Isolating Cause From Effect
A true experiment is the only research design that can establish causality — that changing one thing genuinely causes a change in another. It does so through manipulation: the researcher directly controls an independent variable (IV), the factor being tested, and measures its effect on a dependent variable (DV), the outcome of interest. A study testing whether a new therapy (IV) reduces anxiety (DV) is a classic experimental structure.
The feature that makes an experiment a TRUE experiment, rather than merely suggestive, is random assignment: each participant has an equal chance of being placed in the experimental group (receiving the manipulated treatment) or the control group (not receiving it, serving as a baseline for comparison). Random assignment is what allows researchers to rule out confounding variables — other differences between groups that could explain the result instead of the IV — because, on average, randomization spreads any such differences evenly across both groups.
| Term | Role |
|---|---|
| Independent variable (IV) | The factor the researcher manipulates |
| Dependent variable (DV) | The outcome being measured |
| Control group | Baseline group, no manipulation |
| Confounding variable | An uncontrolled factor that could explain the result instead |
Several further safeguards protect experimental rigor. Operational definitions specify precisely how abstract concepts (like "anxiety") will be measured, so the study is replicable. A placebo controls for the psychological effect of simply believing you are receiving treatment. A double-blind procedure, where neither the participant nor the researcher interacting with them knows who is in which group, guards against both participant expectation effects and experimenter bias — the researcher's own expectations subtly influencing the results or their interpretation.
Common pitfall: assuming ANY study comparing two groups counts as a true experiment. Without genuine random ASSIGNMENT by the researcher (not just comparing groups that already differed beforehand), a study cannot rule out confounding variables and so cannot support a strong causal claim — that distinction is exactly what separates a true experiment from a quasi-experiment.
A two-arm flowchart: a pool of participant icons; a coin-flip icon randomly sorts them into two arms labelled "experimental" and "control"; an accent treatment icon applies only to the experimental arm.