Strategies for construction hazard recognition
Case 2 description Construction, renovation and retrofit in a manufacturing plant
Download 2.75 Mb. View original pdf
|
- Navigate this page:
- Case 2 results, analysis and discussions
| Case 2 description Construction, renovation and retrofit in a manufacturing plant The second study was conducted in a manufacturing plant involved in the production of household detergent located in the Mid-Western United States. The long-term contract on average generated $18 Million annually and required an average of 342,250 worker-hours every year. During our visit, eight crews were actively involved in the project. Two millwright crews and one piping crew were selected to participate in the study. Similar to case 1, two highly experienced safety managers with more than 32 total years of safety experience worked along with the researchers in data collection and intervention integration. The size of the selected crews ranged between 6 and 11. Case 2 results, analysis and discussions The results and of Study 2 are presented in Figure 5 and Table 3. As indicated, similar to Study 1, the measured HR index overtime for each crew were mathematically represented appropriately using model II. This suggests that the intervention introduced a level change improvement in the proportion of hazards identified and communicated. Further, all necessary assumptions of homoscedasticity and normality of error variance required to perform Ordinary Least Squares (OLS) regression procedures were satisfied, indicating robustness of the mathematical model to make valid inferences. The Durbin-Watson test statistic implied that the 145 relationship among the error estimates was equal to zero, eliminating the need for additional autocorrelation parameters. Comparing performance at the baseline and intervention phase, Crew 1 revealed a level-change improvement of 29% (p<0.005) during pre-task planning. An additional 5% of hazards were identified as the tasks were carried out during the execution phase. Similarly Crews 2 and 3 exhibited a level change improvement of 19% (p<0.005) and 23% (p<0.005) in the planning phase. An additional improvement of 9% was observed for Crew 2 and 1% was observed for Crew 3 during execution. In total, Crews 1, 2 and 3 exhibited a level- change improvement of 15% (p<0.005), 28% (p<0.005) and 25% (p<0.005) before any exposure. The weighted overall level change computed using equation 3 was 23% (p<0.005) for the planning phase and 29% (p<0.005) for the execution phase. In the corroborative test with construction photograph images similar to case 1 results, the two sample t- tests revealed that Crews 1, 2 and 3 were able to identify only 49%, 54%, and 52% of hazards respectively. But after the intervention, Crews 1, 2, and 3 were able to identify 84%, 77%, and 87% of hazards. That is, Crews 1, 2 and 3 revealed a statistically significant improvement of 35% (p<0.005), 23% (p<0.005), and 35% (p<0.005), respectively. |