Two competing studies attempting to draw a connection between marijuana legalization and car crashes have created even greater uncertainty on the subject.

The Insurance Institute for Highway Safety released a report on June 22 stating that marijuana legalization in Colorado, Oregon and Washington resulted in collision claim frequencies that are about 3 percent higher than would have been expected without legalization.

The IIHS’s Highway Loss Data Institute report points out that researchers haven’t been able to definitively connect cannabis use with more frequent real-world crashes.

The study looked at insurance claims for vehicle collisions between 2012 and 2016. Colorado legalized recreational marijuana in 2012, which took effect on Jan. 1, 2014. Claims were compared to neighboring states without legal marijuana.

Meanwhile, a second study also published online on June 22 in the American Journal of Public Health found no increase in traffic fatalities in Colorado and Washington after legalization.

Researchers used the U.S. Fatality Analysis Reporting System to determine the annual numbers of vehicle crash fatalities between 2009 and 2015 in Washington, Colorado and eight comparable states. The study compared changes in fatality rates before and after legalization.

The study found that post-legalization changes in fatality rates for Washington and Colorado did not significantly differ from those for the comparable states.

The study also found no association between recreational marijuana legalization and total crash rates when analyzing available state-reported nonfatal crash statistics.

“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,” the study found.

Mason Tvert, spokesman for the pro-legalization Marijuana Policy Project, said the two studies indicate that there are no strong conclusions that can be drawn.

“This area of research is characterized by confounding factors and limited data, so it is not particularly surprising to see different research methods arriving at different results,” Tvert said. “What is clear from both studies is that Colorado’s decision to regulate marijuana for adult use has not resulted in the instant doomsday scenario that some had predicted. This is an important issue deserving of study, and it will surely continue to be examined.”