How to write a research hypothesis

What makes a hypothesis testable

A good hypothesis is a single, specific, falsifiable statement. That means data could show it to be wrong. Compare:

• Not testable: “Social media has an impact on youth.” Which social media? What impact? Which youth? Any result would fit.
• Testable: “Among undergraduate students in Madurai, daily time spent on Instagram is negatively associated with self-reported sleep quality.” It names the variables, the population and the direction, and a survey could show it is false.

Four things to check in every hypothesis:

1. Variables are named and measurable. If you cannot say how you will measure a variable, you cannot test the hypothesis.
2. The relationship is stated. A difference between groups, an association, or an effect of one variable on another.
3. The population or context is stated. Who or what the claim applies to.
4. It is one claim. “X affects Y and Z” is two hypotheses. Split it, because the data may support one and not the other.

A hypothesis is also not the same as an objective. An objective says what you will do (“to examine the effect of X on Y”). The hypothesis says what you expect to find (“X increases Y”).

Null and alternative hypotheses

Statistical testing works with a pair. The null hypothesis (H₀) states that there is no effect or difference. The alternative hypothesis (H₁) states that there is. Your test calculates how likely results at least as extreme as yours would be if the null were true. If that probability (the p-value) is below your chosen significance level, usually 0.05, you reject the null in favour of the alternative.

Two points examiners pick up

You never “accept” or “prove” a hypothesis. You either reject the null or fail to reject it. Failing to reject the null does not show there is no effect; it may mean your sample was too small to detect one. Write “the data did not support H₁” or “H₀ was not rejected”, not “H₀ was accepted” or “it is proved that”.

Statistical significance is not size. With a large sample, a tiny and unimportant difference can be significant. Report the effect size and confidence interval as well, so the reader can judge whether the result matters in practice.

Directional or non-directional?

State a direction only when earlier studies or theory give you a clear reason to expect one. A directional hypothesis is tested one-tailed, which makes it easier to reach significance in that direction but means you cannot claim an effect in the opposite direction. Choosing the direction after seeing the data is not acceptable. When in doubt, use a non-directional hypothesis and a two-tailed test; most examiners and reviewers will not question it.

Deriving hypotheses from your literature

A hypothesis should not come from nowhere. Each one should follow from the literature or from a theory, and your thesis or paper should show the reasoning in a paragraph or two before the hypothesis is stated. In a quantitative thesis, each arrow in your conceptual framework usually becomes one hypothesis. The usual pattern:

1. State what theory or earlier studies say. “The Job Demands–Resources model predicts that job resources reduce burnout, and studies of nurses in several countries support this.”
2. Explain why it should apply to your context, or why it might differ. “Private hospitals in India offer fewer formal job resources, which may make supervisor support more important.”
3. State the hypothesis. “H2: Supervisor support is negatively associated with burnout among nurses in private hospitals in Chennai.”

In management and social science theses using a conceptual model, number the hypotheses (H1, H2, H3a, H3b) and show them as paths on the model diagram. Each arrow should match exactly one hypothesis, and each hypothesis exactly one arrow or set of arrows.

How many hypotheses?

As many as your objectives and model need, and no more. A PhD with five objectives might have eight to fifteen hypotheses. A thesis with forty hypotheses, most of them demographic comparisons (“there is no significant difference in X by age, gender, income, education…”), is common in some Indian departments and is weaker for it. Those comparisons rarely answer the research question. Keep them for a descriptive section if your department expects them, and focus the hypotheses on the relationships your study is really about.

Testing many hypotheses also raises the chance that some will be significant by accident. With 20 independent tests at the 0.05 level, you would expect about one false positive even if no real effects existed. Where you run many related tests, consider a correction and discuss it with your statistician or supervisor.

Examples of research hypotheses from several fields

Each row shows how a research question becomes a testable hypothesis, and the kind of test that would usually evaluate it. The right test depends on your data and design; the statistical test guide matches tests to data types.

Every hypothesis above names a population, a measurable variable and a relationship. The test follows from the hypothesis and the data type, and the sample size follows from the test. Our guide to sample size calculation covers power analysis, which needs you to know your hypothesis and test before you start collecting data.

When you do not need a hypothesis

Not every study tests hypotheses, and forcing one onto the wrong kind of study makes the work weaker.

• Descriptive studies that estimate how common something is (“the level of awareness of crop insurance among farmers”) need research questions, not hypotheses.
• Qualitative studies explore meanings and processes. They are guided by research questions, and theory may emerge from the data rather than being tested.
• Exploratory studies in new areas may not have enough prior work to justify a prediction.
• Design and development work in engineering usually states performance objectives (“to achieve accuracy above the current baseline”) rather than formal hypotheses, although comparisons with baselines can be stated and tested as hypotheses.

If your committee asks for hypotheses in a study like these, explain the design and propose research questions instead. Most will accept a clear rationale. Where the synopsis sits in this is covered in our PhD synopsis guide.

Hypothesis checklist and common mistakes

Check each hypothesis against this list:

• It names the variables, and each one can be measured with the data you will actually have.
• It states the population or context.
• It makes a prediction that the data could show to be wrong.
• It is one claim, not two joined with “and”.
• It follows from your literature review or theory, not from a hunch.
• The direction, if stated, is justified by previous studies.
• You know which statistical test will evaluate it before you collect data.

Common mistakes

• Writing the hypothesis after the analysis to match whatever turned out significant. This is sometimes called HARKing (hypothesising after the results are known), and examiners who notice it take it seriously.
• Stating only null hypotheses, in a list, with no reasoning. Many Indian theses do this by convention. If your department expects it, fine, but give the alternative and its rationale too.
• Hypotheses about variables the questionnaire does not measure.
• Reporting “H₀ accepted” or “hypothesis proved”.
• Treating “no significant difference” as proof of no difference.

Where hypotheses fit in the methods chapter as a whole is covered in our guide to writing the research methodology chapter. If you would like a methodologist to review your model and hypotheses before data collection, our research methodology consulting does that.