I don't think that's right. In Poland the theoretical test is just as stupid as OP said, with a lot of "gotcha" questions designed specifically to catch you out, and yet people pass without bribery and then drive like bellends on the road. Having said that, people drive very aggressively and frequently are inconsiderate, but I'd say they absolutely know what the rules are - they just choose to ignore them because they think they can.
In my US state (which is nationally renown for its authoritarianism) you only ever have "gotcha" questions for stupid stuff that doesn't actually matter on the road. Like "what is the maximum sentence someone with one DUI 366 days, but not more than 24mo ago can receive", as if that actually matters on the road. It would be a lot more palatable if it was some obscure questions about a 5-way intersection with a light only controlling a fraction of the streets or something about a rail crossing and a drawbridge.
A different country, similar test. I remember you would just memorize the questions and their right answers. There was a limited set of questions, publicly available.
Most definitely, which would be the same if something like that had been enforced in the US. We'd have some "Driving Schools", which would "guarantee" your license or something like that. Though people would have likely voted the government out first, but looking at the building codes I am not 100% confident.
Quality depends heavily on how corrupt your government is — in Germany, these strict tests are actually that strict and work well. But obviously not everyone can get a license due to this. (e.g., I opted not to spend the 2000€ that a license would cost).
More the latter.
There are mandatory hours of training in specific conditions before being allowed to take the actual exam.
Driving hours with the instructor aren't cheap.
Don't know about russia, but in my country plenty of people driving without a license or with a suspended one. Also, no limit on how many times you can take the written exam, so someone maybe had to take it ten times, got lucky with the questions or memorised all intersections and passed. The more academic folk ace the written exam, but then struggle with the driving part. And when you fail the driving part, you don't need to revisit the quiz portion, just the driving. Someone who passed their driving test after 5-6 attempts rarely tends to be a good driver.
Is there actually a strong correlation between what people intuitively think of as a good or bad driver and that driver's accident rate or accident severity?
Outside the obvious extremely bad drivers (old people who are practically blind, people who habitually text and drive, people who drive at huge speeds above other traffic even in traffic to dense to really be doing that effectively, etc), probably not.