- Psychology: research methods and statistics
What is the difference between an experiment and a correlational study?
The difference between an experiment and a correlational study is control. In an experiment the researcher changes one variable, allocates people to conditions at random and measures the effect, so it can show cause. A correlational study measures two variables as they already occur and shows only whether they go together, because a third variable or the reverse direction may explain the link.
The topic explained
Two ways to study the same question
Suppose a psychologist wants to know whether sleep affects memory. In an experiment the researcher decides who sleeps how much. Forty volunteers are split by chance into two groups; one sleeps eight hours, the other five, and the next morning both take the same memory test. The amount of sleep is the independent variable (IV), the one the researcher changes, and the test score is the dependent variable (DV), the one that is measured.
In a correlational study the researcher changes nothing. She asks a few hundred students how many hours they usually sleep and collects their grades, then checks whether the two measures go up and down together. The result is a correlation coefficient, a number from minus one to plus one that says how closely the two variables are related. Both designs ask about sleep and memory, so the difference between an experiment and a correlational study lies not in the topic but in who controls the variable.
Why only the experiment can show cause
The key step in the experiment is random allocation: each volunteer has the same chance of landing in either group, for example by a coin toss. With enough people, the groups end up alike on average in everything else, their age, their usual sleep, their stress and their ability. If the five hour group then scores lower, the only systematic difference between the groups is the sleep, so the sleep caused the drop. The other conditions, the same test, room and time, are kept constant for the same reason.
A correlational study can find a strong link without any of that. Students who sleep more may have higher grades, with a coefficient close to plus one, yet the people who sleep more were never made to do so. They chose their own sleep, and people who choose differently also differ in many other ways.
Two traps in reading a correlation
The first trap is the third variable: something else that drives both measures. Stress might shorten sleep and also lower grades, so the two move together although neither causes the other. The second is direction: perhaps low grades cause worry that keeps students awake, which runs the arrow the other way. The most common mistake in an exam answer is to write that a correlation "proves" one thing affects the other. The better answer says the variables are related and names a third variable or the reverse direction that could explain it.
Medicine gives a famous case. Observational studies had suggested that hormone therapy protected women after menopause from heart disease. In the Women's Health Initiative, 16,608 women were allocated at random to the hormones or to a placebo, and in 2002 the trial was stopped early: the hormones did not protect the heart, and the treated group had slightly more heart disease. The women who chose hormones in the earlier studies had differed in other ways.
When a correlational study is the right choice
Many questions cannot be put into an experiment. A researcher cannot allocate children to divorced parents, assign teenagers to years of heavy phone use or make people grow up poor. Here a correlational study is the honest choice: it measures life as it is, can include thousands of people, and points to links worth testing. A good answer names the method, says whether the IV was changed, and states what the method can and cannot conclude.
Practice: experiments and correlational studies
For 10th grade psychology students. Answer each question from the page, then check your work against the answers.
Free, no sign-up. The answer key prints on the second page. Regenerating a new worksheet on the same topic needs a free account.
In the sleep experiment, name the independent variable and the dependent variable.
Fill in the missing step of the experiment.
Decide whether each study is an experiment or a correlational study.
Towns that sell more ice cream also have more drownings. Suggest the third variable.
A study finds that teenagers who use their phones late at night report a lower mood. Give two explanations other than "phones cause low mood".
Why does random allocation allow a researcher to conclude cause and effect?
A researcher wants to know whether growing up with a pet makes children more empathetic. Which method would you choose, and what could it conclude?
Common questions
Can a correlational study ever show cause?
Not on its own. It can rule some explanations out, for example by measuring possible third variables, and a link found again and again in different groups becomes more convincing, but only changing the variable under control shows cause directly.
What is a quasi-experiment?
A study that compares groups the researcher did not create at random, such as boys and girls, or two schools. It looks like an experiment, but because the groups were not formed by chance, they may differ in other ways, and the conclusion about cause is weaker.
Is an experiment always better than a correlational study?
No. An experiment is the only way to show cause, but it can be artificial, small and limited by ethics. A correlational study can measure real life in thousands of people and ask questions no experiment could, such as the effects of poverty or divorce.
What does a negative correlation mean?
That as one variable rises the other tends to fall, for example more hours of night shifts and fewer hours of sleep. Negative describes the direction, not a bad result; a correlation of minus 0.8 is as strong as plus 0.8.
Is a survey an experiment?
No. A survey asks people about what they do or think, and nothing is changed by the researcher. Its answers are usually analyzed for correlations between the questions.

Still have a question? Ask Lessi
Any question about this subject. Lessi answers right here, step by step.
Try asking
Asking needs a free account; your question is kept while you sign up. Answers are written by AI and can contain mistakes. Lessi doesn’t give medical or legal advice.
Learn it with a teacher
Learn this topic live with Lessi
Lessi, the AI teacher, explains it out loud on a board, with short quizzes, at your pace. The first 60 minutes are free.