Using Structural Equation Modeling to Investigate the Causal Ordering of Job Satisfaction and Organizational Commitment Among Staff Accountants

Document Type



Published by the American Accounting Association in Behavioral Research in Accounting, volume 9, 1997. Bryant users may access this article here.


American Accounting Association

Publication Source

Behavioral Research in Accounting


The significant results reported using job satisfaction and organizational commitment as antecedents to turnover intentions, and their extensive use in behavioral research, make the causal relationship between them a subject of interest in accounting and other fields. This study addresses the causal ordering of these two constructs. Using staff accountants as subjects, and structural equation modeling as the statistical technique, the results indicate a causal ordering from organizational commitment to job satisfaction.

Public accounting firms experience annual non-partner turnover rates as high as 45 percent, primarily at the entry-level staff accountant position (Bao et al. 1986). Other studies have shown turnover rates to be as high as 85 percent for initial hires over a ten-year period (Rhode et al. 1976) to annual turnover rates of 23.9% for staff accountants who had one- to three-years experience with a firm (Lampe and Earnest 1984).

Job satisfaction and organizational commitment are two constructs often considered when examining the turnover of practicing accountants. However, few studies have investigated the causal ordering of these constructs. Gregson (1992) reported the results of a study where job satisfaction is antecedent to organizational commitment in a turnover model of practicing accountants. However, there are some limitations to his study. Due to identification problems, a model with a reciprocal relationship between satisfaction and commitment could not be tested. Lack of identification can be caused by the rank condition, order condition, or both conditions not being met. Without proper identification, parameter estimates are arbitrary and the interpretation of these parameters becomes meaningless (Long 1986).

A second reason for investigating the causal ordering of these two constructs is their numerous usage in behavioral research. The incorrect classification of either as an exogenous or endogenous construct to the other, will lead to incorrect models which can extend to mistaken conclusions.

To test the causal relationship between organizational commitment and job satisfaction, this study uses sequential Chi-square difference tests (SCDTs) and other analyses to determine the "best" fitting model. The reciprocal relationship that could not be tested in the Gregson (1992) study is the focus of this study.

The next section presents the theory for the study, followed by a section describing the models to be tested. Subsequent sections present the method employed, results and discussion. Lastly, limitations and implications for future research are presented.