With the aging population in North America, one of the most cited issues is about the U shaped trend of crash fatalities compared to age. One of the papers that highlights this was back in 1993 American Journal of Epidemiology by Leonard Evans where younger drivers had a higher crash risk that decreased with age and increasing again as people got older. Although this is not saying older drivers are more unsafe but other factors such as being more frail and impacts that a younger person would walk away from it would be fatal for an older person.
PER_TYP: Role of the person in crash, e.g. driver, passenger, pedestrian, etc.
In order to read the data, the foreign package will be used and more precisely the function read.dbf to import the data.
Census
Population data was taken from CDC Bridge-Race Population Estimates. This provides yearly population data of from the United States based on Age and Gender.
The files are formatted as tab deliminated and have some extra information at the bottom of the file that needs to be removed.
Crashes Per Million
To calculate crash per million by age, a histogram of all crashes is made first. After this the total population is divided into it and multiplied by a million.
The census data groups people 85 and over into one category, because of this anyone 85 and over will not be counted.
Using the functions above, the data for crashes per million can be calculated.
Graphing
To graph the data the ggplot2 package will be used with age on x-axis and crashes per million on y-axis.
Conclusion
As was the case in the paper, the U shaped trend continues in recent years.