Random Yoking: An Alternative to Feedback Procedures for Preventing Superstition in the Human “Learned Helplessness” Paradigm

The “learned helplessness” model of human depression requires that humans demonstrate deficits similar to animals following exposure to noncontingent events. However, the feedback procedure usually employed in the triadic instrumental induction phase represents a confound in studies of the interference effect in humans. Matute (1994) concluded that the feedback procedure is necessary for the interference effect, which is thus due to feedback induced failure rather than learned helplessness. As an alternative, we hypothesize that feedback alerts participants to noncontingency, such that subsequent interference is not inconsistent with learned helplessness theory. The present study evaluates these competing claims by incorporating a novel manipulation designed to promote the perception of noncontingency in Matute's (1994) triadic no-feedback-procedure induction. A second noncontingent yoked group received the same tones as the usual direct yoked group, but in random order so as to disrupt the “late trials” distribution of short-latency tones which promotes superstitious responding. As predicted, the random-yoking procedure inhibited superstition. The interference effect was observed in the random-yoked but not the direct-yoked triad. Thus random-yoked participants may have developed the expectation of noncontingency which is critical to learned helplessness. It is concluded that the confounded feedback procedure is not necessary for the interference effect and should be avoided in future research.