Color change as a causal agent revisited

Michotte (1946/1963, p. 235) concluded that color "has no bearing whatever on the question of qualitative causality". Surprisingly, this claim has received little empirical investigation in the 60 years since its publication. In 2 experiments, 2 balls struck a cylinder and changed color (either continuously or in a stepped progression) for 2 s. After the 2-s interval, the cylinder disgorged a purple substance. Participants chose which of the 2 balls was most likely to have caused the disgorging effect. An object that changed color was favored as the cause much more often than one that did not, and participants generally preferred an object that reached its terminal color immediately before the effect over objects that reached their terminal colors earlier. When participants performed a causal ratings task, color change produced moderate judgments of causation, with little response differentiation as a function of color dynamics.