structural equations model of organizational commitment, job satisfaction as its covariate, and their antecedent variables

This academic exercise in organizational commitment research explores the causal relationships between organizational commitment, job satisfaction as its covariate latent exogenous variable and nine manifest endogenous variables: (a) organizational conflict, (b) personal values/role conflict, (c) cognitive/affective attitude, (d) intention to leave the organization, (e) job security, (f) role ambiguity, (g) job satisfaction, (h) organizational commitment and, (i) the deficiency in meeting basic Maslovian needs. This data sample of 400 (faculty members from four public community college districts in south-central Florida) represented urban and non-urban community colleges, large and small campuses, east- and west-coast community colleges, and both full- and part-time faculty. The covariance matrix based on complete data for the sample (n = 400), along with information of the estimated reliabilities, were analyzed using EzPATH, the SYSTAT implementation of structured equations modeling. GFI criteria accompanied diagnostic output and are provided. The path coefficients of the primary model and their respective standard errors complied with accepted statistical methods in most cases. Also, the root mean square residual of the primary model was .0180. The p-value for the primary model was 0.4506 which was supported by a chi-square (χ2) divided by DOF equating to 1.00513. The determined AGFI and GFI indices were .9755 and .9897 respectively. Only the psychological attitudes of conflict, either conflicting standards of the organization or personal values with the organization, and role ambiguity, showed any direct influence by satisfaction and non-significant influence by commitment. Only the psychological attitudes of cognitive/affective attitude, intention to leave the organization, and job security showed any direct influence by commitment and non-significant influence by satisfaction. In this study, job satisfaction and commitment show a mutual causal relationship with each other (represented by the high mutual causation coefficient of 0.94556 SE = 0.02163). The primary model was modified to address both the path of satisfaction being a precedent of commitment (SAT --> COM) and commitment being a precedent of satisfaction (COM --> SAT). Neither of these two models demonstrated goodness of fit criteria as strong as the primary model. This dissertation has shown that by using carefully collected data, it is possible to construct structural equation models consisting entirely of psychological-social variables in a model as large as nine manifest endogenous variables.