On the difference between making these generalizations along racial lines vs other things - it mostly comes down to the fact that our ancestors ruined the fun.

On the difference between making these generalizations along racial lines vs other things - it mostly comes down to the fact that our ancestors ruined the fun. After humanity had used certain type of statements to justify mass dehumanization, slavery, Holocaust, etc., people don't want to see that happen again. Then it's just a question of how big a red circle we want to draw around that kind of talk. Most people found a simple ban on racial generalizations relatively intuitive and practical. I'd prefer a more freewheeling culture but it's above my pay grade at this point. Side point - in general an important thing to remember here is the big, big difference between P(A|B) and P(B|A). It could be the case that all aggressive drivers I ever encounter are BMW drivers, and yet the rate of aggressive BMW drivers could be very low and of little practical predictive value on any individual BMW driver (e.g. 1% of BMW drivers versus 0% for non-BMW drivers). People are already bad at distinguishing an implication and its converse, and they're quite a bit worse at the statistical version of the same. Arguably this is a reason we should keep our generalizations to ourselves, even if they're accurate. A lot of people will take them to mean a lot more than they do. It's not my style, but I do recognize the risk.