Memory for impasses during problem solving

Three experiments were conducted to investigate the relative memorability of solved versus unsolved problems in long-term memory. In each experiment, subjects worked on a set of potentially solvable word problems, with the time spent on each problem held constant. Problem memorability was then measured with a free-recall task. In Experiment 1, in which a majority of problems were solved, unsolved problems were better remembered. In Experiments 2 and 3, we expanded on these results by manipulating problem difficulty and thus the ratio of solved to unsolved problems. When unsolved problems were as frequent as or more frequent than solved problems, no memory differences were found. Across all three experiments, the ratio of solved to unsolved problems was found to be a significant predictor of unsolved-problem memorability, but was not significantly related to the memorability of solved problems. The results illustrate that when impasses in problem solving are infrequent, they are more available in memory than are solved problems. It is speculated that this memory phenomenon may facilitate the recognition of opportunities to return to problems that have been terminated short of solution.