Examining the impact of organizational and individual characteristics on forensic scientists’ job stress and satisfaction

Abstract

Research on job stress and satisfaction among police and correctional officers has dramatically improved our knowledge of the experiences of criminal justice system employees. There is, however, minimal research on the experiences of individuals whose work directly informs criminal justice practice, most notably forensic scientists who collect and analyze evidence in support of criminal investigations. This study is one of the first to address the gap in our knowledge using survey responses collected from a sample of 670 forensic scientists operating in local and state laboratories in 25 states across the US. Regression models demonstrate that scientists who report higher stress were females who worked more hours, who had a poor relationship with court actors, minimal managerial support, and role ambiguities that made it difficult to do their jobs. Those with greater job satisfaction were unmarried, highly educated individuals with positive attitudes toward their work, greater managerial support, and few problems concerning their roles in the workplace.

Keywords:

• Forensic science
• occupational reactions
• work stress
• job satisfaction

Acknowledgments

The points of view or opinions expressed in this study are those of the authors and do not necessarily reflect the opinions of the National Institute of Justice or the US Department of Justice.

Notes

1. A comparison of the various populations sampled may be instructive to understand differences in the experiences of scientists working in private laboratories relative to public-funded agencies. This is, however, beyond the scope of this study and its focus on those scientists working most closely with the criminal justice system.

2. The measures in the occupational stress scale used here replicate those used in most research on criminal justice system employees. Organizational psychology prefers to use only the three measures specifically related to stress rather than including items that may otherwise be conflated with measures of individual burnout. To ensure the validity of the measure used, a factor analysis was conducted using oblique rotation which found that all five items loaded onto a single construct. Furthermore, all composite measures presented throughout this analysis loaded on to a single factor, except supervisory support which is discussed separately.

3. Given the distribution of the two dependent variables, equal-width binning was used to identify the high and low categories rather than a single summative scale (see Kotsiantis and Kanellopoulos 2006). Remaining values composed the moderate category.

4. Years spent in forensic science was measured as a categorical variable in the survey rather than an open-ended continuous measure in order to simplify the question for respondents, and is in keeping with other measures used in analyses of stress and satisfaction generally (e.g., Cullen et al. 1985; Holt and Blevins 2011).

5. A factor analysis was conducted to assess whether these items all load onto a single measure. The results suggest a two-factor solution, though the second factor consisted of only one measure: the people I work with often have the importance of their job stressed to them by their supervisors. The findings in the regression models were consistent whether or not this item was used in the supervisory support scale. Ultimately, the item was left in the composite measure because of the stable results and the frequent use of this item in supervisory support scales in the larger literature.

6. The largest variance inflation factor was 2.619 for the stress model and 2.263 for the satisfaction mode, indicating that multicollinearity was not a problem for either model. The multivariate ordinal regression assumption of proportional odds was confirmed by the non-significant test of parallel lines for each model.

7. Separate regression models were created to examine the influence of only individual characteristics and only work-related variables on stress. The combined model was the best fit (pseudo R² = 0.029 for individual characteristics and 0.480 for work-related variables).

8. Separate models were created to regress job satisfaction on only individual characteristics and only work-related variables. The combined model was the best fit (pseudo R² = 0.015 for individual characteristics and 0.422 for work-related variables).