Prediction of Performance in Motor Skills Training

Abstract

Williges, Williges, and Savage (1979) demonstrated the utility of a multiple-regression approach for assigning individuals to training alternatives. Average training time savings of 50% were obtained when the lower predicted score was used to assign students to fixed-difficulty or adaptive training conditions to learn a two-dimensional tracking task. The present study extends this work by developing prediction models for a real-world training task. Regression models were developed to predict performance on several flight manuevers using the ATC-610 Flight Trainer. One hundred VPI undergraduates and 100 USAF cadets, with an equal number of males and females, served as subjects.