The Validity of Police Reported Accident Data?-

Information theory and signal detection theory techniques were used to assess the validity of police reported traffic accident data. The validity criteria were the data and conclusions of muiti-disciplinary accident investigation teams who investigated the same traffic accidents. The results indicated that the accident level variables reported by the police with least reliability were vertical road character, accident severity, and road surface composition. The most reliably reported data were those concerned with the accident location, date, and number of drivers, passengers, and vehicles. The informativeness of the police reports with respect to driver/vehicle characterist@ was practically nil, with the exception of driver age, sex and vehicle model for which the police were correct most of the time (but not errorless). It was afso found that police reports provided very little information regarding the presence of different human conditions and states, vehicle defects and environmental/road deficiencies. The sensitivity of police investigators to all accident causes was low. When causes were categorized into human direct, human indirect (conditions and states) vehicle, and environmental, police were the most reliable with respect to human direct causes and the least reliable with respect to environmental and human indirect causes. Implications for improvement and use of police data are discussed. Police accident reports are probably the most ubiquitous source of traffic accident data analysis. While the primary purpose of such reports is to provide both summary descriptive statistics on accidents and information that might later be used for litigation purposes, very often data from these reports are taken at face value for inferential analysis, most notably in the area of traffic safety improvement programs. Thus, many safety programs are evaluated on the basis of whether or not they yield a reduction in accidents as reported by the police. In conducting such analyses, one must be aware that at least as far as rigorous scientific procedures are concerned, this approach is questionable. This is because in any scientific data gathering effort, the nature of the data collection process is often dependent on the objectives of the program. In the case of analyzing police data, however, the objectives of the researcher may be totally different from those of the policeman who is collecting the data at the scene. Thus, while police reports may be a useful source of information for the evaluation of various safety improvement programs, they are, as indicated by the research described below, often biased and~or incompIete. Various studies have demonstrated that even at the level of reporting accident frequencies, sources other than police reports may be more complete. With the possible exception of alcohol reiated accidents, driver self-reports typically reveal more accidents than police reports [McGuire, 1973, 19761. Accident reporting is also less complete the less severe the accident, Comparison of police records with hospital records in England showed that over 30% of injury-producing accidents were not reported to the police at all [Bull and Roberts, 1973; tThis work was conducted as part of a “Tri Level Study of the Causes of Traffic Accidents Modification, Task 3: Validity Assessment of Police Reported Accident Data” performed by the Indiana University Institute for Research in Public Safety for the National Highway Traffic Safety Administration, Contract DOT-HS-O34->535. Thanks are due to Nicholas S. Tumbas, who was responsible for the mapping of accident data from IRPS to police categories and Ronald W. Drahos who reviewed the accident cases. fRequests for reprints should be sent to David Shinar. AAP Vol. IS. No. 3-A 175