Strategies for construction hazard recognition
Table 3: Results of Case 1- Multiple baseline study on modular construction project
Download 2.75 Mb. View original pdf
|
| Table 3: Results of Case 1- Multiple baseline study on modular construction project Predictor Coefficient Std. Error t value p value Model test (Fcritical = 3.682) r 2 D-W test p H-M test (α = 0.05) p Levene's (α = 0.05) p A-D (α = 0.05) Crew Structural Constant 27.063 1.935 13.987 0.000 Model Iii Fiiobt = 12.849 0.990 2.874 0.287 0.091 0.169 Time 1.853 0.497 3.729 0.003 D 26.006 2.288 11.366 0.000 SC -1.067 0.547 -1.950 0.075 Crew 2: Electrical Constant 38.806 0.904 42.941 0.000 Model II Fobt = 0.187 0.978 2.550 0.669 0.915 0.783 D 31.155 1.278 24.377 0.000 Crew 3: Piping Constant 38.623 0.537 71.973 0.000 Model II Fobt = 1.383 0.993 2.386 0.640 0.578 0.833 Db 0 20 40 60 80 100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Ha za rd re co gn iti on le ve l ( %) Work Period (Time) --> Crew Crew Crew Intervention 100 Comparing the baseline and treatment phase data, crew 1 shows a level-change (D) improvement of 26% (p<0.001), occurring just after the intervention was introduced. This value signifies the difference between the projected value of the baseline phase data in the absence of the intervention for the seventh time period with the value observed just after the intervention was introduced (T=7). That is, the projected regression from the baseline data for the seventh period is 40% (b 0 +b 1 (T)), while the value for the treatment phase is 66% (b 0 + b 1 (T) + b 2 (D) + b 3 (SC)), the difference being equal to the level change coefficient of 26% (66% - 40%). Similarly, level change coefficients may be determined for various points in time using the regression equations. For example, crew 1, using the same regression equation of the intervention phase, it can be inferred that the crew identified about 73% of the hazards in their work environment in the sixteenth time period. Also, the p-value (see table 3) clearly implies a statistically significant level change improvement in hazard recognition and communication for crew 1. In the same manner, crew 2 and 3 exhibited a level-change improvement of 31% and 38%, respectively. The slope change coefficient of -1.067 for crew 1 represents the change in slope between the baseline and intervention phase. Hence, the slope in the intervention phase is the difference between the slopes of the two phases which is equal to 0.79 (-1.067+1.853). Although there is a decrease in the value of slope from 1.85 to 0.79, the data implies the presence of an average increase in the hazard recognition and communication level of 0.79% per work period. However, the p-value (p=0.075) associated with the change in slope does not suggest statistical significance. For the other two crews, since the null hypothesis, β 1 = β 3 = 0 was accepted yielding to the choice of Model II, the slopes in both the A and B phases are assumed to be zero. The overall level change, statistic summarizing the net improvement encompassing the three crews 101 was computed based on equation 3. Accordingly, the reciprocal of error variance estimate showed about 35% increases in hazard recognition and communication. The corroborative test using photographs representing construction scenarios strongly suggested a statistically significant improvement in hazard recognition for crew 1, 2 and 3 of 43%, 52%, and 41%, respectively. The p-value for all comparisons was less than 0.01, indicating highly significant results. Download 2.75 Mb. Share with your friends:
|