Abstract
Pools of paired associates (double or single letters paired with one- or two-digit numbers) were constructed so that all pairs within a pool had approximately the same mediation value (MV), this latter value being defined as the ease with which a mediator can be found within a pair. These pools were then used in a drop-out experiment where a list of 8 pairs were learned to a criterion of one correct repetition. The main result was that rate of presentation, mean MV of the pool and drop-out/constant procedure were significant. That the variable drop-out/constant procedure proved significant is contrary to the results obtained by Rock, and this is discussed in relation to the question of one-trial versus incremental learning.
Get full access to this article
View all access options for this article.
