded by someone else prior to (and for purposes other than) the current project.
Secondary data usually are historical and already assembled. They require no
access to respondent s or subjects. munotes.in
Page 86
86 RESEARCH IN COMPUTING
5.2 Secondary Data
Secondary data is the data that has already been collected through primary sources
and made readily available for researchers to use for their own research. It is a type
of data that has already been collected in the past.
A researcher may have collected the data for a particular project, then made it
available to be used by another researcher. The data may also have been collected
for general use with no specific research purpose like in the case of the national
census.
A data classified as secondary for a particular research may be said to be primary
for another research. This is the case when a data is being reused, making it
a primary data for the first research and secondary data for the second research it is
being used for.
5.3 Sources of Secondary Data
Sources of secondary data includes books, personal sources, journal, newspaper,
website, government record etc. Secondary data are known to be readily a vailable
compared to that of primary data. It requires very little research and need for
manpower to use these sources.
With the advent of electronic media and the internet, secondary data sources have
become more easily accessible. Some of these sources a re highlighted below.
Books
Books are one of the most traditional ways of collecting data. Today, there are
books available for all topics you can think of. When carrying out research, all you
have to do is look for a book on the topic being researched on , then select from the
available repository of books in that area. Books, when carefully chosen are an
authentic source of authentic data and can be useful in preparing a literature review.
Published Sources
There are a variety of published sources availab le for different research topics. The
authenticity of the data generated from these sources depends majorly on the writer
and publishing company.
Published sources may be printed or electronic as the case may be. They may be
paid or free depending on the writer and publishing company's decision.
munotes.in
Page 87
87Chapter 5: Secondary Data Research in a Digital Age
Unpublished Personal Sources
This may not be readily available and easily accessible compared to the published sources. They only become accessible if the researcher shares with another researcher who is not allowed to share it with a third party.
For example, the product management team of an organization may need data on customer feedback to assess what customers think about their product and improvement suggestions. They will need to c ollect the data from the customer
service department, which primarily collected the data to improve customer
service.
Journal
Journals are gradually becoming more important than books these days when data
collection is concerned. This is because journals a re updated regularly with new
publications on a periodic basis, therefore giving to date information.
Also, journals are usually more specific when it comes to research. For example,
we can have a journal on, "Secondary data collection for quantitative data " while a
book will simply be titled, "Secondary data collection".
Newspapers
In most cases, the information passed through a newspaper is usually very reliable.
Hence, making it one of the most authentic sources of collecting secondary data.
The kind of data commonly shared in newspapers is usually more political,
economic, and educational than scientific. Therefore, newspapers may not be the
best source for scientific data collectio n.
Websites
The information shared on websites are mostly not regulated and as such may not
be trusted compared to other sources. However, there are some regulated websites
that only share authentic data and can be trusted by researchers.
Most of these web sites are usually government websites or private organizations
that are paid, data collectors.
Blogs
Blogs are one of the most common online sources for data and may even be less
authentic than websites. These days, practically everyone owns a blog and a l ot of
people use these blogs to drive traffic to their website or make money through paid
ads. munotes.in
Page 88
88 RESEARCH IN COMPUTING
Therefore, they cannot always be trusted. For example, a blogger may write good
things about a product because he or she was paid to do so by the manufacturer
even though these things are not true.
Diaries
They are personal records and as such rarely used for data collection by researchers.
Also, diaries are usually personal, except for these days when people now share
public diaries containing specific events in their life.
A common example of this is Anne Frank's diary which contained an accurate
record of the Nazi wars.
Government Records
Government records are a very important and authentic source of secondary data.
They contain information useful in marketing, management, humanities, and social
science research.
Some of these records include; census data, health records, education institute
records, etc. They are usually collected to aid proper planning, allocation of funds,
and prioritizing of projec ts.
Podcasts
Podcasts are gradually becoming very common these days, and a lot of people listen
to them as an alternative to radio. They are more or less like online radio stations
and are generating increasing popularity.
Information is usually shared dur ing podcasts, and listeners can use it as a source
of data collection.
Advantages of Secondary Data
The primary advantage of secondary data is their availability. Obtaining secondary
data is almost always faster and less expensive than acquiring primary d ata. This is
particularly true when researchers use electronic retrieval to access data stored
digitally. In many situations, collecting secondary data is instantaneous. Consider
the money and time saved by researchers who obtained updated population
estimates for a town during the interim between the 2000 and 2010 censuses.
Instead of doing the fieldwork themselves, researchers could acquire estimates
from a firm dealing in demographic information or from so urces such as Claritas
or PCensus. As in this example, the use of secondary data eliminates many of the
activities normally associated with primary data collection, such as sampling and
data processing. Secondary data are essential in instances when data c annot be
obtained using primary data collection procedures. For example, a manufacturer of munotes.in
Page 89
89Chapter 5: Secondary Data Research in a Digital Age
farm implements could not duplicate the information in the Census of Agriculture
because much of the information there (for example, amount of taxes paid) might
not be accessible to a private firm.
Disadvantages of Secondary Data
An inherent disadvantage of secondary data is that they were not designed specifically to meet the researchers’ needs. Thus, researchers must ask how
pertinent the data are to their particula r project. To evaluate secondary data,
researchers should ask questions such as these:
• Is the subject matter consistent with our problem definition?
• Do the data apply to the population of interest?
• Do the data apply to the time period of interest?
• Do the secondary data appear in the correct units of measurement?
• Do the data cover the subject of interest in adequate detail?
Even when secondary information is available, it can be inadequate. Consider the
following typical situations:
A researcher inte rested in forklift trucks finds that the secondary data on the
subject are included in a broader, less pertinent category encompassing all
industrial trucks and tractors. Furthermore, the data were collected five years
earlier.
An investigator who wishes t o study individuals earning more than $100,000
per year finds the top category in a secondary study reported at $75,000 or more
per year.
A brewery that wishes to compare its per -barrel advertising expenditures with
those of competitors finds that the units of measurement differ because some
report point -of-purchase expenditures with advertising and others do not.
Data from a previous warran ty card study show where consumers prefer to
purchase the product but provide no reasons why. The most common reasons
why secondary data do not adequately satisfy research needs are (1) outdated
information, (2) variation in definition of terms, (3) different units of measurement, and (4) lack of information to verify the data’s accuracy.
Furthermore, in our rapidly changing environment, information quickly becomes outdated. Because the purpose of most studies is to predict the future,
secondary data must b e timely to be useful. Every primary researcher has the
right to define the terms or concepts under investigation to satisfy the purpose
of his or her primary investigation. This practice provides little solace, munotes.in
Page 90
90 RESEARCH IN COMPUTING
5.4 Typical Objectives for Secondary -Data Re search Designs
It would be impossible to identify all the purposes of research using secondary data.
However, some common business and marketing problems that can be addressed
with secondary research designs are useful. Exhibit below shows three general
categories of research objectives: fact finding, model building, and database
marketing.
Figure a: Common Research Objectives for Secondary - Data Studies
Fact -Finding
The simplest form of secondary -data research is fact-finding. A restaurant serving
breakfast might be interested in knowing what new products are likely to entice
consumers. Secondary data available from National Eating Trends, a service of the
NPD Group, show that the most potential may be in menu ite ms customers can eat
on the go. According to data from the survey of eating trends, take -out breakfasts
have doubled over the past few years, and they have continued to surpass dine -in
breakfast sales for over a decade. These trends make smoothies and brea kfast
sandwiches sound like a good bet for a breakfast menu. Also, NPD found that 41
percent of breakfast sandwiches are consumed by people in their cars and 24
percent of people polled take them to work. These findings suggest that the
sandwiches should b e easy to handle.
But what to put on the biscuit or bun? Another research firm, Market Facts, says
almost half of consumers say they would pay extra for cheese. These simple facts
would interest a researcher who was investigating the market for take -out
breakfasts. Fact -finding can serve more complex purposes as well. In the digital
age we live in, the use of music as a means to notify users of a call is commonplace.
The Research Snapshot on the next page gives some of the amazing growth facts
predicted in this industry.
munotes.in
Page 91
91Chapter 5: Secondary Data Research in a Digital Age
5.5 Identification of Consumer Behavior for a Product Category
A typical objective for a secondary research study might be to uncover all available
information about consumption patterns for a particular product category or to
identify demographic trends that affect an industry. For example, a company called
Servigistics offers software that will scan a company’s own parts inventory data
and compare it with marketing objectives and competitors’ prices to evaluate
whether the company should adjust prices for its parts. Kia Motors tried using this
service in place of the usual method of marking up cost by a set fraction. By
considering seconda ry data including internal inventory data and external data
about competitors’ prices, it was able to make service parts a more profitable
segment of its business.3 This example illustrates the wealth of factual information
about consumption and behavior p atterns that can be obtained by carefully collecting and analyzing secondary data.
5.6 Trend Analysis
Business researchers are challenged to constantly watch for trends in the
marketplace and the environment. Market tracking is the observation and analysis
of trends in industry volume and brand share over time. Scanner research services
and other organizations provide facts about sales volume to support this work.
Almost every large consumer goods company routinely investigates b rand and
product category sales volume using secondary data. This type of analysis typically
involves comparisons with competitors’ sales or with the company’s own sales in
comparable time periods. It also involves industry comparisons among different
geog raphic areas. Exhibit below on the next page shows the trend in cola market
share relative to the total carbonated soft -drink industry.
1. Model Building
The second general objective for secondary research, model building, is more
complicated than simple fact -finding. Model building involves specifying relationships between two or more variables, perhaps extending to the development
of descriptive or predictive equations, a technique that is used by the Nielsen
Claritas Company routinely to add value to their secondary data. Models need not
include complicated mathematics, though. In fact, decision makers often prefer
simple models that everyone can readily understand over complex models that are
difficult to comprehend. For example, market share is company sal es divided by
industry sales. Although some may not think of this simple calculation as a model,
it represents a mathematical model of a basic relationship. We will illustrate model building by discussing three common objectives that can be satisfied with secondary research: estimating market potential, forecasting sales, and selecting
potential facility or expansion sites. munotes.in
Page 92
92 RESEARCH IN COMPUTING
2. Estimating Market Potential for Geographic Areas Business researchers often estimate their company’s market potential using secondary data. In many cases exact figures may be published by a trade association
or another source. However, when the desired information is unavailable, the
researcher may estimate market potential by transforming secondary data from two
or more source s. For example, managers may find secondary data about market
potential for a country or other large geographic area, but this information may not
be broken down into smaller geographical areas, such as by metropolitan area, or
in terms unique to the compa ny, such as sales territory. In this type of situation,
researchers often need to make projections for the geographic area of interest.
3. Forecasting Sales
For any project, such as forecasting sales, you need information about the future.
You will need to kn ow what company sales will be next year and in future time
periods. Sales forecasting is the process of predicting sales totals over a specific
time period. Accurate sales forecasts, especially for products in mature, stable
markets, frequently come from s econdary -data research that identifies trends and
extrapolates past performance into the future. Researchers often use internal
company sales records to project sales. A rudimentary model would multiply past
sales volume by an expected growth rate. A resea rcher might investigate a
secondary source and find that industry sales are expected to grow by 10 percent;
multiplying company sales volume by 10 percent would give a basic sales forecast.
5.7 Analysis of Trade Areas And Sites
Managers routinely examine t rade areas and use site analysis techniques to select
the best locations for retail or wholesale operations. Secondary -data research helps
managers make these site selection decisions. Some organizations, especially
franchisers, have developed special computer software based on analytical models
to select sites for re tail outlets. The researcher must obtain the appropriate
secondary data for analysis with the computer software. The index of retail
saturation offers one way to investigate retail sites and to describe the relationship between retail demand and supply.7 It is easy to calculate once the
appropriate secondary data are obtained:
Index of retail saturation= ݈݈ܿܽ ݉ܽݐ݁݇ݎ ݐ݁݊ݐ݅ܽ(݀݁݉ܽ݊݀)
݈݈ܿܽ ݉ܽݐ݁݇ݎ ݎ݁ݐ݈ܽ݅݅݊݃ ݏ݁ܿܽ
munotes.in
Page 93
93Chapter 5: Secondary Data Research in a Digital Age
Data Mining
Large corporations’ decision support systems often contain millions or even
hundreds of millions of records of data. These complex data volumes are too large
to be understood by managers. Two points about data volume are important to keep
in mind. First, relevant data are often in independent and unrelated files. Second,
the number of distinct pieces of information each data record contains is often large.
When the number of distinct pieces of information contained in each data record
and data volume grows too large, end users don’t have the capacity to make sense
of it all. Data mining helps clarify the underlying meaning of the data.
The term data mining refers to the use of powerful computers to dig through
volumes of data to discover patterns about an organization’s customers and products. As seen in the Research Snapshot on the next page, this can even apply
to Internet content from blogs. It is a broad term that applies to many different
forms of analysis. For ex ample, neural networks are a form of artificial
intelligence in which a computer is programmed to mimic the way that human
brains process information.
Market -basket analysis is a form of data mining that analyzes anonymous point -
of-sale transaction databases to identify coinciding purchases or relationships
between products purchased and other retail shopping information.10 Consider this
example about patterns in customer purchas es: Osco Drugs mined its databases
provided by checkout scanners and found that when men go to its drugstores to buy
diapers in the evening between 6:00 p.m. and 8:00 p.m., they sometimes walk out
with a six -pack of beer as well. Knowing this behavioral pa ttern, supermarket
managers may consider laying out their stores so that these items are closer together
A data -mining application of interest to some researchers is known as customer
discovery , which involves mining data to look for patterns identifying w ho is likely
to be a valuable customer. For example, a larger provider of business services
wanted to sell a new product to its existing customers, but it knew that only some
of them would be interested. The company had to adapt each product offering to
each customer’s individual needs, so it wanted to save money by identifying the
best prospects. It contracted with a research provider called DataMind to mine its
data on sales, responses to marketing, and customer service to look for the
customers most likely to be interested in the new product. DataMind assigned each
of the company’s customers an index number indicating their expected interest
level, and the selling effort was much more efficient as a result.
munotes.in
Page 94
94 RESEARCH IN COMPUTING
When a company knows the identity of the customer who makes repeated purchases from the same organization, an analysis can be made of sequences of
purchases. The use of data mining to detect sequence patterns is a popular
application among direct marketers, such as catalog retailers. A catalog merchant
has information for each customer, revealing the sets of products that the customer
buys in every purchase order. A sequence detection function can then be used to
discover the set of purchases that frequently precedes the purchase of, say, a
microwave ov en. As another example, a sequence of insurance claims could lead
to the identification of frequently occurring medical procedures performed on
patients, which in turn could be used to detect cases of medical fraud.
Data mining requires sophisticated computer resources, and it is expensive. That’s
why companies like DataMind, IBM, Oracle, Information Builders, and Acxiom
Corporation offer data -mining services. Customers send the databases they want
analyzed and let the data -mining company do the “number crunching.”
Database Marketing and Customer Relationship Management
CRM (customer relationship management) systems are a decision support system
that manage the interactions between an organization and its customers. A CRM
maintains customer databases containing customers’ names, addresses, phone
numbers, past purchases, responses to past promotional offers, and other relevant
data such as demographic and financial data. Database marketing is the practice
of using CRM d atabases to develop one -to-one relationships and precisely targeted
promotional efforts with individual customers. For example, a fruit catalog
company CRM contains a database of previous customers, including what purchases they made during the Christmas holidays. Each year the company sends
last year’s gift list to customers to help them send the same gifts to their friends and
relatives.
Because database marketing requires vast amounts of CRM data compiled from
numerous sources, secondary data are often acquired for the exclusive purpose of
developing or enhancing databases. The transaction record, which often lists the
item purchased, its value, customer name, addr ess, and zip code, is the building
block for many databases. This may be supplemented with data customers provide
directly, such as data on a warranty card, and by secondary data purchased from
third parties. For example, credit services may sell databases about applications for
loans, credit card payment history, and other financial data. Several companies, such as Donnelley Marketing (with its BusinessContentFile and ConsumerContentFile services) and Claritas (with PRIZM), collect primary data
and then se ll demographic data that can be related to small geographic areas, such
as those with a certain zip code. (Remember that when the vendor collects the data, munotes.in
Page 95
95Chapter 5: Secondary Data Research in a Digital Age
they are primary data, but when the database marketer incorporates the data into
his or her database , they are secondary data.) Now that some of the purposes of
secondary -data analysis have been addressed, we turn to a discussion of the sources
of secondary data.
Other Sources to obtain Secondary Data
Secondary data can be classified as either internal to the organization or external.
Modern information technology makes this distinction seem somewhat simplistic.
Some accounting documents are indisputably internal records of the organization.
Researchers in another organization cannot have access to them. Clearly, a book
published by the federal government and located at a public library is external to
the company. However, in today’s world of electronic data interchange, the data
that appear in a book published by the federal government may also be purcha sed
from an online information vendor for instantaneous access and subsequently
stored in a company’s decision support system. Internal data should be defined as
data that originated in the organization, or data created, recorded, or generated by
the organ ization. Internal and proprietary data is perhaps a more descriptive
term.
Sources of Internal and Proprietary Data
Most organizations routinely gather, record, and store internal data to help them
solve future problems. An organization’s accounting system can usually provide a
wealth of information. Routine documents such as sales invoices allow external
financial reporting, which in turn can be a source of data for further analysis. If the
data are properly coded into a modular database i n the accounting system, the
researcher may be able to conduct more detailed analysis using the decision support
system. Sales information can be broken down by account or by product and region;
information related to orders received, back orders, and unfi lled orders can be
identified; sales can be forecast on the basis of past data. Other useful sources of internal data include salespeople’s call reports, customer complaints, service records, warranty card returns, and other records. Researchers frequently aggregate
or disaggregate internal data. For example, a computer service firm used internal
secondary data to analyze sales over the previous three years, categorizing business
by industry, product, purchase level, and so on. The company discovered that 6 0
percent of its customers represented only 2 percent of its business and that nearly
all of these customers came through telephone directory advertising. This simple
investigation of internal records showed that, in effect, the firm was paying to
attract customers it did not want. Internet technology is making it easier to research
internal and proprietary data. Often companies set up intranets so that employees
can use Web tools to store and share data within the organization. And just as
Google’s search software lets people search the entire World Wide Web, Google is munotes.in
Page 96
96 RESEARCH IN COMPUTING
offering the enterprise search, which is essentially the same technology in a version
that searches a corporate intranet. The enterprise search considers not only how
often a particular docum ent has been viewed but also the history of the user’s past
search patterns, such as how that user has looked at particular documents and for
how long. In addition, other companies have purchased specialized software, such
as Autonomy, which searches internal sources plus such external sources as news
government Web sites
5.8 External Data: The Distribution System
External data are generated or recorded by an entity other than the researcher’s
organization. The government, newspapers and journ als, trade associations, and
other organizations create or produce information. Traditionally, this information has been in published form, perhaps available from a public library, trade association, or government agency. Today, however, computerized data archives
and electronic data interchange make external data as accessible as internal data.
Exhibit below illustrates some traditional and some modern ways of distributing
information.
5.9 Information as a Product and Its Distribution Channels
Because seco ndary data have value, they can be bought and sold like other products.
And just as bottles of perfume or plumbers’ wrenches may be distributed in many
ways, secondary data also flow through various channels of distribution. Many
users, such as the Fortune 500 corporations, purchase documents and computerized
census data directly from the government. However, many small companies get
census data from a library or another intermediary or vendor of secondary
information. 1. Libraries
Traditionally, libraries’ va st storehouses of information have served as a bridge
between users and producers of secondary data. The library staff deals directly with
the creators of information, such as the federal government, and intermediate
distributors of information, such as ab stracting and indexing services. The user
need only locate the appropriate secondary data on the library shelves. Libraries
provide collections of books, journals, newspapers, and so on for reading and
reference. They also stock many bibliographies, abstra cts, guides, directories, and
indexes, as well as offer access to basic databases.
munotes.in
Page 97
97Chapter 5: Secondary Data Research in a Digital Age2. The Internet
Today, of course, much secondary data is conveniently available over the Internet.
Its creation has added an international dimension to the acquisition of secondary
data. For example, Library Spot, at http://www.libraryspot.com , provides links to
online libraries, including law libraries, medical libraries, and music libraries. Its
reference desk features links to calendars, dictionaries, encyclopedias, maps, and
other sources typically found at a traditional library’s reference desk. 3. Vendors
The information age offers many channels besides libraries through which to access
data. Many external producers make secondary data available directly from the
organizations that produce the data or through intermediaries, which are often
called vendors. Vendors such as Factiva now allow managers to access thousands
of external databases via desktop computers and telecommunications systems.
Hoovers ( http://www.hoovers.com ) specializes in providing information about
thousands of companies’ financial situati ons and operations. 4. Producers
Classifying external secondary data by the nature of the producer of information
yields five basic sources: publishers of books and periodicals, government sources,
media sources, trade association sources, and commercial sour ces. The following
section discusses each type of secondary data source.
1) Books and Periodicals
Some researchers consider books and periodicals found in a library to be the
quintessential secondary data source. A researcher who finds books on a topic
of int erest obviously is off to a good start. Professional journals, such as the
Journal of Marketing, Journal of Management, Journal of the Academy of
Marketing Science, The Journal of Business Research, Journal of Advertising
Research, American Demographics, and The Public Opinion Quarterly, as well
as commercial business periodicals such as the Wall Street Journal, Fortune,
and BusinessWeek, contain much useful material.
2) Government Sources
Government agencies produce data prolifically. Most of the data publish ed by the federal government can be counted on for accuracy and quality of investigation. Most students are familiar with the U.S. Census of Population,
which provides a wealth of data.
3) Media Sources
Information on a broad range of subjects is available from broadcast and print
media. CNN Financial News and BusinessWeek are valuable sources for munotes.in
Page 98
98 RESEARCH IN COMPUTINGinformation on the economy and many industries. Media frequently commission research studies about various as pects of Americans’ lives, such
as financial affairs, and make reports of survey findings available to potential
advertisers free of charge. Data about the readers of magazines and the audiences for broadcast media typically are profiled in media kits and advertisements.
4) Trade Association Sources
Trade associations, such as the Food Marketing Institute or the American
Petroleum Institute, serve the information needs of a particular industry. The
trade association collects data on a number of topics of specific interest to
firms, especially data on market size and market trends. Association members
have a source of information that is particularly germane to their industry
questions.
5) Commercial Sources
Numerous firms specialize in selling and/or publishing information. For example, the Polk Company publishes information on the automotive field,
such as average car values and new -car purchase rates by zip code. Many of
these organizations offer information in published formats and as CD -ROM or
Internet dat abases.
Market -Share Data. A number of syndicated services supply either wholesale or retail sales volume data based on product movement. Information Resources, Inc., collects market -share data using Universal
Product Codes (UPC) and optical scanning at re tail store checkouts.
Demographic and Census Updates. A number of firms, such as CACI
Marketing Systems and Urban Information Systems, offer computerized U.S. census files and updates of these data broken down by small geographic areas, such as zip codes. Many of these research suppliers provide indepth information on minority customers and other market segments.
Consumer Attitude and Public Opinion Research. Many research firms
offer specialized syndicated services that report findings from attitude
resear ch and opinion polls. For example, Yankelovich provides custom
research, tailored for specific projects, and several syndicated services.
Consumption and Purchase Behavior Data. NPD’s National Eating
Trends (NET) is the most detailed database available on consumption
patterns and trends for more than 4,000 food and beverage products. This
is a syndicated source of data about the types of meals people eat and when
and how they eat them. The data, called diary panel data, are based on
records of meals and di aries kept by a group of households that have agreed
to record their consumption behavior over an extended period of time. munotes.in
Page 99
99Chapter 5: Secondary Data Research in a Digital Age
Advertising Research. Advertisers can purchase readership and audience
data from a number of firms. W. R. Simmons and Associates measures
magazine audiences; Arbitron measures radio audiences; ACNielsen Media
Measurement estimates television audience ratings. By specializing in
collecting and selling audience information on a continuing basis, these
commercial sources pro vide a valuable service to their subscribers.
5.10 Single -Source Data -Integrated Information
ACNielsen Company offers data from both its television meters and scanner
operations. The integration
of these two types of data helps marketers investigate the im pact of television
advertising on retail sales. In other ways as well, users of data find that merging
two or more diverse types of data into a single database offers many advantages.
The data and information industry uses the term single -source data for d iverse
types of data offered by a single company.
Sources for Global Research
As business has become more global, so has the secondary data industry. The Japan
Management Association Research Institute, Japan’s largest provider of secondary
research data to government and industry, maintains an office in San Diego. The
Institute’s goal is to help U.S. firms access its enormous store of data about Japan
to develop and plan their business there. The office in San Diego provides
translators and acts as an intermediary between Japanese researchers and U.S.
clients.
Secondary data compiled outside the United States have the same limitations as
domestic secondary data. However, international researchers should watch for
certain pitfalls that frequently are associated with foreign data and cross -cultural
research . First, data may simply be unavailable in certain countries. Second, the
accuracy of some data may be called into question. This is especially likely with
official statistics that may be adjusted for the political purposes of foreign governments. Finally, although economic terminology may be standardized, various countries use different definitions and accounting and recording practices
for many economic concepts. For example, different countries may measu re
disposable personal income in radically different ways. International researchers
should take extra care to investigate the comparability of data among countries.
The U.S. government and other organizations compile databases that may aid
international secondary data needs. For example, The European Union in the U .S.
(http://www.eurunion.org ) reports on historical and current activity in the
European Union providing a comprehensive reference guide to information about
laws and regulations. The European U nion in the U.S. profiles in detail each
European Union member state, investment opportunities, sources of grants and
other funding, and other information about business resources. munotes.in
Page 100
100 RESEARCH IN COMPUTING
Summary
A researcher may have collected the data for a particular project, then made it
available to be used by another researcher. Secondary data are known to be readily
available compared to that of primary data. It requires very little research and need
for manpower to use these sources. Secondary data are known to be readily
available compared to that of primary data. It requires very little research and need
for manpower to use these sources. Modern information technology makes this distinction seem somewhat simplistic. Most organizations routinely gather, record, and store internal data to help them solve future problems. Secondary data
compiled outside the United States have the same limitations as domestic
secondary data. International researchers should take extra care to investigate the
comparability of data among countr ies.
Questions
1. What is secondary data? Explain its need in research.
2. Discuss any 4 sources of secondary data.
3. Discuss advantages and dis -advantages of secondary data.
4. Explain brief about data mining with reference to analysis.
5. Discuss the main objectives o f secondary research.
6. Explain in brief about the sources of global research.
7. Write a note on trend analysis.
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• 2. Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable,
• Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN : 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016,
• Business Research Methods, Edition 8, Cengage Publication
munotes.in
Page 101
UNIT 3
6 OBSERVATION METHODS AND
EXPERIMENTAL RESEARCH
Unit Structure
6.0 Objectives
6.1 Introduction
6.2 Observation in Business Research
6.3 The Nature of Observation Studies
6.4 Direct Observation
6.5 Observation of Physical Objects
6.6 Content Analysis
6.7 Mechanical Observation
6.7.1 Television Monitoring
6.7.2 Monitoring Web Site Traffic
6.7.3 Scanner -Based Research
6.7.4 Measuring Physiological Reactions
6.8 Experimental Research
6.9 Creating an Experiment
6.10 Manipulation of the Independent Variable
6.10.1 Experimental and Control Groups
6.10.2 Several Experimental Treatment Levels
6.10.3 More Than One Independent Variable
6.10.4 Repeated Measures
6.11 Selection and Assignment of Test Units
6.11.1 Sample selection and random sampling errors
6.11.2 Randomization
101munotes.in
Page 102
102 RESEARCH IN COMPUTING
6.11.3 Matching
6.11.4 Control over extraneous variables
6.11.5 Experimental confounds
6.11.5.1 Extraneous variables
6.12 Demand Characteristics
6.12.1 What Are Demand Characteristics ?
6.12.2 Experimenter Bias and Demand Effects
6.12.3 Hawthorne Effect
6.12.4 Reducing Demand Characteristics
6.13 Establishing Control
6.14 Practical Experimental Design Issues
6.14.1 Basic versus Factorial Experimental Designs
6.14.2 Laboratory Experiments
6.14.3 Field Experiments
6.14.4 Within -Subjects and Between -Subjects Designs
6.15 Issues of Experimental Validity
6.15.1 Internal Validity
6.15.2 Ext ernal Validity
6.16 Classification of Experimental Designs
6.17 Time Series Designs
6.17.1 Complex Experimental Designs
6.17.2 Completely Randomized Design
6.17.3 Randomized -block Design
6.17.4 Factorial Designs
6.18 Summary
6.19 Model Questions
munotes.in
Page 103
103Chapter 6: Observation Methods and Experimental Research
6.0 Objectives
1. Discuss the role of observation as a business research method
2. Describe the use of direct observation and contrived observation
3. Identify ethical issues in observation studies
4. Explain the observation of physical objects and message content
6.1 Introduction
While survey data can provide some insight into future or past behavior, one can
hardly argue with the power of data representing actual behavior.
6.2 Observation in Business Research
The systematic process of recording the behavioral patterns of people, objects, and
occurrences as they are witnessed i s called observation.
Observation becomes a tool for scientific inquiry when it meets several conditions:
ͻ The observation serves a formulated research purpose.
ͻ The observation is planned systematically.
ͻ The observation is recorded systematically and relate d to general propositions,
rather than simply reflecting a set of interesting curiosities.
ͻ The observation is subjected to checks or controls on validity and reliability.
6.3 The Nature of Observation Studies
Business researchers can observe people, objects, events, or other phenomena using either human observers or machines designed for specific observation tasks. Human observation best suits a situation or behavior that is not easily predictable
in advance of the research. Mechanical observation, as performed by supermarket
scanners or traffic counters, can very accurately record situations or types of
behavior that are routine, repetitive, or programmatic.
A situation in which an observer’s presence is known to the subject involves visible
observati on.
A situation in which a subject is unaware that observation is taking place is hidden
observation . munotes.in
Page 104
104 RESEARCH IN COMPUTING
Hidden, unobtrusive observation minimizes respondent error. Asking subjects to
participate in the research is not required when they are unaware that they are being
observed.
6.4 Direct Observation
Direct observation can produce detailed records of what people actually do during
an event. The observer plays a passive role, making no attempt to control or
manipulate a situation, instead merely recor ding what occurs.
Errors Associated with Direct Observation
A distortion of measurement resulting from the cognitive behavior or actions of the
witnessing observer is called observer bias. Interpretation of observation data is
another potential source of error. Facial expressions and other nonverbal communication may have several meanings
Scientifically Contrived Observation
Most observation takes place in a natural setting, but sometimes the investigator
intervenes to create an artificial environment in order to test a hypothesis. This
approach is called contrived observation. Contrived observation can increase the
frequency of occurrence of certain behavior patterns, such as employee responses
to complaints .
6.5 Observation of Physical Obj ects
Physical phenomena may be the subject of observation study. Physical -trace
evidence is a visible mark of some past event or occurrence. For example, the wear
on library books indirectly indicates which books are actually read (handled most)
when check ed out.
An observer can record physical -trace data to discover information a respondent
could not recall accurately. For example, measuring the number of ounces of a
liquid bleach used during a test provides precise physical -trace evidence without
relying on the respondent’s memory. The accuracy of respondents’ memories is not
a problem for the firm that conducts a pantry audit. The pantry audit requires an
inventory of the brands, quantities, and package sizes in a consumer’s home rather
than responses fro m individuals. The problem of untruthfulness or some other form
of response bias is avoided. For example, the pantry audit prevents the possible
problem of respondents erroneously claiming to have purchased prestige brands. munotes.in
Page 105
105Chapter 6: Observation Methods and Experimental Research
6.6 Content Analysis
Content analysis, which obtains data by observing and analyzing the contents or
messages of advertisements, newspaper articles, television programs, letters, and
the like. This method involves systematic analysis as well as observation to identify
the spec ific information content and other characteristics of the messages.
Content analysis studies the message itself and involves the design of a systematic
observation and recording procedure for quantitative description of the manifest
content of communicatio n.
Content analysis might be used to investigate questions such as whether some
advertisers use certain themes, appeals, claims, or deceptive practices more than
others or whether recent consumer -oriented actions by the Federal Trade Commission have influe nced the contents of advertising.
6.7 Mechanical Observation
In many situations, the primary —and sometimes the only —means of observation
is mechanical rather than human. Video cameras, traffic counters, and other
machines help observe and record behavior. Some unusual observation studies
have used motion -picture cameras and time -lapse photography. An early application of this observation technique photographed train passengers and determined their levels of comfort by observing how they sat and mo ved in their
seats.
6.7.1 Television Monitoring
Computerized mechanical observation used to obtain television ratings. The
Nielsen People Meter gathers data on what each television in a household is playing
and who is watching it at the time. Researchers att ach electronic boxes to television
sets and remote controls to capture information on program choices and the length
of viewing time. Nielsen matches the signals captured through these devices with
its database of network broadcast and cable program schedu les so that it can
identify the specific programs being viewed.
6.7.2 Monitoring Web Site Traffic
Computer technology makes gathering detailed data about online behavior easy and inexpensive. The greater challenges are to identify which measures are meanin gful and to interpret the data correctly. For instance, most organizations munotes.in
Page 106
106 RESEARCH IN COMPUTING
record the number of hits at their Web sites —mouse clicks on a single page of a
Web site.
A click -through rate (CTR) is the percentage of people who are exposed to an
advertisement who actually click on the corresponding hyperlink which takes them
to the company’s Web site.
A more refined count is the number of unique visitors to a Web site. This
measurement counts the initial access to the site but not multiple hits on the site by
the same visitor during the same day or week. Operators of Web sites can collect
the data by attaching small files, called cookies, to the computers of visitors to their
sites and then tracking those cookies to see whether the same visitors return.
6.7.3 Scanner -Based Research
It is a type of consumer panel in which participants’ purchasing habits are recorded
with a laser scanner rather than a purchase diary.
Data from scanner research parallel data provided by a standard mail diary panel,
with some importa nt improvements:
1. The data measure observed (actual) purchase behavior rather than reported
behavior (recorded later in a diary).
2. Substituting mechanical for human record -keeping improves accuracy.
3. Measures are unobtrusive, eliminating interview ing and the possibility of
social desirability or other bias on the part of respondents.
4. More extensive purchase data can be collected, because all UPC categories are
measured. In a mail diary, respondents could not possibly reliably record all
items t hey purchased. Because all UPC -coded items are measured in the panel,
users can investigate many product categories to determine loyalty, switching
rates, and so on for their own brands as well as for other companies’ products
and locate product categories for possible market entry.
5. The data collected from computerized checkout scanners can be combined with
data about the timing of advertising, price changes, displays, and special sales
promotions. Researchers can scrutinize them with powerful analytical software
provided by the scanner data providers.
6.7.4 Measuring Physiological Reactions
Researchers have used a number of mechanical devices to evaluate physical and physiological reactions to advertising copy, packaging, and other stimuli. munotes.in
Page 107
107Chapter 6: Observation Methods and Experimental Research
Researchers use such means when they believe consumers are unaware of their own
reactions to stimuli such as advertising or that consumers will not provide honest
responses. Recent research approaches use devices to monitor and measure brain
activity as desc ribed in the Research Snapshot above. Four major categories of
mechanical devices are used to measure physiological reactions:
(1) eye-tracking monitors: A mechanical device used to observe eye movements;
some eye monitors use infrared light beams to measure unconscious eye
movements.
(2) pupilometers: A mechanical device used to observe and record changes in the
diameter of a subject’s pupils.
(3) psych ogalvanometers : A device that measures galvanic skin response, a
measure of involuntary changes in the electrical resistance of the skin.
(4) voice -pitch analyzers : A physiological measurement technique that records
abnormal frequencies in the voice that a re supposed to reflect emotional
reactions to various stimuli.
6.8 Experimental Research
The term experiment typically conjures up an image of a chemist surrounded by
bubbling test tubes and Bunsen burners. Behavioral and physical scientists have
used expe rimentation far longer than have business researchers. Nevertheless, both social scientists and physical scientists use experiments for much the same purpose —to assess cause and effect relationships.
6.9 Creating an Experiment
Experimental research allows a researcher to control the research situation so that causal relationships among variables may be evaluated. The experimenter manipulates one or more independent variables and holds constant all other possible independent variables while observing effects on dependent variable(s).
Events may be controlled in an experiment to a degree that is simply not possible
in a survey.
Experimental design is a major research topic. In fact, there are courses and books
devoted only to that topic.3 Here, an intro duction into experimental design is
provided. A student should be able to design and implement basic experimental munotes.in
Page 108
108 RESEARCH IN COMPUTING
designs with this introduction. Fortunately, most experimental designs for business
research are relatively simple.
Experimental designs invol ve no less than four important design elements. These
issues include
(1) Manipulation of the independent variable(s)
(2) Selection and measurement of the dependent variable(s)
(3) Selection and assignment of experimental subjects
(4) Control over extraneous variables.
6.10 Manipulation of the Independent Variable
The thing that makes independent variables special in experimentation is that the researcher actually creates his or her values. This is how the researcher manipulates, and therefore controls, indepe ndent variables.
An experimental treatment is the term referring to the way an experimental variable
is manipulated.
6.10.1 Experimental and Control Groups
In perhaps the simplest experiment, an independent variable is manipulated over
two treatment levels resulting in two groups, an experimental group and a control
group. An experimental group is one in which an experimental treatment is
administered. A control group is one in which no experimental treatment is
administered.
6.10.2 Several Experimental Tre atment Levels
An experiment with one experimental and one control group may not tell a manager
everything he or she wishes to know. By analyzing more groups each with a
different treatment level, a more precise result may be obtained than in a simple
exper imental group –control group experiment. This design, only manipulating the
level of advertising, can produce only a main effect.
6.10.3 More Than One Independent Variable
An experiment can also be made more complicated by including the effect of
another ex perimental variable. Our extended example of the self -efficacy
experiment would typify a still relatively simple two -variable experiment. Since
there are two variables, each with two different levels, four experimental groups munotes.in
Page 109
109Chapter 6: Observation Methods and Experimental Research
are obtained. Often, the term cell is used to refer to a treatment combination within
an experiment.
6.10.4 Repeated Measures
Experiments in which an individual subject is exposed to more than one level of an
experimental treatment are referred to as repeated measures designs. Although this
approach has advantages, including being more economical since the same subject
provides more data than otherwise, it has several drawbacks that can limit its
usefulness.
6.11 Selection and Assignment of Test Units
Test units are the subjects or entities whose responses to the experimental treatment
are measured or observed. Individual consumers, employees, organizational units,
sales territories, market segments, or other entities may be the test units. People,
whether as customers or employe es, are the most common test units in most
organizational behavior, human resources, and marketing experiments.
6.11.1 Sample selection and random sampling errors
Systematic or nonsampling error may occur if the sampling units in an experimental
cell are somehow different than the units in another cell, and this difference affects
the dependent variable.
6.11.2 Randomization
The random assignment of subject and treatments to groups —is one device for
equally distributing the effects of extraneous variables to all conditions. These
nuisance variables, items that may affect the dependent measure but are not of
primary interest, often cannot be eliminated.
6.11.3 Matching
Random assignment of subjects to the various experimental groups is the most
common technique used to prevent test units from differing from each other on key
variables; it assumes that all characteristics of the subjects have been likewise randomized. Matching the respondents on the basis of pertinent background informati on is another technique for controlling systematic error by assigning
subjects in a way that their characteristics are the same in each group. This is best
thought of in terms of demographic characteristics. If a subject’s sex is expected to
influence depe ndent variable responses, as in a taste test, then the researcher may
make sure that there are equal numbers of men and women in each experimental munotes.in
Page 110
110 RESEARCH IN COMPUTING
cell. In general, if a researcher believes that certain extraneous variables may affect
the dependent variabl e, he or she can make sure that the subjects in each group are
the same on these characteristics.
6.11.4 Control over extraneous variables
The fourth decision about the basic elements of an experiment concerns control
over extraneous variables. This is rel ated to the various types of experimental error.
In an earlier chapter, we classified total survey error into two basic categories:
random sampling error and systematic error. The same dichotomy applies to all
research designs, but the terms random (sampli ng) error and systematic error are
more frequently used when discussing experiments.
6.11.5 Experimental confounds
A confound means that there is an alternative explanation beyond the experimental
variables for any observed differences in the dependent var iable. Once a potential
confound is identified, the validity of the experiment is severely questioned. In a
simple experimental group –control group experiment, if subjects in the experimental group are always administered treatment in the morning and subje cts
in the control group always receive the treatment in the afternoon, a systematic
error occurs.
6.11.5.1 Extraneous variables
Most business students realize that the marketing mix variables —price, product,
promotion, and distribution —interact with uncontrollable forces in the market,
such as economic variables, competitor activities, and consumer trends. Thus,
many marketing experiments are subject to the effect of extraneous variables. Since
extraneous variables can produ ce confounded results, they must be identified
before the experiment if at all possible.
6.12 Demand Characteristics
6.12.1 What Are Demand Characteristics?
The term demand characteristic refers to an experimental design element that
unintentionally provides subjects with hints about the research hypothesis. Researchers cannot reveal the research hypotheses to subjects before the experiment or else they can create a confounding effect. Think about the self -
efficacy experiment. If the subjects learned that t hey were being intentionally given
positive feedback to enhance their confidence and attitudes toward their job, the
researcher would never be sure if their responses to the dependent variable were munotes.in
Page 111
111Chapter 6: Observation Methods and Experimental Research
really due to the differences in the experimental stimuli or due to the fact that the
subjects were trying to provide a “correct” response.
Once subjects know the hypotheses, there is little hope that they will respond naturally. A confound may be created by knowledge of the experimental hypothesis. This particul ar type of confound is known as a demand effect. Demand
characteristics make demand effects very likely.
6.12.2 Experimenter Bias and Demand Effects
Demand characteristics are aspects of an experiment that demand (encourage) that
the subjects respond in a particular way. Hence, they are a source of systematic
error. If participants recognize the experimenter’s expectation or demand, they are
likely to act in a manner consistent with the experimental treatment. Even slight
nonverbal cues may influence their reactions. Prominent demand characteristics are often presented by the person administering experimental procedures. If an experimenter’s presence, actions, or comments influence the subjects’ behavior or
sway the subjects to slant their answers to coopera te with the experimenter, the
experiment has introduced experimenter bias.
6.12.3 Hawthorne Effect
A famous management experiment illustrates a common demand characteristic. Researchers were attempting to study the effects on productivity of various working conditions, such as hours of work, rest periods, lighting, and methods of
pay, at the Western Ele ctric Hawthorne plant in Cicero, Illinois. The researchers found that workers’ productivity increased whether the work hours were lengthened or shortened, whether lighting was very bright or very dim, and so on.
The surprised investigators realized that the workers’ morale was higher because
they were aware of being part of a special experimental group. This totally
unintended effect is now known as the Hawthorne effect because researchers
realize that people will perform differently when th ey know they are experimental
subjects.
6.12.4 Reducing Demand Characteristics
Although it is practically impossible to eliminate demand characteristics from
experiments, there are steps that can be taken to reduce them. Many of these steps
make it difficu lt for subjects to know what the researcher is trying to find out. Some
or all of these may be appropriate in a given experiment.
1. Use an experimental disguise.
2. Isolate experimental subjects. munotes.in
Page 112
112 RESEARCH IN COMPUTING
3. Use a “blind” experimental administrator.
4. Administer only one exp erimental treatment level to each subject.
Experimental disguise
A placebo is an experimental deception involving a false treatment. A placebo
effect refers to the corresponding effect in a dependent variable that is due to the
psychological impact that goes along with knowledge of the treatment.
Isolate experimental subjects
Researchers should minimize the extent to which subjects are able to talk about the
experimental procedures with each other. Although it may be unintentional,
discussion among s ubjects may lead them to guess the experimental hypotheses.
Use a “blind” experimental administrator
When possible, the people actually administering the experiment may not be told
the experimental hypotheses. The advantage is that if they do not know what
exactly is being studied, then they are less likely to give off clues that result in
demand effects.
Administer only one experimental condition per subject
When subjects know more than one experimental treatment condition, they are
much more likely to guess the experimental hypothesis. So, even though there are
cost advantages to administering multiple treatment levels to the same subject, it
should be avoided when possible.
6.13 Establishing Control
When extraneous variables cannot be eliminated, expe rimenters may strive for
constancy of conditions. This means that subjects in all experimental groups are
exposed to identical conditions except for the differing experimental treatments.
Random assignment and the principle of matching discussed earlier he lp make sure
that constancy is achieved.
If an experimental method requires that the same subjects be exposed to two or
more experimental treatments, an error may occur due to the order of presentation.
For instance, if subjects are examining the effects o f different levels of graphical
interface on video game enjoyment, and they are asked to view each of four
different levels, the order in which they are presented may influence enjoyment.
Subjects might prefer one level simply because it follows a very poo r level. munotes.in
Page 113
113Chapter 6: Observation Methods and Experimental ResearchCounterbalancing attempts to eliminate the confounding effects of order of presentation by requiring that one -fourth of the subjects be exposed to treatment A
first, one -fourth to treatment B first, one -fourth to treatment C first, and finally on e-
fourth to treatment D first. Likewise, the other levels are counterbalanced so that
the order of presentation is rotated among subjects. It is easy to see where
counterbalancing is particularly important for experiments such as taste tests,
where the order of presentation may have significant effects on consumer preference.
6.14 Practical Experimental Design Issues
6.14.1 Basic versus Factorial Experimental Designs
In basic experimental designs a single independent variable is manipulated to
observe its effect on a single dependent variable. The simultaneous change in
independent variables such as price and advertising may have a greater influence
on sales than if either variable is changed alone. In job satisfaction studies, we know
that no o ne thing totally determines job satisfaction.
6.14.2 Laboratory Experiments
In a laboratory experiment the researcher has more complete control over the
research setting and extraneous variables. Our example of the financial protocol
experiment illustrates the benefits of a laboratory setting. The researchers were able
to control for many factors, such as the size of the data file, the models of the
computers, the Internet line, and so forth. This enhanced their confidence in
establishing that the differenc es noted in speed were due to the different protocols.
However, the researchers were not able to determine how the protocols compared
when used in the field, on various computers, with a variety of file sizes, and under
differing “real -world” circumstances .
6.14.3 Field Experiments
Field experiments are research projects involving experimental manipulations that
are implemented in a natural environment. They can be useful in fine -tuning
managerial strategies and tactical decisions. Experiments vary in their degree of artificiality and control. In field experiments, a researcher manipulates experimental variables but cannot possibly control all the extraneous variables.
6.14.4 Within -Subjects and Between -Subjects Designs
Field experiments involving new p roducts or promotions are often conducted in a
retail store. These are known as controlled store tests. The products are put into munotes.in
Page 114
114 RESEARCH IN COMPUTING
stores in a number of small cities or into selected supermarket chains. Product
deliveries are made not through the traditiona l warehouse but by the research
agency, so product information remains confidential.
6.15 Issues of Experimental Validity
An experiment’s quality is judged by two types of validity. These are known as
internal and external validity. Internal validity exists to the extent that an experimental variable is truly responsible for any variance in the dependent
variable. In other words, does the experimental manipulation truly cause changes
in the specific outcome of interest? If the observed results were influenced or
confounded by extraneous factors, the researcher will have problems making valid
conclusions about the relationship between the experimental treatment and the
dependent variable.
6.15.1 Internal Validity
Internal validity exists to the extent that an experimental variable is truly responsible for any variance in the dependent variable. In other words, does the
experimental manipulation truly cause changes in the specific outcome of interest?
If the observed results were influenced or con founded by extraneous factors, the
researcher will have problems making valid conclusions about the relationship
between the experimental treatment and the dependent variable.
Manipulation checks:
The validity of manipulations can often be determined with a manipulation check.
If a drug is administered in different dosages that should affect blood sugar levels,
the researcher could actually measure blood sugar level after administering the drug
to make sure that the dosages were different enough to produce a change in blood
sugar. In business research, the manipulation check is often conducted by asking a
survey question or two. Extraneous variables can jeopardize internal validity. The
six major ones are history, maturation, testing, instrumentation, select ion, and
mortality.
History :
Occurs when some change other than the experimental treatment occurs during the
course of an experiment that affects the dependent variable. A special case of the
history effect is the cohort effect, which refers to a change in the dependent variable
that occurs because members of one experimental group experienced different
historical situations than members of other experimental groups. munotes.in
Page 115
115Chapter 6: Observation Methods and Experimental Research
Maturation
Maturation effects are effects that are a function of time and the natu rally occurring
events that coincide with growth and experience. Experiments taking place over
longer time spans may see lower internal validity as subjects simply grow older or
more experienced.
Testing
Testing effects are also called pretesting effects because the initial measurement or
test alerts or primes subjects in a way that affects their response to the experimental
treatments. Testing effects only occur in a before -and-after study. A before -and-
after study is one requiring an initial baseline measure be taken before an experimental treatment is administered. So, before and- after experiments are a special case of a repeated measures design. For example, students taking standardized achievement and intelligence tests for the second time usually do
better than those taking the tests for the first time.
Instrumentation
A change in the wording of questions, a change in interviewers, or a change in other
procedures used to measure the dependent variable causes an instrumentation
effect, which may jeopard ize internal validity. Sometimes instrumentation effects
are difficult to control.
Selection
The selection effect is a sample bias that results from differential selection of
respondents for the comparison groups, or sample selection error, discussed earlier.
Researchers must make sure the characteristics of the research subjects accurately
reflect the population of relevance.
Mortality
If an experiment is conducted over a period of a few weeks or more, some sample
bias may occur due to the mortality effect (sample attrition). Sample attrition
occurs when some subjects withdraw from the experiment before it is completed.
Mortality effects may occur if subjects drop from one experimental treatment group
disproportionately than from other group s. munotes.in
Page 116
116 RESEARCH IN COMPUTING
6.15.2 Ext ernal Validity External validity is the accuracy with which experimental results can be generalized beyond the experimental subjects. External validity is increased when
the subjects comprising the sample truly represent the population of interest and
when the results extend to other market segments or groups of people. The higher
the external validity, the more researchers and managers can count on the fact that
any results observed in an experiment will also be seen in the “ real world”
(financial market, workplace, sales floor, and so on).
6.16 Classification of Experimental Designs
An experimental design may be compared to an architect’s plans for a building.
The basic requirements for the structure are given to the architec t by the prospective
owner. Several different plans may be drawn up as options for meeting the basic
requirements. Some may be more costly than others. One may offer potential
advantages that another does not. There are various types of experimental design s.
If only one variable is manipulated, the experiment has a basic experimental design.
If the experimenter wishes to investigate several levels of the independent variable
(for example, four different employee salary levels) or to investigate the interact ion
effects of two or more independent variables (salary level and retirement package),
the experiment requires a complex, or statistical, experimental design.
Symbolism for Diagramming Experimental Designs
The work of Campbell and Stanley has helped many students master the subject of
basic experimental designs.
The following symbols will be used in describing the various experimental designs:
X = exposure of a group to an experimental treatment
O = observation or measurement of the dependent variable ; if more than
one observation or measurement is taken, subscripts (that is, O1, O2, etc.)
indicate temporal order
R = random assignment of test units; R symbolizes that individuals selected
as subjects for the experiment are randomly assigned to the expe rimental
groups
The diagrams of experimental designs that follow assume a time flow from left to
right. Our first example will make this clearer. munotes.in
Page 117
117Chapter 6: Observation Methods and Experimental Research
6.17 Time Series Designs
Many experiments may be conducted in a short period of time (a few hours, a week,
or a month). However, a business experiment investigating long -term strategic
and/or structural changes may require a time series design. Time series designs are
quasi -experimental because they generally do not allow the researcher full control
over the treatment exposure or influence of extraneous variables. When experiments are conducted over long periods of time, they are most vulnerable to
history effects due to changes in population, attitudes, economic patterns, and the
like. Although seasonal patterns a nd other exogenous influences may be noted, the
experimenter can do little about them when time is a major factor in the design.
6.17.1 Complex Experimental Design s
The previous discussion focused on simple experimental designs —experiments
manipulating a single variable. Here, the focus shifts to more complex experimental
designs involving multiple experimental variables. Complex experimental designs
are statistical designs that isolate the effects of confounding extraneo us variables
or allow for manipulation of more than one independent variable in the experiment.
Completely randomized designs, randomized block designs, and factorial designs
are covered in the following section.
6.17.2 Completely Randomized Design
A completely randomized design is an experimental design that uses a random process to assign subjects to treatment levels of an experimental variable. Randomization of experimental units is the researcher’s attempt to control extraneous variables while mani pulating potential causes. A one -variable experimental design can be completely randomized, so long as subjects are assigned in a random way to a particular experimental treatment level.
6.17.3 Randomized -block Design
The randomized -block design is an exte nsion of the completely randomized design.
A form of randomization is used to control for most extraneous variation; however,
the researcher has identified a single extraneous variable that might affect subjects’
responses systematically. The researcher wi ll attempt to isolate the effects of this
single variable by blocking out its effects. The term randomized block originated
in agricultural research that applied several levels of a treatment variable to each of
several blocks of land. Systematic differenc es in agricultural yields due to the
quality of the blocks of land may be controlled in the randomized -block design. munotes.in
Page 118
118 RESEARCH IN COMPUTING
6.17.4 Factorial Designs
A factorial design allows for the testing of the effects of two or more treatments
(factors) at various levels.
Summary
1. Discuss the role of observation as a business research method. Observation is
a powerful tool for the business researcher. Scientific observation is the
systematic process of recording the behavioral patterns of people, objects, and
occurrences a s they are witnessed. Questioning or otherwise communicating
with subjects does not occur. A wide variety of information about the behavior of people and objects can be observed. Seven kinds of phenomena are observable: physical actions, verbal behavior, e xpressive behavior, spatial
relations and locations, temporal patterns, physical objects, and verbal and
pictorial records. Thus, both verbal and nonverbal behavior may be observed.
2. Describe the use of direct observation and contrived observation. Huma n
observation, whether direct or contrived, is commonly used when the situation
or behavior to be recorded is not easily predictable in advance of the research.
It may be unobtrusive, and many types of data can be obtained more accurately
through direct ob servation than by questioning respondents. Direct observation
involves watching and recording what naturally occurs, without creating an
artificial situation. For some data, observation is the most direct or the only
method of collection. For example, rese archers can measure response latency,
the time it takes individuals to choose between alternatives. Observation can
also be contrived by creating the situations to be observed, such as with a
mystery shopper or a research laboratory. This can reduce the ti me and expense
of obtaining reactions to certain circumstances.
3. Identify ethical issues in observation studies. Contrived observation, hidden
observation, and other observation research designs have the potential to
involve deception. For this reason, these methods often raise ethical concerns
about subjects’ right to privacy and right to be informed. We mentioned three
questions to help determine the ethicality of observation: (1) is the behavior
being observed commonly performed in public where others can observe it, (2)
is anonymity of the subject assured, and (3) has the subject agreed to be
observed? If the answers to 1 and 2 are “yes,” or if the answer to 3 is “yes,’ the
observation is likely ethical. munotes.in
Page 119
119Chapter 6: Observation Methods and Experimental Research
4. Explain the observation of physical object s and message content. Physical -
trace evidence serves as a visible record of past events. Researchers may
examine whatever evidence provides such a record, including inventory levels,
the contents of garbage cans, or the items in a consumer’s pantry. Conte nt
analysis obtains data by observing and analyzing the contents of the messages
in written or spoken communications.
5. Describe major types of mechanical observation. Mechanical observation uses
a variety of devices to record behavior directly. It may b e an efficient and
accurate choice when the situation or behavior to be recorded is routine,
repetitive, or programmatic. National television audience ratings are based on
mechanical observation (for example, Nielsen’s People Meters) and computerized data c ollection. Web site traffic may be measured electronically.
Scanner -based research provides product category sales data recorded by laser scanners in retail stores. Many syndicated services offer secondary data collected through scanner systems.
6. Summarize techniques for measuring physiological reactions. Physiological
reactions, such as arousal or eye movement patterns, may be observed using a
number of mechanical devices. Eye -tracking monitors identify the direction of
a person’s gaze, and a pup ilometer observes and records changes in the
diameter of the pupils of subjects’ eyes, based on the assumption that a larger
pupil signifies a positive attitude. A psychogalvanometer measures galvanic
skin response as a signal of a person’s emotional react ions. Voice -pitch
analysis measures changes in a person’s voice and associates the changes with
emotional response.
Questions
1. What are the advantages and disadvantages of observation studies relative to
surveys?
2. Under what conditions are observation studies most appropriate?
3. What is a scanner -based consumer panel?
4. What are the major types of mechanical observation?
5. What is a psychogalvanometer?
6. Define experimental condition, experimental treatment, and experimental
group. How are these rel ated to the implementation of a valid manipulation? munotes.in
Page 120
120 RESEARCH IN COMPUTING
7. What is the difference between a main effect and an interaction in an
experiment?
8. In what ways might the design in question 2 yield systematic or non sampling
error?
9. What purpose does the random assignmen t of subjects serve?
10. Why is an experimental confound so damaging to the conclusions drawn
from an experiment?
11. What are demand characteristics? How can they be minimized?
12. What is a manipulation check? How does it relate to internal validity?
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Me thods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN: 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016, Business Research Methods, Edition 8, Cengage Publication
munotes.in
Page 121
121Unit 3
7 SURVEY RESEARCH
Unit Structure
7.0 Objectives
7.1 Introduction
7.2 Using Surveys
7.3 Advantages and disadvantages of Surveys
7.4 Errors in Survey Research
7.4.1 Random sampling error:
7.4.2 Systematic error
7.4.3 Sample bias:
7.4.4 Respondent error:
7.4.5 Self -selection bias:
7.4.6 Response bias:
7.4.7 Deliberate Falsification
7.4.8 Unconscious Misrepresentation
7.5 Types of Response Bias
7.5.1 Acquiescence Bias.
7.5.2 Extremity Bias.
7.5.3 Interviewer Bias.
7.5.4 Social Desirability Bias.
7.6 Administrative Error
7.6.1 Data -processing error
7.6.2 Sample selection error
7.6.3 Interviewer error
7.6.4 Interviewer cheating munotes.in
Page 122
122 RESEARCH IN COMPUTING
7.7 Classifying Survey Research Methods
7.7.1 Structured/Unstructured and Disguised/ Undisguised Questionnaires
7.7.2 Temporal Classification
7.7.3 Interviews as Interactive Communication
7.8 Personal Interviews
7.9 Door -to-Door Interviews and Shopping Mall Intercepts
7.10 Telephone Interviews
7.11 Self-Administered Questionnaires
7.12 Self-Administered Questionnaires Using Other Forms of Distribution
7.13 Pretesting
7.14 Summary
7.15 Model Questions
7.0 Objectives
1. Define surveys and explain their advantages
2. Describe the type of information t hat may be gathered in a survey
3. Identify sources of error in survey research
4. Distinguish among the various categories of surveys
5. Discuss the importance of survey research to total quality management programs
6. Summarize ways researchers gather information th rough interviews
7. Discuss the importance of pretesting questionnaires
8. Describe ethical issues that arise in survey research
7.1 Introduction
Often research entails asking people —called respondents —to provide answers to
written or spoken questions. These interviews or questionnaires collect data through the mail, on the telephone, online, or face -to-face. Thus, a survey is defined as a method of collecting primary data based on communication with a representative sample of individuals. Surveys provid e a snapshot at a given point
in time. The more formal term, sample survey, emphasizes that the purpose of
contacting respondents is to obtain a representative sample, or subset, of the target
population. munotes.in
Page 123
123Chapter 7: Survey Research
7.2 Using Surveys
Surveys attempt to describe what is happening or to learn the reasons for a particular business activity. The term survey is most often associated with quantitative findings. Although most surveys are conducted to quantify certain
factual information, some aspects of surveys may also be q ualitative.
7.3 Advantages and disadvantages of Surveys
Surveys provide a quick, inexpensive, efficient, and accurate means of assessing
information about a population. When properly conducted, surveys offer managers
many advantages. However, they can also be used poorly when researchers do not
follow research principles, such as careful survey and sample design. Sometimes
even a welldesigned and carefully executed survey is not helpful because the results
are delivered too late to inform decisions.
7.4 Errors in Survey Research
7.4.1 Random sampling error:
A statistical fluctuation that occurs because of chance variation in the elements
selected for a sample
7.4.2 Systematic error:
Error resulting from some imperfect aspect of the research design that cau ses
respondent error or from a mistake in the execution of the research.
7.4.3 Sample bias:
A persistent tendency for the results of a sample to deviate in one direction from
the true value of the population parameter.
7.4.4 Respondent error:
A category of sample bias resulting from some respondent action or inaction such
as nonresponse or response bias.
Nonresponse error: The statistical differences between a survey that includes only
those who responded and a perfect surve y that would also include those who failed
to respond.
Nonrespondents: People who are not contacted or who refuse to cooperate in the
research. munotes.in
Page 124
124 RESEARCH IN COMPUTING
No contacts: People who are not at home or who are otherwise inaccessible on the
first and second contact.
Refusals: People who are unwilling to participate in a research project.
7.4.5 Self-selection bias:
A bias that occurs because people who feel strongly about a subject are more likely
to respond to survey questions than people who feel indifferent about i t.
7.4.6 Response bias:
A bias that occurs when respondents either consciously or unconsciously tend to
answer questions with a certain slant that misrepresents the truth.
7.4.7 Deliberate Falsification
Occasionally people deliberately give false answers. It is difficult to assess why
people knowingly misrepresent answers. A response bias may occur when people
misrepresent answers to appear intelligent, conceal personal information, avoid
embarrassment, a nd so on.
7.4.8 Unconscious Misrepresentation
Even when a respondent is consciously trying to be truthful and cooperative,
response bias can arise from the question format, the question content, or some
other stimulus.
7.5 Types of Response Bias
7.5.1 Acquiescence Bias
Some respondents are very agreeable. They seem to agree to practically every
statement they are asked about. A tendency to agree (or disagree) with all or most
questions is known as acquiescence bias.
7.5.2 Extremity Bias
Some individuals te nd to use extremes when responding to questions. For example,
they may choose only “1” or “10” on a ten -point scale. Others consistently refuse
to use extreme positions and tend to respond more neutrally —“I never give a 10
because nothing is really perfect .” Response styles vary from person to person, and
extreme responses may cause an extremity bias in the data.
7.5.3 Interviewer Bias
Response bias may arise from the interplay between interviewer and respondent. If
the interviewer’s presence influences res pondents to give untrue or modified
answers, the survey will be marred by interviewer bias. munotes.in
Page 125
125Chapter 7: Survey Research
7.5.4 Social Desirability Bias
Social desirability bias may occur either consciously or unconsciously because the
respondent wishes to create a favorable impression or save face in the presence of
an interviewer.
7.6 Administrative Error
The result of improper administration or executi on of the research task is called an
administrative error.
Administrative errors are caused by carelessness, confusion, neglect, omission, or
some other blunder.
Four types of administrative error are :
7.6.1 Data -processing error
A category of administr ative error that occurs because of incorrect data entry,
incorrect computer programming, or other procedural errors during data analysis.
7.6.2 Sample selection error
An administrative error caused by improper sample design or sampling procedure
execution.
7.6.3 Interviewer error
Mistakes made by interviewers failing to record survey responses correctly.
7.6.4 Interviewer cheating
The practice of filling in fake answers or falsifying questionnaires while working
as an interviewer.
7.7 Classifying Survey Research Methods
Surveys may be classified based on the method of communication, the degrees of
structure and disguise in the questionnaire,method of communicating with the respondent and the time frame in which the data are gathered (temporal class ification).
7.7.1 Structured/Unstructured and Disguised/ Undisguised Questionnaires
In designing a questionnaire (or an interview schedule), the researcher must decide
how much structure or standardization is needed.11 A structured question limits
the numb er of allowable responses. munotes.in
Page 126
126 RESEARCH IN COMPUTING
An unstructured question does not restrict the respondent’s answers. An open -
ended, unstructured question such as “Why do you shop at Wal -Mart?” allows the
respondent considerable freedom in answering.
The researcher must also de cide whether to use undisguised questions or disguised
questions. A straightforward, or undisguised, question such as “Do you have dandruff problems?” assumes that the respondent is willing to reveal the information. However, researchers know that some que stions are threatening to a
person’s ego, prestige, or self -concept. So, they have designed a number of indirect
techniques of questioning to disguise the purpose of the study.
7.7.2 Temporal Classification
Although most surveys are for individual research projects conducted only once
over a short time period, other projects require multiple surveys over a long period.
Thus, surveys can be classified on a temporal basis
• Cross -sectional study
A study in which various segments of a populat ion are sampled and data are
collected at a single moment in time.
• Longitudinal study
A survey of respondents at different times, thus allowing analysis of response
continuity and changes over time.
• Tracking study
A type of longitudinal study thatuses successive samples to compare trends
and identify changes in variables such as consumer satisfaction, brand image,
or advertising awareness.
• Consumer panel
A longitudinal survey of the same sample of individuals or households to
record their atti tudes, behavior, or purchasing habits over time.
7.7.3 Interviews as Interactive Communication
When two people engage in a conversation, human interaction takes place. Human
interactive media are a personal form of communication. One human being directs
a message to and interacts with another individual (or a small group). When most
people think of interviewing, they envision two people engaged in a face -to-face
dialogue or a conversation on the telephone. munotes.in
Page 127
127Chapter 7: Survey Research
Noninteractive Media
The traditional questionnaire received by mail and completed by the respondent
does not allow a dialogue or an exchange of information providing immediate
feedback. So, from our perspective, self -administered questionnaires printed on
paper are noninteractive. This fact does not mean t hat they are without merit, just that this type of survey is less flexible than surveys using interactive communication media.
7.8 Personal Interviews
Personal interview is a form of direct communication in which an interviewer asks
respondents questions face -to-face. This versatile and flexible method is a two -way
conversation between interviewer and respondent.
Advantages of Personal Interviews
Business researchers find that personal interviews offer many unique advantages.
One of the most impo rtant is the opportunity for detailed feedback.
• Opportunity for Feedback:
Personal interviews, similar to those mentioned in the Research Snapshot on
the next page, provide the opportunity for feedback and clarification.
• Probing Complex Answers:
Anothe r important characteristic of personal interviews is the opportunity to
follow up by probing. If a respondent’s answer is too brief or unclear, the
researcher may request a more comprehensive or clearer explanation.
• Length of Interview:
If the research objective requires an extremely lengthy questionnaire, personal interviews may be the only option. A general rule of thumb on mail
surveys is that they should not exceed six pages, and telephone interviews
typically last less than ten minutes. In contrast, a personal interview can be
much longer, perhaps an hour and a half
• Completeness of Questionnaire:
The social interaction between a well -trained interviewer and a respondent in
a personal interview increases the likelihood that the respondent will answer
all the items on the questionnaire.
Item nonresponse —failure to provide an answer to a question —is least likely
to occur when an experienced interviewer asks questions directly. munotes.in
Page 128
128 RESEARCH IN COMPUTING
• Props and Visual Aids:
Interviewing respondents face -to-face allows the investigator to show them
new product samples, sketches of proposed advertising, or other visual aids. Research that uses visual aids has become increasingly popular with researchers who investigate film concepts, advertising problems, and movie goers’ awareness of performers.
• High Participation:
Although some people are reluctant to participate in a survey, the presence
of an interviewer generally increases the percentage of people willing to
complete the interview.
Disadvantages of Personal Interviews Personal interviews also have some disadvantages. Respondents are not anonymous and as a result may be reluctant to provide confidential information to
another person
• Interviewer Influence :
Some evidence suggests that demographic char acteristics of the interviewer
influence respondents’ answers.
• Lack of Anonymity of Respondent:
Because a respondent in a personal interview is not anonymous and may be
reluctant to provide confidential information to another person, researchers
often sp end considerable time and effort to phrase sensitive questions to
avoid social desirability bias
• Cost :
Personal interviews are expensive, generally substantially more costly than mail, Internet, or telephone surveys. The geographic proximity of respondents, the length and complexity of the questionnaire, and the number
of people who are nonrespondents becau se they could not be contacted (not -
at-homes) will all influence the cost of the personal interview
7.9 Door -to-Door Interviews and Shopping Mall Intercepts
Personal interviews may be conducted at the respondents’ homes or offices or in
many other places. Increasingly, personal interviews are being conducted in
shopping malls. Mall intercept interviews allow many interviews to be conducted munotes.in
Page 129
129Chapter 7: Survey Research
quickly. Often, resp ondents are intercepted in public areas of shopping malls and
then asked to come to a permanent research facility to taste new food items or to
view advertisements.
Door to Door Interviews
The presence of an interviewer at the door generally increases the likelihood that a
person will be willing to complete an interview. Because door -to-door interviews
increase the participation rate, they provide a more representative sample of the
population than mail questionnaires.
Callbacks
When a person selected to be in the sample cannot be contacted on the first visit, a
systematic procedure is normally initiated to call back at another time. Callbacks,
or attempts to recontact individuals selected for the sample, are the major means of
reducing nonresponse error. Ca lling back a sampling unit is more expensive than
interviewing the person the first time around, because subjects who initially were
not at home generally are more widely dispersed geographically than the original
sample units.
Mall Intercept Interviews
Personal interviews conducted in shopping malls are referred to as mall intercept
interviews, or shopping center sampling. Interviewers typically intercept shoppers
at a central point within the mall or at an entrance. The main reason mall intercept
intervie ws are conducted is because their costs are lower. No travel is required to
the respondent’s home; instead, the respondent comes to the interviewer, and many
interviews can be conducted quickly in this way.
7.10 Telephone Interviews
For several decades, la ndline telephone interviews have been the mainstay of commercial survey research. The quality of data obtained by telephone is potentially comparable to the quality of data collected face -to-face. Respondents
are more willing to provide detailed and reliab le information on a variety of
personal topics over the phone while in the privacy of their own homes than when
answering questions face -to-face.
Mobile Phone Interviews
Mobile phone interviews differ from landline phones most obviously because they
are di rected toward a mobile (i.e., cell) phone number. munotes.in
Page 130
130 RESEARCH IN COMPUTING
Phone Interview Characteristics
Phone interviews in general have several distinctive characteristics that set them apart from other survey techniques. These characteristics present significant advantages a nd disadvantages for the researcher.
• Speed:
One advantage of telephone interviewing is the speed of data collection.
While data collection with mail or personal interviews can take several
weeks, hundreds of telephone interviews can be conducted literally overnight. When the interviewer enters the respondents’ answers directly into
a computerized system, the data processing speeds up even more.
• Cost :
As the cost of personal interviews continues to increase, telephone interviews
are bec oming relatively inexpensive. The cost of telephone interviews is
estimated to be less than 2 3.2 percent of the cost of door -to-door personal
interviews. Travel time and costs are eliminated. However, the typical
Internet survey is less expensive than a te lephone survey.
• Absence of Face to Face Contact:
Telephone interviews are more impersonal than face -to-face interviews. Respondents may answer embarrassing or confidential questions more willingly in a telephone interview than in a personal interview.
• Cooperation:
One trend is very clear. In the last few decades, telephone response rates have
fallen. One way researchers can try to improve response rates is to leave a
message on the household’stelephone answering machine or voice mail.
Howe ver, many people will not return a call to help someone conduct a
survey.
• Incentives to Repond:
Respondents should receive some incentive to respond. Research addresses
different types of incentives. For telephone interviews, test -marketing
involving different types of survey introductions suggests that not all introductions are equally effective. munotes.in
Page 131
131Chapter 7: Survey Research
• Representative Samples:
Practical difficulties complicate obtaining representative samples based on
listings in the telephone book. The problem of unlisted ph one numbers can
be partially resolved through the use of random digit dialing. Random digit
dialing eliminates the counting of names in a list (for example, calling every
fiftieth name in a column) and subjectively determining whether a directory
listing i s a business, institution, or legitimate household. In the simplest form
of random digit dialing, telephone exchanges (prefixes) for the geographic
areas in the sample are obtained. Using a table of random numbers, the last
four digits of the telephone num ber are selected.
• Callbacks :
An unanswered call, a busy signal, or a respondent who is not at home
requires a callback. Telephone callbacks are much easier to make than
callbacks in personal interviews. However, as mentioned, the ownership of
telephone a nswering machines is growing, and their effects on callbacks need
to be studied.
• Limited Duration:
Respondents who run out of patience with the interview can merely hang up.
To encourage participation, interviews should be relatively short. The length
of the telephone interview is definitely limited.
• Lack of Visual Medium:
Because visual aids cannot be used in telephone interviews, this method is
not appropriate for packaging research, copy testing of television and print
advertising, and conce pt tests that require visual materials.
7.11 Self-Administered Questionnaires
Surveys in which the respondent takes the responsibility for reading and answering
the questions is called self -administered questionnaires.Self -administered
questionnaires prese nt a challenge to the researcher because they rely on the clarity
of the written word rather than on the skills of the interviewer. The nature of self
administered questionnaires is best illustrated by explaining mail questionnaires.
Mail Questionnaires
A mail survey is a self -administered questionnaire sent to respondents through the
mail. This paper -and-pencil method has several advantages and disadvantages munotes.in
Page 132
132 RESEARCH IN COMPUTING
• Geographic Flexibility:
Mail questionnaires can reach a geographically dispersed sample
simultaneously because interviewers are not required.
• Cost:
Mail questionnaires are relatively inexpensive compared with personal interviews, though they are not cheap.
• Respondent Convenience:
Mail surveys and other self-administered questionnaires can be filled out
when the respondents have time, so respondents are more likely to take time
to think about their replies. Many hardto -reach respondents place a high value
on convenience and thus are best contacted by mail .
• Anonymity of Respondent:
In the cover letter that accompanies a mail or self -administered questionnaire, researchers almost always state that the respondents’ answers will be confidential. Respondents are more likely to provide sensitive or embarrassin g information when they can remain anonymous
• Absence of Interviewer:
Although the absence of an interviewer can induce respondents to reveal
sensitive or socially undesirable information, this lack of personal contact can
also be a disadvantage. Once the respondent receives the questionnaire, the
questioning process is beyond the researcher’s control.
• Standardized Questions:
Mail questionnaires typically are highly standardized, and the questions are
quite structured. Questions and instructions must be clear -cut and straightforward.
• Time is Money:
If time is a factor in management’s interest in the research results, or if
attitudes are rapidly changing (for example, toward a political event), mail
surveys may not be the best communication medium. A mini mum of two or
three weeks is necessary for receiving the majority of the responses. munotes.in
Page 133
133Chapter 7: Survey Research
• Length Of Mail Questionnaire:
Mail questionnaires vary considerably in length, ranging from extremely
short postcard questionnaires to multipage booklets that require respondents
to fill in thousands of answers. A general rule of thumb is that a mail
questionnaire should not exceed six pages in length.
Response rate
The number of questionnaires returned or completed divided by the number of
eligible people who w ere asked to participate in the survey.
Increasing Response Rates for Mail Surveys
Nonresponse error is always a potential problem with mail surveys. Individuals
who are interested in the general subject of the survey are more likely to respond
than those with less interest or little experience.
7.12 Self-Administered Questionnaires Using Other Forms of Distribution
Drop -off method:
A survey method that requires the interviewer to travel to the respondent’s location
to drop off questionnaires that will be picked up later.
Fax survey:
A survey that uses fax machines as a way for respondents to receive and return
questionnaires.
E-Mail Surveys:
Questionnaires can be distributed via e -mail, but researchers mus t remember that
some individuals cannot be reached this way. Certain projects do lend themselves
to e-mail surveys, such as internal surveys of employees or satisfaction surveys of
retail buyers who regularly deal with an organization via e -mail. The benef its of
incorporating a questionnaire in an e -mail include the speed of distribution, lower
distribution and processing costs, faster turnaround time, more flexibility, and less
handling of paper questionnaires. The speed of e -mail distribution and the quic k
response time can be major advantages for surveys dealing with time -sensitive
issues.
Internet Surveys:
An Internet survey is a self -administered questionnaire posted on a Web site.
Respondents provide answers to questions displayed onscreen by highlight ing a
phrase, clicking an icon, or keying in an answer. munotes.in
Page 134
134 RESEARCH IN COMPUTING
Kiosk Interactive Surveys
A computer with a touch screen may be installed in a kiosk at a trade show, at a
professional conference, in an airport, or in any other high -traffic location to
administer an interactive survey.
7.13 Pretesting
Pretesting involves a trial run with a group of respondents to iron out fundamental
problems in the instructions or design of a questionnaire. The researcher looks for
such obstacles as the point at which r espondent fatigue sets in and whether there
are any particular places in the questionnaire where respondents tend to terminate.
Unfortunately, this stage of research is sometimes eliminated because of costs or
time pressures.
Broadly speaking, three basic ways to pretest exist. The first two involve screening
the questionnaire with other research professionals, and the third —the one most
often called pretesting — is a trial run with a group of respondents. When screening
the questionnaire with other researc h professionals, the investigator asks them to
look for such problems as difficulties with question wording, leading questions,
and bias due to question order. An alternative type of screening might involve a
client or the research manager who ordered the research. Often, managers ask
researchers to collect information, but when they see the questionnaire, they find
that it does not really meet their needs. Only by checking with the individual who
has requested the questionnaire does the researcher know for sure that the information needed will be provided.
Once the researcher has decided on the final questionnaire, data should be collected
with a small number of respondents (perhaps 100) to determine whether the
questionnaire needs refinement.
Summary
1. Define surveys and explain their advantages. The survey is a common tool for
asking respondents questions. Surveys can provide quick, inexpensive, and
accurate information for a variety of objectives. The term sample survey is
often used because a survey i s expected to obtain a representative sample of
the target population. munotes.in
Page 135
135Chapter 7: Survey Research
2. Describe the type of information that may be gathered in a survey. The typical survey is a descriptive research study with the objective of measuring awareness, knowledge, behavior, opinions and attitudes, both inside and outside of the organization. Common survey populations including customers,
employees, suppliers and distributors.
3. Identify sources of error in survey research. Two major forms of error are
common in survey rese arch. The first, random sampling error, is caused by
chance variation and results in a sample that is not absolutely representative of
the target population. Such errors are inevitable, but they can be predicted using
the statistical methods discussed in l ater chapters on sampling. The second
major category of error, systematic error, takes several forms. Nonresponse
error is caused by subjects’ failing to respond to a survey. This type of error
can be identified by comparing the demographics of the sample population with
those of the target population and reduced by making a special effort to contact
underrepresented groups. In addition, response bias occurs when a response to a questionnaire is falsified or misrepresented, either intentionally or inadverte ntly. There are four specific categories of response bias: acquiescence
bias, extremity bias, interviewer bias, and social desirability bias. An additional source of survey error comes from administrative problems such as inconsistencies in interviewers’ a bilities, cheating, coding mistakes, and so
forth.
4. Distinguish among the various categories of surveys. Surveys may be classified
according to methods of communication, by the degrees of structure and
disguise in the questionnaires, and on a temporal b asis. Questionnaires may be
structured, with limited choices of responses, or unstructured, to allow open -
ended responses. Disguised questions camouflage the real purpose and may be
used to probe sensitive topics. Surveys may consider the population at a g iven
moment or follow trends over a period of time. The first approach, the cross -
sectional study, usually is intended to separate the population into meaningful subgroups. The second type of study, the longitudinal study, can reveal important population c hanges over time. Longitudinal studies may involve
contacting different sets of respondents or the same ones repeatedly. One form
of longitudinal study is the consumer panel. Consumer panels are expensive to
conduct, so firms often hire contractors who provide services to many companies, thus spreading costs over many clients
munotes.in
Page 136
136 RESEARCH IN COMPUTING
Questions
1. Name several nonbusiness applications of survey research.
2. What is self -selection bias? How might we avoid this?
3. Do surveys tend to gather qualitative or quantitative data? Wh at types of
information are commonly measured with surveys?
4. Give an example of each type of error
5. In a survey, chief executive officers (CEOs) indicated that they would prefer
to relocate their businesses to Atlanta (first choice), San Diego, Tampa, Los
Angeles, or Boston. The CEOs who said they planned on building new office
space in the following year were asked where they were going to build. They
indicated they were going to build in New York, Los Angeles, San Francisco,
or Chicago. Explain the differe nce between these two responses
6. What type of communication medium would you use to conduct the following surveys? Why?
a. Survey of the buying motives of industrial engineers
b. Survey of the satisfaction levels of hourly support staff
c. Survey of television commercial advertising awareness
d. Survey of top corporate executives
7. A publisher offers college professors one of four best -selling mass -market
books as an incentive for filling out a 10 -page mail questionnaire about a new
textbook. What advantag es and disadvantages does this incentive have?
8. “Individuals are less willing to cooperate with surveys today than they were
3.20 years ago.” Comment on this statement.
9. What do you think should be the maximum length of a self administered
e-mail questionnaire?
10. Do most surveys use a single communication mode (for example, the telephone), as most textbooks suggest?
11. A survey researcher reports that “20 3.2 usable questionnaires out of 942
questionnaires delivered in our mail survey converts to a 21.7 percent
response rate.” What are the subtle implications of this statement? munotes.in
Page 137
137Chapter 7: Survey Research
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN: 97881313 16054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016, Business Research Methods, Edition 8, Cengage Publication
munotes.in
Page 138
138 RESEARCH IN COMPUTING
138UNIT 4
8 MEASUREMENT CONCEPTS, SAMPLING
AND FIELD WORK
A. LEVELS OF SCALE MEASUREMENT
Unit Structure
8A.0 Objectives
8A.1 Introduction
8A.2 Levels of Measurement
8A.1.1 Nominal Scale
8A.1.2 Ordinal Scale
8A.1.3 Interval Scale
8A.1.4 Ratio
8A.3 Analysis of Scales
8A.3.1 Discrete Measures
8A.3.2 Continuous Measures
8A.4 Index measures
8A.4.1 Computing Scale Values
8A.5 Criteria for Good Measurement
8A.5.1 Reliability
8A.5.1.1 Types of Reliability
8A.5.2 Validity
8A.5.2.1 Types of Validity
8A.5.2. 2 Reliability vs Validity
8A.5.3 Sensitivity
8A.6 Summary
8A.7 Questions
8A.8 References munotes.in
Page 139
139Chapter 8: Measurement Concepts, Sampling and Field Work
8A.0 Objectives
Following are the objectives of this unit:
9 To understand the levels of measurement
9 To differentiate between different scales of measurement
9 To analyse scales
9 To understand index measures
9 To implement reliability
9 To analyse validity
9 To differentiate between reliability and validity
9 To understand sensitivity
8A.1 Introduction
Measurement is a procedure of allocating numerical value to some characteristics
or variables or events according to scientific rules. It is the process observing and
recording the observations which are collected as part of a research effort. Measurement means the description of data in terms of numbers – accuracy;
objectivity and communication. The combined form of these three is the actual
measurement.
In this unit, we will understand different levels of measurement and see their types.
Definitio n:
‘Measurement is the process of observing and recording the observations that are
collected as part of a research effort. ’
‘Measurement is a process of describing some property of a phenomenon of
interest, usually by assigning numbers in a reliable and valid way. ’
The decision statement, correspondi ng research questions, and research hypotheses
can be used to decide what concepts need to be measured in a given project.
Measurement is the process of describing some property of a phenomenon of
interest, usually by assigning numbers in a reliable and va lid way. The numbers
convey information about the property being measured. When numbers are used,
the researcher must have a rule for assigning a number to an observation in a way
that provides an accurate description. Measurement can be illustrated by thi nking
about the way instructors assign students’ grades. munotes.in
Page 140
140 RESEARCH IN COMPUTING
Example :
A – (above 60% score)
B – (between 50 – 60% score)
C – (between 40 – 50 % score
Here A, B, C can also be termed as scales of measurement. Some scales may better
classify the data and e ach scale has the potential of producing error or some lack of
validity
8A.2 Levels of Measurement
Level of measurement refers to the relationship among the values that are assigned
to the attributes for a variable. It is important because –
9 Knowing the level of measurement helps you decide how to interpret the data
from that variable
9 Knowing that a measure is nominal, then you know that the numerical values
are just short codes for the longer names.
9 Knowing the level of measurement helps you decide what statistical analysis
is appropriate on the values that were assigned.
If a measure is nominal, then you know that you would never average the data
values or do a t -test on the data. There are four distinguish levels of measurement.
The levels are –
9 Nominal
9 Ordinal
9 Interval
9 Ratio
Levels of measurement are important for two reasons.
i) First, they emphasize the generality of the concept of measurement. Although
people do not normally think of categorizing or ranking individuals as
measurement, in fact they are as long as they are done so that they represent
some characteristic of the ind ividuals.
ii) Second, the levels of measurement can serve as a rough guide to the statistical
procedures that can be used with the data and the conclusions that can be
drawn from them. munotes.in
Page 141
141Chapter 8: Measurement Concepts, Sampling and Field Work
8A.2.1 Nominal Scale
The nominal scale (called as dummy coding) simply places people, events, perceptions, etc. into categories based on some common trait. Some data are
naturally suited to the nominal scale such as males vs. females, white vs. black vs.
blue, and American vs. Asian. The nominal scale fo rms the basis for such analyses
as Analysis of Variance (ANOVA) because those analyses require that some
category is compared to at least one other category.
The nominal scale is the lowest form of measurement because it doesn’t capture
information about the focal object other than whether the object belongs or doesn’t
belong to a category; either you are a smoker or not a smoker, you attended
university or you didn’t, a subject has some experience with computers, an average
amount of experience with compu ters, or extensive experience with computers.
No data is captured that can place the measured object on any kind of scale say, for
example, on a continuum from one to ten. Coding of nominal scale data can be
accomplished using numbers, letters, labels, or any symbol that represents a
category into which an object can either belong or not belong. In research activities
a Yes/No scale is nominal. It has no order and there is no distance between Yes and
No.
The statistics which can be used with nominal scales are in the non -parametric
group. The most likely ones would be - mode; crosstabulation - with chi -square.
There are also highly sophisticated modelling techniques available for nominal
data.
8A.2.2 Ordinal Scale
An ordinal level of measurement uses symbo ls to classify observations into
categories that are not only mutually exclusive and exhaustive; in addition, the
categories have some explicit relationship among them. For example, observations
may be classified into categories such as taller and shorter, greater and lesser, faster
and slower, harder and easier, and so forth.
However, each observation must still fall into one of the categories (the categories
are exhaustive) but no more than one (the categories are mutually exclusive). Most
of the commonl y used questions which ask about job satisfaction use the ordinal
level of measurement.
For example , asking whether one is very satisfied, satisfied, neutral, dissatisfied,
or very dissatisfied with one’s job is using an ordinal scale of measurement. The
simplest ordinal scale is a ranking. munotes.in
Page 142
142 RESEARCH IN COMPUTING
Ordinal data would use non -parametric statistics. These would include - median
and mode; rank order correlation; non -parametric analysis of variance. Modelling
techniques can also be used with ordinal data.
8A.2.3 Interval Scale
An interval level of measurement classifies observations into categories that are not
only mutually exclusive and exhaustive, and have some explicit relationship among
them, but the relationship between the categories is known and exact. This is the
first quantitative application of numbers. In the interval level, a common and
constant unit o f measurement has been established between the categories.
For example , the commonly used measures of temperature are interval level scales.
We know that a temperature of 75 degrees is one degree warmer than a temperature
of 74 degrees .
Numbers may be as signed to the observations because the relationship between the
categories is assumed to be the same as the relationship between numbers in the
number system.
For example , 74+1= 75 and 41+1= 42. The intervals between categories are equal,
but they origina te from some arbitrary origin, that is, there is no meaningful zero
point on an interval scale. The standard survey rating scale is an interval scale.
When you are asked to rate your satisfaction with a piece of software on a 7 point
scale, from Dissatisfied to Satisfied, you are using an interval scale. Interval scale
data would use parametric statistical techniques - Mean and standard deviation;
Correl ation; Regression; Analysis of variance; Factor analysis; and whole range of
advanced multivariate and modelling techniques.
8A.2.4 Ratio
The ratio level of measurement is the same as the interval level, with the addition
of a meaningful zero point. There is a meaningful and non -arbitrary zero point from
which the equal intervals between categories originate.
For example , weight, area, spee d, and velocity are measured on a ratio level scale. In public policy and administration, budgets and the number of program participants are measured on ratio scales. In many cases, interval and ratio scales
are treated alike in terms of the statistical t ests that are applied. A ratio scale is the
top level of measurement and is not often available in social research. The factor
which clearly defines a ratio scale is that it has a true zero point.
The simple way to understand the levels of measurement or to select a measurement
scale is as follows – munotes.in
Page 143
143Chapter 8: Measurement Concepts, Sampling and Field Work
9 If one object is different from another, then we use a nominal scale.
9 If one object is bigger or better or more of anything than another, then we use
an ordinal scale.
9 If one object is so many units (degrees , inches, etc.) more than another, then
we use an interval scale.
9 If one object is certain times as big or bright or tall or heavy as another, then
we use a ratio scale.
The following criteria should be considered in the selection of the measurement
scale for variables in a study. Researcher should consider the scale that will be most suitable for each variable under study. Important points in the selection of measurement scale for a variable are :
9 Scale selected should be appropriate for the variables one wishes to categorise.
9 It should be of practical use.
9 It should be clearly defined.
9 The number of categories created (when necessary) should cover all possible
values.
9 The number of categories created (when necessary) should not overlap, i.e.,
it shoul d be mutually exclusive.
9 The scale should be sufficiently powerful. Variables measured at a higher
level can always be converted to a lower level, but not vice versa.
For example , observations of actual age (ratio scale) can be converted to categories
of older and younger (ordinal scale), but age measured as simply older or younger
cannot be converted to measures of actual age.
The four levels of measurement discussed above have an important impact on how
you collect data and how you analyze them later. Collect at the wrong level, and
you will end of having to adjust your research, your design, and your analyzes.
Make sure you consider carefully the level at which you collect your data,
especially in light of what statistical procedures you intend to use once you have
the data in hand.
8A.3 Analysis of Scales
Mathematical operations can be performed with numbers from nominal scales, the
result may not have a great deal of meaning . Although you can put numbers into munotes.in
Page 144
144 RESEARCH IN COMPUTINGformulas and perform calculations with almost any numbers, the researcher has to know the meaning behind the numbers before meaningful conclusions can be
drawn .
8A.3.1 Discrete Measures
Discrete measures are those that take on only one of a finite number of values. A
discrete scale is most often used to represent a classification variable. Therefore,
discrete scales do not represent intensity of measures, only membership. Common
discrete scales include any yes -or-no response, matching, colour choices, or
practically any scale that involves selecting from among a small number of
categories. Thus, when someone is asked to choose from the following responses :
9 Disagree
9 Neutral
9 Agree
the result is a discrete value that can be coded 1, 2, or 3, respectively. This is also an ordinal scale to the extent that it represents an ordered arrangement of agreement. Nominal and ordinal scales are discrete measures.
8A.3.2 Continuous Measures
Continuous measures are those assignin g values anywhere along some scale range
in a place that corresponds to the intensity of some concept. Ratio measures are
continuous measures. Thus, when we measure sales for each salesperson using the
money (every rupee) amount sold, he is assigning a con tinuous measure. A number
line could be constructed ranging from the least amount sold to the most, and a spot
on the line would correspond exactly to a salesperson’s performance.
Table 8A.1: Example of Continuous scales Question/ Rating Strongly Agree Agree Neutral Disagree Strongly Disagree I learned a lot from this study material. 5 4 3 2 1
This is a discrete scale because only the values 1, 2, 3, 4, or 5 can be assigned.
Moreover , it is an ordinal scale because it only orders based on agreement. We
really have no way of knowing that the difference in agreement of somebody
marking a 5 instead of a 4 is the same as the difference in agreement of somebody
marking a 2 instead of a 1. munotes.in
Page 145
145Chapter 8: Measurement Concepts, Sampling and Field Work
(For calculation purpose: Only the mean is not an appropriate way of stating central
tendency and, technically, we really shouldn’t use many common statistics on th ese
responses. )
A scaled response of this type (refer table 1.1) takes on more values, the error
introduced by assuming that the differences between the discrete points are equal
becomes smaller. This may be seen by imagining a Likert scale (the traditiona l business research agreement scale shown above) with a thousand levels of agreement rather than three.
The differences between the different levels become so small with a thousand levels
that only tiny errors could be introduced by assuming each interval is the same.
Therefore, business researchers generally treat interval scales containing five or
more categories of response as interval. (They are commonly called 5 -point Likert
scale; 7 -point Likert scale and so on)
When fewer than five categories are us ed, this assumption is inappropriate. The
researcher should keep in mind, however, the distinction between ratio and interval
measures. Errors in judgment can be made when interval measures are treated as
ratio.
8A.4 Index Measures
Multi -item instruments for measuring a construct are called index measures, or
composite measures. An index measure assigns a value based on how much of the
concept being measured is associated with an observation. Indexes often are formed
by putting sever al variables together.
For example , a social class index might be based on three weighted variables:
occupation, education, and area of residence. Usually, occupation is seen as the
single best indicator and would be weighted highest. With an index, the d ifferent
attributes may not be strongly correlated with each other.
A person’s education does not always relate strongly to their area of residence. The
Consumer Satisfaction Index shows how satisfied consumers are based on an index
of satisfaction scores . Readers are likely not surprised to know that certain
consumers appear more satisfied with soft drinks than they are with cable TV
companies based on this index.
Composite measures also assign a value based on a mathematical derivation of
multiple varia bles. munotes.in
Page 146
146 RESEARCH IN COMPUTING
For example , salesperson satisfaction may be measured by combining questions
such as “How satisfied are you with your job? How satisfied are you with your
territory? How satisfied are you with the opportunity your job offers?” For most
practical app lications, composite measures and indexes are computed in the same
way.
Definition s:
Index Measure : An index assigns a value based on how much of the concept being
measured is associated with an observation. Indexes often are formed by putting
several variables together.
Attribute : A single characteristic or fundamental feature of an object, person,
situation, or issue.
Composite Measures : Assign a value to an observation based on a mathematical
derivation of multiple variables.
8A.4.1 Computing Scale Values
The below stated example is a computation of the data collected using Likert Scale.
For this scale, the value of Strongly Agree (SA) is 5, Agree (A) is 4, Neutral (N) is
3, Disagree (D) is 2 and Strongly Disagree (SD) is 1. For the total scor e obtained
for these segments of questions is 5 + 2 + 3 + 4 = 14
Figure 1.1: Sample of Likert Scale
Such scales are also called as Summated scales.
munotes.in
Page 147
147Chapter 8: Measurement Concepts, Sampling and Field Work
Definition :
Summated Scale: A scale created by simply summing (adding together) the response to each item making up the composite measure.
Sometimes, a response may need to be reverse -coded before computing a summated or averaged scale value. Reverse coding means that the value assigned
for a response is treated oppositely from the other items. Thus, on a 5-point scale,
the values are reversed as follows:
• 5 becomes 1
• 4 becomes 2
• 3 stays 3
• 2 becomes 4
• 1 becomes 5
This happens for questions which are negative in nature. An ideal scale must have
60-70 % questions positive in nature and 30 -40 % questions negative in nature.
This is done to ensure that the person filling the questions is not selecting options
randomly.
Example of a negative question: I would not like to decide when I want to study
(based on the questions given in figure 1.1)
8A.5 Criteria For Good Measurement
8A.5.1 Reliability
Reliability refers to the consistency or repeatability of an operationalized measure.
A reliable measure will yield the same results over and over again when applied to
the same thing. It is the degree to which a tes t consistently measures whatever it
measures. If you have a survey question that can be interpreted several different
ways, it is going to be unreliable. One person may interpret it one way and another
may interpret it another way. You do not know which in terpretation people are
taking.
Even answers to questions that are clear may be unreliable, depending on how they
are interpreted. Reliability refers to the consistency of scores obtained by the same
persons when they are re -examined with the same tests o n different occasions, or with different sets of equivalent items, or under other variable examining conditions. munotes.in
Page 148
148 RESEARCH IN COMPUTING
Research requires dependable measurement. Measurements are reliable to the
extent that they are repeatable and that any random influence which tends to make
measurements different from occasion to occasion or circumstance to circumstance
is a source of mea surement error. Errors of measurement that affect reliability are
random errors and errors of measurement that affect validity are systematic or
constant errors. Reliability of any research is the degree to which it gives an
accurate score across a range o f measurement. It can thus be viewed as being
‘repeatability’ or ‘consistency’.
Internal consistency : Different questions, same construct. Test -retest, equivalent
forms and split -half reliability are all determined through correlation. There are a
number of ways of determining the reliability of an instrument. The procedure can
be classified into two groups –
External Consistency Procedures : It compare findings from two independent
processes of data collection with each other as a means of verifying the r eliability
of the measure. For example , test -retest reliability, parallel forms of the same test,
etc.
Internal Consistency Procedures: The idea behind this procedure is that items
measuring the same phenomenon should produce similar results. For example,
split-half technique.
8A.5.1.1 Types of Reliability
1) Test-Retest Reliability
The most obvious method for finding the reliability of test scores is by
repeating the identical test on a second occasion. Test -retest reliability is a
measure of reliability obtained by administering the same test twice over a
period of time to a group of individuals.
For Example - A test desig ned to assess student learning in psychology could
be given to a group of students twice, with the second administration perhaps
coming a week after the first. The obtained correlation coefficient would
indicate the stability of the scores.
2) Split -Half Reliability
Split -half reliability is a subtype of internal consistency reliability. In split
half reliability we randomly divide all items that purport to measure the same
construct into two sets. We administer the entire instrument to a sample of
people and calculate the total score for each randomly divided half. The most
commonly used method to split the test into two is using the odd -even
strategy. munotes.in
Page 149
149Chapter 8: Measurement Concepts, Sampling and Field Work
3) Inter -Rater Reliability
Inter -rater reliability is a measure of reliability used to assess the degree to
which different judges or raters agree in their assessment decisions. Inter -
rater reliability is also known as inter -observer reliability or inter -coder
reliability. Inter -rater reliability is useful because human observers will not
necessarily interpret answers the same way; raters may disagree as to how
well certain responses or material demonstrate knowledge of the construct or
skill being assessed. Inter -rater reliability m ight be employed when different judges are evaluating the degree to which art portfolios meet certain standards. Inter -rater reliability is especially useful when judgments can be
considered relatively subjective.
4) Parallel -Forms Reliability
Parallel forms reliability is a measure of reliability obtained by administering
different versions of an assessment tool to the same group of individuals. The
scores from the two versions can then be correlated in order to evaluate the
consistency of resu lts across alternate versions.
5) &RHIILFLHQWDOSKDĮ ):
It is the most commonly applied estimate of a multiple -item scale’s UHOLDELOLW\&RHIILFLHQWĮUHSUHVHQWVLQWHUQDOFRQVLVWHQF\E\FRPSXWLQJWKH
average of all possible split -half reliabilities for a multiple -item scale. The coefficient demonstrates whether or not the different items converge. $OWKRXJKFRHIILFLHQWĮGRHVQRWDGGUHVVYDOLGLW\PDQ\UHVHDUFKHUVXVHĮDV
WKHVROHLQGLFDWRURIDVFDOH¶VTXDOLW\&RHIILFLHQWDOSKDUDQJHVLQYDOXHIURP
0, meaning no consistency, to 1, meaning complete consistency
8A.5.2 Validity
Validity refers to whether the measure actually measures what it is supposed to
measure. If a measure is unreliable, it is also invalid. That is, if you do not know
what it is measu ring, it certainly cannot be said to be measuring what it is supposed
to be measuring. On the other hand, you can have a consistently unreliable measure.
For example , if we measure income level by asking someone how many years of
formal education they have completed, we will get consistent results, but education
is not income (although they are positively related).
In general, validity is an indication of how sound your research is. More specifically, validity applies to both the design and the methods of your research. munotes.in
Page 150
150 RESEARCH IN COMPUTING
Validity in data collection means that your findings truly represent the phenomenon
you are claiming to measure. Valid claims are solid claims.
There are two main types of validity, internal and external. Internal validity refers
to the validity of the measurement and test itself, whereas external validity refers
to the ability to generalize the findings to the target population.
8A.5.2.1 Types of Valid ity
1) Face Validity
Face validity refers to the degree to which a test appears to measure what it purports to measure. The stakeholders can easily assess face validity. Although this is not a very ‘scientific’ type of validity, it may be an essential
component in enlisting mo tivation of stakeholders. If the stakeholders do not
believe the measure is an accurate assessment of the ability, they may become
disengaged with the task.
For example, if a measure of art appreciation is created all of the items should
be related to th e different components and types of art. If the questions are
regarding historical time periods, with no reference to any artistic movement,
stakeholders may not be motivated to give their best effort or invest in this
measure because they do not believe i t is a true assessment of art appreciation.
2) Predictive Validit y
Predictive validity refers to whether a new measure of something has the
same predictive relationship with something else that the old measure had. In predictive validity, we assess the operationalization’s ability to predict something it should theoretica lly be able to predict.
For example , we might theorize that a measure of math ability should be able
to predict how well a person will do in an engineering -based profession. We
could give our measure to experienced engineers and see if there is a high
correlation between scores on the measure and their salaries as engineers. A
high correlation would provide evidence for predictive validity - it would
show that our measure can correctly predict something that we theoretically
think it should be able to pre dict.
3) Criterion -Related Validity
Criterion validity is a test of a measure when the measure has several different
parts or indicators in it - compound measures. Each part or criterion of the munotes.in
Page 151
151Chapter 8: Measurement Concepts, Sampling and Field Work
measure should have a relationship with all the parts in the measure for the
variable to which the first measure is related in a hypothesis. When you are
expecting a future performance based on the scores obtained currently by the measure, correlate the scores obtained with the performance. The later performance is called the criterion and the current score is the prediction. It
is used to predict future or current performance - it correlates test results with
another criterion of interest.
For example , if a physics program designed a measure to assess cumulative
student learning throughout the major. The new measure could be correlated
with a standardized measure of ability in this discipline, such as GRE subject
test. The higher the correlation between the established measure and new
measure, the more faith stakeholders can have in the new assessment tool.
4) Content Validity
In content validity, you essentially check the operationalization against the
relevant content domain for the construct. This approach assumes that you
have a good detailed description of the content domain, something that’s n ot
always true. In content validity, the criteria are the construct definition itself
- it is a direct comparison. In criterion -related validity, we usually make a
prediction about how the operationalization will perform based on our theory
of the construc t. When we want to find out if the entire content of the
behavior/ construct/ area is represented in the test we compare the test task
with the content of the behavior. This is a logical method, not an empirical
one.
For Example , if we want to test knowledge on Bangladesh Geography it is
not fair to have most questions limited to the geography of Australia.
5) Convergent Validity
Convergent validity refers to whether two different measures of presumably
the same thing are cons istent with each other - whether they converge to give
the same measurement. In convergent validity, we examine the degree to which the operationalization is similar to (converges on) other operationalizations that it theoretically should be similar to.
For example , to show the convergent validity of a test of arithmetic skills, we
might correlate the scores on test with scores on other tests that purport to
measure basic math ability, where high correlations would be evidence of
convergent validity. Or, if SAT scores and GRE scores are convergent, then munotes.in
Page 152
152 RESEARCH IN COMPUTING
someone who scores high on one test should also score high on the other.
Different measures of ideology should classify the same people the same
way. If they do not, then they lack convergent validity.
6) Concurrent Validity :
Concurrent validity is the degree to which the scores on a test are related to
the scores on another already established, test administered at the same time
or to some other valid criterion available at the same time. This compares the
results from a new measurement technique to those of a more established
technique that claims to measure the same variable to see if they are related. In concurrent validity, we assess the operationalization’s ability to distinguish between groups that it should theoretically be able to distinguish
between.
For example , if we come up with a way of assessing manic -depression, our
measure should be able to distinguish between people who are diagnosed
manic -depression and those diagnosed paranoid schizophrenic. If we want to
assess the concurrent validity of a new measure of empowerment, we might
give the measure to both migrant farm workers and to the farm owners,
theorizing that our measure should show that the farm owners are higher in
empowerment. As in any discriminating test, the results are more powerful if
you are a ble to show that you can discriminate between two groups that are
very similar.
7) Construct Validity
Construct validity is used to ensure that the measure is actually measure what
it is intended to measure (i.e. the construct), and not other variables. Using a
panel of ‘experts’ familiar with the construct is a way in which this type of
validity can be assessed. The experts can examine the items and decide what
that specific item is intended to measure. This is whether the measurements
of a variable in a study behave in exactly the same way as the variable itself.
This involves examining past research regarding different aspects of the same variable. It is also the degree to which a test measures an intended hypothetical construct.
For example , if we w ant to validate a measure of anxiety. We have a
hypothesis that anxiety increases when subjects are under the threat of an
electric shock, then the threat of an electric shock should increase anxiety
scores. munotes.in
Page 153
153Chapter 8: Measurement Concepts, Sampling and Field Work
8) Formative Validity
When applied to outcomes assessment it is used to assess how well a measure
is able to provide information to help improve the program under study.
For example - when designing a rubric for history one could assess student’s
knowledge across the discipline. If the measure can p rovide information that
students are lacking knowledge in a certain area, for instance the Civil Rights
Movement, then that assessment tool is providing meaningful information
that can be used to improve the course or program requirements.
9) Sampling Validity
Sampling validity ensures that the measure covers the broad range of areas
within the concept under study. Not everything can be covered, so items need
to be sampled from all of the domains. This may need to be completed using
a panel of ‘experts’ to ensure that the content area is adequately sampled.
Additionally, a panel can help limit ‘expert’ bias .
For example - when designing an assessment of learning in the theatre
department, it would not be sufficient to only cover issues related to acting.
Other areas of theatre such as lighting, sound, functions of stage managers
should all be included. The assessment should reflect the content area in its
entirety.
10) Discriminant Validity
In discriminant validity, we examine the degree to which the operationalization is not similar to (diverges from) other operationalizations
that it theoretically should be not be similar to.
For example , to show the discriminant validity of a Head Start program, we
might gather evidence that shows that the program is not similar to other early
childhood programs that don’t label themselves as Head Start programs.
8A.5.2. 2 Reliability Versus Validity
Reliability is a necessary but not sufficient condition for validity. A reliable scale
may not be valid. For example, a purchase intention measurement technique may
consistently indicate that 20 percent of those sampled are wil ling to purchase a new
product. Whether the measure is valid depends on whether 20 percent of the
population indeed purchases the product. A reliable but invalid instrument will
yield consistently inaccurate results. munotes.in
Page 154
154 RESEARCH IN COMPUTING
8A.5.3 Sensitivity
The sensitivity of a scale is an important measurement concept, particularly when
changes in attitudes or other hypothetical constructs are under investigation.
Sensitivity refers to an instrument’s ability to accurately measure variability in a
concept. A dichotomous respons e category, such as “agree or disagree,” does not
allow the recording of subtle attitude changes. A more sensitive measure with
numerous categories on the scale may be needed.
For example , adding “strongly agree,” “mildly agree,” “neither agree nor disagree,”
“mildly disagree,” and “strongly disagree” will increase the scale’s sensitivity. The
sensitivity of a scale based on a single question or single item can also be increased
by adding questions or items. In other words, because composite measures allow
for a greater range of possible scores, they are more sensitive than single -item
scales. Thus, sensitivity is generally increased by adding more response points or
adding scale items.
Summary
In this unit, we understood what is meant by Measurement when considered in
research. Some key definitions we saw were: Measurement: Measurement is the process of observing and recording the observations that are collected as part of a research effort.
Index Measure: An index assigns a value based on how much o f the concept being
measured is associated with an observation. Indexes often are formed by putting
several variables together.
Attribute: A single characteristic or fundamental feature of an object, person,
situation, or issue.
Composite Measures: Assign a value to an observation based on a mathematical
derivation of multiple variables.
Summated Scale: A scale created by simply summing (adding together) the
response to each item making up the composite measure.
We also understood the important concepts like reliability and validity and understood the difference between them.
munotes.in
Page 155
155Chapter 8: Measurement Concepts, Sampling and Field Work
Questions
1. Discuss different levels of measurement.
2. Using appropriate examples, write notes on:
a. Nominal Scale
b. Ordinal Scale
c. Interval Scale
3. What is meant by analysis of scales?
4. Elaborate on the criteria for good measurement.
5. Write a note on sensitivity as a criteria for measurement.
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation,
ISBN : 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin, 2016,
Business Research Methods, Edition 8, Cengage Publications
munotes.in
Page 156
156 RESEARCH IN COMPUTING
B. ATTITUDE MEASUREMENT & QUESTIONNAIRE DESIGN
Unit Structure
8B.1 Objectives
8B.2 Introduction
8B.3 Attitude Rating Scale
8B.3.1 Category Scale
8B.3.2 Likert Scale
8B.3.3 Composite Scaling
8B.3.4 Semantic Differential
8B.3.5 Numeric Scaling
8B.3.6 Staple Scaling
8B.3.7 Constant Sum Scale
8B.3.8 Graphic Rating Scale
8B.3.9 Thurstone Interval Scale
8B.4 Questionnaire Design
8B.5 Summary
8B.6 Questions
8B.7 References
8B.1 Objectives
Following are the objectives of this unit:
9 To understand the attitude rating scale
9 To categorise different attitude rating scales
9 To understand the development of questionnaire procedure
8B.2 Introduction
Attitude can be defined as a tendency to react favourably, neutrally, or unfavourably toward a particular class of stimuli, such as a custom, institutional
practice, or national group. There are two challenges a researcher faces when
measuring an individual ’s attitude. First, an individual ’s attitude toward an object munotes.in
Page 157
157Chapter 8: Measurement Concepts, Sampling and Field Work
cannot be observed directly but must be inferred from observed behaviour , such as
responses to a questionnaire. And second, there is no inherent scale associated with
the observed behaviour . Techniques for measuring attitude are:
1. Ranking : A measurement task that requires respondents to rank order a small
number of stores, brands, or objects on the basis of overall preference or some
characteristic of the stimulus .
2. Rating : A measurement task that requires respondents to estimate the
magnitude of a characteristic or quality that a brand, store, or object possesses .
3. Sorting : A measurement task that presents a respondent with several objects
or product concepts and requires the respondent to arrange the objects into
piles or classify the product concepts.
Along with attitude, another critical part of the survey is the creation of questions
that must be framed in such a way that it results in obtaining the desired information
from the respondents. There are no scientific principles that assure an ideal
questionna ire and in fact, the questionnaire design is the skill which is learned
through experience. In this unit, we will see Attitude rating scales and Questionnaire designing
8B.3 Attitude Rating Scales
Simple attitude scaling may be used when questionnaires are extremely long, when
respondents have little education, or for other specific reasons. A number of
simplified scales are merely checklists: A respondent indicates past experience,
preference, and the like merely by checking an item.
In many cases the item s are adjectives that describe a particular object. In a survey
of small -business owners and managers, respondents indicated whether they found
working in a small firm more rewarding than working in a large firm, as well as
whether they agreed with a serie s of attitude statements about small businesses. For
example, 77 percent said small and mid -sized businesses “have less bureaucracy,”
and 76 percent said smaller companies “have more flexibility” than large ones.
8B.3.1 Category Scale
A rating scale that consists of several response categories, often providing respondents with alternatives to indicate positions on a continuum.
munotes.in
Page 158
158 RESEARCH IN COMPUTING
Table 8B.1: Sample of category scale How frequently do you use online payment modes?
Never Rarely Sometimes Often Very Often
The simplest rating scale contains only two response categories: agree/disagree.
Expanding the response categories provides the respondent with more flexibility in
the rating task. Even more information is provided if the categories are ordered
according t o a particular descriptive or evaluative dimension.
This category scale is a more sensitive measure than a scale that has only two
response categories. By having more choices for a respondent, the potential exists
to provide more information. However, if the researcher tries to represent something that is truly bipolar (yes/no, female/male, member/non -member, and so
on) with more than two categories, error may be introduced.
8B.3.2 Likert Scale
A Likert scale is a psychometric scale commonly used in questi onnaires, and is the most widely used scale in survey research. When responding to a Likert questionnaire item, respondents specify their level of agreement to a statement. The
scale is named after its inventor, psychologist Rensis Likert. The Likert scale can
also be used to measure attitudes of people. When responding to a Likert
questionnaire item, respondents specify their level of agreement or disagreement
on a symmetric agree -disagree scale for a series of statements. Thus, the range
captures the inte nsity of their feelings for a given item.
Definition :
A measure of attitudes designed to allow respondents to rate how strongly they
agree or disagree with carefully constructed statements, ranging from very positive
to very negative attitudes toward some object. munotes.in
Page 159
159Chapter 8: Measurement Concepts, Sampling and Field Work
Figure 8B.1: Sample of 5 -point Likert Scale Reverse Recording The statement given in this example (figure 8B.1) is positively framed. If a statement is framed negatively (such as “I dislike online learning as it provides richer instructional content”), the numerical scores would need to be reversed. This is done by reverse recoding the negative item so that a strong agreement really indicates an unfavourable response rather than a favourable attitude. In the case of a five-point scale, the recoding is done as follows:
Table 8B.2: Reverse Scoring of Likert Scale Old Value New Value 1 2
3
4
5 5 4
3
2
1
munotes.in
Page 160
160 RESEARCH IN COMPUTING
8B.3.3 Composite Scaling
A Likert scale may include several scale items to form a composite scale. Each
statement is assumed to represent an aspect of a common attitudinal domain. The
total score is the summation of the numerical scores assigned to an individual’ s
responses. Based on the example given in figure 1.1, twe the maximum possible
score for the composite would be 20 if a 5 were assigned to “strongly agree”
responses for each of the positively worded statements and a 5 to “strongly
disagree” responses for the negative statement. Item 3 is negatively worded and
therefore it is reverse coded, prior to being used to create the composite scale.
Definition :
Composite Scaling: A way of representing a latent construct by summing or
averaging respondents’ reaction s to multiple items each assumed to indicate the
latent construct.
8B.3.4 Semantic Differential
The semantic differential is actually a series of attitude scales. This popular attitude
measurement technique consists of getting respondents to react to some concept
using a series of seven -point bipolar rating scales. Bipolar adjectives —such as
“good” and “bad,” “modern” and “old fashioned,” or “clean” and “dirty” —anchor
the beginning and the end (or poles) of the scale. The subject makes repeated
judgmen ts about the concept under investigation on each of the scales.
Definition
Semantic Scaling: It’s a measure of attitudes that consists of a series of seven point
rating scales that use bipolar adjectives to anchor the beginning and end of each
scale .
The scoring of the semantic differential can be illustrated using the scale bounded
by the anchors “modern” and “old -fashioned.” Respondents are instructed to check
the place that indicates the nearest appropriate adjective. From left to right, the
scale intervals are interpreted as “extremely modern,” “very modern,” “slightly
modern,” “both modern and old -fashioned,” “slightly old -fashioned,” “very old -
fashioned,” and “extremely old -fashioned”:
Modern -------------------------------------------------- Old-fashioned
8B.3.5 Numeric Scaling
A numerical scale simply provides numbers rather than a semantic space or verbal
descriptions to identify response options or categories (response positions). For munotes.in
Page 161
161Chapter 8: Measurement Concepts, Sampling and Field Work
example, a scale using five response positions is called a five -point numerical scale.
A six -point scale has six positions and a seven -point scale seven positions, and so
on. Consider the following numerical scale:
Now that you’ve had your laptop for about one year, please tell us how satisfied
you are.
Extremely Dissatisfied 1 2 3 4 5 6 7 Extremely Satisfied
This numerical scale uses bipolar adjectives in the same manner as the semantic
differential. In practice, researchers have found that a scale with numerical labels
for intermediate points on the scale is as effective a measur e as the true semantic
differential.
Definition
Numeric Scaling: An attitude rating scale similar to a semantic differential except
that it uses numbers, instead of verbal descriptions, as response options to identify
response positions .
8B.3.6 Stapel Scale
The Stapel scale, is used to measure simultaneously the direction and intensity of
an attitude. Modern versions of the scale, with a single adjective, are used as a
substitute for the semantic differential when it is difficult to create pairs of bipolar
adjectives. The modified Stapel scale places a single adjective in the center of an
even number of numerical values (ranging, perhaps, from +3 to –3). The scale
measures how close to or distant from the adjective a given stimulus is perceived
to be.
The advantages and disadvantages of the Stapel scale are very similar to those of the semantic differential. However, the Stapel scale is markedly easier to administer, especially over the telephone. Because the Stapel scale does not require
bipolar adjectives, it is easier to construct than the semantic differential.
Definition
Staple Scale: A measure of attitudes that consists of a single adjective in the center
of an even number of numerical values.
8B.3.7 Constant Sum Scale
A constant sum scale is a type of question used in a market research survey in
which respondents are required to divide a specific number of points or percent as
part of a total sum. The allocation of points is divided to detail the variance and
weight of each category. A constant -sum scale requires respondents to divide a munotes.in
Page 162
162 RESEARCH IN COMPUTING
fixed number of points among several attributes corresponding to their relative
importance or weight.
Definition
Constant Sum Scale: A measure of attitudes in which respondents are asked to
divide a constant sum to indicate the relative importance of attributes; respondents
often sort cards, but the task may also be a rating task.
For Example :
Question: Using 100 points, please apply a number of points to each factor based
on how important each is to you when buying a home. You must total 100 points
divided among the factors.
Answer : Price, Location, School District, Inside Features, etc.
The respondent is given 100 points. They may choose to apply 80 to price, 15 to
location, and spread out the remaining 5 points among other factors. When you
analyze this data set, the differentiation between factors becomes evident. Most
survey software will automatical ly tally and sum the point values to ensure they
add to a constant sum of 100.
This constant sum scale adds another layer of analytical thinking for the respondent
rather than just selecting one, running through a checklist of choices, or selecting
from a grid or scaling question. It forces respondents to slow down and understand
the relative value of each factor and compare the importance of one over another.
It maximizes the chances of creating differentiation between your choices.
8B.3.8 Graphic Rating Scale
A graphic rating scale lists the traits each employee should have and rates workers
on a numbered scale for each trait. The scores are meant to separate employees into
tiers of performers, which can play a role in determining promotions and sa lary
adjustments. A graphic rating scale presents respondents with a graphic continuum.
The respondents are allowed to choose any point on the continuum to indicate their
attitude.
Definition
Graphic Rating Scale: A measure of attitude that allows responde nts to rate an
object by choosing any point along a graphic continuum.
munotes.in
Page 163
163Chapter 8: Measurement Concepts, Sampling and Field Work
For Example:
Table 8B.3: Example of Analyse s performance of employees working on Project
A from April to June 2020 Extremely Poor Bad Average Good Excellent Attention to detail Knowledge Teamwork Initiative Creative A variation of the graphic ratings scale is the ladder scale. This scale also includes
numerical options .
Example : This ladder scale represents the “ladder of life.” As you see, it is a ladder
with eleven rungs numbered 0 to 10. Let’s suppose the top of the ladder represents
the best possible life for you as you describe it, and the bottom rung represents the
worst possible life for you as you describe it.
On which rung of the ladder do you feel your life is today?
0 1 2 3 4 5 6 7 8 9 10
Figure 8B.2 Ladder Scale example
munotes.in
Page 164
164 RESEARCH IN COMPUTING
8B.3.9 THURSTONE INTERVAL SCALE
It is an attitude scale in which judges assign scale values to attitudinal statements
and subjects are asked to respond to these statements. This was developed because
the attitudes vary along continua and should be measured accordingly. A Thurstone
scale has a number of “agree” or “disagree” statements. It is a unidimensional
scale to measure attitudes towards people. The construction of a Thurstone scale
is a fairly complex process that requires two stages. The first stage is a ranking
operation, performed by judges who assign scale values to attitudinal statements. The second stage consists of asking subjects to respond to the attitudinal statements.
8B.4 Questionnaire Design
Questionnaire is a systematic, data collection t echnique consists of a series of
questions required to be answered by the respondents to identify their attitude,
experience, and behavior towards the subject of research.
The following steps are involved in the questionnaire design process:
1. Specify the Information Needed: The first and the foremost step in designing the questionnaire is to specify the information needed from the
respondents such that the objective of the survey is fulfilled. The researcher
must completely review the components of the pro blem, particularly the
hypothesis, research questions, and the information needed.
2. Define the Target Respondent: At the very outset, the researcher must
identify the target respondent from whom the information is to be collected.
The questions must be desi gned keeping in mind the type of respondents
under study. Such as, the questions that are appropriate for serviceman might
not be appropriate for a businessman. The less diversified respondent group
shall be selected because the more diversified the group is, the more difficult
it will be to design a single questionnaire that is appropriate for the entire
group.
3. Specify the type of Interviewing Method: The next step is to identify
the way in which the respondents are reached. In personal interviews, the
respondent is presented with a questionnaire and interacts face -to-face with
the interviewer. Thus, lengthy, complex and varied questions can be asked
using the personal interview method. In telephone interviews, the respondent
is required to give answers to the questions over the telephone. Here the munotes.in
Page 165
165Chapter 8: Measurement Concepts, Sampling and Field Work
respondent cannot see the questionnaire and hence this method restricts the
use of small, simple and precise questions.
The questionnaire can be sent through mail or post. It should be self -
explanatory and contain all the important information such that the respondent is able to understand every question and gives a complete response. The electronic questionnaires are sent directly to the mail ids of the
respondents and are required to give answers online.
4. Determi ne the Content of Individual Questions: Once the information
needed is specified and the interviewing methods are determined, the next
step is to decide the content of the question. The researcher must decide on
what should be included in the question such that it contributes to the
information needed or serve some specific purpose.
In some situations, the indirect questions which are not directly related to the
information needed may be asked. It is useful to ask neutral questions at the
beginning of a questionnaire with intent to establish respondent’s involvement and rapport. This is mainly done when the subject of a questionnaire is sensitive or controversial. The researcher must try to avoid
the use of double -barrelled questions. A question that talks about two issues
simultaneously, such as Is the Real juice tasty and a refreshing health drink?
5. Overcome Respondent’s Inability and Unwillingness to Answer: The
researcher should not presume that the respondent can provide accurate
responses to all the que stions. He must attempt to overcome the respondent’s
inability to answer. The questions must be designed in a simple and easy
language such that it is easily understood by each respondent. In situations,
where the respondent is not at all informed about th e topic of interest, then
the researcher may ask the filter questions, an initial question asked in the
questionnaire to identify the prospective respondents to ensure that they fulfil
the requirements of the sample.
Despite being able to answer the quest ion, the respondent is unwilling to devote time in providing information. The researcher must attempt to understand the reason behind such unwillingness and design the questionnaire in such a way that it helps in retaining the respondent’s
attention.
6. Decid e on the Question Structure: The researcher must decide on the
structure of questions to be included in the questionnaire. The question can munotes.in
Page 166
166 RESEARCH IN COMPUTING
be structured or unstructured. The unstructured questions are the open -ended
questions which are answered by the res pondents in their own words. These
questions are also called as a free-response or free-answer questions.
While, the structured questions are called as closed -ended questions that pre -
specify the response alternatives. These questions could be a multiple -choice
question, dichotomous (yes or no) or a scale.
7. Determine the Question Wording: The desired question content and structure must be translated into words which are easily understood by the
respondents. At this step, the researcher must translate the qu estions in easy
words such that the information received from the respondents is similar to
what was intended.
In case the question is written poorly, then the respondent might refuse to
answer it or might give a wrong answer. In case, the respondent is r eluctant
to give answers, then “nonresponse ” arises which increases the complexity
of data analysis. On the other hand, if the wrong information is given,
then “response error” arises due to which the result is biased.
8. Determine the Order of Questions: At this step, the researcher must decide
the sequence in which the questions are to be asked. The opening questions are crucial in establishing respondent’s involvement and rapport, and therefore, these questions must be interesting, non -threatening and easy.
Usually, the open -ended questions which ask respondents for their opinions
are considered as good opening questions, because people like to express
their opinions.
9. Identify the Form and Layout: The format, positioning and spacing of
questions has a signif icant effect on the results. The layout of a questionnaire
is specifically important for the self -administered questionnaires. The questionnaires must be divided into several parts, and each part shall be
numbered accurately to clearly define the branches of a question.
10. Reproduction of Questionnaire: Here, we talk about the appearance of the
questionnaire, i.e. the quality of paper on which the questionnaire is either
written or printed. In case, the questionnaire is reproduced on a poor -quality
paper; then the respondent might feel the research is unimportant due to
which the quality of response gets adversely affected.
Thus, it is recommended to reproduce the questionnaire on a good -quality
paper having a professional appearance. In case, the questionnaire has several munotes.in
Page 167
167Chapter 8: Measurement Concepts, Sampling and Field Work
pages, then it should be presented in the form of a booklet rather than the
sheets clipped or stapled to gether.
11. Pretesting: Pretesting means testing the questionnaires on a few selected
respondents or a small sample of actual respondents with a purpose of
improving the questionnaire by identifying and eliminating the potential
problems. All the aspects of th e questionnaire must be tested such as question
content, structure, wording, sequence, form and layout, instructions, and
question difficulty. The researcher must ensure that the respondents in the
pre-test should be similar to those who are to be finally surveyed.
Thus, the questionnaire design is a multistage process that requires the researcher’s
attention to many details.
Summary
In this unit, we saw the attitude of measurement and process of designing a
questionnaire. Attitude can be defined as a tendency to react favourably, neutrally,
or unfavourably toward a particular class of stimuli, such as a custom, institutional
practice, or national group. A questionnaire is a research instrument consisting of
a series of questions for the purpose of gath ering information from respondents.
Questionnaires can be thought of as a kind of written interview. ... Often a
questionnaire uses both open and closed questions to collect data.
Some of the key definitions in this section include:
Likert Scale: A measure of attitudes designed to allow respondents to rate how
strongly they agree or disagree with carefully constructed statements, ranging from
very positive to very negative attitudes toward some object.
Composite Scaling: A way of representing a latent const ruct by summing or
averaging respondents’ reactions to multiple items each assumed to indicate the
latent construct.
Numeric Scaling: An attitude rating scale similar to a semantic differential except
that it uses numbers, instead of verbal descriptions, a s response options to identify
response positions.
Graphic Rating Scale: A measure of attitude that allows respondents to rate an
object by choosing any point along a graphic continuum.
munotes.in
Page 168
168 RESEARCH IN COMPUTING
Questions
1 What is attitude rating scale? Discuss any two scales using appropriate
examples.
2 Discuss the structure of Likert scale.
3 What is understood by Numeric scaling?
4 Discuss in details the steps of Questionnaire Design.
5 Write a note on Graphic rating scale.
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research M ethods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation,
ISBN: 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin, 2016,
Business Research Methods, Edition 8, Cengage Publications
munotes.in
Page 169
169Chapter 8: Measurement Concepts, Sampling and Field Work
C. SAMPLING DESIGNS AND PROCEDURES
D. DETERMINATION OF SAMPLE SIZE
Unit Structure
8C.0 Objectives
8C.1 Introduction
8C.2 Sampling Terminologies
8C.3 Purpose of Sampling
8C.4 Stages of Sampling
8C.5 Techniques of Sampling
8C.5.1 Probability Sampling
8C.5.1.1 Simple Random Sampling
8C.5.1.2 Systematic Random Sampling
8C.5.1.3 Stratified Random Sampling
8C.5.1.4 Cluster / Multistage Sampling
8C.5.2 Non-Probability Sampling
8C.5.2.1 Convenience / Accidental Sampling
8C.5.2.2 Quota Sampling
8C.5.2.3 Judgement Sampling
8C.5.2.4 Snowball Sampling
8D Determination of Sample Size
Summary
Questions
References
8C.0 Objectives
Following are the objectives of this unit:
9 To get well versed with Sampling terminologies
9 To understand the purpose of sampling
9 To analyse the stages of sampling munotes.in
Page 170
170 RESEARCH IN COMPUTING
9 To understand the techniques of sampling
9 To differentiate between different types of sampling techniques
9 To calculate the sample size
8C.1 Introduction
Sampling is a familiar part of daily life. A customer in a bookstore picks up a book,
looks at the cover, and skims a few pages to get a sense of the writing style and
content before deciding whether to buy. A high school student visits a college
classroom to listen to a professor’s lecture. Selecting a university on the basis of
one classroom visit may not be scientific sampling, but in a personal situation, it may be a practical sampling experience. When measuring every item in a population is impossible, inconvenient, or too expensive, we intuitively take a
sample. Although sampling is commonplace in daily activities, these familiar
samples are seldom scientific. For researchers, the process of sampling can be quite
complex. Sampling is a central aspect o f business research, requiring in -depth
examination.
Formally defining s ampling : It is a process used in statistical analysis in which a
predetermined number of observations are taken from a larger population. The
methodology used to sample from a larger population depends on the type of
analysis being performed, but it may include simple random sampling or
systematic sampling.
To choose the right size of sample is the next section of this unit titled as
determination of sample size. Sample size determin ation is the act of choosing
the number of observations or replicates to include in a statistical sample.
The sample size is an important feature of any empirical study in which the goal is
to make inferences about a population from a sample.
This unit explains the nature of sampling and ways to determine the appropriate
sample design.
8C.2 Sampling Terminologies
9 Population : Total of items about which information is desired. It can be
classified into two categories - finite and infinite. The population is said to be
finite if it consists of a fixed number of elements so that it is possible to
enumerate in its totality. munotes.in
Page 171
171Chapter 8: Measurement Concepts, Sampling and Field Work
Examples of finite population are the populations of a city, the number of
workers in a factory, etc.
An infinite population is that population in which it is theoretically impossible to observe all the elements. In an infinite population the number
of items is infinite.
Example of infinite population is the number of stars in sky. From practical
consideration, we use the term infinite population for a population that cannot
be enumerated in a reasonable period of time.
9 Sample : It is part of the population that represents the ch aracteristics of the
population.
9 Population Sample Sampling: It is the process of selecting the sample for
estimating the population characteristics. In other words, it is the process of
obtaining information about an entire population by examining only a part of
it.
9 Sampling Unit: Elementary units or group of such units which besides being
clearly defined, identifiable and observable, are convenient for purpose of
sampling are called sampling units. For instance, in a family budget enquiry,
usually a fam ily is considered as the sampling unit since it is found to be
convenient for sampling and for ascertaining the required information. In a
crop survey, a farm or a group of farms owned or operated by a household
may be considered as the sampling unit.
9 Sam pling Frame: A list containing all sampling units is known as sampling
frame. Sampling frame consists of a list of items from which the sample is to
be drawn. Sample Survey: An investigation in which elaborate information
is collected on a sample basis is known as sample survey.
9 Statistic: Characteristics of the sample. For example, sample Mean, proportion, etc.
9 Parameter: Characteristics of the population. For example, population Mean, proportion, etc.
9 Target Population: A target population is the entir e group about which
information is desired and conclusion is made.
9 Sampled Population : The population, which we actually sample, is the
sampled population. It is also called survey population. munotes.in
Page 172
172 RESEARCH IN COMPUTING9 Sampling With and Without Replacement: Sampling schemes may be without replacement ('WOR' - no element can be selected more than once in
the same sample) or with replacement ('WR' - an element may appear
multiple times in the one sample). For example, if we catch fish, measure
them, and immediately return them to the water before continuing with the sample, this is a WR design, because we might end up catching and measuring the same fish more than once. However, if we do not return the
fish to the water (e.g. if we eat the fish), this becomes a WOR design.
9 Sample Desig n: Sample design refers to the plans and methods to be
followed in selecting sample from the target population and the estimation
technique formula for computing the sample statistics. These statistics are the
estimates used to infer the population paramet ers.
8C.3 Purpose of Sampling
The basic purpose of sampling is to provide an estimate of the population parameter
and to test the hypothesis. Advantages of sampling are –
9 Save time and money
9 Enable collection of comprehensive data
9 Enable more accurate measurement as it conducted by trained and experienced investigators
9 Sampling remains the only way when population contains infinitely many
members
9 In certain situation, sampling is the only way of data collection. For example,
in testing the pathological status of blood, boiling status o f rice, etc
9 It provides a valid estimation of sampling error
8C.4 Stages of Sampling Process
The sampling process comprises several stages
1. Define the population.
2. Specifying the sampling frame.
3. Specifying the sampling unit.
4. Selection of the sampling method.
5. Determination of sample size.
6. Specifying the sampling plan.
7. Selecting the sample. munotes.in
Page 173
173Chapter 8: Measurement Concepts, Sampling and Field Work
Define the Population: Population must be defined in terms of elements, sampling
units, extent and time. Because there is very rarely en ough time or money to gather
information from everyone or everything in a population, the goal becomes finding
a representative sample (or subset) of that population.
Sampling Frame: As a remedy, we seek a sampling frame which has the property
that we can identify every single element and include any in our sample. The most
straightforward type of frame is a list of elements of the population (preferably the
entire population) with appropriate contact information. A sampling frame may be
a telephone book, a city directory, an employee roster, a listing of all students
attending a university, or a list of all possible phone numbers.
Sampling Unit: A sampling unit is a basic unit that contains a single element or a
group of elements of the population to be sampled. The sampling unit selected is
often dependent upon the sampling frame. If a relatively complete and accurate
listing of elements is avai lable (e.g. register of purchasing agents) one may well
want to sample them directly. If no such register is available, one may need to
sample companies as the basic sampling unit.
Sampling Method: The sampling method outlines the way in which the sample
units are to be selected. The choice of the sampling method is influenced by the
objectives of the research, availability of financial resources, time constraints, and
the nature of the problem to be investigated. All sampling methods can be grouped
under two distinct heads, that is, probability and non -probability sampling.
Sample Size: The sample size calculation depends primarily on the type of
sampling designs used. However, for all sampling designs, the estimates for the
expected sample characteristic s (e.g. mean, proportion or total) desired level of
certainty, and the level of precision must be clearly specified in advanced. The
statement of the precision desired might be made by giving the amount of error that
we are willing to tolerate in the resul ting estimates. Common levels of precisions
are 5% and 10%.
Sampling Plan: In this step, the specifications and decisions regarding the implementation of the research process are outlined. As the interviewers and their
co-workers will be on field duty of most of the time, a proper specification of the
sampling plans would make their work easy and they would not have to reverting
operational problems.
Select the Sample: The final step in the sampling process is the actual selection of
the sample elements. This requires a substantial amount of office and fieldwork,
particularly if personal interviews are involved. munotes.in
Page 174
174 RESEARCH IN COMPUTING
8C.5 Techniques of Sampling
There are two basic approaches to sampling: Probability Sampling and Non -
probability Sampling.
8C.5.1 Probability Sampling
Probability sampling is also known as random sampling or chance sampling. In
this, sample is taken in such a manner that each and e very unit of the population
has an equal and positive chance of being selected. In this way, it is ensured that
the sample would truly represent the overall population. Probability sampling can
be achieved by random selection of the sample among all the un its of the
population.
Major random sampling procedures are –
9 Simple Random Sample
9 Systematic Random Sample
9 Stratified Random Sample
9 Cluster/ Multistage Sample
8C.5.1.1 Simple Random Sample
For this, each member of the population is numbered. Then, a given size of the
sample is drawn with the help of a random number chart. The other way is to do a
lottery. Write all the numbers on small, uniform pieces of paper, fold the papers,
put them in a container and take out the required lot in a random manner from the
container as is done in the kitty parties. It is relatively simple to implement but the
final sample may miss out small sub groups.
Figure 3.1: Example of Simple Random Sampling
munotes.in
Page 175
175Chapter 8: Measurement Concepts, Sampling and Field Work
A simple random sample is chosen in such a way that every se t of individuals has
an equal chance to be in the selected sample.
For example , if you wanted to study all the adults in the India who had high
cholesterol, the list would be practically impossible to get unless you surveyed
every person in the country. T herefore , other sampling methods would probably be
better suited to that particular experiment.
Advantages :
The sample will be free from Bias (i.e. it’s random!).
Disadvantages :
Difficult to obtain
Due to its very randomness, “freak” results can sometime s be obtained that are not
representative of the population. In addition, these freak results may be difficult to
spot. Increasing the sample size is the best way to eradicate this problem.
8C.5.1.2 Systematic Random Sample
Systematic sampling is one method in the broader category of random sampling
(for this reason, it requires precise control of the sampling frame of selectable
individuals and of the probability that they will be selected). It involves choosing a
first individual at random from the p opulation, then selecting every following nth
individual within the sampling frame to make up the sample.
Systematic sampling is a very simple process that requires choosing only one
individual at random. The rest of the process is fast and easy. As with s imple
random sampling, the results that we obtain are representative of the population,
provided that there is no factor intrinsic to the individuals selected that regularly repeats certain characteristics of the population every certain number of individu als—which is very rarely the case.
The process for conducting systematic sampling is as follows:
1. We prepare an ordered list of N individuals in the population; this will be our
sampling frame.
2. We divide the sampling frame into n fragments, where n is our desired sample
size. The size of these fragments will be
K=N/n
where K is called the interval or lift coefficient. munotes.in
Page 176
176 RESEARCH IN COMPUTING
3. The initial number: we randomly obtain a whole number A, which is less than
or equal to the interval. The numbe r corresponds to the first subject who we
select for the sample within the first fragment into which we have divided the
population.
4. Selection of the remaining n -1 individuals: We select the subsequent individuals based on where they fall, in simple ar ithmetic succession, after
the randomly selected individual, selecting individuals who occupy the same
position as the initial subject in the rest of the fragment into which we have
divided the sample. This is the equivalent of saying that we will select t he
individuals
A, A + K, A + 2K, A + 3K, ...., A + (n -1)K
Example
Let’s say that our sampling frame includes 5,000 individuals and we want a 100 -
individual sample. First, we divide the sampling frame into 100 fragments of 50
individuals. Then, we randomly select one number between 1 and 50 in order to
randomly select the first individual for the first fragment. Let's say we select 24.
The sample is defined by this individual; we select the remaining individuals from
the list at intervals of 50 unit s: 24, 74, 124, 174, ..., 4.974
It also requires numbering the entire population. Then every nth number (say every
5th or 10th number, as the case may be) is selected to constitute the sample. It is
easier and more likely to represent different subgroups.
Figure 3. 2: Example of Systematic Random Sampling
Advantages
Can eliminate other sources of bias.
munotes.in
Page 177
177Chapter 8: Measurement Concepts, Sampling and Field Work
Disadvantages
Can introduce bias where the pattern used for the samples coincides with a pattern
in the population.
8C.5.1.3 Stratified Random Sample
At first, the population is first divided into groups or strata each of which is
homogeneous with respect to the given characteristic feature. From each strata,
then, samples are drawn at random. This is called stratified random sampling.
For example , with respect to the level of socio -economic status, the population may
first be grouped in such strata as high, middle, low and very low socio -economic
levels as per pre -determined criteria, and random sample drawn from each group.
The sample size for each s ub-group can be fixed to get representative sample. This
way, it is possible that different categories in the population are fairly represented
in the sample, which could have been left out otherwise in simple random sample.
As with stratified samples, the population is broken down into different categories.
However, the size of the sample of each category does not reflect the population as
a whole. The Quota sampling technique can be used where an unrepresentative
sample is desirable (e.g. you might want t o interview more children than adults for
a survey on computer games), or where it would be too difficult to undertake a
stratified sample.
Figure 3. 3: Example of Stratified Random Sampling
Advantages
Yields more accurate results than simple random sampling
Can show different tendencies within each category (e.g. men and women)
munotes.in
Page 178
178 RESEARCH IN COMPUTING
Disadvantage
Nil
8C.5.1.4 Cluster/ Multistage Sampl ing
In some cases, the selection of units may pass through various stages, before you
finally reach your sample of study. For this, a State, for example, may be divided
into districts, districts into blocks, blocks into villages, and villages into identifiable
groups of people, and then taking the random or quota sample from each group.
For example , taking a random selection of 3 out of 15 districts of a State, 6 blocks
from each selected district, 10 villages from each selected block and 20 households
from ea ch selected village, totalling 3600 respondents. This design is used for
large -scale surveys spread over large areas.
The advantage is that it needs detailed sampling frame for selected clusters only
rather than for the entire target area. There are saving s in travel costs and time as
well. However, there is a risk of missing on important sub -groups and not having
complete representation of the target population.
Figure 3.4: Example of Cluster Sampling
Advantages
Less expensive and time consuming than a fully random sample.
Can show ‘regional’ variations.
Disadvantages
Not a genuine random sample.
Likely to yield a biased result (especially if only a few clusters are sampled).
munotes.in
Page 179
179Chapter 8: Measurement Concepts, Sampling and Field Work
8C.5.2 Non-Probability Sampling
Non-probability sampling is any sampling meth od where some elements of the
population have no chance of selection (these are sometimes referred to as 'out of
coverage'/'under covered'), or where the probability of selection can't be accurately
determined. It involves the selection of elements based o n assumptions regarding
the population of interest, which forms the criteria for selection.
Hence, because the selection of elements is non -random, non -probability sampling
does not allow the estimation of sampling errors. Non -probability sampling is a
non-random and subjective method of sampling where the selection of the population elements comprising the sample depends on the personal judgment or
the discretion of the sampler.
Non-probability sampling includes :
9 Accidental/ Convenience Sampling
9 Quota Sampling
9 Judgment/ Subjective/ Purposive Sampling
9 Snowball Sampling
8C.5.2.1 Convenience/ Accidental Sampling
Accidental sampling is a type of non -probability sampling which involves the
sample being drawn from that part of the population which is close to hand. That
is, a sample population selected because it is readily available and convenient.
The researcher u sing such a sample cannot scientifically make generalizations
about the total population from this sample because it would not be representative
enough.
For example , if the interviewer was to conduct such a survey at a shopping center
early in the morning on a given day, the people that s/he could interview would be
limited to those given there at that given time, which would not represent the views
of other members of society in such an area, if the survey was to be conducted at
different times of day and several times per week. This type of sampling is most
useful for pilot testing.
The primary problem with availability sampling is that you can never be certain
what population the participants in the study represent. The population is unknown,
the method for selecting cases is haphazard, and the cases studied probably don't
represent any population you could come up with. However, there are some
situations in which this kind of design has advantages. munotes.in
Page 180
180 RESEARCH IN COMPUTING
For example, survey designers often want to have some pe ople respond to their
survey before it is given out in the ‘real’ research setting as a way of making certain
the questions make sense to respondents. For this purpose, availability sampling is
not a bad way to get a group to take a survey, though in this case researchers care
less about the specific responses given than whether the instrument is confusing or
makes people feel bad.
Advantages
Expedited data collection: When time is of the essence, many researchers turn to
convenience sampling for data colle ction, as they can swiftly gather data and begin
their calculations.
Ease of research : For researchers who are not looking for an accurate sampling,
they can simply collect their information and move on to other aspects of their
study. This type of samplin g can be done by simply creating a questionnaire and
distributing it to their targeted group.
Ready availability : Since most convenience sampling is collected with the populations on hand, the data is readily available for the researcher to collect .
Cost effectiveness: One of the most important aspects of convenience sampling is
its cost effectiveness. This method allows for funds to be distributed to other
aspects of the project.
Disadvantages
Bias: The results of the convenience sampling cannot be generalized to the target
population because of the potential bias of the sampling technique due to under -
representation of subgroups in the sample in comparison to the population of
interest.
Power: Convenience sampling is characterized with insufficient power to identify
differences of population subgroups
8C.5.2.2 Quota Sampling
In quota sampling, the population is first segmented into mutually exclusive sub -
groups, just as in stratified sampling. Then judgment is used to select the subjects
or units from each segment based on a specified proportion. For example, an
interviewer ma y be told to sample 200 females and 300 males between the age of
45 and 60. In quota sampling the selection of the sample is non -random.
For example: interviewers might be tempted to interview those who look most
helpful. The problem is that these samples may be biased because not everyone
gets a chance of selection. This random element is its greatest weakness and quota
versus probability has been a matter of controversy for many years. munotes.in
Page 181
181Chapter 8: Measurement Concepts, Sampling and Field Work
Process
In quota sampling, a population is first segmented into mutua lly exclusive sub -
groups, just as in stratified sampling. Then judgment is used to select the subjects
or units from each segment based on a specified proportion. For example, an
interviewer may be told to sample 200 females and 300 males between the age o f
45 and 60. This means that individuals can put a demand on who they want to
sample (targeting).
This second step makes the technique non -probability sampling. In quota sampling,
there is non -random sample selection and this can be unreliable. For example ,
interviewers might be tempted to interview those people in the street who look most
helpful, or may choose to use accidental sampling to question those closest to them,
to save time. The problem is that these samples may be biased because not everyone
gets a chance of selection, whereas in stratified sampling (its probabilistic version),
the chance of any unit of the population is the same as 1/n (n= number of units in
the population). This non -random element is a source of uncertainty about the
nature of the actual sample and quota versus probability has been a matter of
controversy for many years.
8C.5.2.3 Subjective or Purposive or Judgment Sampling
In this sampling, the sample is selected with definite purpose in view and the
choice of the sampling units depends entirely on the discretion and judgment of the
investigator. This sampling suffers from drawbacks of favouritism and nepotism
depending upo n the beliefs and prejudices of the investigator and thus does not give
a representative sample of the population. This sampling method is seldom used
and cannot be recommended for general use since it is often biased due to element
of subjectivity on the part of the investigator. However, if the investigator is
experienced and skilled and this sampling is carefully applied, then judgment
samples may yield valuable results. Some purposive sampling strategies that can
be used in qualitative studies are given below. Each strategy serves a particular data
gathering and analysis purpose. Extreme Case Sampling: It focuses on cases that
are rich in information because they are unusual or special in some way. e.g. the
only community in a region that prohibits felli ng of trees.
8C.5.2.4 Snowball Sampling
Snowball sampling is a method in which a researcher identifies one member of
some population of interest, speaks to him/her, and then asks that person to identify
others in the population that the researcher might sp eak to. This person is then asked
to refer the researcher to yet another person, and so on. munotes.in
Page 182
182 RESEARCH IN COMPUTINGThis sampling technique is used against low incidence or rare populations. Sampling is a big problem in this case, as the defined population from which the
sample can be drawn is not available. Therefore, the process sampling depends on
the chain system of referrals. Although small sample sizes and low costs are the
clear advantages of snowball sampling, bias is one of its disadvantages.
The referral names obtained from those sampled in the initial stages may be similar
to those initially sampled. Therefore, the sample may not represent a cross -section
of the total population. It may also happen that visitors to the site or interviewers
may refuse to disclose the names of those whom they know.
Snowball sampling uses a small pool of initial informants to nominate, through
their social networks, other participants who meet the eligibility criteria and could
potentially contribute to a specific study. The term "snowball sampling" reflects an
analogy to a snowball increasing in size as it rolls downhill.
Method
9 Draft a participation program (likely to be subject to change, but indicative).
9 Approach stakeholders and ask for contacts.
9 Gain contacts and ask them to participate.
9 Community issues groups may emerge that can be included in the participation program.
9 Continue the snowballing with contacts to gain more stakeholders if necessary.
9 Ensure a diversity of contacts by widening the profil e of persons involved in
the snowballing exercise.
Advantages
9 Locate hidden populations: It is possible for the surveyors to include people
in the survey that they would not have known but, through the use of social
network.
9 Locating people of a specific p opulation: There are no lists or other obvious
sources for locating members of the population .
9 Methodology: As subjects are used to locate the hidden population, the
researcher invests less money and time in sampling. Snowball sampling
method does not require complex planning and the staffing required is
considerably smaller in comparison to other sampling methods . munotes.in
Page 183
183Chapter 8: Measurement Concepts, Sampling and Field Work
Disadvantages
9 Community bias: The first participants will have a strong impact on the
sample.
9 Non-random: Snowball sampling contravenes many of the assumptions supporting conventional notions of random selection and representativeness .
9 Unknown sampling population size: There is no way to know the total size
of the overall population.
9 Anchoring: Lack of definite knowledge as to whether or not the sample is an
accurate reading of the target population.
9 Lack of control over sampling method: As the subjects locate the hidden
population, the research has very little control over the sampling method,
which becomes mainly dependent on the original a nd subsequent subjects,
who may add to the known sampling pool using a method outside of the
researcher's control.
*(Some more types of Non - Probability Sampling Methods are listed after the
references of this chapter to get a through understanding of this topic)
8D. Determination of Sample Size
Determination of sample size is probably one of the most important phases in the
sampling process. Generally the larger the sample size, the better is the estimation. But always larger sample sizes cannot be used in view of time and budget constraints. Moreover, when a probability sample reaches a certain size the precision of an estimator cannot be significantly increased by increasing the sample
size any further. Indeed, for a large population the precision of an estimator depends
on the sample size, not on what proportion of the population has been sampled. It
can be stated that whenever a sample study is made, there arises some sampling
error which can be controlled by selecting a sample of adequate size.
Sample size is a frequently -used term in statisti cs and market research, and one that
inevitably comes up whenever you’re surveying a large population of respondents.
It relates to the way research is conducted on large populations.
When you survey a large population of respondents, you’re interested in the entire
group, but it’s not realistically possible to get answers or results from absolutely
everyone. So you take a random sample of individuals which represents the
population as a whole. munotes.in
Page 184
184 RESEARCH IN COMPUTING
The size of the sample is very important for getting accurate, statistically significant
results and running your study successfully.
If your sample is too small, you may include a disproportionate number of
individuals which are outliers and anomalies. These skew the results and you don’t
get a fair picture of the wh ole population.
If the sample is too big, the whole study becomes complex, expensive and time -
consuming to run, and although the results are more accurate, the benefits don’t
outweigh the costs.
If you’ve already worked out your variables you can get to th e right sample size
quickly with the online sample size calculator below:
Confidence Level:
95%
Population Size:
10000
Margin of Error:
5%
Ideal Sample Size:
370
If you want to start from scratch in determining the right sample size for your
market research, let us walk you through the steps.
Learn how to determine sample size
To choose the correct sample size, you need to consider a few different factors that
affect your research, and gain a basic understanding of the statistics involved.
You’ll then be able to use a sample size formula to bring everything together and
sample confidently, knowing that there is a high probability that your survey is
statistically accurate.
The steps that follow are suitable for finding a sample size for cont inuous data –
i.e. data that is counted numerically. It doesn’t apply to categorical data – i.e. put
into categories like green, blue, male, female etc.
STAGE 1: Consider your sample size variables
Before you can calculate a sample size, you need to determ ine a few things about
the target population and the level of accuracy you need:
munotes.in
Page 185
185Chapter 8: Measurement Concepts, Sampling and Field Work
1. Population size
How many people are you talking about in total? To find this out, you need
to be clear about who does and doesn’t fit into your group. For example, if
you want to know about dog owners, you’ll include everyone who has at
some point owned at least one dog. (You may include or exclude those who
owned a dog in the past, depending on your research goals.) Don’t worry if
you’re unable to calculate the exact num ber. It’s common to have an
unknown number or an estimated range.
2. Margin of error (confidence interval)
Errors are inevitable – the question is how much error you’ll allow. The
margin of error, AKA confidence interval, is expressed in terms of mean
numbers. You can set how much difference you’ll allow between the mean
number of your sample and the mean number of your population. If you’ve
ever seen a political poll on the news, you’ve seen a confidence interval and
how it’s expressed. It will look some thing like this: “68% of voters said yes
to Proposition Z, with a margin of error of +/ - 5%.”
3. Confidence level
This is a separate step to the similarly -named confidence interval in step 2. It
deals with how confident you want to be that the actual mea n falls within
your margin of error. The most common confidence intervals are 90%
confident, 95% confident, and 99% confident.
4. Standard deviation
This step asks you to estimate how much the responses you receive will vary
from each other and from the mean number. A low standard deviation means
that all the values will be clustered around the mean number, whereas a high
standard deviation means they are spread out across a much wider range with
very small and very large outlying figures. Since you haven ’t yet run your
survey, a safe choice is a standard deviation of .5 which will help make sure
your sample size is large enough.
STAGE 2: Calculate sample size
Now that you’ve got answers for steps 1 – 4, you’re ready to calculate the sample
size you need. This can be done using an online sample size calculator or with paper
and pencil.
munotes.in
Page 186
186 RESEARCH IN COMPUTING
5. Find your Z -score
Next, you need to turn your confidence level into a Z -score. Here are the Z -
scores for the most common confidence levels:
90% – Z Score = 1.645
95% – Z Score = 1.96
99% – Z Score = 2.576
6. Use the sample size formula
Plug in your Z -score, standard of deviation, and confidence interval into the
sample size calculator or use this sample size formula to work it out yourself:
Necessary Sample Size =
ሺܼെ݁ݎܿݏሻଶൈܦ݀ݐܵ݁ݒൈሺͳെܦ݀ݐܵ݁ݒሻ
ሺܽܯݎ݂݊݅݃ݎݎݎ݁ሻଶ
This equation is for an unknown population size or a very large population
size. If your population is smaller and known, just use the sample size
calculator.
Solved example :
assuming you chose a 95% confidence level, 0.5 standard deviation, and a margin
of error (confidence interval) of +/ - 5%.
= ((1.96)2 x .5(.5)) / (.05)2
= (3.8416 x .25) / .0025
= 0.9604 / .0025
= 384.16
385 respondents are needed
Summary
In this unit we saw sampling which is a process used in statistical analysis in which
a predetermined number of observations are taken from a larger population. The
methodology used to sample from a larger population depends on the type of
analysis being performed, but it may include simple random sampling or systematic
sampling. munotes.in
Page 187
187Chapter 8: Measurement Concepts, Sampling and Field WorkWe also saw the terminologies used along with sampling which includes population, sample, parameter, target population, sampled population etc. We
further analysed the purpose of sampling along with understating the stages of
sampling which included 7 steps. Later, in this unit, we saw the techniques of
sampling (Probability Sampling and non -Probability Sampling techniques)
Later we understood the concept of determining the size of sample. To choose the
right size of sample is the next section of this unit titled as determination of sample
size. Sample size determination is the act of choosing the number of observations
or replicates to include in a statistical sample. The sample size is an important
feature of any empirical study in which the goal is to make inferences about a
population from a sample.
Questions
1. Write a note on purpose of sampling.
2. Discuss the seven stages of sampling.
3. What are sampling techniques? Why do we need them?
4. Elaborate on Probabili ty sampling and state its 3 types.
5. Elaborate on Non -Probabili ty sampling and state its 3 types.
6. Why do we need to determine a sample size? Justify your answer.
7. Explain with appropriate example, snowball sampling.
8. Explain with ap propriate figure and example, simple random sampling.
9. Discuss the steps for determining of the sample size.
References
1 Albright Winson, 2015, Business Analytics, 5th Edition,
2 Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
3 Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
4 Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
5 Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation,
ISBN : 9788131316054, 813131605X munotes.in
Page 188
188 RESEARCH IN COMPUTING
6 William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin, 2016,
Business Research Methods, Edition 8, Cengage Publications
7 Images: https://www.wikiwand.com/en/Sampling_(statistics) creative commons
* Some more types of Non -Probability Sampling Techniques
8C.5 .2.5 Maximum Variation Sampling
Aims at capturing the central themes that cut across participant variations. e.g.
persons of different age, gender, religion and marital status in an area protest ing
against child marriage. Homogeneous Sampling: Picks up a small sample with
similar characteristics to describe some particular sub -group in depth. e.g. firewood
cutters or snake charmers or bonded laborers.
8C.5 .2.6 Typical Case Sampling
Uses one or more typical cases (individuals, families / households) to provide a
local profile. The typical cases are carefully selected with the co -operation of the
local people/ extension workers.
8C.5. 2.7 Critical Case Sampling
Looks for critical cases that can make a point quite dramatically. e.g. farmers who
have set up an unusually high yield record of a crop. Chain Sampling: Begins by
asking people, ‘who knows a lot about ________’. By asking a number of people,
you can identify specific kinds of ca ses e.g. critical, typical, extreme etc.
8C.5 .2.8 Criterion Sampling
Reviews and studies cases that meet some pre -set criterion of importance e.g.
farming households where women take the decisions. In short, purposive sampling
is best used with small numb ers of individuals/groups which may well be sufficient
for understanding human perceptions, problems, needs, behaviours and contexts,
which are the main justification for a qualitative audience research.
8C.5 .2.9 Matched Random Sampling
A method of assigni ng participants to groups in which pairs of participants are first
matched on some characteristic and then individually assigned randomly to groups.
The Procedure for Matched random sampling can be briefed with the following
contexts - (a) Two samples in wh ich the members are clearly paired, or are matched
explicitly by the researcher. For example, IQ measurements or pairs of identical
twins. (b) Those samples in which the same attribute, or variable, is measured twice
on each subject, under different circumstances. Commonly called repeated measures. munotes.in
Page 189
189Chapter 8: Measurement Concepts, Sampling and Field Work
8C.5 .2.10 Mechanical Sampling
Mechanical sampling is typically used in sampling solids, liquids and gases, using
devices such as grabs, scoops; thief probes etc. Care is needed in ensuring that the
sample is repr esentative of the frame.
8C.5 .2.11 Line -Intercept Sampling
Line-intercept sampling is a method of sampling elements in a region whereby an
element is sampled if a chosen line segment, called a ‘transect’, intersects the
element.
8C.5 .2.12 Panel Sampling
Panel sampling is the method of first selecting a group of participants through a
random sampling method and then asking that group for the same information again
several times over a period of time. Therefore, each participant is given the same
survey or interview at two or more time points; each period of data collection is
called a ‘wave’. This sampling methodology is often chosen for large scale or
nation -wide studies in order to gauge changes in the population with regard to any
number of var iables from chronic illness to job stress to weekly food expenditures.
Panel sampling can also be used to inform researchers about within -person health
changes due to age or help explain changes in continuous dependent variables such
as spousal interaction .
8C.5 .2.13 Rank Sampling
A non -probability sample is drawn and ranked. The highest value is chosen as the
first value of the targeted sample. Another sample is drawn and ranked; the second
highest value is chosen for the targeted sample. The process is repeated until the
lowest va lue of the targeted sample is chosen. This sampling method can be used
in forestry to measure the average diameter of the trees.
8C.5 .2.14 Voluntary Sample
A voluntary sample is made up of people who self -select into the survey. Often,
these folks have a strong interest in the main topic of the survey. Suppose, for
example, that a news show asks viewers to participate in an on -line poll. This would
be a volunteer sample. The sample is chosen by the viewers, not by the survey
administrator.
munotes.in
Page 190
190 RESEARCH IN COMPUTING
UNIT 5
9 DATA ANALYIS AND PRESENTATION
A. EDITING AND CODING
Unit Structure
9A.0 Objectives
9A.1 Introduction
9A.2 Stages of Data Analysis
9A.2.1 The Editing Phase
9A.2.2 The Coding Phase
9A.2.3 The Data File Phase
9A.3 Code Construction
9A.4 Precoding Fixed -Alternative Questions
9A.5 Code Book
9A.6 Editing and Coding Combined
9A.7 Computerized Survey Data Processing
9A.8 Error Checking
9A.9 Summary
9A.10 Questions
9A.11 References munotes.in
Page 191
191Chapter 9: Data Analyis and Presentation
9A.0 Objectives
Following are the objectives of this unit:
• To understand Data Analysis and its phases
• To gain an understanding of how the data are represented in a data file
• To know when a response is an error and how it should be edited
• To implement Computers for Data P rocessing
• To understand Error Checking
9A.1 Introduction
Once the data from the survey is collected which is raw in nature, the next step to
be conducted is that the data must be edited. “The data editing process consists of
reviewing the data and making adjustments to the data collected.” The Data editing
helps the Researcher to have a consistent, clear and integrated data. This data helps
us to generate accurate statistics for which the rese arch is conducted .
The reasons to carry Data Editing are as following:
• To find the errors in the data
• To validate the data.
• To provide accurate data.
• To find if there is any inconsistency in the data and if found then it is
removed .
9A.2 Stages of Data Analysis
The data analysis process used by researchers is to reduce the raw data into
meaningful, integrated and consistent chunks of data which is effectively used for
application area of research.
The data whether it is quantitative or qualitative will not be of any use, unless and
until it is not analyzed or interpreted by using various scientific methods. The stages
of the Data Analysis process focuses on finding the errors in the data. If the error
remains in the data then the results generated will be more risky and generating the
conclusions of the research will be difficult. munotes.in
Page 192
192 RESEARCH IN COMPUTING
Stages of Data Analysis
9A.2.1 Stage 1: Data Editing
The Data Editing is the process in which the checking and making adjustments to
the data like omitted data, making the data consistent; modifications in the errors
are done. Once the editing process is done, the data is in integrated and consistent
form wh ich can be electronically used for analysis. The data which is collected
lacks in uniformity. For example: The data which is collected from questionnaires
does not have answers ticked at the right places or sometimes we also see that the
answers are left blank. There are times at which the data is collected in monthly
form but later needs to be converted to an annual form. The Researcher needs to
make the decisions to transform the data according to the objectivity of the
Research.
The Data Editing is cate gorized into the following forms :
9A.2 .1.1 Field Editing
The field editing is done by the field supervisor and the field editing is done the
same day when the interview is conducted to collect the data. The main objective
is :
• To find the technical omissions in the form.
• All the responses which are inconsistent or in conceptual in nature are clarified.
• Handwriting legibility is checked. Raw Data collected Step 1: Editing Phase Step 2 : Coding Phase Step 3 : Data File Phase
Data
Editing Field
Editing In-House
Editing munotes.in
Page 193
193Chapter 9: Data Analyis and Presentation
There are situations when there are gaps present from the conducting of interviews
then in such case it is important that instead of making a guess work, a call back
should be made and find out what response is by the respondent. The main purpose
of field editing is to control the quality of the existing data.
Example: Blank page on an interview form.
9A.2 .1.2 In-House Editing
In the process of in -house editing, the results are investigated from the results of
data which is collected and various coding functions and editing process is carried
out.
Example: If age is not indicated, then the respondent will be ca lled to ensure the
information.
9A.2 .2 STAGE 2: CODING PHASE
Coding is defined as the process in which the assigning of numbers or symbols is
done to the answers so that the responses can be grouped into a limited number of
classes or categories. The codi ng process helps the researcher to reduce the thousands of answers to a limited number of categories which contains only the
relevant and critical information for that particular question which is asked to
collect the data. When the questionnaire is being prepared, numerical coding
depending upon the nature of the data to be collected is used. This process is also
called as Pre -coding. The Pre-coding can be applied only to those questions where
we know what all will be the categories of the answers for e.g. sex, religion,Status
etc. The questions which are answered are called as Post -coding. When the
qualitative research is carried out the data is collected from interviews,
questionnaires or observations made. The main objective of the coding is to give
mean ing to the data which the respondent has given. The data coder extracts the
codes which are preliminary in nature from the data which is observed and then the
preliminary data is filtered and refining is done so that accurate, précised and
concise codes ar e generated. The objective is not to eliminate the excess amount of
data but also summarize it in such a way that it becomes meaningful. The data coder
sees that during the coding process the important points are not lost.
9A.2 .3 STAGE 3: DATA FILE PHASE
Once the coding phase is done, the next step is to enter the coded information into
a file which can be stored on a disc, tape or any other storage media. The data is
stored in the form of a matrix which is more like a spreadsheet . It is represented in
the form of rows and columns. Each row represents the score of a variable given munotes.in
Page 194
194 RESEARCH IN COMPUTING
by the respondents and the row represents a variable for which a value is given by
the respondent’s. A very common example to store a data file used is S preadsheet
like Excel. Now more advanced programs like SAS , SPSS and many more which
works well with the Excel Spreadsheet. The most important thing which needs to
be considered is the construction of data files.
9A.3 Code Construction
Code construction s hould be done keeping in mind two rules which are as
following:
1) The coding category should be there for all the possible responses. The coder
has to see that the responses should be made available for the entire categori cal
variable such as sex and also for the categories where the response falls into a
class which is not found. The missing data should also be represented with
acode. A blank response now these days is considered as a missing data.
2) The categories of the coding are independent and mutually exclusive. It is to
be ensured that the response can be placed in only one category among the
specified.
9A.4 Precoding Fixed Alternat ive Questions
Whenever a survey is conducted, in Precoded Fixed Alternative Questions there
are several responses are given for a question and the participant is asked to pick
up the correct or the response which is best suited to the question.
For exampl e:
1) Have you conducted a research earlier?
R Yes
R No
2) Which of the following most closely corresponds to your salary per month?
R 10,000 or lesser than it
R 11000 to 19000
R 20000 to 39000
R 40,000 to 49000
R 50,000 or more than munotes.in
Page 195
195Chapter 9: Data Analyis and Presentation
9A.5 Code Book
The Codebooks are used by the researchers because of the two main reasons:
1) Codebooks are used as a guide for coding response
2) Codebooks are used as a documentation which defines the layout and the code
definitions which are used in a data file.
Codebooks are used to document the values associated with the answer options for
a given survey question by the respondent. Every answer category is given a unique
numeric value, and these unique numeric values are later used for the conclusions
of the res earch.
9A.6 Editing and Coding Combined
The person who sometimes who codes the questionnaire also does the editing
functions. For example: Suppose the respondent is asked for description of job and
the respondent instead of indicating “teaching” as their occupation they write
teaching in defenc e,or teaching in Social Science. So in such situations the coder is
provided with the instructions that they must do the editing function.
9A.7 Computerized Survey Data Processing
The results which are collected from the Survey plays a very important role. The
survey data which is collected requires time to enter into the system and also for
the processing process. The research where the sample size id very large ,computers
can be used for the processing of the data. The Computerized Survey data processing consists of the following phases:
1) Survey Designing
2) Data entry and data capture
3) Data quality assurance and analysis
4) Reporting and Data Tabulation
The data entry process consists of transferring the entire research data project to
computers. The optical scanning systems are also used o read the data directly into munotes.in
Page 196
196 RESEARCH IN COMPUTING
the memory of the computer from the mark sensed questionnaires. Computer
assisted teleph one interviewing or a self administered internet questionnaire is used
where the responses from the respondent are automatically stored and presented in
the reports in the tabular form. Use of Computer’s have become a growing
phenomenon in the Business Research field.
9A.8 Error Checking
The final stage in the coding process is error checking and verification, or data
cleaning, to ensure that all codes are legitimate. If any mistake is fou nd in the data
then the adjustments are made and data is represented in the correct form.
Summary
In this chapter we understood the critical key steps of the data entry process and
coding. The detailed attention is required for the data editing process an d coding.
The complete data analysis process with its detailed phases underlines how
important it is in the research process. The different methods in which the data can
be represented in the data file is discussed. The importance of error checking and
validation was discussed in the coding phase. Evaluating the response as an error
and how it should be edited is explained.
Questions
1. Explain the objective of Editing process?
2. Justify when the raw data from the respondent be alters by a data editor.
3. Write a short note on Code Book.
4. Explain how Computers plays a role in Survey Data Processing.
5. Discuss the Data Analysis process in detail.
munotes.in
Page 197
197Chapter 9: Data Analyis and Presentation
References
R Albright Winson, 2015, Business Analytics, 5th Edition,
R Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
R Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable,
Reliability, Validity, and Norm.
R Mark Saunders, 2011, Research Methods for Business Students, 5th
Edition
R Shefali Pandya, 2012, Research Methodology, APH Publishing
Corporation, ISBN: 97881313 16054, 813131605X
R William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin, 2016, Business Research Methods, Edition 8, Cengage
Publications
munotes.in
Page 198
198 RESEARCH IN COMPUTING
198B. BASIC DATA ANALYSIS : DESCRIPTIVE STATISTICS
Unit Structure
9B.0 Objectives
9B.1 Introduction
9B.2 Nature of Descriptive Analysis
9B.3 Tabulation
9B.3.1 Objectives of Tabulation
9B.4 Cross Tabulation Contingency Tables
9B.4.1 Percentage Cross -Tabulations
9B.4.2 Elaboration and Refinement
9B.4.3 Quadrant Analysis
9B.5 Data Transformation
9B.5.1 Simple Transformations
9B.5.2 Benefits of Data Transformation
9B.5.3 Problems with Data Transformations
9B.5.4 Index Number
9B.6 Calculating Rank Order
9B.7 Tabular and Graphic Methods of Displaying Data
9B.8 Computer Programs for Analysis
9B.8.1 Statistical Packages
9B.8.2 Benefits
9B.8.3 Commonly used Tools for Analysis
9B.8.3.1 SPSS
9B.8.3.2 SAS/STAT
9B.8.3.3 Computer Graphics and Computer Mapping
9B.9 Interpretation
9B.10 Summary
9B.11 Questions
9B.12 References munotes.in
Page 199
199Chapter 9: Data Analyis and Presentation
9B.0 Objectives
• To understand what Descriptive analysis is and the reason why are they use.
• To study the need and objectives of Tabulation
• To explore the various methods used for Displaying the data
• To study the various Statistical tools used for analysis
• To understand how the data is interpreted
9B.1 Introduction
Descriptive analysis is a statistical analysis which is used to organize and summarize the characteristics of a data set. Descriptive statistics is used for the
following objectives:
1) to provide basic information about variables in a dataset and
2) to highlight potential relationships between variables.
It summarizes the response which is received from large number of respondents in
a simple form of statistics.
9B.2 The Nature of Descriptive Analysis
Descriptive Analysis is the process in which the elementary transformation of data is done. This process describes the basic characteristics such as variability, distribution and central tendency.
9B.3 Tabulation
Tabulation is define d as the systematic & logical presentation of numeric data. The
data is represented in the form of rows and columns so that comparison and
statistical analysis can be performed. The comparison of data is performed by
bringing information which is related c lose to each other and then statistical
analysis and interpretation is performed.
It is also defined as one of the methods of placing organized data into a tabular
form is called as tabulation.
When we are counting the different ways in which the responden ts have answered
the question and we arrange them in a simple tabular form it is called as a Frequency
Table. The tabulation also tells the researcher how frequently each response occurs. munotes.in
Page 200
200 RESEARCH IN COMPUTING
9B.3.1 Objectives Of Tabulation :
(1) To Simplify the Complex Data
(2) To Bring Out Essential Features of the Data
(3) To Facilitate Comparison
(4) To Facilitate Statistical Analysis
(5) Saving of Space
Example: A simple tabulation of Grades obtained by Students obtained in Class X
having a strength of 100. Grades obtained by Students of Class X No. of Students A+(above 80) 60 A (70 to 80) 25 B (60 to 70) 5 C (50 to 60) 3 D (40 to 50) 5 Below 40 2 9B.4 Cross – Tabulation
During the analysis of survey, another method of quantitative research method
which is used is called as Cross tabulation. This methodology is used to analyze
the relationship between two or more variables. It also helps to analyze and
compare the results of one or more variables with the results of another.
Some of the surve y results are presented in aggregate only – meaning, the data
tables are based on the entire group of survey respondents. Cross tabulations are
simply data tables that present the results of the entire group of respondents as well
as results from sub -group s of survey respondents. Cross tabulations enables us to
examine relationships within the data that might not be readily apparent when
analyzing total survey responses.
9B.4 .1 Contingency Tables
This is one of the tool which is used by the statisticians when the data has more
than one variable. These are also called as Cross tabulation tables or cross tab.
These tables are displayed in the form of grid or matrix. The numbers displayed in
the table represent the frequenc y of each data point. The contingency table munotes.in
Page 201
201Chapter 9: Data Analyis and Presentation
represents the better understanding of the data using probability and relative
frequencies.
Example:
Here we can see that we have data of 100 people who have pets were polled to see
if there was a correlation be tween gender and whether they had a dog or a cat. This
is a contingency table outlining the data. Dog Cat Female 32 20 Male 10 38 Total 42 58 The number of males, females, dog owners and cat owners are called marginal
totals. The total number of people involved in the study is called the grand total .
By placing the data in a table, some conclusions can be drawn. The user can see
that there seems to be a correlation between gender and pet ownership. If there is a
correlation between the s ets of data (in this example, the strong correlation between
gender and pet ownership) the data is said to be contingent , or dependent. If there is not a relationship between the data, then the data is not contingent, or independent.
The contingency tables can have any number of rows and columns depending on
the amount of data.
9B.4.2 Percentage Cross -Tabulations
When the data which is available from the survey and it is represented in the
percentage cross tabulation form, it becomes easy for the researcher to understand
the relationship by making the comparisons in an simpler way. The total number is
used as a sta tistical base for calculating the percentage in the cell. The conventional
rule determines the direction of percentages. The marginal total of the independent
variable should be used as a base for calculating the percentages.
Example: Dog Cat Total Female 32 20 52 Male 10 38 48 Total 42 58 100 munotes.in
Page 202
202 RESEARCH IN COMPUTING
9B.4.3 Elaboration and Refinement
Sometimes the researcher after examining the relationship between two variables
wants to investigate the relationship under a variety of more conditions. A third
variable is introduced into the analysis so that the research can be made more
refined and also have a deeper understanding of the conditions under which the
relationship between the first two variables is strongest and weakest.
“Elaboration analysis invol ves the basic cross tabulation within various subgroups
of the sample”. The researcher breaks down the analysis for each level of another
variable. For example: If the researcher has cross -tabulated travelling trips
preference by gender and wishes to inve stigate another variable (say, marital
status), a more elaborate analysis may be conducted Single Married Men Women Men Women Do you wish to travel trips? Yes 54% 80% 87% 80% No 46% 20% 13% 20% The data analyzed from the above table shows that the women whether married or
single have the same preference for travelling trips. Whereas the married men like
more to travel for trips as compared to men who are single. The analysis says that
the relationship between the travelling and the gender b ehavior for women has to
be retained. Married men more frequently travelling trips at Target than do single
men.
The combination of the two variables, gender and marital status, is associated with
differences in the dependent variable. Interactions between variables examine
moderating variables. A moderator variable is a third variable that changes the
nature of a relationship between the original independent and dependent variables.
Marital status is a moderator variable in this case. The interaction effec t suggests that marriage changes the relationship between gender and travelling trips preference.
9B.4.4 Quadrant Analysis
In this technique the responses which are made to two rating scale questions are
plotted in the four quadrants of a two-dimensional table. A common quadrant
analysis in business research portrays or plots relationships between average responses about a product attribute’s importance and average ratings of a munotes.in
Page 203
203Chapter 9: Data Analyis and Presentation
company’s performance on that product feature. The term importa nce-performance
analysis is sometimes used because consumer’s rate perceived importance of
several attributes and rate how well the company’s brand performs on that attribute.
The business would like to end up in the quadrant indicating high performance on
an important attribute.
9B.5 Data Transformation
9B.5.1 Simple Transformations
Whenever we analyze the information, it requires analyzing it in such a way that the structured and accessible data is available for the best results. Data transformation enables organizations to alter the structure and format of raw data
as needed.
“Data transformation is the process of changing the format, structure, or values of
data.“
9B.5.2 Benefits of Data Transformations: • Data is transformed to make i t better -organized. • Properly formatted and validated data improves data quality and protects applications from potential landmines such as null values, unexpected duplicates,
incorrect indexing, and incompatible formats. • Data transformation facilitates co mpatibility between applications, systems, and
types of data.
9B.5.3 Problems with Data Transformations: • Data transformation can be expensive. The cost is dependent on the specific
infrastructure, software, and tools used to process data. Expenses may include those
related to licensing, computing resources, and hiring necessary personnel. • Data transformation processes can be resource -intensive. Performing transformations in an on -premises data warehouse after loading, or transforming
data before feeding it into applications, can create a computational burden that
slows down other operations. If you use a cloud -based data warehouse, you can do
the transformations after loading because the platform can scale up to meet demand. • Lack of expertise and careles sness can introduce problems during transformation.
Data analysts without appropriate subject matter expertise are less likely to notice
typos or incorrect data because they are less familiar with the range of accurate and munotes.in
Page 204
204 RESEARCH IN COMPUTING
permissible values. For example, someone working on medical data who is
unfamiliar with relevant terms might fail to flag disease names that should be
mapped to a singular value or notice misspellings.
9B.5.4 Index Numbers
It is one of the methods which measure the changes in a variable or the group of
variables with respect to certain characteristics. To name a few are consumer price
index and wholesale price index are the secondary data sources which are used by
the business researchers. Price indexes, like other index numbers, represent simple
data transformations that allow researchers to track a variable’s value over time and
compare a variable(s) with other variables.
9B.6 Calculating Rank Order
The respondents of the survey are asked to rank their preference of some items .For example consumers are sometimes asked to rank there favourite brands or sometimes employees provide Ranking data can be summarized by performing a
data transformation. The transformation involves multiplying the frequency by the
ranking score for each choice to result in a new scale. For example, suppose a
Manager has 10 team leaders rank their preferences for locations in which to hold
the company’s annual conference. Team Leaders Chennai Mumbai Delhi Kolkatta 1 1 2 4 3 2 1 3 4 2 3 2 1 3 4 4 2 4 3 1 5 2 1 3 4 6 3 4 3 2 7 2 3 3 4 8 1 4 1 3 9 4 3 1 1 10 2 1 2 1
9B.7 Tabular and Graphic Method of Displaying Data
Another way of analyzing numerical data is Graphical representation. A graph is a
chart through which statistical data are represented in the form of lines or curves
drawn across the coordinated points plotted on its surface. Graphs help us to munotes.in
Page 205
205Chapter 9: Data Analyis and Presentation
understand t he cause and effect relationship between two variables. Graphs help to
measure the extent of change in one variable when another variable changes by a
certain amount. Researcher uses many convenient tools to quickly produce charts,
graphs, or tables. Even common programs such as Excel and Word include chart functions that can construct the chart within the text document. Bar charts (histograms), pie charts, curve/line diagrams, and scatter plots are among the most
widely used tools. Some choices match well with certain types of data and analyses.
Bar charts and pie charts are very effective in communicating frequency tabulations
and simple cross -tabulations. A pie chart is a circular statistical graphic, which is
divided into slices to illustrate numerical p roportion. In a pie chart, the arc length
of each slice is proportional to the quantity it represents.
9B.8 Computer Programs for Analysis
9B.8.1 Statistical Analysis
Statistical software or statistical analysis software is one of the tools that helps
in the statistics -based collection and helps in analysis of data. The basic statistical
features are now menu driven, reducing the need to memorize function labels.
Spreadsheet packages like Excel continue to evolve and become more viable for
performing many basic statistical analyses. Despite the advances in spreadshee t applications, commercialized statistical software packages remain extremely popular among researchers. They continue to become easier to use and more compatible with other data interface tools including spreadsheets and word processors. Like any speciali zed tool, statistical packages are more tailored to the
types of analyses performed by statistical analysts, including business researchers.
munotes.in
Page 206
206 RESEARCH IN COMPUTING
Thus, any serious business or social science researcher should still become familiar
with at least one general comp uter software package.
9B.8.2 Benefits
• Increases efficiency from streamlined and automated business data analysis
workflows
• Returns more accurate predictions based on machine learning, statistical
algorithms and hypothesis testing
• Easy customization allow s you to ensure the software correctly processes the
data and results you want
• Grants access to larger databases which reduces sampling error and enables
more precise conclusions
• Empowers you to make data -driven decisions with confidence
9B.8.3 Commonly used tools which are used for statistical analysis are as
following :
9B.8.3.1 SPSS Statistics
SPSS Statistics is statistical software from IBM that can quickly crunch large data
sets to provide insights for decision -making and research. Accor ding to IBM’s
website, 81% of reviewers rank SPSS as easy to use, making it a good choice for
novice users as well as expert statisticians. It also can estimate and uncover missing
values in data sets, allowing for more accurate reports. Scalable and agil e, SPSS
Statistics is built to work with large volumes of data with as many user licenses as
needed, performing anything from descriptive analytics to advanced statistics
simulations.
9B.8.3.2 SAS/STAT
SAS/STAT is a cloud -based platform that allows users to harness tools and
procedures for statistical analysis and data visualization. Designed to address both
specialized and enterprise -wide analytics needs, it is used by business analysts,
statisticians, data scientists, researchers and engineers prim arily for statistical
modeling, observing trends and patterns in data and aiding in decision -making. Its
procedures are multithreaded, performing multiple operations at once, increasing
the efficiency and stability of the program. Users can create hundreds of built -in,
customizable statistical charts and graphs.
SAS has an established reputation in the industry for reliable results and ensures
that code produced with SAS/STAT is documented and verified to meet corporate munotes.in
Page 207
207Chapter 9: Data Analyis and Presentation
and governmental compliance requireme nts. An open -source analytics platform,
SAS allows users the freedom to experiment and program in either the interface or
the coding language of their choice.
9B.8.4 Computer Graphics and Computer Mapping Graphic aids which are prepared by the computers have replaced graphic presentation aids which are drawn by artists. Computer graphics are extremely
useful for descriptive analysis. Decision support systems can generate two - or
three -dimensional computer maps to portray data about sales, demographics, lifestyles, retail stores, and other features. Computer mapping uses the speed and
versatility of computer graphics to display spatial data.
9B.9 Interpretation
Interpretation is the process of reviewing dat a through some predefined processes
which will help assign some meaning to the data and arrive at a relevant conclusion.
It involves taking the result of data analysis, making inferences on the relations
studied, and using them to conclude. Interpretation is concerned with relationships
within the collected data. “Interpretation is drawing inferences from the analysis of
results. Each statistical analysis produces the results which are interpreted so that
we can come to a particular decision. The logical an d statistical analysis is done to
make conclusion of the research. When we talk about the perspective of the
management, the qualitative meaning of the data and there implications are an
important aspect of process of Interpretation.
Summary
In this chapter we studied about Descriptive analysis which is defined as is a
statistical analysis which is used to organize and summarize the characteristics of
a data set. Statistical Tabulation is a technique with the help of which the counting
of numb er of observations in each response was done is explained. The complete
understanding of cross -tabulations is explained to display the relationship. Data transformations were used to assist in data analysis. The different computer software products are exp lained which helps in descriptive analysis. The researcher
role is well explained in the interpretation of the data.
munotes.in
Page 208
208 RESEARCH IN COMPUTING
Questions
1. Describe the three basic data transformations.
2. Write short note on SPSS.
3. Explain the role of researcher in Interpretation of data.
4. Justify how computer software products help in descriptive statistical analysis
5. What is Cross Tabulation? How does it reveals relationships?
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN: 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016, Business Research Methods, Editi on 8, Cengage Publication
munotes.in
Page 209
209Chapter 9: Data Analyis and Presentation
C. UNIVARIATE STATISTICAL ANALYSIS
Unit Structure
9C.0 Objective
9C.1 Introduction
9C.2 Hypothesis Testing
9C.2.1 Hypothesis Testing Procedure
9C.2.2 Steps involved in hypothesis testing
9C.2.3 Significance levels and p -values
9C.2.4 Type I Aand Type II Errors
9C.3.4.1 Type I Error
9C.3.4.2 Type II Error
9C.3 Choosing the Appropriate Statistical Technique
9C.3.1 The type of question to be answered
9C.3.2 Number of Variables
9C.3.3 Level of Scale of Measurement
9C.3.4 Parametric Versus Nonparametric Hypothesis Tests
9C.4 The t -Distribution
9C.5 The Chi -Square Test for Goodness of Fit
9C.6 Hypothesis Test of a Proportion
9C.7 Additional Application of Hypothesis Testing
9C.8 Summary
9C.9 Questions
9C.10 References
9C.0 Objective
• To learn the hypothesis -testing procedure
• To understand Type I and Type II Errors
• To choose the Appropriate Statistical Technique
• To learn the Chi -square test for Goodness of Fit
• To analyze Hypothesis test of a Proportion munotes.in
Page 210
210 RESEARCH IN COMPUTING
9C.1 Introduction
Univariate analysis is one of the simplest form of analyzing data. In this analysis
the data which is analyzed contains only one variable. Since it's a single variable it
doesn't deal with causes or relationships. The main objective of univariate
analysis is to describe the data and find patterns that exist within it . Statistical
analysis can be divided into the following groups:
9C.2 Hypothesis Testing
Hypothesis testing is a formal procedure for investigating the data of the research .
Hypotheses are formal statements of explanations stated in a testable form. Most
of the times the hypotheses should be stated in concrete manner so that the method
of empirical testing can be done
The different types of hypotheses test which are commonly c onducted in the
business research are as following:
1. Relational hypotheses
2. Hypotheses about differences between groups
3. Hypotheses about differences from some standard
The various factors which are considered while formulating hypothesis are as :
• Hypothesis should be clear and precise.
• Hypothesis should be capable of being tested.
• Hypothesis should state relationship between variables.
• Hypothesis should be limited in scope and must be specific.
• Hypothesis should be stated as far as possible in mos t simple terms so that
the same is easily understandable by all concerned. Statistical
Analysis Univariate
Statistical
Analysis Bivariate
Statistical
Analysis Multivariate
Statistical
Analysis
munotes.in
Page 211
211Chapter 9: Data Analyis and Presentation
• Hypothesis should be amenable to testing within a reasonable time.
• Hypothesis must explain empirical reference.
9C.2.1 The Hypothesis –Testing Procedure
The Procedure which is carried out to conduct hypothesis testing consists of all
those steps that we undertake for making a choice between the two actions i.e.,
rejection and acceptance of a null hypothesis.
9C.2.2 The various steps involved in hypothesis testing are stated below:
1. Making a formal statement:
The step consists in making a formal statement of the null hypothesis (H0) and
also of the alternative hypothesis (Ha or H1). This means that hypotheses
should b e clearly stated, considering the nature of the research problem.
2. Selecting a significance level:
The hypotheses are tested on a pre -determined level of significance and as such
the same should be specified. Generally, in practice, either 5% level or 1%
level is adopted for the purpose.
3. Deciding the distribution to use:
After deciding the level of significance, the next step in hypothesis testing is to determine the appropriate sampling distribution. The choice generally remains between normal distribution and the t -distribution.
4. Selecting a random sample and computing a n appropriate value:
Another step is to select a random sample(s) and compute an appropriate value from the sample data concerning the test statistic utilizing the relevant distribution. In other words, draw a sample to furnish empirical data.
5. Calculatio n of the probability:
One has then to calculate the probability that the sample result would diverge
as widely as it has from expectations, if the null hypothesis were in fact true.
6. Comparing the probability and Decision making:
Yet another step consist s in comparing the probability thus calculated with the
VSHFLILHGYDOXHIRUĮWKHVLJQLILFDQFHOHYHO,IWKHFDOFXODWHGSUREDELOLW\LV
HTXDOWRRUVPDOOHUWKDQWKHĮYDOXHLQFDVHRIRQH -WDLOHGWHVWDQGĮLQFDVH
of two -tailed test), then reject the n ull hypothesis (i.e., accept the alternative
hypothesis), but if the calculated probability is greater, then accept the null
hypothesis. munotes.in
Page 212
212 RESEARCH IN COMPUTING
9C.2.3 Significance level AND p -Values
A significance level is a critical probability associated with a statistical hyp othesis
test that indicates how likely it is that an inference supporting a difference between
an observed value and some statistical expectation is true. The term p -value stands
for probability -value and is essentially another name for an observed or comp uted
significance level.
The P value is defined as the probability under the assumption of no effect or no
difference (null hypothesis), of obtaining a result equal to or more extreme than
what was actually observed. The P stands for probability and measures how likely
it is that any observed difference between groups is due to chance. Being a
probability, P can take any value between 0 and 1. Values close to 0 indicate that
the observed difference is unlikely to be due to chance, whereas a P value close to
1 suggests no difference between the groups other than due to chance. efore the
advent of computers and statistical software, researchers depended on tabulated
values of P to make decisions. This practice is now obsolete and the use of exact P
value is much preferred. Statistical software can give the exact P value and allows
appreciation of the range of values that P can take up between 0 and 1.
9C.2.4 Type I and Type II E rrors in Hypothesis Testing
There are basically two types of errors in hypothesis testing. Creatively, they call
these errors Type I and Type II errors. Both types of error relate to incorrect
conclusions about the null hypothesis.
9C.2.4.1 Type I Errors
These are the errors which results in when the null hypothesis is rejected when it is
true. and has a probability of alpha. A Type I error occurs when the researcher
concludes that a relationship or difference exists whereas in reality it doe s not exist.
9C.2.4.2 Type II Errors
When we failed to reject the null hypothesis and the alternative hypothesis is true
,in such situation these types of errors occur. It has a probability of beta. The
researcher in Type II Error makes a conclusion that no relationship or difference
exists whereas actually one does exist.
9C.3 Choosing the Appropriate Statistical Technique
There are different statistical techniques which help the researcher in interpre tation
of the data. Whenever we are considering which statistical technique to select
following are the factors which needs to be considered: munotes.in
Page 213
213Chapter 9: Data Analyis and Presentation
1. What is the type or nature of question to be answered
2. What are the number of variables involved
3. How much is the levelof scale measurement
9C.3.1 Type of question to be answere d
The nature or type of question the researcher is attempting to answer has to be
considered while making a choice of selecting a
statistical technique. For example, a researcher may be concerned simply with the
central tendency of a variable or with the distribution of a variable. The Comparison
of different business divisions’ sales results with some target level will require a
one-sample t-test. Comparison of two sale speople’s average monthly sales will
require a t-test of two means, but a comparison of quarterly sales distributions will
require a chi -square test. The researcher should consider the method of statistical
analysis before choosing the research design and b efore determining the type of
data to collect.
9C.3.2 Number of Variables
A primary consideration in the choice of statistical technique. Is what is the number
of variables which will be simultaneously investigated For example researcher who
is interested only in the average number of times a prospective home buyer visits
financ ial institutions to shop for interest rates can concentrateon investigating only
one variable at a time. In case the researcher is trying to measure multiple complex
organizational variable s cannot do the same. The statistical procedure such as
univariate, bivariate, and multivariate are differentiated only on the basis of the
number of variables involved in an Analysis process.
9C.3.3 The level of Scale of Measurement
This is one of the factor which helps us to select the best statistical techniques and
the most appropriate empirical operations. Testing a hypothesis about a mean, is
best used for interval scaled or ratio scaled data. For example a researcher is
working with a nominal scale
which identifies users versus nonusers of bank credit cards. Because of the type of
scale, the researcher may use only the mode as a measure of central tendency. In
other situations, where data are measured on an ordinal scale, the median may be
used as the ave rage or a percentile may be used as a measure of dispersion. The
Nominal and ordinal data are analyzed using frequencies or cross -tabulation.
munotes.in
Page 214
214 RESEARCH IN COMPUTING
9C.3.4 Parametric versus Nonparametric Hypothesis Tests
The statistical procedures can be divided into two categories which are as following:
• Parametric statistics
• Nonparame tric statistics
One of the basic differences between them is the underlyin g assumptions about the
data which has to be analyzed. Parametric statistics involve numbers with known,
continuous d istributions. When the data is in interval or ratio scaled and the sample
size is large, in such cases the parametric statistical procedures are appropriate.
Nonparame tric statistics works best in the scenarios when the numbers do not
conform to a known distribution. Parametric statistics are based on the assumption
that the data in the study are drawn from a populat ion with a normal distribution or
normal sampling distribution. For example, if an investigator has two interval -
scaled measures, such as gross national product (GNP) and industry sales volume,
parametric tests are appropriate. The statistical tests might include product -moment
correlation analysis, analysis of variance, regression, or a t -test for a hypothesis
about a mean. Nonparam etric methods are used when the researcher does not know
how the data are distributed. Making the assumption that the population distribution
or sampling distribution is normal generally is inappropriate when data are either
ordinal or nominal. Thus, nonpa rametric statistics are known as distribution free .
9C.4 The t -Distribution
The univariate t -test is used for testing hypotheses when we have observed mean
against some specified value. The t -distribution is a symmetrical, bell -shaped
distribution having mean of 0 and a standard deviation of 1.0. When sample size
(n) is larger than 30, the t -distribution and Z -distribution are identical. When the t -
test when used involves small sample sizes with unknown standard deviations, then
the apply the t -test for co mparisons involving the mean of an interval or ratio
measure. The precise height and shape of the t -distribution vary with sample size.
The shape of the t -distribution is related with the degrees of freedom (df). The
degrees of freedom is determined by the number of distinct calculations. In the case
of a univariate t -test, the degrees of freedom are equal to the sample size (n) minus
one.
munotes.in
Page 215
215Chapter 9: Data Analyis and Presentation
9C.5 The Chi -Square Test for Goodness of Fit
Chi-Square goodness of fit test is a non -parametric test which is used to find out
how the observed value of a given phenomena is significantly different from the
expected value. In Chi -Square goodness of fit test, the term goodness of fit is used
to compare the observed sample distribution with the expected probabil ity
distribution. Chi-Square goodness of fit test determines how well theoretical
distribution (such as normal, binomial, or Poisson) fits the empirical distribution.
In Chi -Square goodness of fit test, sample data is divided into intervals. Then the
numb ers of points that fall into the interval are compared, with the expected
numbers of points in each interval.
Procedure for Chi -Square Goodness of Fit Test:
Set up the hypothesis for Chi -Square goodness of fit test:
A. Null hypothesis: In Chi -Square goodne ss of fit test, the null hypothesis assumes
that there is no significant difference between the observed and the expected value.
B. Alternative hypothesis: In Chi -Square goodness of fit test, the alternative
hypothesis assumes that there is a significant d ifference between the observed and
the expected value. • Compute the value of Chi -Square goodness of fit test using the following formula:
Ȥ 2Lí(L
Ei
Where, Ȥ= Chi-Square goodness of fit test
Oi = observed value
Ei = expected value
9C.6 Hypothesis Test of a Proportion
The univariate statistical hypotheses about population proportions. The population
proportion ( ʌ) can be estimated on the basis of an observed sample proportion ( p).
Conducting a hypothesis test of a proportion is conceptually similar to hypothesis
testing when the mean is the characteristic of interest. Mathematically the formulation of the standard error of the proportion differs somew hat, though. munotes.in
Page 216
216 RESEARCH IN COMPUTING
Consider the following example. A state legislature is considering a proposed right -
to-work law. One legislator has hypothesized that more than 50 percent of the
state’s labor force is unionized. In other words, the hypothesis to be tested is that
the proportion of union workers in the state is greater than 0.5.The researcher
formulates the hypothesis that the population proportion ( ʌ) exceeds 50 percent(0.5):
HBʌ> 0.5
9C.7 Additional Applications of Hypothesis Testing
The concept of statistical in ference is restricted to examining the difference
between an observed sample mean and a population or pre -speFLILHGPHDQDȤ
test examines the difference between an observed frequency and the expected
frequency for a given distribution and Z -tests to test hypotheses about sample
proportions when sample sizes are large. Other hypothesis tests for population
parameters estimated from sample st atistics exist but are not mentioned here. Many
of these tests are no different conceptually in their methods of testing.
Summary
Univariate analyses are used extensively in quality of life research. Univariate
analysis is defined as analysis carried out on only one (“uni”) variable (“variate”)
to summarize or describe the variable .Descriptive statistics describe and summarize data. Univariate descriptive statistics describe individual variables. The
implementation of the hypothesis testing procedure was d one using one variable.
The use of p -values were done to test the statistical significance. The differentiation
was done between Type I and Type II Errors and these errors are very sensitive to
the sample size.
Questions
1) Explain the objective of a stat istical hypothesis.
2) Define Significance level? How does a researcher selects a significance level?
3) Differentiate between Type I and Type II Errors.
4) Discuss the factors which are considered while making a choice of the
appropriate statistical technique?
5) How is a p -value used to test a hypothesis? munotes.in
Page 217
217Chapter 9: Data Analyis and Presentation
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN: 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016, Business Research Methods, Edition 8, Cengage Publication
munotes.in
Page 218
218 RESEARCH IN COMPUTING
218D. BIVARIATE STATISTICAL ANALYSIS : DIFFERENCE
BETWEEN TWO VARIABLES
Unit Structure
9D.0 Objective
9D.1 Introduction
9D.2 What is the appropriate Test of Difference?
9D.3 &URVV7DEXODWLRQ7DEOHV7KHȤ7HVWIRU*RRGQHVV -of-Fit
9D.4 The t -Test for Comparing Two Means
9D.4.1 Independent Samples t -Test
9D.4.2 Paired -Samples t -Test
9D.5 The Z -Test for Comparing Two Proportions
9D.6 Analysis of Variance(ANOVA)
9D.6.1 Pa rtitioning Variance in ANOVA
9D.6.1.1 Total Variability
9D.6.1.2 Between Groups Variance
9D.6.1.3 Within -Group Error
9D.6.2 F-Test
9D.7 Summary
9D.8 Questions
9D.9 References
9D.0 Objective
x To learn the Bivariate Statistical Analysis
x To analyze the comparison of two proportions using the Z -Test
x To understand the concept of ANOVA
x To learn the use of Cross -Tabulation Tables munotes.in
Page 219
219Chapter 9: Data Analyis and Presentation
9D.1 Introduction
Bivariate analysis means the analysis of bivariate data. It is one of the simplest
forms of statistical analysis, used to find out if there is a relationship between two
sets of values. It usually involves the variables X and Y. The test of difference is
an investigation of a hypothesis stating that two or more groups differentiate on the
basis of variables with respect to measurement.
9D.2 What is the appropriate Test of Difference?
.It is commonly seen that the researchers are interested in testing differences in mean scores between groups or in comparing how two groups’ scores are distributed across possible response categories. The Construction of contingency
WDEOHVIRUȤDQDO\VLVJLYHVDSURFHGXUHI or comparing observed frequencies of one
group with the frequencies of another group. This is a good starting point from
which to discuss testing of differences.
9D.3 &URVV7DEXODWLRQ7DEOHV7KHȋ7HVW for Goodness -of-Fit
Cross -tabulation is one of the most widely used statistical techniques among
business researchers. Cross -tabulations are intuitive, easily understood, and lend
themselves well to graphical analysis using tools like bar charts. Cross -tabs are
appropriate wh en the variables of interest are less -than interval in nature. A cross -
tabulation, or contingency table, is a joint frequency distribution of observations on
two or more variables. Researchers generally rely on two -variable cross -tabulations
the most since the results can be easily communicated. Cross -tabulations are much
like tallying. When two variables exist, each with two categories, four cells result.
7KHȤGLVWULEXWLRQSURYLGHVDPHDQVIRUWHVWLQJWKHVWDWLVWLFDOVLJQLILFDQFHRID
contingency table. ,Q RWKHU ZRUGV WKH ELYDULDWH Ȥ WHVW H[DPLQHV WKH
statistical significance of relationships between two less -than interval variables.
7KHȤWHVWIRUDFRQWLQJHQF\WDEOHLQYROYHVFRPSDULQJWKHREVHUYHGIUHTXHQFLHV
(Oi ) with the expected frequencies (Ei ) in each cell of the table. The goodness -
(or closeness -) of-fit of the observed distribution with the expected distribution is
captured by this statistic. Remember that the convention is that the row variable is
considered the independent variable and th e column variable is considered the
dependent variable.
munotes.in
Page 220
220 RESEARCH IN COMPUTING
9D.4 The T -Test for Comparing Two Means
Cross -WDEXODWLRQVDQGWKHȤWHVWDUHDSSURSULDWHZKHQERWKYDULDEOHVDUHOHVV -than
interval level. However, researchers often want to compare one interval or ratio
level variable across categories of respondents. The Chapter Vignette describes
such a situation. The researchers are interested in comparing the ethical perceptions
between genders. When a researcher needs to compare means for a variable
grouped into two categories based on some less -than interval variable, a t -test is
appropriate. One way to think about this is testing the way a dichotomous (two -
level) independent variable is associated with changes in a continuous dependent
variable. Se veral variations of the t -test exist.
9D.4.1 Independent Samples T -Test
The researcher will apply the independent samples t -test, which tests the differences between means taken from two independent samples or groups. So, for
example, if we measure the pri ce for some designer jeans at 30 different retail
stores, of which 15 are Internet -only stores (pure clicks) and 15 are traditional
stores, we can test whether or not the prices are different based on store type with
an independent samples t -test. The t -test for difference of means assumes the two samples (one Internet and one traditional store) are drawn from normal distributions and that the variances of the two populations are approximately equal.
9D.4.2 Paired -Samples T -Test
What happens when means nee d to be compared that are not from independent
samples? Such might be the case when the same respondent is measured twice —
for instance, when the respondent is asked to rate both how much he or she likes
shopping on the Internet and how much he or she likes shopping in traditional
stores. Since the liking scores are both provided by the same person, the assumption
that they are independent is not realistic. Additionally, if one compares the prices
the same retailers charge in their stores with the prices the y charge on their Web sites, the samples cannot be considered independent because each pair of observations is from the same sampling unit. A paired -samples t -test is appropriate
in this situation. Researchers also can compute the paired samples t -test usi ng
statistical software. For example, using SPSS, the click -through sequence would
be:
$QDO\]Hĺ&RPSDUH0HDQVĺ3DLUHG -Samples t -test munotes.in
Page 221
221Chapter 9: Data Analyis and Presentation
A dialog box then appears in which the “paired variables” should be entered. When
a paired - samples t -test is appropriate, the two numbers being compared are usually
scored as separate variables.
9D.5 The Z -Test for Comparing Two Proportions
What type of statistical comparison can be made when the observed statistics are
proportions? Suppose a researcher wishes t o test the hypothesis that wholesalers in
the northern and southern United States differ in the proportion of sales they make
to discount retailers. Testing whether the population proportion for group 1 (p1)
equals the population proportion for group 2 (p2 ) conceptually the same as the t -
test of two means.
9D.6 Analysis Of Variance (ANOVA)
What is ANOVA?
Considering the scenario that if we want to test and see if employee turnover differs
across our five production plants? When the means of more than two gr oups or
populations are to be compared, one -way analysis of variance (ANOVA) is the
appropriate statistical tool. ANOVA involving only one grouping variable is often
referred to as one -way ANOVA because only one independent variable is involved.
Another wa y to define ANOVA is as the appropriate statistical technique to
examine the effect of a less -than interval independent variable on an at -least interval dependent variable. Thus, a categorical independent variable and a continuous dependent variable are in volved. An independent samples t -test can be
thought of as a special case of ANOVA in which the independent variable has only
two levels. When more levels exist, the t -test alone cannot handle the problem. The
statistical null hypothesis for ANOVA is state d as follows: µ1 = µ2 = µ3 = µ4 =
………= µn The symbol k is the number of groups or categories for an independent
variable. In other words, all group means are equal. The substantive hypothesis
tested in ANOVA is At least one group mean is not equal to anoth er group mean. As the term analysis of variance suggests, the problem requires comparing variances to make inferences about the mean
9D.6.1 Partitioning Variance in ANOVA
9D.6.1.1 Total Variability
An implicit question with the use of ANOVA is, “How can the dependent variable
best be predicted?” Absent any additional information, the error in predicting an
observation is minimized by choosing the central tendency, or mean for an interval munotes.in
Page 222
222 RESEARCH IN COMPUTING
variable. Fo r the coffee example, if no information was available about the work
shift of each respondent, the best guess for coffee drinking consumption would be
four cups. The Sum of Squares Total (SST) or variability that would result from
using the grand mean, mea ning the mean over all observations, can be thought of
as
SST = Total of (Observed Value – Grand Mean) 2
Although the term error is used, this really represents how much total variation
exists among the measures. Using the first observation, the error of observation
would be (1 cup - 4 cups) 2 = 9 The same squared error could be computed for each
observation and these squared errors totaled to give SST.
9D.6.1.2 Between -Groups Variance
ANOVA tests whether “grouping” observations explains variance in the dependent
variable. Given this additional information about which shift a respondent works,
the prediction changes. Now, instead of guessing the grand mean, the group mean
would be used. So, once we know that someone works the day shift, the prediction
would be that he or she consumes 2 cups of coffee per day. Similarly, the second
and night -shift predictions would be 4 and 6 cups, respectively. Thus, the between -
groups variance or Sum of Squares Between -groups (SSB) can be found by taking
the total sum of the weighted difference between group means and the overall mean
as shown:
SSB = Total of ngroup (Group Me an - Grand Mean)2
The weighting factor (ngroup ) is the specific group sample size. Let’s consider the
first observation once again. Since this observation is in the day shift, we predict 2
cups of coffee will be consumed. Looking at the day shift group o bservations, the
new error in prediction would be
(2 cups - 4 cups)2 = (2)2 = 4
The error in prediction has been reduced from 3 using the grand mean to 2 using
the group mean. This squared difference would be weighted by the group sample
size of 5, to yie ld a contribution to SSB of 20. Next, the same process could be
followed for the other groups yielding two more contributions to SSB. Because the
second shift group mean is the same as the grand mean, that group’s contribution
to SSB is 0. Notice that the night -shift group mean is also 2 different than the grand
mean, like the day shift, so this group’s contribution to SSB is likewise 20. The total SSB then represents the variation explained by the experimental or independent variable. In this case, total S SB is 40. munotes.in
Page 223
223Chapter 9: Data Analyis and Presentation
9D.6.1.3 Within -Group Error
Finally, error within each group would remain. Whereas the group means explain
the variation between the total mean and the group mean, the distance from the
group mean and each individual observation remains unexplai ned. This distance is
called within group error or variance or the Sum of Squares Error (SSE). The values
for each observation can be found by
SSE = Total of (Observed Mean - Group Mean) 2
Again, looking at the first observation, the SSE component would b e
SSE = (1 cup - 2 cups) 2 = 1 cup
This process could be computed for all observations and then totaled. The result
would be the total error variance —a name used to refer to SSE since it is variability
not accounted for by the group means. These three co mponents are used in
determining how well an ANOVA model explains a dependent variable.
9D.6.2 The F -Test
he F-test is the key statistical test for an ANOVA model. The F -test determines
whether there is more variability in the scores of one sample than in the scores of
another sample. The key question is whether the two sample variances are different
from each other or whether they are from the same population. Thus, the test breaks
down the variance in a total sample and illustrates why ANOVA is analysis of
variance. The F -statistic (or F -ratio) can be obtained by taking the larger sample
variance and dividing by the smaller sample variance. It is much like using the
tables of the Z - and t -distributions that we have previously examined. These tables
portra y the F -distribution, which is a probability distribution of the ratios of sample
variances. These tables indicate that the distribution of F is actually a family of
distributions that change quite drastically with changes in sample sizes. Thus,
degrees of freedom must be specified. Inspection of an F -table allows the researcher
to determine the probability of finding an F as large as a calculated F.
Summary
Bivariate data is when you are studying two variables . For example, if you are
studying a group of college students to find out their average SAT
score and their age, you have two pieces of the puzzle to find (SAT score and age).
Or if you want to find out the weights and heights of college students, then you also
have bivariate data. Bivariate data could also be two sets of items that are dependent munotes.in
Page 224
224 RESEARCH IN COMPUTING
on each other. It helps you to recognize when the bivariate statistical test is right to
use. The calculation and interpretation of an independent sample t -test which
compares two mean values. ANNOVA is one of the statistical technique which is
used to examine the effect of a less*than interval independent variable with three
or more categories on an at least interval independent variable.
Questions
1) Explain the t -Test which is used for comparing t wo means.
2) Write short note on ANOVA.
3) Discuss how does F -Test is the key statistical test for ANOVA Model.
4) Explain Total Variability.
5) Describe the significance of Paired -Samples t -Test.
References
x Albright Winson, 2015, Business Analytics, 5th Edition,
x Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
x Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
x Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
x Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN: 9788131316054, 813131605X
x William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016, Business Research Methods, Edition 8, Cengage Publication
munotes.in
Page 225
225Chapter 9: Data Analyis and Presentation
225E. MULTIVARIATE STATISTICAL ANALYSIS
Unit Structure
9E.0 Objective
9E.1 Introduction
9E.1.1 The “Variate” in Multivariate
9E.2 Classifying Multivariate Techniques
9E.2.1 Dependence Techniques
9E.2.2 Interdependence Techniques
9E.3 Analysis of Dependence
9E.3.1 Multiple Regression Analysis
9E.3.2 Steps in interpreting a multiple regression model
9E.3.3 ANOVA (n -Way) and MANOVA
9E.4.3.1 N-WAY (UNIVARIATE) ANOVA
9E.4.3.2 Interpreting MANOVA
9E.3.4 Discriminant Analysis
9E.4 Analysis of Interdependence
9E.4.1 Factor Analysis -
9E.4.1.1 How many factors
9E.4.1.2 Factor loadings
9E.4.1.3 Factor rotation
9E.4.1.4 Data reduction technique
9E.4.1.5 Creating composite scales with factor results -
9E.4.1.6 Communality
9E.4.1.7 Total Variance Explained
9E.5 Cluster Analysis
9E.6 Multidimensional Scaling
9E.7 Summary
9E.8 Questions
9E.9 References
munotes.in
Page 226
226 RESEARCH IN COMPUTING
9E.0 Objectives
• To understand what Multivariate Statistical Technique is and its types
• To learn how cluster analysis can identify market segments.
• To analyze and Interpret results from multivariate analysis of variance
(MANOVA)
• To study how Cluster Analysis classifies multiple observations into a smaller
number of mutually exclusive and exhaustive groups
• To learn the significance of Multidimensional Scaling
9E.1 Introduction Multivariate statistical analysis refers to multiple advanced techniques for examining relationships among multiple variable s at the same time . Multivariate
procedures are used in research where there are more than one dependent variable,
more than one independent variable or both. Research areas which involve multivariate data analysis including most employee motivation rese arch and
research that seeks to identify viable market segments.
9E.1.1 The “Variate” in Multivariate
To list another characteristic of multivariate analysis is the variate. The variate is a
mathematical way in which a set of variables can be represented with one e quation.
A variate is formed as a linear combination of variables, each contributing to the
overall meaning of the variate based upon an empirically derived weight. To define
it in terms of mathematics the variate is a function of the measured variables
involved in an analysis:
Vk = f (X1, X2, . . . , Xm)
Vk is the k th variate.
Every analysis could involve multiple sets of variables, each represented by a
variate. X1 to Xm represent the measured variables. Suppose we measured nostalgia with five question s during the survey process . With these five variables,
a variate of the following form could be created:
Vk = L1X1 + L2X2 + L3X3 + L4X4 + L5X5 munotes.in
Page 227
227Chapter 9: Data Analyis and Presentation
Vk represents the score for nostalgia, X1 to X5 represent the observed scores on
the five scale items (survey questions) that are expected to indicate nostalgia, and
L1 to L5 are parameter estimates much like regression weights that suggest how
highly related each variable is to the overall nostalgia score.
9E.2 Classifying Multivariate Techniques
There are two groups of multivariate techniques :
1) Dependence methods
2) Interdependence methods.
9E.2.1 Dependence Techniques When hypotheses involve distinction between independent and dependent variables, dependence techniques are needed. For instance, w hen we hypothesize
that nostalgia is related positively to purchase intentions, nostalgia takes on the
character of an independent variable and purchase intentions take on the character
of a dependent variable.
9E.2.2 Interdependence Techniques
When rese archers examine questions that do not distinguish between independent
and dependent variables, interdependence techniques are used. No one variable or
variable subset is to be predicted from or explained by the others. The most
common interdependence methods are factor analysis, cluster analysis, and multidimensional scaling.
9E.3 Analysis of Dependence
Multivariate dependence techniques are variants of the general linear model
(GLM) . Simply, the GLM is a way of modeling some process based on how different variables cause fluctuations from the average dependent variable. Fluctuations can come in the form of group means that differ from the overall mean
as in ANOVA or in the form of a significant slope coefficient as in regression. The
basic idea ca n be thought of as follows:
Yˆi = ȝǻX ǻF ǻXF
Here, ȝrepresents a constant, which can be thought of as the overall mean of the
dependent YDULDEOHǻ X DQGǻF represent changes due to main effect independent
variables (such as experimental variables) and blocking independent variables munotes.in
Page 228
228 RESEARCH IN COMPUTING
VXFKDVFRYDULDWHVRUJURXSLQJYDULDEOHVUHVSHFWLYHO\DQGǻ XF represents the
change due to the combination (interaction effect) of those variables. Realize that
Yi in this case could represent multiple dependent variab les, just as X and F could
represent multiple independent variables.
9E.3.1 Multiple Regression Analysis
Multiple regression analysis is an extension of simple regression analysis allowing a metric dependent variable to be predicted by multiple independent
variables. Thus, one dependent variable (sales volume) is explained by one independent variable (number of building permits). The independent variables
include price, seasonality, interest rates, advertising intensity, consumer income,
and other economic factors in the area. The simple regression equation can be
expanded to represent multiple regression analysis:
Yi = b0 + b1X1 + b2X2 + b3X3 + . . . + bnXn + ei
Thus, as a form of the GLM, dependent variable predictions ( Yˆ ) are made by
adjusting the constant (bo), which would be equal to the mean if all slope
coefficients are 0, based on the slope coefficients associated with each independent
variable ( b1, b2, . . . , bn).Less -than interval (nonmetric) independent variables can
be used in multiple regression. This can be done by implementing dummy variable
coding. A dummy variable is a variable that uses a 0 and a 1 to code the different
levels of dichotomous variable. Multiple dummy variables can be included in a
regression model. For example, dummy coding is appropriate when data from two
countries are being compared. Suppose the average labor rate for automobile
production is included in a sample taken from respondents in the United States and
in South Korea. A response fro m the United States could be assigned a 0 and
responses from South Korea could be assigned a 1 to create a country variable
appropriate for use with multiple regression.
9E.3.2 Steps in Interpreting a Multiple Regression Model
Multiple regression models often are used to test some proposed theoretical model.
A researcher may aske to develop and test a model explaining business unit
performance. Why do some business units outperform others? Multiple regression
models can be inter preted using these steps:
1. Examine the model F-test. If the test result is not significant, the model should
be dismissed and there is no need to proceed to further steps.
2. Examine the individual statistical tests for each parameter estimate. An independent variable with significant results can be considered a significant munotes.in
Page 229
229Chapter 9: Data Analyis and Presentation
explanatory variable. If an independent variable is not significant, the model
should be run again with no significant predictors deleted.
3. Examine the model R2. No cutoff values exist that can distinguish an acceptable amount of explained variation across a ll regression models. T he
absolute value of R2 is more important when the researcher is more interested
in prediction than in explanation.
4. Examine collinearity diagn ostics. Multicollinearity in regression analysis
refers to how strongly interrelated the independent variables in a model are.
When multicollinearity is too high, the individual parameter estimates become
difficult to interpret.
9E.3.3 ANOVA (n -Way) and MANOVA
Regression is a form of the GLM with a single continuous dependent measure and continuous independent measure(s). An ANOVA or MANOVA model also represents a form of the GLM. ANOVA can be extended beyond one -way ANOVA
to predict a continuous dependent variable with multiple categorical independent variables. Multivariate analysis of variance (MANOVA), is a multivariate technique that predicts multiple continuous dependent variables with multiple
independent variables. The independent variables are categorical, although a continuous control variable can be included in the form of a covariate.
9E.3.3.1 N-Way (Univariate ) ANOVA
The interpretation of an n -way ANOVA model follows closely from the regression
results. The steps involved a re essentially the same with the addition of interpreting
differences between means:
1. Examine the overall model F -test result. If significant, proceed.
2. Examine individual F -tests for each individual independent variable.
3. For each significant categorical independent variable, interpret the effect by
examining the group means.
4. For each significant continuous variable (covariate), inte rpret the par ameter
estimate .
5. For each significant interaction, interpret the means for each combination.
9E.3.3.2 Interpreting Manova
Compared to ANOVA, a MANOVA model produces an additional layer of testing.
The first layer of testing involves the multivariate F -test, which is based on a
VWDWLVWLF FDOOHG :LONH¶V /DPEGD ȁ 7KLV WHVW H[DPLQHV ZKHWKHU RU QRW DQmunotes.in
Page 230
230 RESEARCH IN COMPUTING
independent variable expla ins significant variation among the dependent variables
within the model. If this test is significant, then the F -test results from individual
univariate regression models nested within the MANOVA model are interpreted.
The rest of the interpretation resul ts follow from the one -way ANOVA or multiple
regression model results above.
9E.3.4 Discriminant Analysis
During the research process the Researchers often need to produce a classification
of sampling units. This process may involve using a set of indepe ndent variables to
decide if a sampling unit belongs in one group or another. A physician might record
a person’s blood pressure, weight, and blood cholesterol level and then categorize
that person as having a high or low probability of a heart attack. A r esearcher
interested in retailing failures might be able to group firms as to whether they
eventually failed or did not fail on the basis of independent variables such as location, financial ratios, or management changes. A bank might want to discriminate between potentially successful and unsuccessful sites for electronic
fund transfer system machines . The challenge is to find the discriminating variables
to use in a predictive equation that will produce better than chance assignment of
the individuals to t he correct group. Discriminant analysis is a multivariate
technique that predicts a categorical dependent variable (rather than a continuous,
interval -scaled variable, as in multiple regression) based on a linear combination
of independent variables. In ea ch problem above, the researcher determines which
variables explain why an observation falls into one of two or more groups. A linear
combination of independent variables that explains group memberships is known
as a discriminant function. Discriminant ana lysis is a statistical tool for determining
such linear combinations. The researcher’s task is to derive the coefficients of the
discriminant function (a straight line). We will consider an example of the two -
group discriminant analysis problem where the d ependent variable, Y, is measured
on a nominal scale. (Although n -way discriminant analysis is possible, it is beyond
the scope of this discussion.) Suppose a personnel manager for an electrical
wholesaler has been keeping records on successful versus unsu ccessful sales
employees. The personnel manager believes it is possible to predict whether an
applicant will succeed on the basis of age, sales aptitude test scores, and mechanical
ability scores. As stated the problem is to find a linear function of the i ndependent
variables that shows large differences in group means. The first task is to estimate
the coefficients of the applicant’s discriminant function. To calculate the
individuals’ discriminant scores, the following linear function is used:
Zi = b1X1i + b2X2i + · · · + bnXni munotes.in
Page 231
231Chapter 9: Data Analyis and Presentation
where
Zi = ith applicant’s discriminant score
bn = discriminant coefficient for the nth variable
Xni = ith applicant’s value on the nth independent variable
Using scores for all the individuals in the sample, a discriminant function is
determined based on the criterion that the groups be maximally differentiated on
the set of independent variables.
9E.4 Analysis of Interdependence
9E.4.1 Factor Analysis
Factor analysis is a prototypical multivariate, interdependence technique. Factor
analysis is a technique of statistically identifying a reduced number of factors from
a larger number of measured variables. Factor analysis can be divided into two
types:
1. Exploratory factor analysis (EFA) —performed when the researcher is uncertain about how many factors may exist among a set of variables. The
discussion here concentrates primarily on EFA.
2. Confirmatory factor analysis (CFA) —performed when the researcher has strong theoretical expectations about the factor structure (number of factors and
which variables relate to each factor) before performing the analysis. CFA is a
good tool for assessing construct validity because it provides a test of how well
the researcher’s “theory” about the factor structure fits the actual observations.
Many books exist on CFA alone and the reader is advised to refer to any of
those sources for more on CFA. Suppose a researcher is asked to examine how
feelings of nostalgia in a restaurant influence customer loyalty. Three hundred
fifty customers at themed restaurants around the country are interviewed and
asked to respond to the following
Likert scales (1 = Strongly Disagree to 7 = Strongly Agree):
X1—I feel a strong connection to the past when I am in this place.
X2—This place evokes memories of the past.
X3—I feel a yearning to relive past experiences when I dine here.
X4—This place looks like a page out of the past. munotes.in
Page 232
232 RESEARCH IN COMPUTING
X5—I am willing to pay more to dine in this restaurant.
X6—I feel very loyal to this establishment.
X7—I would recommend this place to others.
X8—I will go out of my way to dine here.
Factor analysis can summarize the information in the eight variables in a smaller
number of variables. How many dimensions, or groups of variables, are likely
present in this case? More than one technique exists for estimating the variates that
form the factors. However, the general idea is to mathematically produce variates
that explain the greatest total variance among the set of vari ables being analyzed.
Thus, EFA provides two important pieces of information:
1. How many factors exist among a set of variables?
2. What variables are related to or “load on” which factors?
9E.4.1.1 How Many Factors
One of the first questions the researcher asks is, “How many factors will exist
among a large number of variables?” While a detailed discussion is beyond the
scope of this text, the question is usually addressed based on the eigenvalues for a
factor solution. Eigen values are a measure of how much variance is explained by
each factor. The most common rule —and the default for most statistical programs —is to base the number of factors on the number of eigenvalues greater
than 1.0. The basic thought is that a factor with an Eigen value of 1 .0 has the same
total variance as one variable.
9E.4.1.2 Factor Loadings
Each arrow connecting a factor (represented by an oval) to a variable (represented
by a box) is associated with a factor loading. A factor loading indicates how
strongly correlated a measured variable is with that factor. Loading estimates are
provided by factor analysis programs.
9E.4.1.3 Factor Rotation -
Factor rotation is a mathematical way of simplifying factor results. The most
common type of factor rotation is a process calle d varimax. Rotation “clears things
up” by producing more obvious patterns of loadings. Users can observe this by
looking at the unrotated and rotated solutions in the factor analysis output.
9E.4.1.4 Data Reduction Technique
Factor analysis is considered a data reduction technique. Data reduction techniques
allow a researcher to summarize information from many variables into a reduced munotes.in
Page 233
233Chapter 9: Data Analyis and Presentation
set of variates or composite variables. Data reduction is advantageous for many
reasons. Fact or analysis accomplishes data reduction by capturing variance from
many variables with a single variate. Data reduction is also a way of identifying
which variables among a large set might be important in some analysis. Thus, data
reduction simplifies deci sion making .
9E.4.1.5 Creating Composite Scales with Factor Results
When a clear pattern of loadings exists as in this case, the researcher may take a
simpler approach. F1 could be created simply by summing the four variables with
high loadings and creatin g a summated scale representing nostalgia. F2 could be
created by summing the second four variables (those loading highly on F2) and
creating a second summated variable. While not necessary, it is often wise to divide
these summated scales by the number of items so the scale of the factor is the same
as the original items. For example,
F1 would be ((X1 + X2 + X3 + X4 )/4)
The result provides a composite score on the 1 –7 scale. The composite score
approach would introduce very little error given the patter n of loadings. In other
words, very low loadings suggest a variable does not contribute much to the factor.
The reliability of each summated scale can be tested by computing a coefficient
alpha estimate. Then, the researcher could conduct a bivariate regre ssion analysis
that would test how much nostalgia contributed to loyalty.
9E.4.1.6 Communality
While factor loadings show the relationship between a variable and each factor, a
researcher may also wish to know how much a single variable has in common wit h
all factors. Communality is a measure of the percentage of a variable’s variation
that is explained by the factors. A relatively high communality indicates that a
variable has much in common with the other variables taken as a group. A low
communality me ans that the variable does not have a strong relationship with the
other variables.
9E.4.1.7 Total Variance Explained
Along with the factor loadings, the percentage of total variance of original variables
explained by the factors can be useful. Recall that common variance is correlation
squared. Thus, if each loading is squared and totaled, that total divided by the
numbe r of factors provides an estimate of the variance in a set of variables
explained by a factor. This explanation of variance is much the same as R 2 in
multiple regression. Again, these values are computed by the statistics program so
there is seldom a need to compute them manually. In this case, though, the variance
accounted for among the eight variables by the nostalgia factor is 0.36 and the
variance among the eight variables explained by the loyalty factor is 0.35. Thus, munotes.in
Page 234
234 RESEARCH IN COMPUTING
the two factors explain 71 perce nt of the variance in the eight variables: 0.36 + 0.35
= 0.71 In other words, the researcher has 71% of the information in two factors that
are in the original eight items, another example of the rule of parsimony.
9E.5 Cluster Analysis
Cluster analysis i s a multivariate approach for identifying objects or individuals
that are similar to one another in some respect. Cluster analysis classifies
individuals or objects into a small number of mutually exclusive and exhaustive
groups. Objects or individuals are assigned to groups so that there is great similarity
within groups and much less similarity between groups. The cluster should have
high internal (within -cluster) homogeneity and high external (between -cluster)
heterogeneity. Cluster analysis is an import ant tool for the business researcher. For
example, an organization may want to group its employees based on their insurance
or retirement needs, or on job performance dimensions. Similarly, a business may
wish to identify market segments by identifying sub jects or individuals who have
similar needs, lifestyles, or responses to marketing promotions. Clusters, or
subgroups, of recreational vehicle owners may be identified on the basis of their
similarity with respect to recreational vehicle usage and the bene fits they want from
recreational vehicles.
9E.6 Multidimensional Scaling
Multidimensional scaling provides a means for placing objects in multidimensional
space on the basis of respondents’ judgments of the similarity of objects. The
perceptual differenc e among objects is reflected in the relative distance among objects in the multidimensional space. In the most common form of multidimensional scaling, subjects are asked to evaluate an object’s similarity to
other objects.
Summary
Multivariate statistic al methods analyze multiple variables or even multiple sets of
Variables simultaneously. They are particularly useful for identifying latent constructs using multiple individual measures. Multiple regression analysis predicts a continuous dependent variable with multiple independent variables.
MANOVA is an extension of ANOVA i nvolving multiple related dependent
variables. Discriminant analysis uses multiple independent variables to classify
observations into one of a set of mutually exclusive categories. Cluster analysis
classifies multiple observations into a smaller number of mutually exclusive and
exhaustive groups. munotes.in
Page 235
235Chapter 9: Data Analyis and Presentation
Questions
1) Define multivariate statistical analysis.
2) What is the variate in multivariate? What is an example of a variate in
multiple regressions and in factor analysis?
3) What is the distinction between dependence techniques and interdependence
techniques?
4) What is GLM? How can multiple regression and n -way ANOVA be
described as GLM approaches?
5) What are the steps in interpreting a multiple regression analysis result?
References
• Albright Winson, 2015, Business Analytics, 5th Edition,
• Hair, 2014, Multivariate data Analysis, 7th Edition, Pearson Publication
• Kabir, Syed Muhammad. (2016). Measurement Concepts: Variable, Reliability, Validity, and Norm.
• Mark Saunders, 2011, Research Methods for Business Students, 5th Edition
• Shefali Pandya, 2012, Research Methodology, APH Publishing Corporation, ISBN: 9788131316054, 813131605X
• William G.Zikmund, B.J Babin, J.C. Carr, Atanu Adhikari, M.Griffin,
2016, Business Research Methods, Edition 8, Cengage Publication
munotes.in