LLM Foundation Course Legal Education and Research Methodology 2019 Question Paper with Solutions

For full marks, cover: the difference between a topic and a problem; the steps of identification and then of formulation, kept separate because the question separates them; the criteria a formulated problem must satisfy; then the second limb on induction and deduction, which is shorter but must be answered; and a conclusion joining the two.

A research problem is a stated difficulty or unanswered question about which existing knowledge is unsatisfactory and which can be investigated with available methods. The distinction from a topic decides the answer. "Legal aid" is a topic. "What proportion of persons granted free legal services under section 12 of the Legal Services Authorities Act, 1987, in a given district withdrew before the matter was decided, and why" is a problem, because it names a difficulty, identifies the material and admits of an answer that could come out either way.

The steps of identification come first.

4. Discuss it with those who know the field, that is the supervisor, practitioners and experienced researchers. A question nobody in the field regards as open is usually not open.
5. Make a preliminary survey of the material available, to find out whether the sources exist and can be reached. A problem for which no data can be obtained is not a problem the researcher can work on.

The steps of formulation follow, and they are a different exercise.

1. State the problem in a single interrogative sentence. If it cannot be put as one question, it is more than one problem.
2. Narrow the scope by fixing the population, the place and the period. Every unqualified problem is too large.
3. Define the key terms operationally, that is in terms of what will actually be observed or counted. "Effective legal aid" must become something measurable before anything can be measured.
4. State the objectives as a short list of what the study will establish.
5. Frame the hypothesis, or, in doctrinal and exploratory work, the research questions.

Deduction moves from a general rule to a particular conclusion. In law it takes the form of the syllogism: the rule is the major premise, the found fact the minor, and the conclusion follows necessarily. Its strengths are certainty, economy and testability. In research it supplies the hypothesis, by deducing from an accepted theory a consequence that ought to be observable. Its limits are that it produces nothing the premise did not contain, that it depends entirely on the soundness of the major premise, and that it cannot decide a case the rule does not reach.

Induction moves from particular instances to a general proposition. It is how a principle is stated from a line of decisions, and how a finding is stated from a body of data. The Supreme Court reasoned inductively in M.C. Mehta v. Union of India, (1987) 1 SCC 395, reviewing the authorities on strict liability and stating the wider rule of absolute liability for hazardous enterprise. Its strengths are that it generates new propositions and keeps law tied to fact; its weakness is that the conclusion is probable and never certain, and is only as good as the sample, which in legal research is usually reported cases.

The topic is default bail. The difficulty noticed on reading is that the Supreme Court has repeatedly described the right to default bail as indefeasible once the investigation period expires, and nobody has examined what the trial courts actually do on the day it expires. The problem stated is: in the court of a named Chief Judicial Magistrate, in matters instituted between two stated dates under the Bharatiya Nagarik Suraksha Sanhita, 2023, in what proportion of cases in which the accused became entitled to default bail was the entitlement given effect to at the first hearing after the period expired, and what explains the cases in which it was not? The variables are then defined operationally, "given effect to" meaning an order releasing the accused on bail on that date, and the hypothesis follows: the entitlement is given effect to less often where the accused was unrepresented at that hearing than where the accused was represented.

Notice what the narrowing has achieved. The universe is defined so a sample can be drawn; the outcome is defined so two researchers would count the same events; the source of the data is identified so feasibility can be tested in advance; and the proposition could fail, which is what makes it research.

Is it a question? If it cannot be written as a single interrogative sentence it is more than one problem, and it should be split and one of them chosen.

Is the population named? Every unqualified problem is too large. "Aided litigants" becomes "applicants granted legal services by the District Legal Services Authority of a named district in a named year".

Are the terms operationally defined? Not "effective" but "disposed of on merits within eighteen months". Until the concepts are countable nothing can be tested.

Can the data be obtained, lawfully and in time? Court records, prisons and police stations need permissions, and personal data now attracts the Digital Personal Data Protection Act, 2023, read against Justice K.S. Puttaswamy v. Union of India, (2017) 10 SCC 1.

Is the answer unknown? This is settled by the literature review and by nothing else.

Could it come out either way? A problem whose answer is already contained in its wording is a conclusion.

The one test that disposes of a bad problem

Ask what finding would make the researcher abandon the proposition. If nothing comes readily to mind, the problem is not researchable, however interesting it is, and the study will produce a defence of a belief rather than a test of a claim. That single question catches the topic dressed as a question, the evaluative question no observation would settle, and the loaded question framed so that only one answer is respectable, and it catches all three before a single source has been collected.

Conclusion. Identifying a research problem is a matter of reading until a difficulty is recognised, and formulating it is a matter of reducing that difficulty to one answerable question with defined terms, a fixed population and honest limits. The two methods of reasoning run through both stages: induction produces the sense that there is a problem and the tentative principle behind it, deduction converts that principle into a prediction that the study can test. A problem formulated without that discipline yields a dissertation that describes a subject, and a subject is not a problem.

Continuing legal education is the structured professional learning a lawyer or a law teacher undertakes after qualification, for as long as they practise or teach. It is distinct from the law degree, which is preparation for entry, and from ordinary experience, which is unstructured. In most mature professions it is compulsory and audited: medicine requires continuing medical education credits, and in several common law jurisdictions a practising certificate depends on a minimum number of hours a year.

Prof. N.R. Madhava Menon, the founding Director of the National Law School of India University established at Bangalore in 1987, is the natural author of the criticism in the question, because his own work was directed at exactly this gap, and the criticism is justified for four reasons that should be stated before any reform is proposed.

First, there is no general obligation. The Advocates Act, 1961 conditions enrolment on a degree and, since the Constitution Bench decision in Bar Council of India v. Bonnie Foi Law College, 2023 SCC OnLine SC 130, decided 10 February 2023, on passing the All India Bar Examination. Nothing in the Act requires an advocate to learn anything after that. An advocate enrolled in 1990 may practise in 2026 without ever having attended a single organised programme.

The three grounds the question supplies are worth developing in a line each. For professional development, continuing education is the only mechanism by which a practitioner acquires competence in a field entered after qualification. For better delivery of legal services, an incompetent lawyer harms a client who has no way of measuring competence before instructing, which is the classic justification for professional regulation. For accountability, a profession that regulates entry and discipline but never re-examines competence is asking the public to take continued competence on trust.

The reforms should be concrete, and this is where the marks are.

1. Make it compulsory and measured, in hours or credits. A minimum annual requirement, tied to the renewal of the certificate of practice, is the only reform that changes behaviour. Any credit scheme must recognise the difference between a rural practitioner and a metropolitan one, and the requirement should be modest enough to be met.

5. Fix the content to what has actually changed. Priority modules on the three new criminal statutes and the transition between the old and new provisions; on mediation practice under the Mediation Act, 2023; on data protection compliance; on electronic evidence; on court technology and e-filing; and on the use and the dangers of artificial intelligence tools in legal work.

6. Include ethics and mental health as standing components, since the disciplinary jurisdiction under Chapter V of the Advocates Act deals with the consequences and continuing education is the only place the causes can be addressed.

7. Build a parallel scheme for law teachers, run by the national law universities and the Indian Law Institute and directed at legal pedagogy: clinical supervision, the case method, assessment design and research supervision.

8. Tie a part of it to legal aid. Credit for supervised legal aid work under the Legal Services Authorities Act, 1987 would serve continuing education and Article 39A at once.

The natural statutory home is the Advocates Act, 1961 itself. Section 7(1)(h) charges the Bar Council of India with promoting legal education and laying down standards, and section 49(1) carries the general rule-making power. The extent of that power was contested and is now settled in the Council's favour: in V. Sudeer v. Bar Council of India, (1999) 3 SCC 176, the Court held that the Council could not add pre-enrolment training as a condition of enrolment because section 24(1) is exhaustive, and in Bar Council of India v. Bonnie Foi Law College, 2023 SCC OnLine SC 130, decided 10 February 2023, a Constitution Bench overruled V. Sudeer and upheld the Council's power to prescribe an examination for entry. A continuing education requirement attached to the renewal of the certificate of practice, rather than to enrolment, sits more comfortably within section 49(1)(ah), which speaks to the conditions subject to which an advocate shall have the right to practise, and that is the drafting point on which a scheme would stand or fall.

Three objections must be met, because the Bar will raise them.

What other professions do, and the content this scheme would have to carry

A reforms answer is stronger for showing that the model exists elsewhere and for saying what the modules would actually contain.

Continuing professional development is compulsory and audited in most regulated professions, medicine being the obvious Indian comparison, where practitioners are expected to earn credits for accredited programmes as a condition of continued registration. Several common law jurisdictions attach a minimum number of hours a year to the renewal of a practising certificate for lawyers. The Indian legal profession is the outlier in requiring nothing at all after enrolment, and after Bar Council of India v. Bonnie Foi Law College, 2023 SCC OnLine SC 130, upheld an examination at the point of entry, the absence of any requirement after it is harder to defend than it was.

The content is where a reform proposal becomes concrete, and the priorities in India are unusually easy to identify because the statute book has moved.

And artificial intelligence, both its use and its danger, since generative tools produce fluent text and will cite decisions that do not exist.

For law teachers a parallel scheme is needed in legal pedagogy, that is clinical supervision, the case method, assessment design and research supervision, because UGC orientation and refresher courses are generic and rarely address any of them.

Conclusion. The criticism in the question is accurate: Indian legal education ends at the degree and, since 2023, at the Bar examination, and nothing after that requires a practitioner or a teacher to learn anything again. The case for changing that is strongest now, because the criminal law, the law of dispute resolution and the law of data have all been rewritten within the last three years. The reform that matters is a compulsory, credited, accredited and hybrid scheme with a statutory basis in the Advocates Act, priced and delivered so that the district practitioner can meet it, with a parallel scheme for law teachers. Anything voluntary will continue to be attended by those who need it least.

• Descriptive hypothesis. Asserts the existence, size or distribution of a single variable. Example: more than half of the applications filed under the Protection of Women from Domestic Violence Act, 2005, in a given court are filed together with proceedings under another enactment.
• Relational hypothesis. Asserts an association between two variables. Example: the time taken to dispose of a maintenance application is associated with whether the applicant was legally represented.
• Causal hypothesis. Asserts that one variable produces a change in another, which requires temporal order and the elimination of competing explanations, not merely association. Example: the introduction of a statutory time limit for filing a written statement reduced the average time to framing of issues.
• Null hypothesis. Asserts that no relationship exists, and it is the form in which statistical testing is actually carried out, because a null can be rejected by evidence while a positive proposition can only be supported. Example: there is no difference in the rate of compliance with maintenance orders between salaried and self employed respondents.
• Alternative hypothesis. What stands if the null is rejected.

In empirical legal research the formulation follows the ordinary social science route, adapted to legal material. The researcher narrows the problem until it names a population, a place and a period. The literature is reviewed, both for what is established and for the variables other researchers have found to matter. The variables are then defined operationally, which in legal research is the step most often skipped: "delay", "access to justice" and "effective legal aid" cannot be tested until they are converted into something countable, such as days from institution to disposal, or the proportion of applicants who obtained a lawyer within thirty days. The relation is then stated, first in working form and then in null form.

The sources from which a legal hypothesis is properly derived should be named, because a hypothesis produced from nothing is a guess. They are: existing theory and juristic writing; the reasoning of decided cases, particularly where the courts have assumed a factual proposition; Law Commission and committee reports, which characteristically identify what has not been examined; official data such as the National Judicial Data Grid, the National Crime Records Bureau and the annual reports of the legal services authorities; a pilot study; and the researcher's own practical observation, which is legitimate provided it is then tested rather than assumed.

Whichever kind of research it is, the formulated hypothesis is checked against the characteristics of a workable one. It must be conceptually clear, with every term defined; empirically testable, so that some evidence would defeat it; specific as to population, place and period; feasible with the data and techniques available, which in India now includes whether personal data may lawfully be collected under the Digital Personal Data Protection Act, 2023; grounded in a body of theory, so the result can be connected to what is known; simple enough to be tested by one researcher; consistent with established facts unless a reason is given; capable of statement in null form where the study is quantitative; and neutral in wording, because a hypothesis written so that only one answer is respectable will produce that answer.

The exception completes the answer. Exploratory or formulative studies of a field about which too little is known to predict anything, and purely doctrinal expositions, may properly proceed without a hypothesis at all. The discipline the hypothesis would have supplied must then come from a clearly stated research question, and the report should say so plainly rather than manufacture a hypothesis to satisfy a chapter heading.

Stating a proposition no observation could contradict. "Legal education should serve social justice" is an important claim and is not a hypothesis.

Revising the hypothesis after the data has been seen so that it fits the result. This is fitting, not testing, and the safeguard is that the hypothesis is recorded in the research proposal before collection begins.

One further point completes the answer. A hypothesis must be tested at a level of significance fixed in advance, conventionally five per cent, and the two errors it controls should be named: a Type I error is the rejection of a true null, that is finding a difference that is not there, and a Type II error is the failure to reject a false null, that is missing a difference that is there. The first is controlled by the significance level and the second chiefly by sample size, and because reducing one increases the other the choice between them is a judgment the researcher should make explicitly.

Where a hypothesis legitimately comes from, and the pilot

The question asks about the manner of formulation, and a manner that does not say where the proposition originates is incomplete.

One Indian illustration shows a court behaving exactly like a researcher. In Imtiyaz Ahmad v. State of Uttar Pradesh, (2012) 2 SCC 688, the Supreme Court, examining criminal matters in which High Courts had granted stays and finding that average pendency worked out at about 7.4 years, wanted to know whether delay was related to judge strength, found no reliable study, and asked the Law Commission of India to examine the question. The Commission's 245th Report, "Arrears and Backlog: Creating Additional Judicial (wo)manpower" (2014), tested precisely that relationship and recommended a rate of disposal method in preference to the judge to population ratio then in use. The hypothesis originated in an assumption the Court was being asked to act upon.

The pilot study is the last step of formulation and not a separate stage. A small trial run of the instrument on a handful of respondents from the same population exposes ambiguous wording, unrealistic assumptions about access and coding categories that do not fit the answers people actually give, all while they can still be changed cheaply. The hypothesis is then restated. A hypothesis that survives a pilot entirely unchanged is unusual, and if one does it should be re-read for vagueness, because a proposition too loose to be contradicted by twenty returns will not be contradicted by two hundred.

For full marks, cover: all three limbs, in order and clearly separated, because the question is a list of three. The third limb, on Indian problems, is the one that distinguishes a good answer and should be the longest.

Limb one: the objectives of research in general. Research is the systematic and objective search for new knowledge, and its objectives are conventionally classified as follows.

• Exploratory or formulative, to gain familiarity with a phenomenon about which little is known and to formulate a problem for more precise investigation.
• Descriptive, to portray accurately the characteristics of a person, situation or group, and to establish frequency and distribution.
• Diagnostic, to determine the frequency with which something occurs and its association with something else.
• Explanatory or causal, to test a hypothesis about a causal relationship.

The factual basis for legislation and reform. The Fourteenth Report of the Law Commission of India (1958) rested on inquiry into how courts actually worked; the Justice J.S. Verma Committee Report (2013) rested on material about the investigation and trial of sexual offences, and the Criminal Law (Amendment) Act, 2013, followed.

Adjudication. Indian constitutional adjudication constantly turns on the effect of a measure on a class of persons, which is an empirical question. In Bandhua Mukti Morcha v. Union of India, (1984) 3 SCC 161, the Supreme Court appointed commissioners to gather facts, and in Hussainara Khatoon v. State of Bihar, (1980) 1 SCC 81, facts about undertrials produced the right to a speedy trial under Article 21.

Indian conditions in particular. Indian law is largely reformist. Statutes abolishing untouchability, prohibiting dowry, restraining child marriage and guaranteeing education were enacted to change behaviour, not to record it, and a statute meant to change behaviour can only be evaluated by looking at the behaviour. Upendra Baxi made this argument in his 1975 study for the Indian Council of Social Science Research, criticising Indian legal scholarship as overwhelmingly doctrinal.

5. Language and translation. A study in Maharashtra will encounter records and respondents in Marathi, and translation is a cost and a source of error, particularly in the wording of a schedule.
6. Non response and unreliable response. Officials decline to answer, litigants cannot be traced after disposal, and respondents give the socially acceptable answer rather than the true one, especially on anything touching corruption, caste or the police.
7. Sampling difficulties. There is often no usable sampling frame. The accessible population, typically reported cases or willing respondents, is systematically unrepresentative of the real one.
8. Scale and diversity. A finding in one district of one State may not hold in the next, so generalisation to India is rarely defensible from a single study, and multi State studies are beyond most researchers' resources.
9. Time. Empirical work does not fit comfortably within the time allowed for a dissertation, which pushes students towards doctrinal topics and perpetuates the imbalance.

Indian constitutional law now requires empirical material as a condition of exercising power. In M. Nagaraj v. Union of India, (2006) 8 SCC 212, a Constitution Bench upheld the amendments enabling reservation in promotion but held that the State must first collect quantifiable data showing the backwardness of the class, the inadequacy of its representation in public employment and the effect on administrative efficiency. In Jarnail Singh v. Lachhmi Narain Gupta, (2018) 10 SCC 396, a five judge Bench revisited Nagaraj, held that no fresh demonstration of the backwardness of Scheduled Castes and Scheduled Tribes was needed, and retained the requirement of data on inadequacy of representation. In Indra Sawhney v. Union of India, 1992 Supp (3) SCC 217, both the identification of backward classes and the exclusion of the creamy layer turned on the adequacy of the material before the government.

A country whose Supreme Court demands quantifiable data of its executive, and whose law faculties produce almost none, has an institutional gap and not merely an academic preference. That is the sharpest available answer to the relevance limb of this question.

In Imtiyaz Ahmad v. State of Uttar Pradesh, (2012) 2 SCC 688, the Court was dealing with criminal matters in which High Courts had granted stays and the proceedings had then remained pending. Its examination of the figures disclosed that about nine per cent of such cases had been pending for more than twenty years since the stay, roughly twenty one per cent for more than ten, and that average pendency was around 7.4 years. The Court then wanted to know whether the delay was related to the strength of the judiciary. No reliable study existed. It therefore requested the Law Commission of India to examine the question, and the Commission produced its 245th Report, "Arrears and Backlog: Creating Additional Judicial (wo)manpower" (2014), which considered the methods by which required judge strength might be calculated and recommended a rate of disposal method rather than the judge to population ratio. In 2017 a three judge Bench directed that judge strength be computed for each State on an interim approach pending the final report of the National Court Management Systems Committee.

Three conclusions follow, and each answers part of this question.

The instruction for the third limb

A list question is marked on coverage, so the twelve problems should be given as a numbered list with a sentence each rather than as prose. An examiner reading a wall of text for a "enlist" instruction has to hunt for the items, and items that have to be hunted for are items that may not be found. The same applies to the first limb: the objectives of research are a recognised list, and setting them out as one demonstrates that the candidate knows it is a list.

Conclusion. The objectives of research run from exploration and description through diagnosis and explanation to prediction and evaluation, and empirical legal research pursues the last of these in relation to law: it asks whether a rule does what it says. Its relevance in India is at its highest because Indian law is chiefly an instrument of intended social change and its characteristic failure is in implementation. The obstacles are not intellectual but practical, being training, money, access, data, language, sampling and time, and they explain why a legal system that needs empirical research more than most produces less of it than most.

3. Finalise the outline into chapters, sections and sub sections, checking that each chapter carries one idea and that the order is the order of the argument, not the order of discovery.
4. Write the first draft straight through without polishing, because the only purpose of a first draft is to find out whether the argument holds.
5. Analyse and interpret. Present the authorities or the data and then say what they show. Description that stops short of analysis is the standard defect of legal dissertations.
6. Rewrite and cut. Remove what is unsupported, make transitions explicit, and check that the conclusion answers the question the introduction posed.
7. Prepare citations and the bibliography in one consistent style, so that every case, statute, book and article can be traced from the footnote.
8. Revise finally and check the mechanics: pagination, table of contents, table of cases, table of statutes, list of abbreviations, appendices, the supervisor's certificate, the declaration of originality and the similarity report.

Its relevance in the research process is fourfold. It is the fastest route into an unfamiliar field, because a good commentary states the scheme of an Act and collects the decisions under each section. It supplies criticism, since a judgment says what was decided and an article says why it is thought to be wrong. It is where new legal concepts are proposed and named before any court adopts them. And it identifies the gap that a research problem must occupy, which is why the literature review precedes the hypothesis.

Its persuasive weight in India is real. Indian courts cite Mulla, Seervai, Basu, Ratanlal and Dhirajlal, and Kanga and Palkhivala, and where the point is novel and the authority thin, the weight of juristic opinion often decides the direction of the argument.

The Indian periodicals a researcher must know are the Journal of the Indian Law Institute, the Annual Survey of Indian Law published by the Indian Law Institute, the National Law School of India Review, the Indian Journal of International Law and the journal sections of the reports. The Annual Survey is the most efficient instrument for locating a gap, because it reviews each branch of law year by year.

Communication. Knowledge becomes knowledge when it enters the common stock. Research that is not reported has changed nothing.

Verification. A reader who is told the problem, the method, the universe, the sample and the response rate can judge how much the conclusion is worth. A reader who is given only conclusions must take them on trust, which is the opposite of scholarship.

Preservation. The report is the permanent record. In law this matters unusually, because a study of an Act as it operated before amendment can never be repeated. Empirical work done on the Code of Criminal Procedure before 1 July 2024 is now irreplaceable evidence about a regime that no longer exists.

Assessment. A dissertation is judged on the report and not on the work behind it, so a good study reported badly fails.

Utility. The report is what a legislature, a court, a Law Commission or another researcher can actually use. The influence of the Fourteenth Report of the Law Commission (1958) came from the fact that the findings were written down in a form that others could act on.

The Indian sources, which should be named. The subscription databases are Manupatra, SCC Online, LexisNexis India, Westlaw India and AIR Online. The free public sources are the Supreme Court's eSCR portal for the official reports, the India Code portal of the Legislative Department for authenticated bare Acts with their amendment history, the National Judicial Data Grid and the eCourts services for case data, and the Gazette of India for commencement notifications. Foreign primary material is reachable through BAILII, AustLII and CanLII.

Currency and verification. A database records not only the judgment but what has happened to it since, marking a decision as followed, distinguished or overruled.

Analysis and writing. Statistical packages make empirical work practicable for one researcher; reference managers such as Zotero maintain citations and generate footnotes; similarity detection software is now standard under the UGC regulations of 2018.

Access. Free portals have narrowed, though not closed, the gap between a national law university and an ordinary affiliated college.

Foreign periodicals and juristic literature is tested on use, not on lists of journals. The four uses are to find a solution already worked out elsewhere, to test whether an Indian rule is peculiar, to construe a provision with a foreign ancestry, and to give effect to an international obligation. The last is the strongest and has an authority: in Vishaka v. State of Rajasthan, (1997) 6 SCC 241, there being no Indian statute on sexual harassment at the workplace, the Supreme Court read the Convention on the Elimination of All Forms of Discrimination against Women into Articles 14, 15, 19(1)(g) and 21 and declared guidelines binding under Article 141 until Parliament legislated, which it did in 2013.

Significance of research report writing is tested on why, not how, and the six reasons are communication, verification, preservation, assessment, utility and the discipline that writing imposes on the argument. The preservation point has a current Indian illustration: empirical work on the working of the Code of Criminal Procedure before 1 July 2024, when the Bharatiya Nagarik Suraksha Sanhita replaced it, can never be repeated.

For the steps in report writing, the specimen is the chapter scheme itself, and reproducing it earns marks because it demonstrates the order the steps produce. Preliminary part: title page; supervisor's certificate; declaration of originality; acknowledgements; table of contents; table of cases; table of statutes; list of abbreviations; similarity certificate. Main body: Chapter I, introduction, stating the problem, objectives, hypothesis or research questions, scope, limitations, research methodology and chapter scheme; Chapter II, review of literature; Chapters III to V, the substantive analysis; Chapter VI, findings; Chapter VII, conclusions and suggestions. Reference part: bibliography divided into primary and secondary sources; appendices carrying the questionnaire or interview schedule; and any statistical tables too bulky for the text.

Conclusion. The four notes describe the two ends of the research process. Report writing, in its steps and in its significance, is how research becomes usable, verifiable and permanent. Juristic literature, foreign periodicals and the electronic sources are how the researcher finds out what is already known and keeps the work current. The connecting discipline is citation: material found electronically is worth nothing to a reader who cannot reach it, and a proposition taken from a commentary is worth nothing until it is traced to the authority behind it.

The questionnaire, a set of written questions completed by the respondent unaided. Cheap, wide in reach, uniform in wording, free of interviewer bias and anonymous. Limited by low and biased response rates, unusable with illiterate respondents, and incapable of clarification or follow up.

The schedule, the same instrument filled in by a trained investigator in the respondent's presence. The practical Indian substitute for the questionnaire where literacy cannot be assumed, with far higher completion rates and the cost and bias that come with an investigator.

The interview, structured, semi structured or unstructured. Its value is depth, the ability to probe, and access to material a form cannot elicit, such as why a litigant abandoned a claim. Its costs are time, skill, interviewer bias and the tendency of respondents to give the socially acceptable answer.

Observation, participant or non participant. Court room observation is its standard legal use: a day in a Magistrate's court, recording matters listed, matters adjourned, time taken and reasons given, produces data no file discloses. It records what people do rather than what they say they do, but it changes what it observes, needs permission and is slow.

Legislation, together with subordinate legislation, notifications and rules, is the primary source of doctrinal legal research and the starting point of most empirical work as well. Its merits are, first, that it is authoritative: the words of the section, as published in the Gazette, are the law and not evidence of it. Second, it is accessible and free, through the India Code portal and the Gazette. Third, it is precise, so that a definition clause or a limitation period is a hard fact and not an estimate. Fourth, it carries its own history, since the amendment footnotes, the repealing provisions and the Statement of Objects and Reasons record what changed and why. Fifth, it is complete on its own subject: the researcher is not sampling, but reading the whole of the material.

Its merits. It gives depth, which no survey can: a single maintenance application followed from filing to disposal shows the adjournment practice, the settlement pressure and the delay that no average conveys. It reveals process and therefore mechanism, which is what a legislature needs in order to fix something. It is the only usable method where the phenomenon is rare, complex or too little understood for a questionnaire to be framed about it. It generates hypotheses for later testing on a larger sample. It produces concrete detail, which is what persuades a court or a law reform body, and is why such reports carry case studies as annexures. And it is feasible for a single researcher with limited resources.

Its limitations. One unit cannot support a generalisation, so the method establishes possibility and not frequency. Selection bias is a standing danger, because the interesting or the accessible unit is chosen. It depends on the skill and neutrality of one researcher and is difficult for another to replicate. It is slow and expensive relative to what it proves. And it produces qualitative material that resists quantification. The standard remedy is to combine it with a survey, using the case study for the mechanism and the survey for the extent.

One source the question does not name deserves a paragraph, because it has transformed what an individual Indian researcher can do. The National Judicial Data Grid and the eCourts services publish institution, pendency and disposal figures down to the level of a single court, and the Supreme Court's eSCR portal publishes the official reports without charge. A study of delay that would once have required a funded team and months of permissions can now begin with published data, and the researcher's effort shifts from obtaining figures to understanding what they mean, which is where it should have been all along.

Applied to the two named sources, the same three concepts sharpen the answer. Legislation is perfectly reliable, since the words do not change between readers, and it has no validity at all as a measure of what happens, since a statute is a statement of what ought to occur. A case study is high in validity for the process it describes and low in reliability, since another researcher might select a different unit and reach a different account. That is precisely why the two are combined, and why a dissertation that rests on one alone is vulnerable at the viva.

Imtiyaz Ahmad v. State of Uttar Pradesh, (2012) 2 SCC 688, is the illustration of the same problem at the highest level. The Supreme Court had pendency figures and wanted to know whether delay was related to judge strength; no study existed, so it asked the Law Commission, which produced the 245th Report (2014) and recommended calculating required strength by rate of disposal rather than by the judge to population ratio. Two measures of one shortage, producing different numbers, is exactly the point that classification is a choice.

Applied to the two sources this question names, the same discipline sharpens the answer. Legislation is perfectly reliable, since its words do not vary between readers, and it has no validity as a measure of what happens, because it states what ought to occur. A case study is high in validity for the process it describes and low in reliability, since another researcher might select a different unit. That is precisely why the two are used together, and why a design resting on either alone is vulnerable at the viva.

For full marks, cover: what the lecture is and why it dominates in India; its merits; its demerits; the critical point about what it structurally cannot do; then the measures, which carry a third of the marks and must be specific enough to act on.

The lecture is continuous exposition by a teacher to an assembled class, and in Indian legal education it is the default and very often the only method. A critical analysis has to begin with why, and the honest answer is that it is the only method that functions at the student to teacher ratios, library provision and syllabus coverage requirements that most affiliated law colleges actually operate under. It survives on economics, not on pedagogy.

The critical point is that last one, and it should be pressed. The four practical papers introduced by the Bar Council of India in 1997 and now carried by the Bar Council of India Rules of Legal Education, 2008, moot court and trial participation, drafting and pleading, professional ethics, and alternative dispute resolution, are skills papers. Where they are taught by lecture and examined by written paper, which is common, the curriculum is complied with in form and defeated in substance. That is not an argument against the lecture; it is an argument against using it for something it structurally cannot do. The same point now has a professional edge, because after Bar Council of India v. Bonnie Foi Law College, 2023 SCC OnLine SC 130, entry to the profession is tested by the All India Bar Examination, and a lecture based curriculum leaves that test to be met by coaching after graduation.

The measures should be concrete.

1. Break the hour into segments. Ten to fifteen minutes of exposition, then a question, a two minute problem or a discussion in pairs. This restores attention and supplies the feedback the method otherwise denies.

8. Use technology where it removes drudgery. A recorded lecture frees the class hour for discussion; a projected judgment puts the whole class on the same paragraph; a live database demonstration teaches research in the time it would take to describe it.

9. Train and evaluate the teacher. Structured student feedback, peer observation and induction training in legal pedagogy are ordinary in other disciplines and rare in Indian law colleges.

Why the measures fail without the examination, and the quality-control gap

A "suggest measures" limb is marked on realism, so the answer should say which measures work and why the others have not.

Method follows assessment. A student assessed by an essay answerable from a memorised note will treat every lecture as dictation, whatever the teacher does with the hour. Segmentation, teaching from the bare Act, distributing reading in advance and opening with facts all raise the cost to the student of not preparing, and all are cancelled by an examination that rewards reproduction. The single most powerful measure is therefore to set problems on unseen facts instead of essay questions, and it is also the cheapest, since it requires no staff, no infrastructure and no regulatory change.

A worked lecture, and the two things it cannot do

A critical analysis is stronger for showing the method at its best before setting out its limits.

A good lecture on the right to legal aid, in one hour. The teacher opens by stating the scheme: the class will see a right built out of Article 21 in four steps. Ten minutes on M.H. Hoskot v. State of Maharashtra, (1978) 3 SCC 544, where a prisoner had no means to engage counsel for his appeal and the Court held free legal services to be an ingredient of the fair procedure Article 21 requires. Ten minutes on Hussainara Khatoon v. State of Bihar, (1980) 1 SCC 81, where newspaper reports about undertrials held longer than the maximum sentence for the offences alleged were treated as a petition and produced a right to a speedy trial. A two minute question to the class on what follows for a person yet to be produced before a Magistrate. Ten minutes on Khatri (II) v. State of Bihar, (1981) 1 SCC 627, holding that the obligation arises at the first production. Ten minutes on Suk Das v. Union Territory of Arunachal Pradesh, (1986) 2 SCC 401, setting aside a conviction because the accused was unrepresented and had never been told he could ask, and identifying legal illiteracy as the obstacle. A short problem on unseen facts. A close stating what is settled and what remains open.

The measure that costs nothing

Of every measure that could be suggested, one requires no additional staff, no infrastructure and no regulatory change: setting problems on unseen facts in place of essay questions. It changes the student's calculation rather than the teacher's workload, because a problem cannot be answered from a memorised note, and it makes every other measure effective: a segmented lecture, a distributed reading list and an open bare Act all become worth attending to once the paper rewards application. It also aligns the university with the external standard, since after Bar Council of India v. Bonnie Foi Law College, 2023 SCC OnLine SC 130, upheld the All India Bar Examination, application is what a graduate is independently tested on. That many colleges have still not made a free reform indicates that the constraint here is inertia rather than resources.

The phases can be stated briefly, because the weight of this question is on the regulators. Ancient and medieval legal learning was textual and produced advisers rather than advocates. Formal legal education began in 1855, and the universities of Calcutta, Bombay and Madras, founded in 1857, began awarding law degrees; the colonial model was a two year part time evening course intended to supply pleaders cheaply. The post independence phase was one of diagnosis, from the Fourteenth Report of the Law Commission of India (1958) onwards. The modern phase began in the 1980s with the five year integrated course and the establishment of the National Law School of India University at Bangalore in 1987 with Prof. N.R. Madhava Menon as founding Director.

The Bar Council of India is the professional regulator, and its powers are statutory. Under section 7(1)(h) of the Advocates Act, 1961 it has the function of promoting legal education and laying down standards of such education in consultation with the universities imparting it and the State Bar Councils. Under section 7(1)(i) it recognises universities whose degree in law qualifies for enrolment, and under section 24(1)(c)(iii) a degree so recognised is a condition of enrolment as an advocate. Section 49(1)(af) and (d) give it rule making power.

The University Grants Commission is the academic regulator. Established by the University Grants Commission Act, 1956, its functions under section 12 are to promote and coordinate university education and to determine and maintain standards of teaching, examination and research. Its instruments are grants, the recognition of universities and degrees under section 22, and regulations on matters such as the minimum qualifications for appointment of teachers, the award of the M.Phil. and Ph.D. degrees, and academic integrity. Its contribution to legal education specifically has come through funding, faculty qualification norms, the promotion of postgraduate study and the UGC (Promotion of Academic Integrity and Prevention of Plagiarism in Higher Educational Institutions) Regulations, 2018, which govern every law dissertation submitted in the country.

The friction between the two is the substance of this question and should be addressed directly. The Bar Council regulates law as a profession; the University Grants Commission regulates law as a university discipline. Their jurisdictions overlap at exactly the point where most students are, that is the law degree awarded by a university and recognised for enrolment. Three consequences follow.

Third, the Law Commission proposed a solution and it was not adopted. The 184th Report of the Law Commission of India (2002) on legal education recommended amendments to both the Advocates Act, 1961, and the University Grants Commission Act, 1956, to define the respective roles and to create a machinery for coordination. The National Knowledge Commission Report on Legal Education (2007) went further and recommended an independent standing committee for legal education. Neither recommendation has been enacted, and the overlap therefore continues.

Two further matters complete the picture. The Supreme Court has held that the State's obligation extends to resourcing: in State of Maharashtra v. Manubhai Pragaji Vashi, (1995) 5 SCC 730, it held that the duty to provide free legal aid under Article 39A read with Article 21 requires an adequate number of law colleges with proper facilities, and directed the extension of grant in aid to recognised private law colleges. And the profession's sensitivity to external regulation was demonstrated recently when the draft Advocates (Amendment) Bill, 2025 was withdrawn on 23 February 2025 after nationwide protests, largely over provisions perceived as reducing the autonomy of the Bar Council.

On the limits of the same power, Indian Council of Legal Aid and Advice v. Bar Council of India, (1995) 1 SCC 732, struck down a rule barring enrolment of persons over forty five as beyond section 24(1) and arbitrary. Together the two lines establish that the Council may test ability and may not invent disqualifications.

On the State's obligation to fund the institutions both regulators supervise, State of Maharashtra v. Manubhai Pragaji Vashi, (1995) 5 SCC 730, held that the duty to provide free legal aid under Article 39A read with Article 21 requires an adequate number of properly equipped law colleges, and directed that grant-in-aid be extended to recognised private law colleges which the State had excluded while aiding other colleges.

Prof. Yashpal v. State of Chhattisgarh, (2005) 5 SCC 420. A State enactment permitted private universities to be created by notification, and more than a hundred came into being, many with no campus, faculty or infrastructure. The Supreme Court struck the Act down, holding that a university cannot be brought into existence by notification alone without provision for standards, and that the University Grants Commission Act, 1956 places the coordination and determination of standards with the Commission, which State legislation cannot bypass. For legal education the principle is that the power to create an institution is not the power to create a degree.

Bar Council of India v. Board of Management, Dayanand College of Law, (2007) 2 SCC 202, fixes the boundary from the professional side, upholding the Council's insistence on the prescribed qualification for the Principal of a law college and confirming that its function under section 7(1)(h) of the Advocates Act, 1961 reaches the conditions in which a recognised degree is taught and not merely the syllabus.

Conclusion. Legal education in India is needed because a codified system, a monopoly profession, a constitutional democracy and a legal aid guarantee each require it, and it developed from a colonial arrangement designed to supply pleaders cheaply into a system that, since 1987, has produced institutions of real quality. It is regulated by two bodies with different objects and overlapping jurisdiction: the Bar Council of India, which controls the curriculum, recognition and now entry to the profession, and the University Grants Commission, which controls degrees, faculty standards and research. The unresolved problem is that nobody has ever settled where one ends and the other begins, the LL.M. dispute of 2021 being the current cost of that failure, and the recommendation of the 184th Report of the Law Commission that both statutes be amended to define the roles remains the most sensible proposal on the table.

The case study method in teaching is the study of decided cases to extract the principle and to see how principles are made and applied. It originated with Langdell at Harvard in 1870 and is built on the case book, a compilation of edited judgments arranged so that a doctrine's development emerges from reading them in order. In class the student states the facts, the procedural history, the issue, the holding and the reasoning; the teacher tests the statement with hypotheticals; and the class distinguishes the case from its neighbours and separates the ratio decidendi from obiter dicta.

Demerits of the case method. It is slow and cannot cover a syllabus. It over emphasises appellate decisions, which say nothing about trial courts, negotiation or drafting. It needs a good case book and a prepared class and fails entirely without either. And in India it strains against the fact that most of our law is codified, so the section, not the case, is the primary material.

The seminar method is a small group meeting at which one or more students present prepared work on a defined question and the group discusses it under the teacher's direction. The teacher fixes a narrow and genuinely arguable question and distributes reading in advance; a student presents; a designated respondent replies; the floor discusses; and the teacher closes by stating what was settled and what remains open.

Demerits of the seminar. It is expensive in teacher time and needs a small group, which is why it is rare at undergraduate level in India. It collapses if the class has not prepared. It can be monopolised by two or three confident speakers. Coverage is slow. And a discussion the teacher does not close leaves the class less certain of the law than it was at the start.

The comparison on stated criteria.

A third worked example, which shows the two methods used in sequence rather than in competition. Take the law of absolute liability. Case method, week one: the class reads Rylands v. Fletcher, (1868) LR 3 HL 330, where a reservoir flooded a neighbouring colliery through disused shafts and the rule of strict liability for the escape of a dangerous thing was stated, and then M.C. Mehta v. Union of India, (1987) 1 SCC 395, where oleum gas escaped from the Shriram plant in Delhi and the Supreme Court refused the Rylands exceptions, laying down an absolute and non-delegable duty with compensation correlated to the size of the enterprise. The class emerges knowing what the rule is and how it changed. Seminar, week two: the question is whether the M.C. Mehta measure of damages, which scales with the capacity of the enterprise, is compensation or punishment, and whether a court may impose the latter in a writ proceeding. The reading is short, the question is genuinely arguable, and it cannot be answered by anyone who did not do week one.

After a term of seminars the student can prepare a position from primary sources and defend it under questioning; can concede a bad point without abandoning the argument; can identify the strongest version of the opposing case before attacking it; and can produce a short written paper with citations that another reader can check. Those are the operations of scholarship and of written advocacy.

The seminar costs contact time and cannot be reused. A group of twelve requires the teacher to read twelve papers, chair the discussion and close it, and none of that preparation transfers to the next group, because the discussion will go elsewhere. That is why the seminar survives at postgraduate level, where numbers are small, and disappears at undergraduate level, where they are not.

Both cost the same thing when they fail, which is the hour. An unprepared class converts either method into a poor lecture, and the teacher discovers this only after the class has begun.

Conclusion. The case study method is the right principal method for a subject built on precedent, because it forces the student into the judgment and teaches reasoning by demonstration, and it is at its most useful in Indian constitutional law and tort, where the doctrine is judge made. The seminar is the right principal method for postgraduate work, because at that level the object is judgment rather than coverage. They fail on the same condition, unprepared students, and both therefore depend on assessment that rewards preparation. In a full curriculum they are stages rather than alternatives: lectures for framework, the case method for reasoning, seminars for argument, and clinical work to connect all three to practice.

The need arises from four specific failures, and stating them as failures is better than asserting a general benefit.

The graduate cannot do the work. A student may carry the whole Code of Civil Procedure in memory and be unable to draft a plaint, because drafting is a skill. The Fourteenth Report of the Law Commission of India (1958) said this, the National Knowledge Commission Report on Legal Education (2007) said it again half a century later, and the persistence of the complaint is itself the evidence.

The regulator does not trust the degree. The Bar Council prescribed pre enrolment training by its Training Rules of 1995; those rules were struck down in V. Sudeer v. Bar Council of India, (1999) 3 SCC 176; and a Constitution Bench overruled V. Sudeer in Bar Council of India v. Bonnie Foi Law College, 2023 SCC OnLine SC 130, decided 10 February 2023, upholding the All India Bar Examination. An examination after the degree is an admission that the degree alone does not establish readiness.

The Indian framework already exists and should be named precisely. The Bar Council of India introduced four practical papers in 1997, now carried by the Bar Council of India Rules of Legal Education, 2008: moot court, pre trial preparation and participation in trial proceedings; drafting, pleading and conveyancing; professional ethics and professional accounting system; and alternative dispute resolution. Alongside them, the National Legal Services Authority (Legal Aid Clinics) Regulations, 2011 provide for legal aid clinics in law colleges and universities, staffed by students under the supervision of a lawyer, which makes the law school clinic part of the statutory legal aid machinery rather than a voluntary good work.

The evidence that the method works is institutional. The National Law School of India University, established at Bangalore in 1987 with Prof. N.R. Madhava Menon as founding Director, built its course on the case method, the compulsory moot, the internship and a working clinic, and the change in what its graduates could do was immediate and has been reproduced by the national law universities that followed. Where the clinical components were adopted the output changed; where only the syllabus was adopted it did not.

The implementation gap is the necessary honest paragraph. In many affiliated colleges the practical papers are taught and examined as theory, the moot is one annual competition for a handful of students, the clinic is a notice board and the internship diary is signed without the internship. The cause is structural: clinical teaching needs a small student to teacher ratio, a supervising practitioner and a teacher prepared to be judged on the student's work.

What has changed recently, and why it sharpens the need

"Contemporary India" is the operative phrase, and it is answered with dates.

Dispute resolution has moved outside the courtroom. The Mediation Act, 2023 places pre-litigation mediation on a statutory footing, and section 89 of the Code of Civil Procedure, whose working was explained in Afcons Infrastructure Ltd. v. Cherian Varkey Construction Co. (P) Ltd., (2010) 8 SCC 24, requires the court to consider settlement and sets out which categories of case are normally suitable for reference. Negotiation and mediation advocacy are skills, and the fourth practical paper is the only place in the curriculum they appear.

Unmet legal need has not fallen. Section 12 of the Legal Services Authorities Act, 1987 entitles a very wide class to free legal services, and Suk Das v. Union Territory of Arunachal Pradesh, (1986) 2 SCC 401, holds that the aid must be offered even to an accused who does not know to ask. The NALSA (Legal Aid Clinics) Regulations, 2011 put the student clinic inside the machinery for meeting that demand.

Each of those five changes rewards a graduate who has done legal work and penalises one who has only read about it, which is why "learning by doing" is a more urgent proposition in 2026 than it was in 1997 when the four practical papers were introduced.

Practical Paper III, professional ethics and professional accounting system. The student studies the standards of professional conduct framed under section 49(1)(c) of the Advocates Act, 1961, the disciplinary jurisdiction in Chapter V and the leading disciplinary decisions. Taught clinically through problems it teaches that an ethical difficulty arrives as a conflict between duties, which is the situation Ex-Capt. Harish Uppal v. Union of India, (2003) 2 SCC 45, presupposes when it holds that lawyers have no right to strike and that a lawyer who abstains may answer to the client.

Practical Paper IV, alternative dispute resolution. The student studies and simulates arbitration, conciliation, mediation, negotiation and the Lok Adalat. Its importance has grown sharply since the Mediation Act, 2023 placed pre-litigation mediation on a statutory footing, and section 89 of the Code of Civil Procedure, construed in Afcons Infrastructure Ltd. v. Cherian Varkey Construction Co. (P) Ltd., (2010) 8 SCC 24, requires the court to consider settlement and identifies the categories normally suitable for reference.

The four together are a fair description of what a junior lawyer actually does, which is the strongest available answer to anyone who treats clinical education as a supplement to the real curriculum rather than as part of it.

For full marks, cover: the definition and what a design actually contains; why it is needed; the types of design; then the second limb, the tools and techniques used to prepare one, which is the part most answers omit; the features of a good design; and a conclusion.

A research design is the advance plan of the whole investigation: the blueprint that states what will be studied, from whom or from what, by what methods, and how the material will be analysed. Kerlinger describes it as the plan, structure and strategy of investigation conceived so as to obtain answers to research questions and to control variance. It is prepared after the problem and the hypothesis are settled and before any material is collected, and its purpose is to ensure that the data gathered will actually bear on the question asked.

• Exploratory or formulative design, used where too little is known to frame a hypothesis. It is flexible, relies on a survey of the literature, on experience surveys and on the study of a few illustrative instances, and its output is a properly formulated problem rather than a finding.
• Descriptive design, used to portray the characteristics of a group or situation accurately. It requires rigid definition of the population, careful sampling and standardised instruments, because accuracy is its whole purpose.
• Diagnostic design, which establishes the frequency of something and its association with something else.
• Experimental or hypothesis testing design, which manipulates a variable and observes the effect. True experiments are rare in legal research for obvious reasons, but quasi experimental designs are available and underused: the before and after study of a legal change, and the comparison of a jurisdiction where a provision applies with one where it does not, are both practicable.

The pilot study is the second and most valuable. A small scale trial run of the instrument on a handful of respondents exposes ambiguous questions, unrealistic assumptions about access and unworkable coding schemes while they can still be changed cheaply.

Operational definition of variables is a technique in its own right: converting concepts such as delay, access or effectiveness into something that can be counted is what makes a design testable.

Sampling design is prepared using the frame, the choice between probability methods, that is simple random, systematic, stratified and cluster sampling, and non probability methods, that is purposive, quota, convenience and snowball sampling, and a reasoned decision on sample size based on the variability of the population and the precision required.

Construction and pre testing of the instrument, whether questionnaire, schedule, interview guide or observation schedule, including the choice between closed and open questions and the design of any scale.

Consultation with experts and with the supervisor, and where the study touches an institution, the early negotiation of access, since a design that assumes permission which is not forthcoming is worthless.

The pilot study is the single most valuable technique in the preparation of a design. A small scale trial run of the instrument on a handful of respondents exposes ambiguous questions, unrealistic assumptions about access and unworkable coding schemes while they can still be changed cheaply. A design that has not been piloted is a hypothesis about the design, and it usually fails in the first week of fieldwork, when failure is most expensive.

Dummy tables are the most effective device for finding what the design has forgotten. Before collecting anything, the researcher draws the empty tables the study intends to fill: mode of disposal against continuity of representation, time to disposal against number of adjournments, and so on. The exercise takes an afternoon and reliably reveals that a variable the analysis requires was never going to be collected, which is a discovery worth months.

The features of a good design, and the design a doctrinal study needs

A design question is finished by stating how a design is judged, because the tools are only means to that.

A good research design is appropriate to the problem rather than to the researcher's preference, which is the commonest failure: a candidate who has learned survey method will design a survey for a question that needed a case study.

It minimises bias and maximises the reliability of the data. Bias is controlled by removing the researcher's discretion from selection, that is by probability sampling where a frame exists, by neutral wording, and by piloting. Reliability is tested by having two coders code the same returns and comparing them.

It yields the smallest experimental error for the resources available, which is the practical meaning of choosing stratified rather than simple random sampling where the strata differ.

It is feasible in time, money and access, and access should be secured in writing before the design is finalised, because a design assuming permission that is refused is worthless.

The instruction that follows

A design question is answered in the exam by describing a design rather than by classifying designs. A candidate who sets out one worked plan, naming the universe, the frame, the sampling method, the instrument, the pilot and the plan of analysis, has demonstrated more than one who lists exploratory, descriptive, diagnostic and experimental designs with a definition of each. The classification is worth a paragraph; the worked plan is worth the answer.

Conclusion. A research design is the advance plan that connects the question to the evidence: it fixes the universe, the sample, the sources, the instruments and the analysis before anything is collected, and it is what makes the difference between a study and a collection of material. It is prepared by literature survey, operational definition, sampling design, instrument construction, pilot testing, expert consultation and the drawing of dummy tables, and it is judged by whether it will actually produce data capable of testing the hypothesis within the resources available. A study that begins collecting before the design is settled almost always finishes by asking a different question from the one it set out to answer.

Step 1: state the problem. Begin from a difficulty rather than a subject. The difficulty here is that legal aid is granted and yet a large proportion of aided matters do not reach a decision. The problem is therefore: why do aided litigants in this district discontinue before disposal?

Step 2: review the literature. Read the annual reports of the State and District Legal Services Authorities, the National Legal Services Authority's own studies, the Law Commission material, and the academic work on legal aid. The reading serves two purposes: it prevents a settled question from being asked again, and it identifies the variables other researchers have found to matter, here typically the quality of the panel lawyer, the distance to the court, the number of hearings and the cost of attending them.

Step 3: identify the variables. The dependent variable is discontinuance. The candidate independent variables are the number of adjournments, the distance from the litigant's residence to the court, whether the panel lawyer changed during the proceeding, the litigant's daily wage loss, and the type of matter.

Step 8: refine after the pilot study. A pilot on twenty files will usually show that a term is ambiguous, that a record does not exist, or that a question is understood differently from how it was meant. The hypothesis is then restated. A hypothesis that survives a pilot unchanged is unusual and is worth checking for vagueness.

The characteristics a formulated hypothesis must satisfy, each illustrated:

• Conceptually clear. Every term defined so that another researcher would apply it identically. "Effectiveness of legal aid" fails this; "discontinuance as defined above" passes.
• Empirically testable. Some obtainable evidence would defeat it. "Legal aid should be a fundamental right" is a normative claim, not a hypothesis.
• Specific. It names the population, the place and the period: aided litigants in this district in matters instituted between two stated dates.
• Feasible with the techniques and data available, including lawful access. Where personal data is collected the Digital Personal Data Protection Act, 2023 applies to its collection and processing.

Two further worked examples show the range of forms. A descriptive hypothesis: more than half the applications under the Protection of Women from Domestic Violence Act, 2005, filed in a given court in a given year are accompanied by proceedings under another enactment. This is tested by counting. A causal hypothesis: the introduction of a statutory time limit for filing a written statement reduced the average time to framing of issues in that court. This requires data from before and after the amendment, which is why a baseline study has a value only realised years later.

The exception must be stated. Doctrinal research and genuinely exploratory studies proceed without a hypothesis and use a clearly stated research question instead. Manufacturing a hypothesis to fill a chapter heading misdescribes the method.

The characteristics tested against the worked example, and the errors

A formulation question is best closed by running the finished hypothesis through the criteria, because that is what a supervisor does.

The two errors of testing must be named. A Type I error is the rejection of a true null hypothesis, that is finding a difference that does not exist, and the level of significance fixed in advance, conventionally five per cent, is the probability of it the researcher accepts. A Type II error is the failure to reject a false null, that is missing a difference that does exist, and it is controlled chiefly by sample size. Reducing one increases the other, so the choice of level is a judgment about which mistake would matter more, and in policy research it should be made explicitly.

Two cautions close the answer. Statistical significance is not practical importance: with a large enough sample a trivial difference becomes significant, so the size of the effect must be reported alongside the p value. And association is not causation: a difference in discontinuance between represented and unrepresented applicants may reflect a difference in the claims they bring, and only a design that controls for that can separate the two.

Where the hypothesis comes from, and the null form worked out

A process answer should say where the proposition originates, because a hypothesis produced from nowhere is a guess in formal dress.

Worked on the running example. Working hypothesis: aided litigants who lose a day's wages to attend court discontinue at a higher rate than those who do not. Null: there is no difference in the rate of discontinuance between the two groups. Alternative: there is a difference. Because a direction is asserted in the working form, a one tailed test is available; if only a difference is asserted, the test is two tailed. Significance level fixed in advance at five per cent. Test chi square, because both variables are categorical. Decision reported whichever way it falls.

Two errors are then named. A Type I error is rejecting a true null, that is finding a difference that is not there, and the significance level is the probability of it the researcher accepts. A Type II error is failing to reject a false null, that is missing a difference that is there, and it is controlled chiefly by sample size. Reducing one increases the other, so the choice is a judgment about which mistake would matter more, and in research bearing on policy it should be stated rather than assumed.

For full marks, cover: both notes fully, since this question offers no choice. For (a), definitions, a comparison on criteria, and why both are needed in India. For (b), what sampling is, the probability and non probability types with legal examples, sample size and the two kinds of error.

(a) Doctrinal and non-doctrinal research

Doctrinal research is research into the authoritative legal materials, that is statutes, subordinate legislation, judicial decisions, treaties and the commentary on them. It asks what the law is, and it answers by analysis, systematisation and criticism of those materials. It is library based and remains the dominant form of legal scholarship everywhere.

Doctrinal research is indispensable because it alone states the law authoritatively; it is quick, inexpensive, replicable by any competent reader, and it produces the systematic account on which practice, teaching and reform all rest. Its weakness is that it can become an internal conversation about texts, unable to say whether anything it describes is obeyed or enforced.

Non-doctrinal research is indispensable in India in particular because Indian law is largely reformist: statutes abolishing untouchability, prohibiting dowry, abolishing bonded labour and guaranteeing education were enacted to change behaviour and can only be evaluated by looking at behaviour. Upendra Baxi argued in his 1975 study for the Indian Council of Social Science Research that Indian legal scholarship was overwhelmingly doctrinal and disconnected from Indian society, and the criticism has not been fully answered. Its costs are money, time, training, access and, since the Digital Personal Data Protection Act, 2023, statutory obligations where personal data is collected.

They are complementary, and the order matters. An empirical study conducted by a researcher who has the doctrinal position wrong measures the wrong thing, so the doctrinal chapter is a precondition of the empirical one, not an alternative to it.

• Stratified sampling: the population is divided into strata that matter to the study and a random sample is drawn from each. Example: separating represented from unrepresented applicants so that a small but important group is not missed by chance.
• Cluster and multi stage sampling: naturally occurring groups are sampled and then units within them. Example: five districts, then three courts in each, then thirty files in each court. This is the practical method for a State wide study.

Non-probability sampling gives an unknown chance of selection and cannot support statistical generalisation, but is often the only method available.

• Purposive or judgment sampling: units chosen as typical or particularly informative, for example the ten advocates in a district who handle most legal aid matters.
• Quota sampling: fixed quotas filled from defined groups.
• Convenience sampling: whoever is accessible. The weakest form, and its limits must be stated openly.
• Snowball sampling: each respondent leads to the next. The standard method for populations that are hard to locate, such as released undertrials.

The cause is incentives rather than conviction. Doctrinal work needs no funding, no permissions and no statistical training; it fits inside the time allowed for a dissertation; and it is easier to place in a journal. Empirical work fails on every one of those counts. Any proposal to rebalance Indian legal scholarship that does not address funding, access to institutional data and methods training is therefore an exhortation and not a plan.

The cost is that the courts now require what the academy does not produce. In M. Nagaraj v. Union of India, (2006) 8 SCC 212, the Supreme Court held that the State must collect quantifiable data on backwardness, inadequacy of representation and administrative efficiency before providing reservation in promotion, and Jarnail Singh v. Lachhmi Narain Gupta, (2018) 10 SCC 396, retained the requirement of data on inadequacy of representation. Where governments must produce data and universities do not teach how to gather it, the data produced will be of whatever quality the executive finds convenient.

The explanation is incentives and not conviction. Doctrinal work needs no funding, no permissions and no statistical training; it can be completed by one person in a library within the time a dissertation allows; and it is easier to place in a journal. Empirical work fails on every one of those counts, needing money for investigators, travel and translation, permissions for court records, prisons and police stations, training the LL.B. does not supply, and more time than a dissertation permits. Any proposal to rebalance Indian legal scholarship that does not address funding, access and methods training is an exhortation rather than a plan, which is why Upendra Baxi's criticism of 1975 has outlived every response to it.

The cost is now demonstrable rather than rhetorical. In Imtiyaz Ahmad v. State of Uttar Pradesh, (2012) 2 SCC 688, the Supreme Court wanted to know whether delay was related to judge strength, found no reliable study, and had to ask the Law Commission of India, which produced the 245th Report (2014). The highest court in the country could not obtain from Indian legal scholarship an answer to a factual question about its own institution.

Conclusion. The two notes answer the two questions a reader asks of any piece of legal research: what kind of question is being answered, and how much the answer is worth. Doctrinal research states the law and cannot evaluate it; non-doctrinal research evaluates it and cannot state it. Where the second is undertaken, the sampling method is what decides whether its findings describe a population or only the respondents who happened to reply, and a report that does not state its universe, frame, method, size and response rate has not told the reader enough to judge it.

Its authority in India is persuasive and real. Indian courts cite Mulla on the Code of Civil Procedure and on Hindu law, Seervai and Basu on constitutional law, Ratanlal and Dhirajlal on the penal law and Kanga and Palkhivala on income tax. Where the point is novel and the authority thin, the weight of juristic opinion frequently decides the direction an argument takes.

Its functions in the research process are five. It is the fastest route into an unfamiliar field, because a good commentary states the scheme of an Act and collects the decisions under each section. It supplies criticism, since a judgment records what was decided and an article explains why it is thought wrong. It identifies the gap a research problem must occupy. It supplies the theoretical framework within which a problem is stated, as Upendra Baxi's 1975 study for the Indian Council of Social Science Research did for a generation of Indian socio-legal work. And it is where new legal concepts are proposed, tested and named before any court adopts them.

(b) Classification and tabulation of data

Classification and tabulation are two successive operations in the processing of data, and they must not be run together. Classification is the arrangement of data into groups according to a common characteristic. Tabulation is the presentation of the classified data in rows and columns. Classification decides what the categories are; tabulation displays them.

They follow editing and coding. Editing checks the returns for completeness, legibility, consistency and accuracy. Coding assigns a numeral or symbol to each answer so that it can be counted. Only then can data be classified.

The kinds of classification are four.

• Qualitative or classification by attribute, where the characteristic cannot be measured, for example the sex of the applicant, the nature of the offence, or whether the party was represented. Where there are two categories it is simple classification; where more, manifold.
• Quantitative or classification by class interval, where the characteristic is measurable, for example time to disposal grouped in bands of six months, or the amount of maintenance awarded in bands.

The parts of a table should be named, because an examiner looks for them: table number, a title stating what, where and when, captions for the columns, stubs for the rows, the body, the unit of measurement, totals, and, below the table, the source note and any footnote explaining an exclusion.

The rules of a good table are that it should answer one question; that it should be small enough to be read at a glance, with large data sets broken into several tables; that percentages must state the base on which they are computed; that totals must be shown and must add up; and that the source of every figure must be given. Non responses and unusable returns must be shown and not silently omitted, because a table that quietly drops them overstates its own base.

Two errors are common and fatal. The first is presenting a table and leaving it to speak for itself: a table is evidence, and the text must say what it shows. The second is percentaging in the wrong direction in a cross table, so that the figures answer a different question from the one the researcher thinks they answer.

Four rules of good tabulation are visible in it. The table answers one question. The base of every percentage is stated, here 100 in each column, so that the columns are comparable. The totals are shown and they add up. And the categories are exhaustive and mutually exclusive, so every case has exactly one place.

Start with a commentary for the scheme of the Act and the decisions collected under each section, which saves weeks and prevents the beginner's error of reading a statute in the order it is printed rather than the order it operates. Move to the Annual Survey of Indian Law, published by the Indian Law Institute, which reviews each branch year by year and therefore shows at a glance where a subject has moved and where it has stood still for a decade. Then read periodical articles, which is where criticism lives and where a writer will state expressly that a question has not been examined; that sentence is the research problem. Then stop and verify: every proposition taken from a commentary must be traced back to the statute or the judgment, because a commentary is dated from the day it is printed, and the replacement of the criminal codes on 1 July 2024 turned a shelf of standard works into legal history.

On classification and tabulation, two errors are fatal and both are invisible to the person who commits them.

The first is presenting a table and leaving it to speak for itself. A table is evidence. The text must state what it shows, and must then resist calling an association a cause: if unrepresented applicants are dismissed for default four times as often as represented ones, that is a finding about association, and the two groups may differ in ways the table does not record.