A parametric analysis of varying levels of consequence-based procedural fidelity errors when learning new tasks.

Procedural fidelity is the extent to which procedures are implemented as prescribed. Fidelity errors involving the delivery of consequences can have detrimental effects during discrimination training. This study consisted of a human-operant evaluation of how varying levels of fidelity errors involving consequences may affect performance on an arbitrary computerized matching-to-sample task. We investigated how implementing the task with varying levels of fidelity may affect learning and how initially exposing learners to varying levels of fidelity affects ongoing performance once errors were corrected for six participants. The results indicated that varying fidelity levels delayed learning for five out of six participants. However, all six participants mastered the task once the errors were corrected. We discuss the implications of the results and provide suggestions for future research.