Some officials have assumed that there is a direct correlation between an increase in police action (more reports) and a decline in the incidence of driving while intoxicated (DWI). The logical conclusion from such an assumption is that the greater the number of police employed to detect DWI, the less people will drink and drive. The situation is much more complex than this, however, because people's decisionmaking, particularly while under the influence of alcohol, involves varying degrees of risk-taking and subjective perceptions of the probability of being observed and stopped by the police while DWI. Further, the situations occasioning a person to drive while under the influence of alcohol will vary. It is much more reasonable to suppose that the subjective perception of high risk-taking in DWI, to the extent of deciding not to do it even while decisionmaking is influenced by alcohol consumption, will only be favorably impacted as the number of police employed to detect DWI reaches a significant level. This discussion provides ground for empirically investigating the consequences of reducing police action against DWI in certain areas, such as the reduction in mobile patrol. A reduction in mobile patrols would not only give the police an opportunity to give more attention to matters other than DWI, it would reduce report processing time and court backlog. The effects of this reduction on DWI would be determined. Roadside surveys should be kept at present levels, and in addition, a publicity campaign should be mounted that will encourage citizens to change drinking patterns if they know they are going to be driving. Figures, footnotes, and seven references are provided.