correlation and causation

Two things varying together does not mean that one causes the other. When you see a relationship there are three explanations: X causes Y; Y causes X; or a third variable causes both. Tools that strengthen causal claims: random assignment, controlling third variables, and following the same people over time.

In Doob and Macdonald’s Toronto study the relationship between television viewing and fear disappeared once the real crime level of the neighbourhood was held constant. In the Payne Fund, Shuttleworth and May equated frequent and infrequent movie-goers and still wrote that they lacked precise experimental controls. In the debate on violent media and aggression, the “effect” and “selection” hypotheses are considered together with Slater’s idea of reinforcing spirals. See experiment (research method).