Recently, a study published by the Highway Loss Data Institute tried to link marijuana and car wrecks. The study examined the number of collision claims in the western states of Colorado, Oregon and Washington where marijuana is legal for recreational adult use. The ultimate goal of the report seemed to be to show that an increase in the number of collision claims can be attributed to legalization. Yet since its publication, many have questioned the authority of the report.
What the study says about marijuana and car wrecks
"The combined-state analysis shows that the first three states to legalize recreational marijuana have experienced more crashes," says Matt Moore, senior vice president of the Highway Loss Data Institute (HLDI).
Colorado apparently had the biggest increase in claim frequency when compared with the different control states. When combined, the three states of Colorado, Oregon, and Washington had about 3% more collision claims than the control states — Idaho, Montana, Nevada, Utah and Wyoming.
The study was reported across numerous American and international media outlets.
If car wrecks totals have increased, at least fatalities haven't
The American Public Health Association (APHA) published a study in the American Journal of Public Health, exploring a nuance overlooked by the HDLI study, which linked marijuana and car wrecks overall. The conclusion of their study is that “three years after recreational marijuana legalization, changes in motor vehicle crash fatality rates for Washington and Colorado were not statistically different from those in similar states without recreational marijuana legalization. Future studies over a longer time remain warranted.”
Notice the “fatality rates” nuance. The APHA findings were focused on the severity of crashes, not their overall number.
Pushback to the HLDI study
The HLDI study made quite a bit of noise, and many industry watchers wondered how to interpret its results. In a truly scientific study, the entire methodology of testing and analysis are explained, and different possible explanations are given for the obtained results. Just take a look at the study published by the HLDI, and you’ll notice that it isn’t clear how they obtained such results. The study conducted by the American Public Health Association seems much more plausible, even for the untrained eye.
Sure, there might be a correlation between the year in which recreational cannabis use was legalized and the number of insurance claims for collisions, but that doesn’t prove cannabis causes more accidents. Causation and correlation are not one in the same.
It’s important to understand the difference between correlation and causation. Especially because it appears the HLDI work might lead lawmakers and voters to believe that correlation is enough to influence cannabis-related legislation.
Correlation is a mutual relation between two or more things. Here’s an example:
When temperatures rise, ice cream sales rise. But imagine, if crime rates also rise when it gets hotter outside, does that mean that ice cream causes people to become criminals because violence and ice cream consumption increase at the same time? No, because there’s no causation.
So, if the number of people sending in a claim to their insurance company after a collision rises, does that mean it’s because their state legalized the recreational use of cannabis? It might be, but it isn’t necessarily the case. After a correlation has been found, much more analysis must be done to see if there’s causation or not.
A lot of other things might cause an increase in collision claims. Who knows, maybe people work longer hours, and it’s the fatigue that causes them to react slower than usual, and hit another car. It doesn’t seem like the HDLI studied these kind of possible causes. This makes the study less trustworthy, and more opinionated. It’s easy to look for a correlation between two things if you want to show your opinion on a subject. It’s a lot harder to prove causation between them in order to prove your point.
So, is there a link between marijuana and car wrecks?
If the statistics presented by HLDI are correct, then yes, there is a correlation. However, we agree with a growing number of voices that more scientific study needs to be conducted before drawing a straight line between the two (aka causation).