The conventional paradigm for testing expectancy theory predictions of work behavior has been to correlate expectancy-value reports with concurrent measures of motivation and performance. Although this static, two-variable approach has typically yielded statistically significant results, correlations have not been sizable. This study, using a sample of design and development engineers over a four-year period, posited and found that expectancy theory predictions were materially strengthened by taking into account the effects of five complicating factors. Specifically, these factors were: (1) time, (2) the initial level of the criterion, (3) the level of rewards, (4) task-specific ability, and (5) control system responsiveness. (Author)