Predicting Goal Choice and Goal Commitment Within a Control Systems Framework

203 p.Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1988.Two laboratory studies were conducted to test specific links in a control systems model of goal choice and goal commitment. Male and female subjects performed five trials on either a paper and pencil text editing task (Study 1) or a text transfer task on a microcomputer (Study 2). The results demonstrated that both performance expectations and previous task performance influenced choice of goal difficulty for subsequent trials. The importance of performance expectations relative to previous trial performance for predicting goal choice diminished as subjects gained experience with the task. The value of goal achievement had minimal influence on goal choice. However, the value of goal achievement was highly and positively related to commitment to achieving goals; performance expectations were not related significantly to commitment.In addition to testing the general model of goal choice and goal commitment, three specific expectancy-value models were tested for predicting goal choice. The three models that were tested were subjective expected utility (SEU) theory, Vroom's valence-instrumentality-expectancy (VIE) theory, and Naylor, Pritchard, and Ilgen's (NPI) theory. The proportion of subjects whose goals matched the goal levels that were predicted by the models was determined for each of the models, for each trial. None of the three models classified correctly more than 34% of one-half of the standard deviation of goals. The SEU model was consistently the best predictor of subjects' goals. The NPI model consistently over-predicted subjects' goals. Explanations for the lack of success of the models for predicting goal choice are offered.U of I OnlyRestricted to the U of I community idenfinitely during batch ingest of legacy ETD