Abstract
Three intra-runway delay durations (0, 15, and 45 sec.) were combined factorially with three training levels (0, 15, and 75 trials) to determine the effect of the level of training at the time of delay introduction on performance. Three models, each with different predictions concerning the effect of training on performance after delay introduction, were evaluated. First, the frustration-drive model was unable to account for either the decremental effects of delay, or the fact that high training groups were more affected by delay and recovered more slowly. Second, delay of reinforcement theory while predicting decremental effects was unable to account for their temporal course and spatial distribution in the runway. Especially detrimental to this theory was complete recovery from the effects of delay in spite of continued delay. Third, the novelty-reactions model, while failing to account for the specific nature of the decremental effects of delay and training level on performance, was able to handle the fact that decrements did take place and recovery did occur and was less rapid for high training groups. Of the three models this latter was considered to be most compatible with the results.