The Work-Expectation-Scale: A New Measure for Research in Human and Social Service Organizations
Methods: The initial scale consisted foremost of items from pre-existing measures with a few new items added. As suggested for smaller variable sets, we applied common factor analysis with squared multiple correlations as initial communality estimates for 3-7 factor models rotated according to varimax, equamax, and promax criteria. Each model was evaluated for its ability to produce dimensions that a) satisfy the minimum constraints for Cattell's scree test, Velicer's minimum-average partialing test, and parallel factoring of normal random variables on the basis of 100 replications; b) retain at least three items with salient loadings, where loadings ≥.40 are considered salient; c) yield high internal consistency (≥.70); d) remain invariant across models; e) produce the highest hyperplane count and f) make theoretical sense in terms of parsimonious coverage (mutually exclusive assignment of items to factors, maximum numbers of items retained).
Results: The four-factor promax (k = 4) model was found superior and satisfied all criteria; whereas the five-factor model produced unreliable factor scores with too few salient markers. Salience was found for 20 of the 25 items. The Work-Expectation Scale consists of four distinct dimensions: Societal Impact (α=.76), Work Conditions (α=.72), Leadership Prospect (α=.74), and Skill Development (α=.72).
Implications and Conclusion: The Work-Expectation-Scale contributes to research on work expectations of employees in social and human service organizations and can be applied to research on diverse populations within those organizations. Future research on work expectations using the Work-Expectation-Scale should investigate the reliability and generality of the factor structure across samples as well as validate the four-factor structure through confirmatory factor analysis.