66 Similar to case 1, two safety managers with 34 total years of experience worked alongside with the researchers in the study. Three crews – electrical, structural and insulation – were selected to participate in the study from a stratified sample. The size of each crew ranged between 7 and 12. The same protocol implemented in case 1 was strictly followed.
Case 2 analysis results Figure 6 represents the
HR index measured overtime, and Table 4 tabulates the analysis results. As can be seen, Model II
was selected to represent the HR performance of crew 1 and 2 based on the model comparison statistic as represented in Equation 5. Likewise, the performance of crew 3 was appropriately represented using Model I. That is, each crew revealed a level change but crew
3 also revealed a slope change in hazard recognition performance. As indicated in Table 4, all assumptions of homoscedasticity and normality of error variance were satisfied with an alpha value of 0.05. The Durbin-Watson test revealed no evidence of autocorrelation for crew 2 and 3, but showed inconclusive results for crew 1. However, the
Huitema-McKean test statistic for autocorrelation indicated that the assumption of independent errors
were valid for all cases, implying the adequacy of the selected models.
67
Figure 6: Results of Case 1- Multiple baseline study on fluff pulp processing facility Table 4: Results of Study 2- Multiple baseline study on fluff pulp processing facility Predictor Coefficient Std.
Error t value p value Model r2
D-W test p
H-M test
(α =
0.05) p
Levene's
(α =
0.05) p
A-D
(α =
0.05)
Crew Electrical Constant
53.670 1.896 28.312 0.000
Model Iii Fiiobt = 2.6711 0.899 1.327*
0.069 0.203 0.349 D
26.794 2.398 11.174 0.000
Crew 2: Iron workers 0.126 0.079 0.611 Constant
38.374 2.069 18.544 0.000
Model II Fobt = 0.786 0.880 1.446 D
30.837 2.927 10.537 0.000
Crew 3: Insulators Constant
47.225 2.747 17.191 0.000
Model Iii Fiiobt = 5.1066 0.918 1.872 0.190 0.116 0.636 Time
-0.314 0.443
-0.708 0.492 D
26.949 4.002 6.734 0.000 SC
3.309 1.058 3.127 0.009 inconclusive using Durbin-Watson test
0 20 40 60 80 100 1
2 3
4 5
6 7
8 9
10 11 12 13 14 15 16
Share with your friends: 
