> People don't go around memorizing statistics of how likely someone who is exhibiting one kind of antisocial behavior (being obviously high, defecating in public, etc.) would be to engage in another kind of antisocial behavior (committing physical violence upon another person)

So you're saying people don't do a rational analysis of the risks... but calling it "irrational" is strange? Lots of human behavior is irrational. It doesn't have to be unusual to call it irrational.

Overestimating the risks of one choice and therefore not getting the benefits of it, and incurring the higher risks of alternatives instead (for the elderly, being socially isolated at home or driving with impaired abilities) is the kind of irrationality that people would benefit from correcting, through education and changing social assumptions.

> So you're saying people don't do a rational analysis of the risks... but calling it "irrational" is strange?

GP's wording implied that rational people take public transit because they've done the risk-benefit analysis, and irrational people don't. I'm not sure if there's actual evidence behind such a risk-benefit analysis.

> people would benefit from correcting, through education and changing social assumptions.

If the goal is to attract more people to use public transit, it would be vastly more effective to police the occurrence of antisocial behavior on it, instead of telling people that it's actually OK, the chance of stepping on a puddle of piss or being shouted at incoherently or being pushed onto an oncoming train is much lower than they think, and is lower than the chance of being injured in a car crash while driving to work. Whether or not that is true (and I wouldn't mind seeing some cold, hard data on it), most people will prefer to roll the dice with their car.

In other words: if you're a social architect keen on implementing your vision of a flourishing public transit system, it would be irrational not to account for people's loss aversion, whether or not that loss aversion itself is irrational.

How many places have public transit systems used by older people, and how many have accomplished this by policing the antisocial behavior away? It seems like it is at least as easy to get older people to use public transit where they might see people with mental illness and addiction than it is to keep mental illness and addiction off of public transit. In countries that value civil liberties it is probably easier. As you observed, fear is driven by perceptions and rare incidents of violence covered in the media, not by statistics, and you can't pull somebody out of a bus or a subway car because they contribute to somebody's fear. You can arrest and remove a person who shits on the floor or threatens somebody, but for every one person like that, there are twenty people who are obviously in bad shape, obviously suffering, and look like they might do something alarming at any moment. You can't arrest and remove those people. You can get people to ride public transit despite the odds of seeing people like that.

People also (rightly or not) feel they can influence the outcome when driving much more than they can on transit (note this doesn't really apply to planes because planes are mass transit that so obviously beat driving after a certain distance).

And they might be right, something like 30% of traffic fatalities involve alcohol, so you can avoid driving after drinking and avoid times when others are drink driving, and improve your odds.