Hypothesis Examples in Psychology With Template

A psychology hypothesis is a testable prediction about how measurable variables relate, written so that a study can support it or fail to support it. This guide gives a fill-in template, four worked examples (directional, non-directional, correlational and group-difference), a table that turns vague claims into testable ones, and a checklist. Every study, participant group and result below is fictional and invented for illustration. For the general theory of hypotheses across disciplines, see how to write a strong research hypothesis.

What makes a psychology hypothesis testable

A testable hypothesis names the variables, says how each will be measured, states the expected relationship, and could be contradicted by the data. In psychology that usually means four things: a population (who), an independent variable or predictor (what is manipulated or compared), a dependent variable or outcome (what is measured), and a direction or pattern (what you expect). Constructs such as stress, memory or motivation are not directly observable, so the hypothesis must tie each one to an operational definition, for example “number of words correctly recalled after a 10-minute delay”.

Most hypotheses are stated as a pair. The null hypothesis (H0) predicts no effect or no difference; the alternative hypothesis (H1) predicts that there is one. A significance test then rejects or fails to reject H0; it does not prove either. See null hypothesis vs. alternative hypothesis for the definitions.

A fill-in template

Copy this frame and replace each bracket:

Alternative (H1): Among [population], [independent variable or group] will be associated with [higher / lower / different] [dependent variable, operationally defined] than [comparison condition].
Null (H0): Among [population], there will be no difference in [dependent variable] between [group or condition A] and [group or condition B].
Direction: directional (one-tailed) or non-directional (two-tailed), chosen before data collection.
Measurement: the instrument or task and the score that will be analysed.

Worked examples (all fictional)

Directional (one-tailed) example

Study idea: a fictional experiment on sleep and recall.

H1: Among university students, those restricted to four hours of sleep will recall fewer words on a 20-word delayed recall task than students who sleep eight hours.
H0: Among university students, there will be no difference in delayed recall scores between the sleep-restricted and the full-sleep groups.

The prediction names a direction (“fewer”), so the test is one-tailed. A one-tailed test looks for an effect in one direction only; see one-tailed vs. two-tailed tests for when that choice is defensible.

Non-directional (two-tailed) example

H1: Among adult volunteers, background music will change the number of arithmetic problems solved in five minutes compared with silence.
H0: Among adult volunteers, background music will not change the number of arithmetic problems solved in five minutes compared with silence.

Here the researcher has no basis for predicting whether music helps or hinders, so the alternative allows either direction.

Correlational example

H1: Among first-year students, weekly hours of social media use will be positively correlated with scores on a self-report loneliness scale.
H0: Among first-year students, there will be no correlation between weekly hours of social media use and loneliness scores.

A correlational hypothesis predicts an association, not a cause. Wording such as “social media use will cause loneliness” would claim more than a correlational design can show.

Group-difference example

H1: Children taught with a fictional “story-based” reading programme will score higher on a reading comprehension test after twelve weeks than children taught with the standard programme.
H0: Children taught with the story-based programme will not differ in reading comprehension scores after twelve weeks from children taught with the standard programme.

Weak versus testable: rewrites

Weak statement	Problem	Testable rewrite (fictional)
Stress affects memory.	No population, no measures, no direction.	Adults who complete a timed arithmetic task under evaluation will recall fewer words on a delayed recall test than adults who complete it without evaluation.
People are happier when they exercise.	“Happier” and “exercise” are not operationally defined.	Among office workers, minutes of moderate exercise per week will be positively correlated with scores on a self-report wellbeing scale.
Therapy works better.	Better than what, and measured how?	Participants who receive eight sessions of a fictional skills-based programme will report lower anxiety scores than a waiting-list group at the end of the programme.
Students may or may not do better with feedback.	Cannot be contradicted by any result.	Students who receive written feedback within 48 hours will score higher on the follow-up quiz than students who receive it after one week.

Hypothesis checklist

It names the population and both variables or groups.
Each construct has an operational definition (the task, scale or count that will be analysed).
The direction, or the decision to leave it open, is stated and matches the planned test.
It is falsifiable: some plausible result would fail to support it.
Correlational wording is not causal wording, and the reverse.
It is written before the data are seen.
Each hypothesis tests one relationship; split compound predictions.

Stating and reporting hypotheses in APA style

This section reflects the APA Publication Manual, 7th edition (2020), the current edition as of 2026-10. APA’s abstract guide lists hypotheses among the content an abstract should cover, alongside the research problem or question, methods, results and implications, and it describes a typical limit of 250 words. In the Results section, APA guidance on verb tense calls for past tense throughout, as in “hypotheses were not supported”. Stating the hypothesis in present or future tense in the Introduction and then reporting it in past tense in Results follows that pattern.

Using the fictional sleep example, a Results sentence would read: “Students in the sleep-restricted group recalled fewer words than students in the full-sleep group, and the hypothesis was supported.” Report the statistics your analysis produced alongside such a sentence; a verdict word alone is not a result. For paper-level formatting, see writing a research paper in APA style.

Preregistering your hypothesis

The Center for Open Science describes preregistration as fixing your plan before the study begins and submitting it to a registry (see Sources). It states that at least one confirmatory test must be specified in each preregistration, and that in a confirmatory study the specific hypothesis is very clear. It also says that once a researcher begins to change slightly the way the hypothesis is tested, the work should be considered exploratory. A hypothesis written with the template above, with variables, measures and direction fixed in advance, is what makes a clean confirmatory test possible. The COS page also notes that analysis plans can be conditional, using IF-THEN decision rules. For the wider picture see preprints and preregistration.

If your hypothesis grows out of a theory, building a theoretical framework from construct to hypothesis shows how to trace the prediction back to its constructs.

Frequently asked questions

What is an example of a hypothesis in psychology?

“Among university students, those restricted to four hours of sleep will recall fewer words on a delayed recall task than students who sleep eight hours” is a fictional example. It names the population, the manipulated variable, an operationally defined outcome and a direction.

What is the difference between a directional and a non-directional hypothesis?

A directional hypothesis predicts which way an effect will go (fewer words, higher scores). A non-directional hypothesis predicts only that a difference or relationship exists. The choice should be made before data collection and match the planned test.

Can a hypothesis be proven?

No. Data can support a hypothesis or fail to support it; a significance test rejects or fails to reject the null hypothesis rather than proving either.

Does every psychology study need a hypothesis?

No. Exploratory and many qualitative studies pose research questions instead. A hypothesis is expected where the study sets out to test a specific prediction.