Validating the Job Satisfaction Survey Scale in Voluntary Child Welfare Agencies

Background and purpose:

Job satisfaction has been linked to workforce retention in child welfare agencies. One of the most widely used measures on job satisfaction is the Job Satisfaction Survey (JSS). Although it has been validated in workers of many kinds of social service agencies, its psychometric properties remain untested in workers of voluntary child welfare agencies. Employing a cross-sectional design, this study aims to examine the reliability and validity of the JSS in frontline child welfare workers in voluntary (i.e. private non-profit) agencies.

Methods:

The sample included a total of 535 direct care and clinical workers in a Northeastern state. Survey data were collected on-site at different sites of the state from 2009 to 2011. Most participants were white (70%), female (63%), and frontline workers (87%), and held a child welfare job for the first time (61%).

Composed of 36 items, the JSS is a multidimensional measure for job satisfaction (e.g., “I like the people with whom I work.”). To assess the criterion validity of the JSS, we used the Intent to Leave Child Welfare Scale to measure workers’ intention to leave (7 items; e.g., “How often have you thought about leaving?”).

The sample was randomly split into two subsamples for factor analysis. Sample I (N = 267) was used to conduct the EFA (maximum likelihood estimation and Geomin rotation). The number of factors was preset as nine, which is consistent with the original validation study of the JSS (Spector, 1985). Only items with a factor loading greater than .40 were retained. Sample II (N = 268) was used to conduct the CFA (weighted least squares means and variance adjusted estimation). We also tested the internal consistency reliability and criterion validity of the JSS using Sample II.

Results:

The EFA yielded an eight-factor structure with a total of 27 items retained (A factor was eliminated because only one item loaded on it). In the subsequent CFA, four more items were dropped in light of theoretical and model specification considerations. The final CFA model contained 23 items loading on seven factors: 1) pay, 2) benefits, 3) promotion, 4) supervision, 5) coworkers, 6) nature of work, and 7) communication. This seven-factor model fit the data well: CFI = .96, TLI = .96, RMSEA = .05. Compared to the original psychometric study, contingent rewards and operating procedures were not validated in our study. All but one factor (communication) had adequate reliability (α ranging from .75 to .85). Three factors including pay, promotion, and nature of work were significantly correlated with workers’ intention to leave, showing marginal criterion validity of the JSS.

Conclusion:  

The original nine dimensions of the JSS were only partially validated in our study. This could be due to the unique characteristics of our sample or the modification to the rating scale of the JSS (changing from a 6-point to a 4-point). Caution may be exercised when using the JSS to assess job satisfaction in voluntary child welfare workers. More validation research is warranted in the future.