Error Trends in Paired-associates Learning After Response Familiarization

A study was run to reveal the course of overt errors over paired-associate trials when response learning was eliminated, and to determine a method of satisfying Ss on such a task for use in treating data from a subsequent motor task learned to the same stimuli. Sixty-eight Ss were given thorough response familiarization, followed by 17 anticipation trials of a paired-associate task using verbal responses to non-verbal stimuli. Subsequently they received 30 trials on the Star Discrimeter. Mean % correct responses and errors were plotted across pretraining trials. Ss were divided into two strata on the basis of pretraining scores and the strata were plotted separately across pretraining and motor trials. A Pearson product-moment between %CR on pretraining and criterion task was .47, a conservative estimate of the curvilinear relationship. Overt errors in pretraining were found to decrease monotonically from the first anticipation trial. Pretraining stratification was found to differentiate Ss by performance on the motor task, showing its effects in the error function