RESERACH METHOD

Research Method

Research Method

Statistical Analysis

The purpose of this investigation was to study rider performance in terms of crossing time among different rider populations. The initial hypotheses was made that, gender does not affect a cyclists' crossing time performance. In order to test the hypotheses, statistical tests were performed on the data and are discussed in the following sections. These statistical tests are the chi-square test for homogeneity and unpaired t-tests on the mean and 85th percentiles crossing times.

The chi-square test for homogeneity is performed to determine if the distributions in crossing times between categorical groups are homogenous; this test gives an indication as to whether or not the crossing time distributions between the groups being studied are significantly different or not. Unpaired t-tests are conducted to determine if a statistically significant difference in mean and 85th percentile crossing times between age and gender groups exists (Field, 2009, 66). The variances of populations represented by the compared samples were assumed to be equal and unknown. Results are discussed for the three measured crossing times (T1, T2 and T3), where the data are binned using crossing time percentile ranges (0-25%, 25-50%, 50-75%, 75-100%). This produces four bins in which the data samples are grouped. The null hypothesis (that the distributions are homogenous, and do not exhibit a statistically significant difference) was tested at a confidence interval of 95%, where the null hypothesis is rejected at an alpha value of 0.05 or less.

Results

At the level intersection during the winter collection, the null hypothesis (that the data are homogenous) was rejected for the male-female comparison for T1, T2 and T3. During the summer collection, the male-female comparison shows a rejection of the null hypothesis for T2 only. This suggests that the distributions between male and female T1 crossing times in the summer collection period are similar. A possible explanation is that, because T1 corresponds to the portion of the intersection where cyclists are accelerating, the reaction time and accelerations for male and female cyclists may be similar. But for crossing T2, males may continue accelerating, while females may have reached peak cruising speed. At the grade intersection, the results for winter and summer are very similar to the level intersection (Vincent, 1999, 44). The null hypothesis was rejected for the male-female comparison for all three measured crossing times, expect for T2 in the winter collection. These results may be explained by male cyclists ...