banner of Methodology for Determining Motorcycle Operator Crash Risk and Alcohol Impairment

3. Detailed Report of Discussion

Geo-General Comparison Data

This methodology provides comparison data by using a roadside survey to collect BACs and other data. Rather than attempting to match crash data by going to the exact sites of previous crashes to collect comparison BACs, this method is a stratified sampling system that samples cases from several locations. Analysis of BAC, location type, and demographic data allows a more accurate determination of probable BACs at crash sites. In the past, data from this type of study have been compared to FARS data, though it may also be possible to find other sources of crash data.

In some cases, this methodology includes asking surveyed riders where they have been, rather than simply recording where they are. This allows researchers to determine not just where the rider is at the time of the survey, but also all the places the rider has been on this trip, which gives a more accurate picture of where riding after drinking is taking place.

In studies conducted with automobile drivers, survey locations are based on road usage data (i.e., data showing where automobile traffic is concentrated). Ideally, rider survey locations would be based on knowledge of where motorcyclists tend to ride. However, panel members felt that road usage data specific to motorcycles were lacking.

Because this methodology involves collecting data for motorcyclists alone, not for automobile drivers, it will probably be necessary to single motorcyclists out for traffic stops. This could be perceived as persecuting, harassing, or profiling of motorcyclists, many of whom already feel as though they are being treated unfairly by the government and by other road users. It would be beneficial to advertise to local riders that a survey is taking place, and to gain the backing of rider organizations, if possible, to gain support among the rider population when rider-only surveys are conducted. An alternative would be to also stop some car drivers, to avoid the appearance of singling out motorcyclists. This would likely add to the cost and complexity of the study with little benefit beyond the public relations benefits. Another possibility would be to obtain motorcyclist data by including riders in the next national roadside survey. However, based on the numbers of drivers stopped in the last national roadside survey, Voas et al. (1998), and the proportion of riders generally found in traffic, it would not be possible to get enough rider data by simply including motorcyclists in a single national roadside survey.

As with all surveys, OMB clearance would be needed with this methodology.


Compared to collecting comparison BACs in the Case Control study, there would be a greater likelihood of finding riders within a reasonable period. The greater number of riders sampled may make it possible to more accurately assess the distribution of BACs in the rider population at large. Based on past experience with this methodology (Zador et al., 2000), there is a high likelihood of obtaining accurate BAC readings on a high percentage of those surveyed. Compared to the Case Control study there is less need to survey at specific (prior crash) sites, some of which may be more dangerous survey sites than those used for the Geo-General Control Data study.


This methodology has disadvantages generally associated with surveys (e.g., a need to stop motorcyclists, need for cooperation from local police agencies, need for cooperation from motorcyclists, potential impacts on traffic flow, safety issues, IRB issues).

Because there are relatively few motorcyclists in the traffic stream, it would take much longer to survey a sufficient number of riders than it takes to find a usable sample of drivers.

Evidence indicates that selecting survey sites randomly may underestimate drinking compared to going to motorcycle crash sites, because random sites may not be where riding takes place or where drinking and riding takes place. As well, the randomly selected sites have no direct relationship to the crash sites, thus confounding the two data sources.



The cost associated with this type of study would mostly be related to roadside surveys. A survey crew would be necessary. If the same crew is used in multiple parts of the country, there will be travel costs. Because of the low frequency of motorcycles in the traffic stream, it will be necessary to spend far more time at the roadside to get the same number of cases collected in surveys of drivers. If the decision is made to survey drivers as well as riders, to address the appearance of discriminating against motorcyclists, this may result in increased workload and, therefore, increased costs. In some cases, offers of payment to survey subjects may be a cost-effective means of ensuring higher response rates, but these need to be figured into the cost of doing the study. It may also be necessary to reimburse law enforcement officials for their participation.