Tax Day – Auto Fatalities – Give Me a Break

The numbers the media cites are wrong the claimed effect is trivial, so why is this news?

Since United States tax returns must be filed by next Tuesday there was the inevitable story this week linking an increased risk a dying in an auto accident to tax day. The source was a “Research Letter” in the Journal of the American Medical Association authored by Dr. Redelmeier titled “Road Crash Fatalities on Tax Day.”
CNN reported:

“The study focused on the 6,783 people who died in car crashes in the U.S. over the last 30 years on April 15, compared to the week preceding tax day and the week that comes after. The results found a nationwide average increase of 13 fatalities on tax day, which is a 6% increase.”

Reporters apparently forgot the very basic step of verifying the numbers they were reporting on. The data used in the study is available from the National Highway Traffic Safety Administration in their Fatality Analysis Reporting System. So verifying the numbers is not a major endeavor. Dr. Redelmeier did not use auto fatalities for his research, but rather persons involved in fatal accidents. Not all people involved in a fatal accident die. So the actual number of fatalities is about half of those cited. The 6,783 deaths on tax day cited by CNN is actually about half that number. Thus the claim of 13 additional deaths on an average tax day is more likely about six additional deaths.

What was the source of the media error? The original item in the JAMA was titled “Road Crash Fatalities on Tax Day.” This may have caused the confusion. The situation is not helped by descriptions in the text of the article which refer to the “number of individuals in fatal crashes.”

To put this discussion in perspective this news item is discussing six additional deaths on Tax Day. At least that is the claim. That is six deaths out of total of 32,000 auto fatalities each year. While I would not want to trivialize six deaths, it does seem that priorities are very much misplaced if much in the way of resources is expended to deal with 0.02% of auto deaths.

The media got carried away with this thought. The Los Angles Times cites a former head of the National Highway Traffic Safety Administration suggesting more research on who are the drivers creating the increase risk.

More research is needed to untangle the link between tax day and traffic deaths, said Dr. Ricardo Martinez, who headed the traffic safety administration in the 1990s. Among the outstanding questions: Does the increased risk come from a small number of really stressed-out drivers — last-minute filers, perhaps — or a large number of slightly edgy ones?

This is looking very much like the full employment act of 2012. Let’s spend money and resources to understand something that has at best a very small impact in the world.

One major methodological issue in the work is the question of the impact of the Easter holiday on the claimed results. Tax day frequently falls in the middle of the Easter celebrations. School students are out on spring break. That effects driving patterns and surely impacts accident probabilities. That was not even considered in the study. And it is only one of many possible confounding variables that should be considered in this type of research.