Test statistics: t-statistics are presented for each of the independent variables in the 6 specifications of the TFP growth equation. The t- statistics for the final (selected) model for the OHS variables are as follows: INSPECTION 0.41 (NS), REFUSAL 1.14 (NS), PROTECTIVE -1.91 (P<0.01), INFRACTION 3.02 (P<0.05), PREVENTION 2.86 (P<0.05).
Facilitators/ Not applicable Barriers
Did the Design Lack Statistical Power?
Using the rule of thumb of 10 observations per predictor variable, the study may be slightly underpowered (i.e., analysis is based on 57 observations and 9 independent variables are in the model).
Were Any Harms of the Intervention Identified? None reported. IWH Reviewers’ Comments:
This is a pooled time series study of the degree of implementation of LSST regulations for health and safety and effect on productivity growth. The equation for TFP was well developed theoretically and was an improvement upon those used in past analyses of OHS regulations in the US. The statistical analyses were quite thorough, however it was unusual that one of the specifications with 3 significant variables removed had a higher R-squared value than one with those variables included. The findings were unique and suggested that there was the potential for positive economic impacts from the implementation of prevention programs. The primary concerns pertained to the level of observation and analysis and the possibility for confounding. The model was based on 57 observations from aggregated industry-level data. Some of the variables were surrogates for constructs in the model so there was an issue of validity of the measures. It was impossible to rule out the effects of variables operating at the firm level or other possible co-intervening factors on the final outcome variable.
Effectiveness of Occupational Health & Safety Management Systems: A Systematic Review