Data were obtained from a sample of 45 officers in a metropolitan police department using data collection periods 90 days apart. Several well-known job attitude instruments, each of which has multiple scales designed to measure various job attitude constructs, were administered. The scales were organized according to the following sources of job stress: (1) work relationships, (2) organization structure and climate, (3) intrinsic job characteristics, (4) organizational role, (5) career development, and (6) perceived stress (an average score of two scales measuring different dimensions of job stress). The two-phase study design resulted in six correlation coefficients: two synchronous correlations, two autocorrelations, and two cross-lagged correlations. By examining the differences in the cross-lagged correlations, an inference regarding reciprocal causality can be formed. Three criteria were established for developing a causal inference regarding any job attitude and perceived stress: (1) statistically significant synchronous correlations, (2) statistically different cross-lagged correlation coefficients as assessed by the Pearson-Filon method, and (3) a statistically significant dynamic correlation coefficient. The only group of job-related stressors where causal inferences were justified was the organization structure and climate measures. Specifically, influence, norm observance, and trust in management measures were negatively related to perceived stress, and the formation of these attitudes precede stress. Administrative actions that could be taken to help reduce stress are identified, and the failure of the data to reveal more causal influences is examined. Tabular data and 25 references are provided.