Revisiting the "strong but wrong" effect

In the 1930s, J. M. Stephens found that strong verbal associations called wrong are more likely to be changed than weak verbal associations called wrong, relative to items with no feedback. The present study produced a much larger "strong but wrong" effect, apparently as a result of defining "strong" in terms of meanings presumably stored in long-term memory. It was also found that such items are followed by high recall of feedback on the next trial, and that, indeed, when feedback recall is statistically controlled, the effect disappears. An explanation of why feedback is better recalled for "strong but wrong" items is offered, and several predictions are generated, some of which are tested (and confirmed) from the present data.