Data collection is such an imperative piece. First of all, it establishes baselines. That's a critical component to determining whether legalization has had an impact. I will tell you that Colorado government agencies struggle to answer this question because we didn't have good baselines.
While alcohol-impaired driving still constitutes the vast majority of cases, alcohol-impaired driving numbers have gone down. Drug-impaired driving numbers are on the rise, and that is multiple categories of drugs. It would appear there was some uptick in marijuana-impaired driving cases, including some evidence of a rise in fatalities, but as I alluded to, we had certain limitations within the data that was collected from crashes, where we didn't know if marijuana was truly the proximate cause. This was in part because when we had fatalities, when people were tested, they were only given a urine test, and urine tests really only tell you whether somebody had marijuana in their system within the last 30 days.
It did appear there was an increase, although that number has levelled off a bit. To give you truly accurate numbers as to what the impact has been on traffic safety in our state, I think we're still struggling to answer that question.