tl;dr: Inattentive driving kills. Alcohol can cause inattentiveness. So can texting. If someone constantly texts while driving except for when they're 0.081, th

tl;dr: Inattentive driving kills. Alcohol can cause inattentiveness. So can texting. If someone constantly texts while driving except for when they're 0.081, then they're probably a safer driver drunk than sober. Which is to say, they shouldn't be allowed to drive in either condition. 1. Regarding the last sentence of my prior post, I think you're also underestimating the extreme danger associated with inattentiveness while sober driving and/or how common distracted sober driving is. I 100% believe an very attentive person at .081 is less dangerous than an idiot texting and messing with the radio while driving. Which is to say, both are extremely dangerous. See e.g. http://distracteddriveraccidents.com/texting-driving-dangerous-drunk-driving/ http://distracteddriveraccidents.com/texting-driving-dangero... 2. You / the grandparent are under-estimating the impact that attentiveness has in drunk driving accidents, and also underestimating the ability of drunk people to be attentive on demand. The VR test is not a good test of impairment because the drunk driver will be artificially attentive during the administration of the test. Driving drunk is dangerous even before motor control is significantly impaired. And the reason for that is mostly psychological -- a bit reduced reaction time, okay, but the reduced attentiveness and judgement are really far more dangerous. The guy flying down the road 40 over "just wasn't paying attention". It's not that he would have noticed but didn't because "delayed reaction time". He never would have noticed because he was drunk and didn't have the good sense to pay attention while driving. The problem with the VR test is that, when put on the spot, a drunk person can focus for a short or intermediate period of time. So that VR test isn't an accurate reflection of the actual typical behavior of an impaired driver. Under-estimating the only-on-demand abilities of drunk people would be a deadly mistake when designing measurements of impairment. Measuring ability, not behavior, is not enough. Drinking changes behavior in ways that are dangerous while driving for for which a drunk person can often compensate (e.g. when confronted with a potential DUI, but often not when driving home on a routine route).