Psychology Experiment Design (Research Guide)
A psychology experiment turns a focused question into a planned test. Define what you will compare and measure, protect participants, and decide how you will analyze results before collecting data. Get your institution’s required ethics review first. Then follow a consistent procedure, report results and limits clearly, and document changes so others can judge what the evidence supports.
A strong experiment is not just an interesting activity. It is a careful plan for testing whether a change in one factor is linked to a measurable difference in another. For a student project, that means choosing a manageable question, checking the rules with your instructor, and resisting the urge to start recruiting before the review process is clear.
I often see research plans become easier to assess once students define exactly what they will do and measure. This guide walks through that process, from question to report. Institutional rules take priority, since a project’s review needs depend on where and how it is conducted.
Research Question, Experimental Design & Ethics Review
An experimental design is the plan for comparing conditions and measuring outcomes. It connects a research question to a testable hypothesis, participant group, procedure, and analysis. Ethical review is a separate check on whether the planned work meets applicable rules for participant protection. Both plans should be settled before recruitment or data collection begins.
Turn a topic into a testable question
Start with a question narrow enough to test. “Does stress affect memory?” needs more detail: What kind of stress? Which memory task? Who will take part, and when will memory be measured?
A hypothesis predicts a result that evidence could support or fail to support. For example: “Students assigned to a brief timed task will recall fewer words than students assigned to an untimed task.” The prediction does not claim that every participant will respond the same way.
Define the variables in plain, observable terms:
- Independent variable: The factor you change or compare, such as task timing.
- Dependent variable: The outcome you measure, such as the number of words recalled.
- Operational definition: The exact way a concept is created or measured. “Memory” might mean the number of correctly recalled words after a set delay.
- Control or comparison condition: The condition used to interpret the result, such as the same task without a time limit.
A clear operational definition helps another researcher understand what your terms mean in practice. It also helps you spot mismatches, such as a hypothesis about long-term memory paired with a measure taken immediately after learning.
Choose a design that fits the question
In an experiment, researchers assign or expose participants to different conditions and compare outcomes. In a correlational study, researchers measure variables without assigning the condition. Correlation can show that variables vary together, but it does not by itself show that one caused the other.
| Design choice | What happens | Example | Main caution |
|---|---|---|---|
| Between-subjects | Each person takes part in one condition | One group has a timed task; another has an untimed task | Groups may differ before the task |
| Within-subjects | Each person takes part in more than one condition | Each person completes both task versions | Order or practice effects may affect results |
| Random assignment | Chance decides who enters each condition | A computer assigns participants to groups | It does not guarantee identical groups, especially in small samples |
| Matched assignment | People are paired on a relevant feature, then split across conditions | Pair participants by baseline score | Matching cannot account for every difference |
Choose a design based on the question, setting, measures, and likely risks. For instance, a within-subjects study may use fewer participants, but the first task could affect performance on the second. Counterbalancing, or varying the order across participants, can help address order effects when suitable.
Get the required ethics review before starting
Ethics review checks whether a study has suitable protections for the people taking part. In the United States, research covered by the Common Rule, 45 CFR 46, must follow its requirements. Other rules may apply, and institutions decide which review path fits a project.
Contact your instructor or research office before recruiting or collecting identifiable or private information. Do not decide on your own that a class project is exempt. Some learning activities may not count as research under the applicable rules, but the institution must confirm the correct pathway.
A consent process should explain the study in understandable language. Depending on the project, it may cover what participants will do, foreseeable risks, possible benefits, privacy, voluntary participation, and how to stop taking part. If deception is proposed, ask the review body whether it is allowed and what debriefing is needed.
Minors and other protected groups may need additional safeguards. Do not assume that a parent’s permission, a consent form, or removal of names replaces required review. Keep written confirmation of the review decision and follow the approved plan.
Methods, Measures, Sample Size & Analysis Requirements
Methods explain how the study will run, what information it will collect, and how results will be tested. A complete plan sets out recruitment, measures, procedure, sample-size reasoning, exclusions, missing-data rules, and analysis before data collection. This makes the study easier to carry out and reduces the risk of changing decisions to fit the results.
Write the procedure so it can be repeated
Describe who may take part and how you will invite them. State where the study will happen, what participants will see or do, how long it will take, and how each condition will differ. Keep instructions as consistent as possible.
A pilot test is a small trial of the procedure, used to find unclear instructions, timing problems, or technical issues. Run one only after the appropriate review or approval. If a pilot leads you to change the study, check whether those changes need review before using the new procedure.
Plan data security in practical terms. Explain what data you will collect, who can access it, where it will be stored, and how long it will be kept or when it will be destroyed. Follow the approved plan and your institution’s rules.
Select measures and justify the sample size
A measure is a tool or task used to record a variable. Check that it fits the question and participant group. If you use a published scale or task, verify its instructions, scoring, and any permission requirements. Avoid changing a measure without recording what changed.
An a priori power analysis estimates the sample size needed for a planned statistical test under stated assumptions. It uses the expected effect size, significance level, desired statistical power, and test. An effect size describes the size of a difference or relationship; power is the chance of detecting an effect of a chosen size if it exists.
A significance level of .05 and power of .80 are common planning choices, not universal rules. The right assumptions depend on the question, method, measure, and available evidence. Ask your instructor or methods advisor for help selecting them. Do not pick an arbitrary participant count or treat a small p-value as proof that an effect matters.
Decide how you will analyze the data
Write down the main analysis before collecting data. Name the outcome, comparison, and statistical test, and explain why the test fits the design and data. Also set rules for incomplete responses, unusual values, and exclusions before you see the results.
| Planning item | Decide in advance | Why it matters |
|---|---|---|
| Primary outcome | The main measure used to test the hypothesis | Keeps the main question clear |
| Exclusions | Which records, if any, will not be analyzed and why | Limits decisions made after viewing results |
| Missing data | How unanswered items or unfinished tasks will be handled | Makes data handling transparent |
| Main analysis | The test and comparison you plan to use | Connects analysis to the research question |
| Deviations | How you will record departures from the plan | Helps readers understand what actually happened |
Do not silently change the analysis if the data look different from what you expected. You may conduct an additional analysis, but label it as exploratory, explain the reason, and distinguish it from the planned test.
Use a hypothetical study to check the plan
Suppose a student wants to test whether background music affects performance on a short attention task. The student defines performance as the number of correct responses in a fixed time, then compares a music condition with a no-music condition.
Before starting, the student would need to settle the assignment method, task instructions, recruitment plan, consent materials, data protection, sample-size reasoning, and analysis. The student must also get the institution’s required review decision before recruitment or data collection. This example is illustrative, not evidence that music changes attention.
Academic Preparation, Research Skills & Student Experience
Preparing an experiment builds skills in reading research, defining variables, planning procedures, and evaluating evidence. These skills matter in an undergraduate methods course or capstone, but the right scope depends on course goals, supervision, time, and institutional rules. A smaller, well-controlled project is often more useful than a broad question the student cannot test well.
Build the proposal in a sensible order
Use this sequence to move from idea to approved protocol:
- Read relevant research and note how earlier studies defined and measured key terms.
- Write one focused question and a hypothesis that could be tested.
- Choose a design, conditions, and measures that match the question.
- Draft recruitment, consent, procedure, privacy, and data-storage plans.
- Set the sample-size rationale, exclusions, missing-data rules, and main analysis.
- Ask your instructor or supervisor to review the plan.
- Submit it through the institutional review process and wait for the required decision.
- Begin only under the confirmed review pathway, then follow the approved protocol.
Preregistration means recording study plans before examining the results, often in a public or controlled registry. It can make planned decisions easier to distinguish from later choices. It does not replace ethics review, and your course or institution may have specific rules about whether or how to use it.
Keep a record of what happens
Use a version-controlled protocol or dated notes to record approved procedures and changes. If a participant experiences an unexpected issue, or the procedure cannot be followed as written, contact the supervisor or review office for guidance. Do not make an unapproved change and assume it is harmless.
Collect only the information needed for the study. Store it as approved, limit access to authorized people, and report results without exposing participants. At the end, follow the approved plan for keeping or disposing of data.
A protocol violation can put the project at risk. Starting before required review, using unapproved recruitment or measures, or changing procedures without permission may lead to data being unusable or other institutional action. Confirm the review pathway in writing before starting.
Career Pathways, Graduate Study & Research Employment Outcomes
A student experiment can provide practice in research tasks used in later academic or research roles. It does not guarantee a job, admission, or a particular career result. Its value depends on what you learn, how carefully you work, and whether you can explain the study’s strengths and limits.
Use the project to develop research skills
A well-documented project can help you discuss skills such as literature review, survey or task administration, data handling, statistical reasoning, and research ethics. If you apply for a research assistant role or graduate study, describe what you did accurately. Be clear about your contribution, the supervisor’s role, and the project’s limits.
Reporting should help readers judge the evidence, not just see whether a result was statistically significant. APA’s JARS–Quant guidance supports clear reporting of quantitative research, including the sample, measures, procedures, analysis, and results. Report effect sizes and uncertainty estimates where appropriate, along with limitations and any departures from the plan.
Before sharing a paper, poster, or dataset, check course, supervisor, ethics, and data-sharing rules. Some information may not be suitable for public release, even if names have been removed. The final report should reflect what the approved study actually did.
Key next step: Ask your instructor or institutional research office what review applies to your project, then build the design and analysis plan around that guidance.
Frequently Asked Questions
What is a psychology experiment?
A psychology experiment tests a question by comparing conditions and measuring an outcome. Researchers assign or expose participants to conditions, then assess whether the outcome differs. The specific design and review needs depend on the question, participants, and setting.
What should a psychology experiment proposal include?
Include the question, hypothesis, variables, design, participants, recruitment, measures, procedure, ethics plan, sample-size rationale, and analysis plan. State how you will handle exclusions, missing data, and deviations from the planned procedure.
Do I need IRB approval for a class project?
You need the review or determination required by your institution. Do not decide on your own that a class project is exempt. Ask your instructor or research office before recruiting participants or collecting identifiable or private data.
Can I collect data while my ethics review is pending?
Do not recruit or collect data until you have the institutional decision required for your project. Starting early may violate institutional rules and could make the data unusable.
How many participants do I need?
There is no single sample size that fits every experiment. Use an a priori power analysis based on the planned test, justified effect size, significance level, and desired power, with guidance from your instructor or methods advisor.
What is the difference between random sampling and random assignment?
Random sampling concerns how people are selected from a population. Random assignment concerns how enrolled participants are placed into study conditions. Assignment can help reduce group differences, but it does not make a sample representative of everyone.
What does statistical significance mean?
Statistical significance describes how compatible data are with a specified statistical model and null hypothesis. It does not show that an effect is large, important, or certainly real. Report effect size and uncertainty as well.
What should I do if I need to change the protocol?
Pause and contact your supervisor or institutional review office. Follow the required process for proposed changes before using them. Record any permitted changes and report departures from the plan in your final write-up.
What should a final psychology experiment report include?
A quantitative report should describe the sample, measures, procedure, analysis, results, effect sizes, uncertainty, and limitations. Use APA JARS–Quant guidance where appropriate, and explain deviations from the planned method.
Does preregistration replace ethics review?
No. Preregistration records study plans; ethics review assesses participant protections and applicable rules. They serve different purposes, and preregistration does not replace institutional review.
(This article was written by one of our staff writers, Alan Westbrook. Visit our Meet the Team page.)
