What a Hypothesis Is

The repair in each case is the same: give the terms an operational definition and state the threshold.

The two forms a hypothesis can take

A statement about a magnitude. That the proportion is below a stated level; that the median cost exceeds a stated sum.

A statement about a relationship. That awareness is lower among women than among men; that the proportion informed falls as the length of the remand list rises.

A relational hypothesis is more informative and harder to establish, because it needs enough cases in each group, chapter 1000.

And a student should usually state one of each: a magnitude hypothesis that the study will certainly be able to test, and a relational one that it may.

Where the hypothesis sits in the design

After the problem and the questions, chapters 1100 and 1110, because it is a proposed answer to a question.

Before the instrument, because the instrument must be capable of testing it. A schedule that cannot produce the number the hypothesis is about is a schedule designed too early.

Before collection, always. A hypothesis formulated after the data is looked at is not a hypothesis; it is a description of the data, and presenting it as though it had been stated in advance is a serious form of dishonesty because it cannot be detected.

And it is stated in the dissertation, in the methodology component MU requires, chapter 1310, so that the reader knows what was expected before anything was found.

MU's own requirement

The Department of Law's notice of 1 July 2026 says that evaluation will be done on the basis of the content of the chapterisation, the proving of the hypothesis, and the suggestive recommendations by the researcher.

Three things follow for the reader of this book.

A dissertation without a hypothesis is missing something MU says it marks.

The word proving must be read sensibly. Research does not prove in the mathematical sense; it supports or fails to support. A candidate should use the University's word and then say what they mean by it, which is that the evidence collected either supported the stated proposition or did not.

And a disproved hypothesis is not a failure, chapter 1170, which is the most important thing in this part of the book and the thing students most often do not believe.

A worked example

The research question, from chapter 1110. In what proportion of remand productions was the accused informed?

Two hypotheses, stated before any observation.

H1, magnitude. In the study court, over the study period, the accused was informed of the right to free legal services, as operationally defined, in fewer than one quarter of remand productions.

H2, relational. The proportion informed falls as the number of persons produced at a sitting rises.

The derivation, which must be stated. H1 rests on the absence of any published record of the practice, on the small number of legal aid assignments recorded at the remand stage in the authority's returns, and on the observation in Khatri itself that the right is illusory unless the court speaks, chapter 630. H2 rests on the ordinary expectation that a longer list produces shorter productions.

What would refute each. H1: an observed proportion of a quarter or more. H2: no fall, or a rise, in the proportion as list length increases.

And what the instrument must therefore capture. Whether information was given, and the number of persons produced at that sitting. Both are on the schedule of chapter 960, which is what it means for the hypothesis to precede the instrument.

Quick revision

A hypothesis is a tentative, reasoned proposition formulated before investigation, offering an explanation or expected relationship, and testable, meaning it specifies what evidence would refute it.

It does four things: focuses the study; makes it falsifiable; connects it to what is known; and structures the writing.

It is not: a guess; the research question; the conclusion; a statement of what the researcher hopes; and it is not compulsory in every kind of research.

Two forms: about a magnitude, and about a relationship. State one of each where possible.

Its place: after the problem and questions, before the instrument, and always before collection. A hypothesis formulated after looking at the data is a description of the data.

MU marks the proving of the hypothesis, so a dissertation without one is missing something the University says it assesses; and a disproved hypothesis is a finding, not a failure.

Test yourself

1. Define a hypothesis and give its defining characteristic. A tentative proposition, formulated before investigation, offering an explanation or an expected relationship and capable of being tested against evidence. Its defining characteristic is testability: it specifies what evidence would establish it and, crucially, what evidence would refute it.

2. Show why two of these are not testable and how each would be repaired: that legal aid in India is inadequate; that the proportion informed is below one quarter. The first is not testable because inadequate has no operational meaning, so nothing would refute it; it is repaired by defining the term operationally and stating a threshold. The second is testable as it stands, since an observed proportion of one quarter or more would refute it.

3. Why must a hypothesis precede the instrument as well as the collection? Because the instrument has to be capable of testing it. If the hypothesis concerns the relationship between list length and whether information was given, the schedule must record both, and a schedule designed before the hypothesis will not.

4. What does MU say it marks, and what follows for a candidate? That evaluation is on the content of the chapterisation, the proving of the hypothesis and the suggestive recommendations. It follows that a dissertation without a hypothesis lacks something the University assesses, that the word proving should be used but explained as meaning supported or not supported by the evidence, and that a disproved hypothesis is a finding rather than a failure.