Research Methods in Psychology Explained

Research methods is the course that turns psychology from a collection of interesting claims into a science. It teaches you how psychologists decide what is true, how they design studies and how to judge the evidence behind any headline that begins with "a new study shows."

This guide walks through the ideas you will meet in a typical research methods course, with plain examples and tips for designing or reading a study.

Why methods matter

Humans are poor judges of cause and effect from everyday life. We notice patterns that are not there, remember hits better than misses and let our expectations shape what we see. Research methods are tools to reduce those errors.

A well-designed study asks a specific question, defines its terms, collects data in a systematic way and analyzes it with methods that are open to checking. When you understand these steps, you can evaluate claims, not just accept them.

From question to hypothesis

Research begins with a question, such as whether sleep affects memory. A hypothesis turns that question into a testable prediction: "Students who sleep eight hours after studying will recall more words than students who stay awake."

A good hypothesis is:

1. Specific: it states what you expect to find.
2. Testable: you can collect data that could support or contradict it.
3. Falsifiable: some result could show it is wrong.
4. Grounded: it follows from prior research or theory.

Before designing a study, researchers read the existing literature so that they build on what is known instead of repeating it.

Variables and how to define them

A variable is a characteristic that can take different values. Two kinds matter most in simple designs.

• Independent variable (IV): the factor you manipulate or compare, such as hours of sleep.
• Dependent variable (DV): the outcome you measure, such as the number of words recalled.

Researchers also give each variable an operational definition, which states exactly how it will be measured. "Stress" is vague, while "score on a standard stress questionnaire" is concrete. Operational definitions let others repeat the study.

Other factors that might influence the outcome are confounding variables. Good designs try to hold these constant or balance them across groups.

Main types of research design

Different questions call for different designs. Most undergraduate courses cover the following.

Experiments

In an experiment, the researcher manipulates the IV and randomly assigns participants to conditions. Random assignment spreads other differences evenly across groups, which allows the researcher to claim that the IV caused a change in the DV.

Experiments often include a control group that does not receive the treatment, and may use blinding so that participants or researchers do not know who is in which group.

Quasi-experiments

When you cannot randomly assign people, as when comparing two existing classes, the design is quasi-experimental. It can show differences between groups but is weaker evidence of cause because the groups may differ in other ways.

Correlational studies

Correlational research measures two or more variables as they occur and describes their relationship. A correlation can be positive, negative or near zero. It cannot show cause. If ice cream sales and sunburns rise together, hot weather is likely the reason for both.

Observational studies

Researchers watch and record behavior, either in natural settings or in the lab. Observation is valuable for new topics and for behavior that is hard to elicit, but it can be affected by observer bias and by participants changing their behavior when watched.

Surveys

Surveys collect self-reports from many people quickly. Their quality depends on question wording, the sampling method and honest responding. Leading or double-barreled questions distort the results.

Case studies

A case study examines one person or a small group in depth. It can reveal rare conditions and generate ideas, but results do not generalize easily.

Longitudinal and cross-sectional designs

A cross-sectional study compares different groups, such as age groups, at one time. A longitudinal study follows the same people over time. Longitudinal studies track change more directly but take longer and lose participants.

Sampling

Researchers rarely study everyone, so they choose a sample from a population. How they choose determines how far the results apply.

• Random sampling gives every member of the population an equal chance of selection.
• Convenience sampling uses whoever is easy to reach, such as students in an introductory course.
• Stratified sampling divides the population into groups and samples from each.

Many classic findings come from university students in a few countries, which limits generalization. A responsible report names its sample and its limits.

Measurement: reliability and validity

Good conclusions need good measurement. Two qualities define it.

Reliability means consistency. A reliable scale gives similar results on repeated use. Researchers check test-retest reliability, internal consistency and agreement between raters.

Validity means the measure captures what it claims to. A bathroom scale that always reads five pounds low is reliable but not accurate. Types of validity include construct validity, criterion validity and content validity.

Studies also face threats to internal validity, whether the IV really caused the change, and external validity, whether the findings apply beyond the study.

Ethics

Psychology research with people is governed by ethical principles. These include informed consent, protection from harm, confidentiality, the right to withdraw and, where deception is used, a debriefing afterward. In the United States, studies usually go through an institutional review board before data collection.

Even for class projects, check your school's rules before you recruit a single participant. Ethics review protects participants and protects you.

Basic statistics in methods

Methods and statistics are closely linked. You will describe data with means, standard deviations and graphs, then use inferential tests to judge whether results are likely due to chance. A p-value is the probability of seeing results at least as extreme if there were no real effect. Many journals now also ask for effect sizes and confidence intervals, which show how big and how precise an effect is.

You do not need to calculate everything by hand, but you do need to understand what each test is for. Statistics for psychology students covers the main tests in plain language.

Planning a simple study

If your course asks you to design a study, use this sequence.

1. Pick a narrow question you can answer with data you can actually collect.
2. Review prior research so your question adds something.
3. State a hypothesis with variables and predictions.
4. Choose a design that fits the question and the limits of your resources.
5. Define your measures and check that they are reliable and valid.
6. Decide on a sample and how you will recruit it.
7. Plan ethics and obtain approval.
8. Plan your analysis before you collect data.
9. Pilot test your procedure with a few people.
10. Collect, analyze and report honestly, including limitations.

Reading a study critically

Use these questions on any article.

• Is the sample large and varied enough for the claim?
• Did the researchers manipulate a variable, or only measure it?
• Could a confounding variable explain the result?
• Are the measures reliable and valid?
• Is the effect large enough to matter, not only statistically significant?
• Have other studies found the same thing?
• Who funded the work, and could that bias the conclusions?

If a study cannot answer these questions, treat its headline with caution.

Open science and replication

Psychology has had a hard look at itself over the past decade. Some famous findings did not replicate when other teams repeated them. In response, many researchers now preregister hypotheses, share data and materials, and report all analyses, not just those that worked. Learning these practices early makes you a better reader and a more trustworthy researcher.

For help writing up a research paper in this style, the guide on psychology research paper help walks through structure and common problems.

Common mistakes

• Treating correlation as cause.
• Writing vague hypotheses that no result could contradict.
• Using a convenience sample without noting the limit.
• Collecting data before planning the analysis.
• Skipping ethics review.
• Reporting only significant results.
• Overgeneralizing from one small study.

Final thoughts

Research methods gives you a way of thinking that lasts beyond the course. By learning to ask how a claim was tested, what could explain the result and how strong the evidence really is, you become a more careful scientist and a more discerning reader of the world around you.