Other important issues related to the data collection for a drinking-riding study are discussed in this section.
Issues to be considered when selecting sites for data collection or areas from which to analyze archival data should include the following:
- The likelihood of finding sufficient motorcycle crashes. Some areas of the country have relatively longer riding seasons that encourage more people to ride motorcycles and to ride more miles per year. States such as Florida, Texas, and California would be ideal for doing research in this respect.
- The likelihood of finding sufficient riders for participation in surveys. This consideration can be addressed by conducting surveys in States with longer riding seasons.
- The generalizability of results to the larger population. Throughout the United States and the world there are physiological, cultural, demographic, and other differences among riding populations, their drinking behavior, their willingness to drink and ride, and the way in which their bodies process alcohol. To the extent that differences exist among rider populations, data showing the prevalence of riding after drinking and the relationship of BAC to impairment in one area may not generalize to another. On the other hand, it may be reasonable to assume that whatever effect the BAC has on motorcycle operation, it is primarily physiological, and would apply to people of similar physiology regardless of location. For that reason, it seems appropriate to conduct research where it can be done most cost-effectively with confidence that the result can be generalized to other areas, provided there are no general differences between the physiologies of riders in the two areas.
- Differences between States in terms of the existence and enforcement of helmet laws will lead to different rates of helmet usage. Different States may also have different levels of usage of “novelty” helmets—helmets which do not meet DOT standards and offer little protection. Because unhelmeted (or “improperly-helmeted”) riders are likely to be more severely injured in a crash, and more severe crashes are more likely to be reported, any methodology that depends on motorcycle crash reporting will likely see a higher proportion of unhelmeted riders than exist in the population at large. Because there is a correlation between riding unhelmeted and riding while impaired (as well as other demographic variables), some State laws and practices may skew results with respect to BACs of the crashing rider sample. For this reason, it may be impossible to generalize results across States.
- Some cities or areas will have different types of roadways that will influence the types of crashes and the outcome severities. For example, cities having the highest numbers of motorcycle fatalities tend to be in the southern and western States where roads tend to be wider and average speeds within the city are faster. To get representative results, it would be best to examine areas with a variety of road types and speeds.
- Consideration of the relationship between the crash data and the comparison data is necessary. This will reduce the number of confounding variables (e.g., environment, roadway type).
- The San Diego Trauma System was specifically mentioned as a site where quality data could be collected from a relatively large sample of motorcycle crash victims.