Table H.1.4. Study #4: Lewchuk, W., Robb, A. L., and Walters, V. The Effectiveness of Bill 70 and Joint Health and Safety Committees in Reducing Injuries in the Workplace: The Case of Ontario, Canadian Public Policy. 1996; 22(3): 225–243.
Did Bill 70, which took effect at the end of 1979, change the work-related injury and illness frequency rate?
POOLED TIMES SERIES WITH COMPARISON GROUP
This pooled time-series cross-sectional design compares the period BEFORE implementation of the bill (1976-1979) to the period AFTER implementation of the bill (1980-1989). The comparison is made in BOTH manufacturing and retail sectors. The authors then contrast the before-after comparisons in the two sectors.
TYPE OF INDUSTRY: 436 were manufacturing and 201 were retail. SAMPLE SIZE: 637 (only 636 participated). SIZE OF WORKPLACES: Manufacturing enterprises had at least 50 employees in 1988, Retail enterprises had at least 20 employees in 1988, There is no indication of the range in sizes of the workplaces, nor is there any indication of the size of the workplaces over time (data was presented for the period 1976 through to 1989, so the number of employees in 1988 may not accurately represent the number of employees in 1976 through to 1987 or in 1989), Retail enterprises had at least 20 employees in 1988, There is no indication of the range in sizes of the workplaces, nor is there any indication of the size of the workplaces over time (data was presented for the period 1976 through to 1989, so the number of employees in 1988 may not accurately represent the number of employees in 1976 through to 1987 or in 1989), Number of Workplaces: 637 (only 636 participated). SAMPLING METHOD: Sample included all those workplaces that had provided at least some information in a 1991 IAPA survey of workplaces (n=497) plus some additional workplaces (n=140), The 140 additional workplaces were all non-respondents to the IAPA survey, Although the authors report that the additional workplaces were included to “maintain the confidentiality of the IAPA survey participants”, no specific criteria is provided with respect to how these workplaces were chosen from all non-responding workplaces, nor is there any indication of the proportion that were chosen from the retail sector versus the proportion chosen from the manufacturing sector, The response rate to the survey was low (539/1032 = 52%), Even with the addition of the 140 non-responding workplaces, the response rate would still be low (679/1032 = 65%), Therefore, sampling bias may affect this study whereby those workplaces which initially participated in the IAPA study and those workplaces that were chosen from the non-respondents do not accurately represent the OHS activities and outcomes of all workplaces, In fact, it may be that the sample represents OHS conscientious workplaces since those likely to respond to a survey about OHS activities and outcomes are likely those workplaces which are active in OHS activities and have positive outcomes, If overrepresentation of OHS conscientious workplaces occurred equally across the sectors (retail and manufacturing), the estimate of the effect of the legislation will be unbiased, however, the results will not be generalizable to workplaces outside of the study.
Maintenance of Sample: Not applicable. Cross-sectional time series.
Effectiveness of Occupational Health & Safety Management Systems: A Systematic Review