Where a Hypothesis Comes From, and What Makes It Usable
Chapter -Nine
Syllabus topic 6, "Identification of Research Problem and formulation of Hypothesis."
Pages 432 to 434 of 543
In one line
A hypothesis is derived rather than invented, and it is usable only if it is clear, specific, testable, related to what is known and within reach of the evidence the study will actually have.
In the wording a student can write in an exam: a hypothesis may be derived from theory, from the findings of earlier research, from the researcher's own observation or professional experience, from analogy with another field or another jurisdiction, from the express or implied assumptions of the law itself, and from the questions left open by decided cases; to be usable it must be conceptually clear, specific rather than general, empirically testable, related to existing knowledge, and capable of being tested by the techniques and data available to the researcher.
The six sources
Theory. A general proposition about how things work, from which a specific expectation follows. That a right without machinery is less effective than a right with it, applied to the fifteen years between Hoskot and the commencement of the Act, chapter 550.
Earlier research. Somebody has found something elsewhere, and the hypothesis is that the same holds here, or that it does not.
Observation and professional experience. A student who has sat in a clinic notices that people arrive without documents, chapter 1080's seventh source.
Analogy. From another field, another jurisdiction or another institution. Chapter 880's comparative method supplies these.
The law's own assumptions. This is the source most useful in legal research and the one most often missed. Every rule assumes something about the world, chapter 900, and each assumption is a hypothesis somebody could test. Section 13(2) assumes that an affidavit as to income is obtainable; regulation 13 of the 2011 Regulations assumes a signboard will be read; Khatri assumes a magistrate can inform an accused within the time a remand production takes.
And open questions in decided cases, chapter 1080's fourth source.
Stating the derivation
A hypothesis must arrive with its reasons. A dissertation that states an expectation without saying where it came from has produced a guess, chapter 1120.
The derivation belongs in the text, in a sentence or two, immediately after the hypothesis: this is expected because of X, Y and Z.
And it does work later. If the hypothesis is disproved, chapter 1170, the derivation tells the reader which of X, Y or Z was wrong, which is where the interesting part of the finding usually lies.
The five characteristics of a usable hypothesis
One: conceptually clear. Every term has a definite meaning, and the key term has an operational definition, chapter 1050. Aware, effective, adequate and satisfactory are not clear until defined.
Two: specific. It says how much, how many, or in what direction. Not that awareness is low, but that fewer than half of the respondents will have heard of free legal services.
Where a Hypothesis Comes From, and What Makes It Usable
Three: testable. Something would refute it, chapter 1120.
Four: related to existing knowledge. It follows from a source and does not contradict established knowledge without saying so.
Five: within reach of the available techniques and data. A hypothesis requiring evidence the researcher cannot obtain is a promise the dissertation will break, chapter 1090.
And a sixth, worth adding for legal work: it must be about something that matters. A precisely testable hypothesis about a trivial magnitude is a well made instrument pointed at nothing.
Turning a vague expectation into a usable hypothesis
Start with what the researcher actually believes. People do not know about legal aid.
Ask: who, where, when? Adults in three named settlements, in the study period.
Ask: how much? Fewer than half.
Ask: what exactly is the thing? Having heard that a lawyer can be provided free of cost by the Government, which is the operational definition and is question 5 of the schedule at chapter 970.
Ask: what would refute it? A proportion of half or more.
And ask: why do I expect it? Because the entitlement under section 12 is very wide and the number of applications recorded by the District Authority is small, chapter 640, and because no publicity of the kind regulation 15 of the 2011 Regulations contemplates was found in these settlements.
The result. Fewer than half of the adult respondents in the three study settlements will report having heard that a lawyer can be provided free of cost by the Government; expected because the statutory entitlement is wide while recorded applications are few, and because no local publicity was observed.
Five questions, and a vague belief has become a usable hypothesis with its derivation attached.
The common defects
Too vague. No operational definition, so nothing can be counted.
Too general. About India, when the study is about one taluk.
Unfalsifiable. No possible result would disturb it.
Value-laden. Contains a word like should, adequate or fair, which imports a standard, chapter 1050.
Untestable with the available data, chapter 1090.
Two propositions in one. That awareness is low and that it is lower among women. Those are two hypotheses and may have different fates, so they must be separated.
And a hypothesis that merely restates a definition, which cannot be false and therefore cannot be tested.
A worked example, three attempts
Attempt one. Legal aid is not reaching the poor. Vague, general, value-laden and unfalsifiable.
Attempt two. Awareness of legal aid is low in rural Maharashtra. Better, but general, and low is undefined.
Attempt three. Fewer than half of the adult respondents in the three study settlements will report having heard that a lawyer can be provided free of cost by the Government.
Where a Hypothesis Comes From, and What Makes It Usable
Check attempt three against the five characteristics. Conceptually clear, with the thing defined by the words of question 5. Specific, giving a threshold. Testable, refuted by a proportion of half or more. Related to existing knowledge, since the derivation cites the width of section 12 against the number of applications. And within reach, since the schedule of chapter 970 produces exactly this number.
And add the relational companion. The proportion reporting awareness will be lower among women respondents than among men. Two propositions, separated, each with its own fate.
Quick revision
Six sources: theory; earlier research; observation and professional experience; analogy, including comparative; the law's own assumptions, which is the source most useful in legal research; and open questions in decided cases.
State the derivation in a sentence after the hypothesis, because if the hypothesis fails the derivation tells the reader which premise was wrong.
Five characteristics: conceptually clear with an operational definition; specific, saying how much or in what direction; testable; related to existing knowledge; and within reach of the available techniques and data. And a sixth for legal work: about something that matters.
Five questions that convert a belief into a hypothesis: who, where and when; how much; what exactly is the thing; what would refute it; and why do I expect it.
Defects: too vague; too general; unfalsifiable; value-laden; untestable with the available data; two propositions in one; and a restatement of a definition.
Test yourself
1. Name the six sources of a hypothesis and say which is most useful in legal research. Theory; the findings of earlier research; the researcher's own observation and professional experience; analogy with another field or jurisdiction; the express or implied assumptions of the law itself; and questions left open by decided cases. The law's own assumptions is the most useful in legal research, because every rule assumes something about the world and each assumption is a testable proposition.
2. Give the five characteristics of a usable hypothesis. Conceptual clarity, with an operational definition of the key term; specificity, stating how much or in what direction; testability, so that something would refute it; relation to existing knowledge, so that it is derived rather than guessed; and testability with the techniques and data actually available to the researcher.
3. Set out the five questions that turn a vague belief into a hypothesis. Who, where and when; how much; what exactly is the thing being measured; what result would refute it; and why the researcher expects it, which supplies the derivation.
4. Why must two propositions never be combined in one hypothesis? Because they may have different fates. If the hypothesis asserts both that awareness is low and that it is lower among women, evidence may support one and refute the other, and a single combined proposition cannot record that result.
The rest of this subject
These notes are cut from the University's printed syllabus. Open the syllabus itself, or the past papers, for the same subject.