Premium
The Distribution of Risks: Vehicle Occupant Fatalities and Time of the Week
Author(s) -
Schwing Richard C.,
Kamerud Dana B.
Publication year - 1988
Publication title -
risk analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.972
H-Index - 130
eISSN - 1539-6924
pISSN - 0272-4332
DOI - 10.1111/j.1539-6924.1988.tb01159.x
Subject(s) - quantile , statistics , risk assessment , econometrics , distribution (mathematics) , mathematics , computer science , computer security , mathematical analysis
Automobile accident risks vary significantly across populations, places, and times. This study describes the time‐varying pattern of societal risk. The relative risks of occupant fatality per person‐mile of travel are estimated here for each hour of the week, using 1983 data. The results exhibit a strong time‐of‐day effect and have a highly skewed frequency distribution, implying wide variations in risk‐taking behavior. Indeed, the 168 hourly estimates ranged from a low of 0.32 times the average around Sunday noon to a high of 43 times the average at 3:00 a.m. on Sunday, i.e., by a factor of 134 from bottom to top. Quantile‐quantile plots or “Lorenz curves,” introduced to display the unequal distribution of risks, show that approximately 34% of the vehicle occupant fatalities occur in hours representing only 5% of the travel. These findings have serious implications for risk analysis. First, when attempting to reconcile objective and subjective risk estimates, risk communicators should carefully control for when and to whom the risk in question is applicable. Second, comparisons of hazards on the basis of average risk are necessarily misleading for risks distributed so unevenly. Third, resource allocation decisions can benefit by knowing how incidence, exposure, and risk vary across time, place, and other relevant variables. Finally, certain cost‐benefit analyses that use average values to estimate risk exposure can be misleading.