A true experiment in psychology is a study where the researcher manipulates one variable, randomly assigns participants to different groups, and measures the effect on an outcome. These three elements, manipulation, random assignment, and a control comparison, are what separate a true experiment from every other type of research design. The American Psychological Association defines it as “a study in which participants are assigned at random to two or more experimentally manipulated treatment conditions or to a treatment group and a control group.”
The Three Requirements
Every true experiment has the same basic structure. First, the researcher changes something on purpose. This is the independent variable, the thing being tested. Second, participants are randomly assigned to groups, so the researcher (not the participant) decides who gets which treatment. Third, there’s a control group that doesn’t receive the treatment, giving the researcher a baseline to compare against.
If any one of these pieces is missing, the study isn’t a true experiment. A survey measuring how much people exercise and how happy they feel lacks manipulation. A study comparing outcomes between men and women lacks random assignment (you can’t randomly assign someone’s sex). These designs can reveal patterns, but they can’t prove that one thing caused another. True experiments can.
Why Random Assignment Matters So Much
Random assignment is the single feature that gives true experiments their power. It works by distributing all the differences between people, their age, personality, background, motivation, evenly across groups before the study even begins. That way, the only systematic difference between the groups is the treatment itself.
This is different from random sampling, and the distinction trips people up. Random sampling is about who you recruit for the study. It determines whether your results apply to the broader population. Random assignment is about what happens after recruitment: which group each person ends up in. Sampling happens first, assignment happens second. You can have one without the other. A study could randomly assign participants to groups (allowing causal conclusions) but recruit only college students (limiting how broadly those conclusions apply).
The practical payoff: when random assignment is done properly, any difference you observe between groups at the end of the study is most likely caused by the treatment, not by some pre-existing difference between the people in each group.
Independent and Dependent Variables
In a true experiment, the independent variable is whatever the researcher deliberately changes. The dependent variable is the outcome being measured, called “dependent” because its value depends on what the researcher did.
Consider a classic example: a researcher wants to know whether listening to classical music improves reading ability in children. She divides kids into two groups. Group A listens to Mozart for one hour daily over a month. Group B avoids classical music entirely for the same period. At the end of the month, everyone takes a reading comprehension test. The independent variable is exposure to Mozart. The dependent variable is the reading test score. The researcher manipulated the music exposure and measured whether it changed reading performance.
What the Control Group Actually Does
The control group exists to answer a simple question: what would have happened without the treatment? If you test a new therapy for depression but have no control group, you can’t tell whether people improved because of the therapy or because time passed, seasons changed, or they simply expected to feel better.
Control groups are kept as natural and unchanged as possible. In a medication study, the control group might receive a sugar pill that looks identical to the real drug. In a behavioral study, the control group might continue with their normal routine. The key is that control participants experience everything the experimental group does, the same setting, the same measurements, the same timeline, minus the active ingredient being tested.
A Classic True Experiment: The Bobo Doll Study
Albert Bandura’s Bobo doll experiments at Stanford University are among the most famous true experiments in psychology. Bandura wanted to know whether children who watched an adult behave aggressively would imitate that aggression. Children were assigned to different conditions: some watched an adult punch, kick, and yell at a large inflatable clown doll for about ten minutes, while others did not observe any aggressive behavior.
Afterward, an experimenter brought each child to a room with attractive toys, then took the toys away, deliberately frustrating the child. The child was then left in a room containing, among other things, a Bobo doll. Children who had watched the aggressive adult were significantly more likely to hit and yell at the doll themselves. Later variations showed that when the adult model was punished for being aggressive, children were less likely to copy the behavior. When the model was rewarded or faced no consequences, children’s aggression increased. The study demonstrated that aggression could be learned simply by watching someone else, a finding with enormous implications for understanding media violence and parenting.
True Experiments vs. Quasi-Experiments
The closest relative to a true experiment is a quasi-experiment. Quasi-experiments also involve manipulating a variable and measuring an outcome, but they lack random assignment. Instead, groups are formed based on pre-existing characteristics or circumstances. A researcher studying whether a new school curriculum improves test scores might compare two schools, one that adopted the curriculum and one that didn’t, but the students weren’t randomly placed in those schools.
This matters because without random assignment, you can’t rule out alternative explanations. Maybe the school that adopted the new curriculum also had more experienced teachers, more funding, or more motivated parents. The University of Florida describes true experiments as the “gold standard” for showing that interventions actually work, precisely because random assignment eliminates these competing explanations. Quasi-experiments can still suggest cause and effect, but the case is always weaker.
Threats to Internal Validity
Even true experiments aren’t bulletproof. Psychologist Donald Campbell identified seven classic threats that can undermine an experiment’s ability to draw accurate causal conclusions. History refers to outside events that occur during the study and affect the outcome. Maturation means participants naturally change over time, growing older, more tired, or more skilled, independent of any treatment. Testing effects occur when taking a pretest influences how people perform on a posttest. Attrition happens when participants drop out unevenly across groups, skewing the comparison.
True experiments are specifically designed to minimize these threats. Random assignment handles the biggest one, selection bias, by ensuring groups start off equivalent. Using a control group accounts for history and maturation, because both groups experience those influences equally. But researchers still need to watch for attrition and other problems that can creep in after the study begins.
The Trade-Off With Real-World Relevance
True experiments excel at internal validity, the confidence that your treatment actually caused the observed effect. But they often sacrifice something in return: external validity, or how well results translate to everyday life. This tension sits at the heart of psychological research.
The concern is straightforward. Laboratory settings are artificial and simplified. A study conducted in a quiet room with a computer screen and a tightly controlled task may not capture how people behave in the noise and complexity of their actual lives. Researchers sometimes call this the “real-world or the lab” dilemma. There is always an inherent trade-off between the control imposed by experimental conditions and the naturalness of the behavior being studied.
This doesn’t make true experiments less valuable. It means their findings sometimes need to be tested again in more realistic settings, and that other research designs (observational studies, field experiments, longitudinal surveys) fill in the gaps that lab-based experiments leave open. The strongest conclusions in psychology come from multiple types of studies converging on the same answer.