How to Write Research Hypotheses: Types, a Formula and Examples
How to write research hypotheses for a seminar paper or thesis - directional and non-directional, relationship, difference, interaction and mediation hypotheses, a formula, how the hypothesis decides the statistical test, and right and wrong examples.
Part of the series: How to write the method chapter of an empirical paper
Hypotheses are the spine of the paper: the review leads to them, the method is built to test them, and the results chapter is organised by them. A well-written hypothesis decides the statistical test before you've collected a single data point.
What makes a good hypothesis?
A research hypothesis is an expected, specific, testable claim about the relationship between variables. It:
- follows from the review - it rests on theory or previous studies;
- names measurable variables - ones that appear in your measures section;
- states a direction when there's a basis for one ("higher", "lower", "positive relationship");
- is falsifiable - you can imagine a result that would contradict it.
The formula
[variable A] + [type of relationship/difference + direction] + [variable B] + [in population/condition]
- "Social support will be positively related to study satisfaction among first-year students."
- "Students who study with retrieval practice will score higher on the test than students who re-read."
Types of hypotheses - and the test each dictates
| Type | Example wording | Typical test |
|---|---|---|
| Relationship | anxiety will be negatively related to grades | Pearson/Spearman correlation |
| Difference, 2 groups | women will report higher anxiety than men | t-test |
| Difference, 3+ groups | burnout will differ between three departments | ANOVA |
| Prediction | support and self-efficacy will predict satisfaction | multiple regression |
| Moderation (interaction) | the stress–burnout link will be stronger among those with low support | moderation / two-way ANOVA |
| Mediation | worry will mediate the link between anxiety and grades | PROCESS |
| Categorical association | dropout rates will differ between study tracks | chi-square |
Unsure which test fits? See How to choose a statistical test.
Directional or non-directional?
- Directional: "a positive relationship", "higher". When the literature points to a direction - which is the case in most papers.
- Non-directional: "there will be a difference", "a relationship". When the literature is contradictory or thin. Then it's often better to write a research question instead: "Is there a difference...".
Right and wrong hypotheses
| ❌ | The problem | ✅ |
|---|---|---|
| "Anxiety affects studying" | Affects what? How is "studying" measured? A correlational study can't show an effect | "Test anxiety will be negatively related to GPA" |
| "The study will examine the relationship between..." | That's an aim, not a hypothesis | "There will be a positive relationship between..." |
| "Happier students will succeed more and sleep more" | Two hypotheses in one | Split into hypotheses 1 and 2 |
| "There will be no difference..." | A null hypothesis - usually not written | State the research hypothesis, or a research question |
Where and how they appear
At the end of the introduction, after the rationale (Writing the introduction and rationale), numbered: Hypothesis 1, Hypothesis 2... Each can be followed by a short sentence recalling what it's based on. The same numbering returns in the results chapter - a paragraph per hypothesis - and in the discussion.
FAQ
How many hypotheses does a seminar paper need? Usually 2–4. More makes the analysis and discussion hard to manage.
Hypothesis vs research question? A hypothesis predicts an outcome ("a positive relationship will be found"); a research question asks without predicting ("is there a relationship?"). Use a question when there's no basis for a direction, and in qualitative research.
My hypothesis wasn't supported - did I fail? No. An unsupported hypothesis is a legitimate finding; report it exactly like a supported one and explain it in the discussion. See How to write a discussion chapter.
Can I change a hypothesis after seeing the results? No - that's HARKing (hypothesising after the results are known) and it undermines credibility. Report an unexpected finding as unexpected.
Not sure your hypotheses fit your data and tests? Tell us about your study - we'll go through the hypotheses and match an analysis to each.
Tell us what you need, and we'll come back with an expert in Research methods & questionnaires and a quote.
Last updated: