Prediction of Complex Task Proficiency by Means of Component Responses

Abstract

The present study evolved from previous work on the relationship between a complex motor skill and its components (1957a). The objective was the fractionation of total-task behavior into individual part skills and the examination of relations between the parts and the total. The analysis leads to the synthesis of the complex skill. That is, a combinational rule is produced where the criterion score for the total task is reconstructed by means of the scores on the parts. Previous successful combinational rules were based upon models of independent probabilities (Bilodeau & Bilodeau, 1954; Bilodeau, 1955) and of multivariate regression . (Bilodeau, 1957b). A third approach (Bilodeau, 1957a), and also the one used here, make use of geometric relationships between the total and part tasks. There are two general types of training situations for which the synthesis can be made. In one, S practices the parts separately and is later tested with the parts simultaneously. In the other practice mode, training is confined to the total task. In the present experiment the training alternated between part and total tasks. S's task was to crank using two hands (total task) or one hand (part-task R and part-task L). The purpose was to show that the scores taken when the hands worked separately could be used to predict the scores when the hands worked together within a 5% margin of error. The practice of the part and whole tasks continued over a period of 20 days in order to ensure the development of more than an ordinary amount of coordination. It will be shown that (a) the problem of synthesizing whole-task performance is much simpler for skilled Ss than for the unskilled, and that (b) the terms of the combinational rule are suggestive of the nature of the developing coordination. METHOD Subjects.--On the basis of a brief foretest, three male graduate srudents were selected from five volunteers. The three Ss were chosen so as to maximize between-subject variance. Ss were paid $1.00 per hour plus a bonus for best and second best performance. Apparatus.-The appararus was Model CMlOlB of the Standard Two-Hand Coordination Test, modified as previously described (1957a) to change the cask from simple pursuit to tracing. Turning the crank for the right hand clockwise (C) or