Improving recall by recoding interfering material at the time of retrieval.

Our experiments demonstrate that interference of an interpolated list of items with recall of an original list can be substantially reduced by forming Ss just before testing how to reorganize and simplify the interpolated material. In Experiments 1 and 2, Ss better recalled an initial serial list of letters when informed at testing that an interpolated list spelled a certain phrase backward. Similarly, in Experiments 3 and 4, Ss better recalled an initial list of cities when told that the interpolated cities were also names of former U.S. presidents. Control experiments rule out several simple explanations. In contrast to an editing hypothesis, the postorganizing clue helped recall even when problems of list differentiation were minimized. Current memory models appear unable to explain this benefit of a postlearning clue that enables Ss to segregate the interpolated material from the to-be-remembered material.