Strategies for construction hazard recognition
Download 2.75 Mb. View original pdf
|
- Navigate this page:
- Case 1 analysis results
| RESULTS AND DISCUSSIONS Case 1: Maintenance and Construction in oil and gas facility We conducted our first study in an oil and gas facility that manufactured oil additives in the southern US. The facility employed more than 650 workers and accumulated an average of 277,400 worker-hours every year. The annual revenue of the facility was not disclosed to the researchers for confidentiality reasons. During our visit, two safety professionals with more than 37 total years of relevant experience were recruited by the site leadership to assist in gathering 62 data and introducing the SAVES training protocol. Three crews – mechanical, civil, and maintenance – were identified from a stratified population to participate in the study. The size of each crew ranged between 5 and 9. Case 1 analysis results The HR index for the duration of the study is presented for all three crews in Figure 5. The respective findings are summarized in Table 3. As indicated in the previous section, for each crew, model I was estimated by regressing the HR index over the three predictor variables (T, D, and SC) and model II was estimated by regressing only on the level change dummy variable (D) see Equation 2 and footnotes. This was followed by the model comparison test to determine the most representative mathematical model for statistical inference. For crew 1, the obtained value of F (F obt = 0.209) was compared to the critical value (F critical = 3.682) using an alpha level of 0.05 and the degree of freedom (df = 2, 12). Clearly, the test indicated that Model II was preferable to Model I (F obt < F critical ). Using the same procedure, the results of crew 2 were appropriately represented using Model II and the results of crew 3 were represented using Model I. Accordingly, results of crew 1 and 2 indicated only level change but the results of crew 3 indicated both level change and slope change. |