Effect of an intervention on a road fatality time series is studied using (i) the usual intervention analysis of Box and Tiao (1975), and (ii) CUSUM charts for the one-step-ahead forecast errors. It is shown that the seat belt and speed limit legislations of the Ontario Government had some impact in bringing down the road toll level.
KEY WORDS CUSUM charts Forecast errors Intervention
Seat belt and speed limit legislations Time series
The Ontario Ministry of Transportation and Communication, in 1975, developed and implemented a program to inform the public of the benefits of seat belts as a road safety countermeasure and in November 1975 the Ontario Government announced mandatory seat belt legislation to take effect from January 1, 1976 onwards. This legislation requires everybody, with few exceptions, in a motor vehicle equipped with seat belts to wear seat belts while travelling on Ontario roads (Snow, 1975). Seat belt wearing may not reduce traffic accidents. However, it could transform deaths to injuries and greatly reduce the severity of some of the injuries. The legislation was expected to reduce the number of deaths due to traffic accidents. It is of interest to study whether or not the expected result materialized. This evaluation is made difficult because of the Government's introduction, at the same time, of another legislation to reduce speed limits in the Ontario road system (Davis, 1975). Hence we are forced to consider the two legislations together as one intervention.
- DATA AND MODELS
2.1 Data
Traffic fatality data usually include the deaths of drivers, passengers, pedestrians, bicyclists, etc. However the deaths of drivers and passengers are more relevant to our discussion. Figure 1 shows the monthly number of drivers and passengers killed (D,) in traffic accidents in Ontario from