Causation means one variable actually produces a change in another. Only an experiment with random assignment can show it, because random assignment spreads out pre-existing differences between groups.
Why it matters
Development and learning research is full of tempting links, like media violence and aggression. You must be able to tell which designs can prove cause and which only show a link.
How the exam tests it
Questions ask which design gives the strongest causal evidence. The answer is the one that randomly assigns teens to play violent or nonviolent games, then measures aggression. Longitudinal surveys, cross-sectional studies, and reviews of correlational data cannot rule out confounding variables, so they cannot prove cause.
Don’t confuse it with
Correlation shows that two variables move together, but it cannot say which one causes the other, or whether a third variable drives both.