ILLUSION OF CONTROL - DETECTING RESPONSE-OUTCOME INDEPENDENCE IN ANALYTIC BUT NOT IN NATURALISTIC CONDITIONS

Experiments in which subjects are asked to analytically assess respons e-outcome relationships have frequently yielded accurate judgments of response-outcome independence, but more naturalistically set experimen ts in which subjects are instructed to obtain the outcome have frequen tly yielded illusions of control. The present research tested the hypo thesis that a differential probability of responding, p(R), between th ese two traditions could be at the basis of these different results. S ubjects received response-independent outcomes and were instructed eit her to obtain the outcome (naturalistic condition) or to behave scient ifically in order to find out how much control over the outcome was po ssible (analytic condition). subjects in the naturalistic condition te nded to respond at almost every opportunity and developed a strong ill usion of control. Subjects in the analytic condition maintained their p(R) at a point close to .5 and made accurate judgments of control. th e illusion of control observed in the naturalistic condition appears t o be a collateral effect of a high tendency to respond in subjects who are trying to obtain an outcome; this tendency to respond prevents th em from learning that the outcome would have occurred with the same pr obability if they had not responded.