An analysis and evaluation of the selective traffic enforcement program in American Samoa

Abstract

The study investigated the relationship among four dependent variables reflecting the accident problem in American Samoa and seven independent variables considered to have potential impact on accidents between July, 1974 and April, 1977. The primary goal was to isolate the effects of the Selective Traffic Enforcement Program (STEP), and determine its impact on the number of accidents. The results of this research were based on data collected from the responsible agencies in American Samoa. The data included the number of traffic fatalities, accidents, injury accidents, and injuries resulting from traffic accidents, as dependent variables. The independents variables examined were the number of citations issued, inches of rainfall, gasoline sales, alcohol sales, traffic violation convictions, and overall effect of time in months. Two statistical analyses were used. Product-Moment correlation was applied to determine the relationship between the variables and extant of the relationship as an indicator for more in-depth analysis. Those variables which correlated to accidents and STEP at the .05 level of significance were subjected to stepwise regression analysis. Both analyses were performed on an IBM System III computer in American Samoa. The correlation analysis at the .05 level of significance indicated four potentially significant independent variables for regression analysis, with accidents as the dependent variable. These variables were alcohol sales, STEP, time, and publication. Regression analysis at the .01 level of significance revealed that STEP, the motivating factor behind this research, did not significantly affect accident reduction. In fact, of the 55% of the variability accounted for, STEP accounted for less than 2%. Alcohol sales also did not have impact at the .01 level although it accounted for 19% of the variability. The variable "time" was identified as a significant impactor on accidents at the .01 level and accounted for 23% of the variability. The variability "publication" was also identified as having significant impact on accidents at the .01 level and accounted for 12% of the total variability. It may be of interest to note that had the results been examined at the .05 level of significance the outcome would have remained the same.