Levine, D. I., Toffel, M. W., & Johnson, M. S. (2012). Randomized government safety inspections reduce worker injuries with no detectable job loss. Science, 336(6083), 907-911.
Random safety inspections to evaluate workplaces' adherence to safety regulations are controversial. Dr. David I. Levine and colleagues (2012) note this controversy in the context of the Occupational Safety and Health Administration (OSHA) in the United States. On the one hand, supporters believe that the regulations and inspections save lives. On the other hand, those opposed believe that the regulations and inspections cost people their jobs.
In the study by Dr. Levine and colleagues, the goal was to conduct a rigorous test of whether safety inspections are beneficial or detrimental. Because not only do anecdotal views vary, but research findings on the topic also show varied results. For example, some prior research suggests no correlation between inspections and injury rates whereas other research suggests a decline in injuries following inspections. Similarly, some research suggests no association between inspections and productivity while other research suggests lower productivity.
The inconsistent findings of prior research are likely in part due to the research designs used. Thus, Dr. Levine and colleagues sought to provide a more rigorous test of the relationship between safety inspections and injury rates.
The state of California provided the perfect conditions for such a test. From 1996-2006, the California division of OSHA conducted random safety inspections in workplaces within high risk industries. This provided conditions for a simulated randomized controlled trial.
The Study:
Levine et al. (2012) took the randomly selected workplaces and matched them with eligible workplaces that were not selected for a safety inspection. Therefore, they had a treatment group (the workplaces randomly selected for inspection) and a control group (workplaces not selected, but were eligible).
Safety outcomes of interest: injury rates, and injury costs.
OSHA mandates injuries be recorded; however, these data are not always kept up-to-date in organizations. This under-reporting is often identified during OSHA inspections. After OSHA inspections, reporting of injuries improves which causes it to appear as though injuries have increased even though this increase is most often due to accurate reporting which was lacking prior to the inspections. As such, for this study, they used workers’ compensation injury data which is less likely to be affected in such a way by OSHA inspections.
Non-safety outcomes of interest: employment, company survival, and compensation. The purpose of these outcomes was to see if there are unintended negative consequences of inspections as some claim there to be.
To be included in the analyses organizations had to meet the following criteria:
Single-establishment firms (because injury data are company level)
No inspection in prior two years (separate from focal inspections in treatment group)
Final sample consisted of 409 matched organizations.
The researchers adjusted for pre-inspection characteristics (this decision was driven by difference in pre-inspection injury rates, although this was likely due to sampling error and not how the OSHA inspectors chose workplaces to inspect).
Findings:
(1) Randomized inspections related to fewer annual injuries:
In inspection year
3 years post-inspection
4 years post-inspection
(no significant effect on injuries 1 and 2 years post-inspection)
(2) Distinguishing between minor and major injuries (based on monetary amount of the workers’ compensation claim cost), the inspections reduced both minor and major injuries.
(3) Randomized inspections associated with lower injury costs:
In inspection year
3 years post-inspection
4 years post-inspection
(no significant effect on injury costs 1 and 2 years post-inspection)
(4) Inspected workplaces vs those not inspected did not differ in employment, total earnings, nor savings.
Declines in these outcomes for inspected workplaces can be ruled out (i.e., there is no evidence for unintended negative consequences of safety inspections).
Conclusion:
Overall, findings from this study suggest safety inspections are beneficial, not detrimental.
That is, safety inspections were related to lower injury rates and injury costs, and were not related to employment, earnings, nor savings.
This study suggests that the opposition view to safety inspections is not supported.
Thank you to Julian Barling for comments on this blog post!
Comments