AP® Psychology · Science Practices: Research Methods and Data
2.B Random Assignment, Control Groups, and Placebos
Random assignment sorts already-chosen participants into groups by chance alone, so the groups start out equivalent; comparing a treatment group to a control or placebo group then isolates the actual effect of the treatment.
An experiment needs to prove that the treatment caused the change, not something else. If researchers let people choose their own group, the groups could differ in ways that mess up the results: maybe the more motivated people picked the real drug. Random assignment fixes this by using chance, a coin flip, a random number generator, to decide who goes where. Because the assignment has nothing to do with any personal trait, age, mood, genetics, expectations, all of it gets scattered roughly evenly across every group. Any leftover differences between groups become random noise, not a systematic bias, which means any gap in outcomes is far more likely to come from the treatment itself.
Random assignment is not random sampling, and mixing them up is the single most common error here. Random sampling happens earlier and answers a different question: how do we pick a sample that represents the whole population, so our results generalize? Random assignment happens after the sample exists and answers: how do we split these people into groups so the groups are comparable? A study can have one, both, or neither. Picture it as a funnel: a large population, a random sample pulled from it, then that sample split by random assignment into an experimental group and a control group.
The control group usually gets a placebo, a fake treatment that looks real, because expecting to improve can itself cause improvement. If both groups improve, some of the experimental group's gain is just that placebo effect, not the drug. Subtracting the placebo group's improvement from the treatment group's improvement strips that shared improvement away, leaving the medication's true, isolated effect.
Key terms
- Random assignment
- Using chance to place already-selected participants into experimental or control groups so the groups start out equivalent.
- Random sampling
- Using chance to select participants from a population so the sample represents that population.
- Control group
- The group that does not receive the treatment being tested, used as a baseline for comparison.
- Placebo
- A fake treatment with no active ingredient, given so expectation effects can be measured and subtracted out.
- Placebo effect
- Improvement that occurs simply because a participant believes they are receiving a real treatment.
Watch it explained
- How it worksHow does random assignment scatter every unmeasured difference evenly across groups?
- Common mistakeWhy isn't random assignment the same thing as random sampling?
- How it worksWalk me through an experiment diagram from population to sample to two randomized groups.
- Going furtherWhat happens to the measured effect when the placebo group also improves?
Check yourself
End of topic.