CIE x050-PO013 PDF

Comparing the counts of events occurring at different times of day is one approach to measuring the impact of ambient light level on events such as crime, road traffic collisions and travel flows. We investigated the choice of case and control hours on the outcome of analysis of cyclist flows for two cities, Arlington (VA, USA) and Birmingham (UK). The analyses revealed some differences in the estimated odds ratios. While all combinations tended to show a significant effect, there was some variation in effect size, and that could affect the determination of practical relevance. We suggest that future work presents mean odds ratio from all combinations of case and control hour, weighted by the number of events in the case hour.