Expectancy predictions of reenlistment in the National Guard

A within-person test of the expectancy model was made in the Army National Guard. Surveys were administered to 1169 Guardsmen in 29 National Guard units. Guardsmen were asked to compare National Guard duty with a civilian work alternative in terms of their instrumentality for various role outcomes. The perceived within-person difference between the two jobs significantly predicted intention to reenlist (r = .31) and actual reenlistment (r = .28). Further, including negative role outcomes in the expectancy model did not enhance its predictive power, and intrinsic role outcomes were not superior to extrinsic outcomes. Using a smaller set of idiosyncratic role outcomes considered important by the respondent resulted in higher validity than using a larger, standard list of outcomes. Also, equally weighting instrumentality perceptions generated higher levels of prediction than weighting instrumentalities differentially by valence measures.