INSTRUCTIONS AND FEEDBACK EFFECTS ON SPEED AND ACCURACY WITH DIFFERENT LEARNING-CURVE MODELS

This study compares powerform versus exponential learning curve models on the first ten trials of a task, using several dependent measures. Also shown are differences in instructions about speed emphasis, accur acy emphasis, and equal speed-accuracy emphasis. Empirical data were c ollected on human subjects performing a devised task on discovering a sequence of switch connections to lights. Learning curves of each mode l were fitted to the serial data and to cumulative average serial data on time, errors, and their product for each individual subject. Diffi culties in fitting these models to data are reported along with suppor ting reasons. The quality of data fits to each model is examined for e ach criterion. It is shown that the exponential model exhibits more di fficulties in being fitted to data, but that it provides a more accura te description of each criterion of learning in this task when the fit is good. Accuracy and speed operating characteristics are shown for e ach group of subjects who are given different criterion-emphasis. Thes e characteristics were consistent with their instructions.