Date of Award

Winter 1989

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Psychology

Program/Concentration

Industrial/Organizational Psychology

Committee Director

Donald D. Davis

Committee Member

Albert S. Glickman

Committee Member

Glynn D. Coates

Committee Member

Paul Champagne

Abstract

This multivariate test of the full Mobley, Griffeth, Hand, and Meglino (1979) model of turnover used 184 volunteer and 202 paid workers employed as firefighters, library workers, and rescue squad workers. It addressed the issue of whether the motives and behaviors of volunteers could be accommodated by the same model that normally would be applied to paid workers. Little research exists on the characteristics of volunteer workers, and their worth and manageability are commonly denigrated. The study of volunteers is similar jobs to paid workers might reveal much about the disengagement process since the need to maintain a source of income will not mask their affective responses. Volunteers represent cost-reduction potential for the cities using them wisely. The Mobley et al. (1979) model included cognitive, affective, and behavioral processes as determinants of turnover. Path analysis, tested with the Specht (1975) procedure for goodness-of-fit, determined that the Mobley model fit the volunteer data as well as it did the paid worker data but with only a low goodness-of-fit index (Q =.20 for both samples). An augmented model, which included Organizational Commitment and Intent to Alter Involvement, was tested for the entire sample, the paid workers, and the volunteer worker sample, but its fit was unsatisfactory. Finally, a trimmed model which included just six outcome variables and 24 paths fit the paid worker data at Q =.77 and the volunteer data at Q =.80. No general conclusions about the validity of the trimmed model can be made prior to cross validation. The disengagement patterns of volunteers could be understood by the same models as were applied to paid workers although the weighting of variables varied somewhat as a function of pay. Fluctuations in the effects of job characteristics were tied to occupational category, demonstrating the pooling information from various occupations in research of this sort leads to the instability of relationships found across studies. Greater precision in specifying variables and paths in models is called for but must be based on occupation-specific empirical data.

DOI

10.25777/qn0q-v427

Share

COinS