| The main purpose of this study is to determine if statistically significant differences exist between the national data sets, the individual state data sets, and the University of Louisville (UofL) state data sets. As described in Chapter five, the analysis was performed by use of a single multiple regression model, which determined slopes and intercepts simultaneously, using a dummy variable. The data sets were also examined graphically by plotting the data points and regression curves. The regression curves (i.e., the national and state data regression curves) developed by the regression analyses were compared by the p-values from SPSS calculations.; For automobiles, approximately 10.9% (7 out of 64 comparisons) of the comparison results are statistically the same between the regression curves, while about 89.1% are significantly different. For medium trucks, approximately 40.6% (26 out of 64 comparisons) of the comparison results are statistically coincident between the regression curves, and more than 59% of the comparison results are significantly different. Finally, for heavy trucks, approximately 15.6% (10 out of 64 comparisons) of the comparison results are statistically the same between the regression curves, with about 84.4% of the comparison results being significantly different.; Therefore, this research work proves that the national and the majority of the individual state data sets are different. This implies that one may use the national data set to estimate sound levels, however one cannot always predict state-specific sound levels accurately. The comparison analysis in this study proves that there are large variations between the national data set and the individual state data sets. Based on the comparison results, this study further supports the theory that state-specific REMEL data may generate more accurate results in current noise prediction models than using national averages in the same models. |
PDF Full Text Request