How to write a research hypothesis: directional, non-directional, null, and alternative hypotheses

To write a research hypothesis, state a clear, testable prediction about the expected relationship between two or more variables. In quantitative research, a good hypothesis identifies the variables, suggests what kind of relationship or difference you expect, and can be tested with data.

Short summary: A research hypothesis turns a broad research interest into a specific claim that can be examined statistically. This article explains how to write a research hypothesis, how it differs from a research question, and how to use directional, non-directional, null, and alternative hypotheses correctly. It also includes worked research hypothesis examples across common student research topics.

What is a research hypothesis?

A research hypothesis is a specific, testable statement about what you expect to find in a study. It is usually used in quantitative research, where the researcher measures variables and tests relationships, differences, or effects.

A hypothesis often answers a question such as:

• Does one variable affect another?
• Are two groups different?
• Are two variables associated?
• Does an intervention change an outcome?
• Is one condition linked to a higher or lower score than another?

For example:

Students who sleep at least 7 hours per night will report lower perceived stress than students who sleep fewer than 7 hours per night.

This statement is testable because it identifies:

• Population: students
• Independent or grouping variable: sleep duration
• Dependent variable: perceived stress
• Expected relationship: more sleep is associated with lower stress

How is a hypothesis different from a research question?

A research question asks what the study aims to find out. A hypothesis predicts the likely answer.

Example:

• Research question: What is the relationship between sleep duration and perceived stress among undergraduate students?
• Research hypothesis: Undergraduate students who report longer sleep duration will report lower perceived stress.

The research question is open-ended. The hypothesis is a testable prediction.

Not every study needs a hypothesis. Exploratory qualitative research often uses research questions rather than hypotheses. However, quantitative studies commonly use hypotheses because they are designed to test measurable relationships or differences.

How to write a research hypothesis in 5 steps

A strong hypothesis usually follows a logical sequence: topic, variables, relationship, population, and testability.

1. Start with a focused research topic

Begin with a topic that is narrow enough to study with available time and data.

Too broad:

Social media and mental health

More focused:

The relationship between daily social media use and anxiety symptoms among first-year university students

A focused topic makes it easier to identify variables and choose a suitable method.

2. Identify the variables

A variable is something that can vary and be measured.

Common types include:

• Independent variable: the predictor, exposure, condition, or factor that may influence another variable
• Dependent variable: the outcome being measured
• Control variable: a factor included to reduce alternative explanations
• Grouping variable: a category used to compare groups, such as year of study, programme type, or intervention group

Example:

Topic: Social media use and anxiety symptoms among first-year university students

Possible variables:

• Independent variable: daily social media use, measured in hours
• Dependent variable: anxiety symptoms, measured using a validated scale
• Population: first-year university students

3. Decide what relationship you expect

A hypothesis should indicate what you expect to happen. Depending on the study, you might predict:

• A positive association
• A negative association
• A difference between groups
• An effect of an intervention
• A relationship without specifying direction

Examples:

• Positive association: as study time increases, exam score increases.
• Negative association: as financial stress increases, academic satisfaction decreases.
• Group difference: students in an online learning group differ from students in an in-person learning group.
• Intervention effect: participants who receive feedback improve more than those who do not.

4. Make the statement testable

A hypothesis must be testable with data. Avoid vague terms that cannot be measured clearly.

Weak:

Better teaching improves student success.

Stronger:

Students who receive weekly formative feedback will achieve higher final assessment scores than students who receive only end-of-module feedback.

The stronger version identifies a comparison, a measurable outcome, and an expected direction.

5. Match the hypothesis to your method

Your hypothesis should fit your research design and planned analysis.

For example:

• If you are comparing two groups, your hypothesis may involve a difference in means.
• If you are examining association, your hypothesis may involve correlation.
• If you are testing prediction, your hypothesis may fit regression analysis.
• If you are evaluating an intervention, your hypothesis may compare pre-test and post-test scores or intervention and control groups.

An AI-powered academic writing assistant can help students organise variables, draft hypothesis options, and check whether a hypothesis aligns with the research question and chapter outline, while the student remains responsible for reviewing, editing, and using the work appropriately.

What are directional and non-directional hypotheses?

A hypothesis can be directional or non-directional, depending on whether it predicts the direction of the relationship or difference.

What is a directional hypothesis?

A directional hypothesis predicts the specific direction of the expected result. It states whether a variable will increase, decrease, be higher, be lower, or have a positive or negative association.

Examples:

• Students who attend more seminars will achieve higher course grades.
• Higher caffeine intake is associated with higher self-reported sleep disturbance.
• Participants who complete a mindfulness programme will report lower stress after the programme than before it.

Directional hypotheses are often used when previous theory or prior research gives a clear reason to expect a particular direction.

What is a non-directional hypothesis?

A non-directional hypothesis predicts that a relationship or difference exists, but it does not state the direction.

Examples:

• There is a relationship between seminar attendance and course grades.
• Caffeine intake is associated with self-reported sleep disturbance.
• Stress scores will differ before and after a mindfulness programme.

Non-directional hypotheses may be suitable when there is limited prior research, mixed evidence, or a reason to avoid predicting a specific direction.

What is the null hypothesis?

The null hypothesis is the statement that there is no effect, no relationship, or no difference in the population being studied. It is commonly written as H0.

The null hypothesis is not usually what the researcher hopes to prove. Instead, it provides a baseline claim that statistical testing evaluates.

Examples:

• There is no relationship between sleep duration and perceived stress among undergraduate students.
• There is no difference in mean exam scores between students who receive weekly feedback and those who do not.
• The mindfulness programme has no effect on participants’ stress scores.

In statistical testing, researchers usually assess whether the data provide enough evidence to reject the null hypothesis. If the evidence is not strong enough, the researcher does not “prove” the null hypothesis; they simply fail to reject it.

What is the alternative hypothesis?

The alternative hypothesis is the statement that an effect, relationship, or difference exists. It is commonly written as H1 or Ha.

The alternative hypothesis usually reflects the researcher’s expected finding.

Examples:

• There is a relationship between sleep duration and perceived stress among undergraduate students.
• Students who receive weekly feedback achieve different exam scores from students who do not.
• Participants’ stress scores decrease after completing a mindfulness programme.

The alternative hypothesis may be directional or non-directional.

Null hypothesis vs. alternative hypothesis

The null and alternative hypotheses are paired statements. They should refer to the same variables, population, and comparison.

Example pair:

• H0: There is no relationship between daily social media use and anxiety symptoms among first-year university students.
• H1: There is a relationship between daily social media use and anxiety symptoms among first-year university students.

Directional version:

• H0: Daily social media use is not associated with anxiety symptoms among first-year university students.
• H1: Higher daily social media use is associated with higher anxiety symptoms among first-year university students.

The wording depends on the statistical approach and your institution’s guidance. In many student projects, the main requirement is that the hypotheses are clear, consistent, and testable.

Research hypothesis examples by study type

The best way to understand hypothesis writing is to see how the wording changes across different research designs.

Example 1: Correlational study

Topic: Sleep and stress among university students

Research question: What is the relationship between sleep duration and perceived stress among undergraduate students?

Directional hypothesis:

Undergraduate students who report longer sleep duration will report lower perceived stress.

Non-directional hypothesis:

There is a relationship between sleep duration and perceived stress among undergraduate students.

Null hypothesis:

There is no relationship between sleep duration and perceived stress among undergraduate students.

Why it works:

• The variables are measurable.
• The population is defined.
• The directional version states the expected direction.
• The null version states no relationship.

Example 2: Group comparison study

Topic: Feedback frequency and assessment performance

Research question: Do students who receive weekly formative feedback achieve different assessment scores from students who receive only final feedback?

Directional hypothesis:

Students who receive weekly formative feedback will achieve higher assessment scores than students who receive only final feedback.

Non-directional hypothesis:

Assessment scores will differ between students who receive weekly formative feedback and students who receive only final feedback.

Null hypothesis:

There will be no difference in assessment scores between students who receive weekly formative feedback and students who receive only final feedback.

Why it works:

• It compares two clearly defined groups.
• The outcome is measurable.
• The directional version predicts which group will score higher.

Example 3: Intervention study

Topic: Mindfulness practice and stress

Research question: Does a four-week mindfulness programme affect perceived stress among postgraduate students?

Directional hypothesis:

Postgraduate students will report lower perceived stress after completing a four-week mindfulness programme than before the programme.

Non-directional hypothesis:

Postgraduate students’ perceived stress scores will differ before and after completing a four-week mindfulness programme.

Null hypothesis:

There will be no difference in postgraduate students’ perceived stress scores before and after completing a four-week mindfulness programme.

Why it works:

• It includes a before-and-after comparison.
• The intervention is defined.
• The outcome is measurable.

Example 4: Prediction study

Topic: Study time and exam performance

Research question: Does weekly study time predict exam performance among first-year students?

Directional hypothesis:

Greater weekly study time will predict higher exam scores among first-year students.

Non-directional hypothesis:

Weekly study time will predict exam scores among first-year students.

Null hypothesis:

Weekly study time will not predict exam scores among first-year students.

Why it works:

• It uses prediction language suitable for regression.
• It identifies the predictor and outcome.
• The directional hypothesis states the expected positive relationship.

Example 5: Association between categorical variables

Topic: Mode of study and course completion

Research question: Is mode of study associated with course completion status among adult learners?

Alternative hypothesis:

Mode of study is associated with course completion status among adult learners.

Null hypothesis:

Mode of study is not associated with course completion status among adult learners.

Why it works:

• The variables are categorical.
• It avoids claiming causation.
• The wording fits an association-based analysis.

What makes a good research hypothesis?

A good research hypothesis is:

• Specific: it identifies the variables and population.
• Testable: it can be examined using data.
• Clear: it avoids vague wording.
• Focused: it does not try to answer too many questions at once.
• Aligned: it matches the research question, literature review, and method.
• Ethical: it does not make harmful assumptions about groups or individuals.
• Appropriately cautious: it predicts a relationship or difference without overstating certainty.

A helpful formula is:

Among [population], [independent variable or condition] is associated with / affects / predicts [dependent variable] in [expected direction].

Example:

Among undergraduate students, higher weekly study time is associated with higher exam performance.

For a group comparison:

[Group 1] will have higher/lower/different [outcome] than [Group 2].

Example:

Students who receive weekly feedback will have higher assessment scores than students who receive only final feedback.

Common mistakes when writing hypotheses

Making the hypothesis too broad

Too broad:

Technology affects education.

Better:

Students who use a course planning app at least three times per week will report higher study organisation scores than students who do not use the app.

Using variables that cannot be measured

Hard to test:

Students with a positive mindset will do better.

Better:

Students with higher academic self-efficacy scores will achieve higher final assessment scores.

Confusing correlation with causation

Careful wording matters. If your study is correlational, avoid causal language such as “causes” or “leads to”.

Problematic:

Social media use causes anxiety among students.

Better for a correlational study:

Higher daily social media use is associated with higher anxiety symptom scores among students.

Writing a hypothesis that does not match the research question

Research question:

What is the relationship between attendance and course satisfaction?

Mismatched hypothesis:

Students who attend more classes will receive higher exam scores.

Better:

Higher class attendance is associated with higher course satisfaction.

Including too many variables in one hypothesis

Overloaded:

Sleep, diet, exercise, social media use, income, and study time will predict stress, grades, attendance, and student satisfaction.

Better:

Higher weekly exercise frequency is associated with lower perceived stress among undergraduate students.

You can include multiple hypotheses in a larger study, but each one should be easy to understand and test.

Should you use a directional or non-directional hypothesis?

Use a directional hypothesis when theory or prior research gives you a clear reason to predict a specific direction. For example, if earlier literature suggests that higher study time is generally linked to higher exam performance, a directional hypothesis may be appropriate.

Use a non-directional hypothesis when:

• Previous findings are mixed.
• The topic is under-researched.
• You expect a difference but cannot justify the direction.
• Your supervisor or course guidance recommends non-directional wording.

If you are unsure, check your module handbook, dissertation guide, or supervisor’s preference. Some institutions expect formal null and alternative hypotheses, while others prefer a research question followed by one or more research hypotheses.

How should hypotheses appear in a dissertation or research paper?

In student research projects, hypotheses often appear near the end of the introduction or literature review, after the rationale for the study has been established.

A typical sequence is:

1. Introduce the topic.
2. Review relevant literature.
3. Identify a gap, limitation, or unresolved question.
4. State the research aim.
5. Present the research question.
6. Present the hypothesis or hypotheses.

Example wording:

Based on the literature on feedback and student performance, this study examines whether weekly formative feedback is associated with assessment outcomes. The research question is: Do students who receive weekly formative feedback achieve higher assessment scores than students who receive only final feedback? The hypothesis is that students who receive weekly formative feedback will achieve higher assessment scores.

This structure helps the reader see why the hypothesis follows logically from the literature.

Quick checklist for writing a research hypothesis

Before finalising your hypothesis, ask:

• Have I identified the population?
• Have I named the independent and dependent variables?
• Is the hypothesis testable with the data I plan to collect?
• Does the wording match my research design?
• Have I avoided unsupported causal language?
• Is the direction justified, if I use a directional hypothesis?
• Is the null hypothesis clearly paired with the alternative hypothesis?
• Would another student understand exactly what I plan to test?

If the answer to any of these questions is no, revise the hypothesis before moving on to data collection or analysis planning.

Summary

A research hypothesis is a clear, testable prediction about a relationship, difference, or effect. To write one well, define your variables, specify your population, decide whether the hypothesis should be directional or non-directional, and make sure it fits your research question and method. The null hypothesis states that there is no relationship, difference, or effect, while the alternative hypothesis states that one exists. Strong hypotheses are specific, measurable, and cautious enough for academic research.

Frequently Asked Questions

What is the easiest way to write a research hypothesis?

Use this structure: “Among [population], [independent variable] is associated with [dependent variable] in [expected direction].” Then check that both variables can be measured.

What is the difference between a null hypothesis and an alternative hypothesis?

The null hypothesis states that there is no relationship, difference, or effect. The alternative hypothesis states that a relationship, difference, or effect exists.

What is a directional hypothesis?

A directional hypothesis predicts the expected direction of the result, such as higher, lower, positive, negative, increase, or decrease.

Can a hypothesis be non-directional?

Yes. A non-directional hypothesis predicts that a relationship or difference exists but does not specify its direction.

Do all dissertations need hypotheses?

No. Quantitative dissertations often use hypotheses, but qualitative or exploratory dissertations may use research questions instead. Always follow your programme guidance.

Recommended internal links

• Recommended internal link: "How to write a research question"
• Recommended internal link: "How to write a literature review"
• Recommended internal link: "How to choose a dissertation topic"
• Recommended internal link: "How to build a dissertation chapter outline"
• Recommended internal link: "How to revise a first draft"