The other half are rich pricks. That also checks out. The kind of privileged douche who would buy a BMW is the same kind that thinks they're above the law.
The Dodge ram is red and I know the guy who drives it and keeps getting DUIs. I've offered to share my Uber with him just so he's not putting people in danger but he just yells "Yee haw, brother", fixes his greasy mullet, and hops in.
It is likely a bit of selection bias by cops as well. They target the rich jerks because of the expensive eye catching cars and the truck assholes because they are big and drien bynyokels.
The drunken Camry crew is out there just blending in.
The real crazy part here is that the s-10 was discontinued in North America in 2004 and still makes this list? Does that tell us that s-10 drivers are wild or that this data is 20 years old?
That was my thought too. Wonder what the timeframe was because if it’s data collected over multiple years you’d expect to see an overrepresentation of vehicles that were sold through that whole period while models that get discontinued, or launched in that timeframe would be underreported. Also maybe some demographics, like was the high number of S-10 while it was available new and presumably driven by people that recently purchased those new vehicles, or is it 10+ years after it stopped being sold when it’s the old farm shitbox or a young guys first truck.
Make and model aside, what I want to know is when the bars close at two and I see numerous clearly drunk drivers on my way to work at five, what have they been doing for the past three hours?
I'd like to see this data against ownership numbers. Like, what percentage of Dodge Ram owners have a DUI.
I don't think that's what this data is showing us. From what I see, it looks like, from a random set of 1000 DUI's, these were the most common vehicles, and/or, from 1000 vehicle drivers, those that got DUIs were driving these cars most often... Sort of thing.
I want to know, how many were purchased, and how many individual drivers were cited for DUI while driving that vehicle.
I'm betting that all the very rare and expensive vehicles would have a huge percentage of DUI drivers. Someone driving around in a hurrican, probably has enough money to afford whatever ticket the law can throw at them, and there's probably so few of them that actually obey the laws that a large portion have been caught DUI.
I like how you explained how the dataset was biased, then instead of describing how to eliminate/reduce that bias, you applied the same underlying tactic to make someone else look bad instead.
Oh, I'm just curious about alternatives. There's a lot of ways to measure this, but only a few that are mostly unbiased.
The results will likely always be skewed one way or another, since this is about punishment, not the number of people committing the crime. It's impossible to know how many people drive while intoxicated, we only know how many get caught doing it. The data is flawed from the start.