Citation needed. I've seen claims like that, but none of them from a source that stands up to scrutiny. Most such claims come from the people promoting self driving cars and so they have reason to "lie with statistics". Those who have unbiased data (ie governments) are not talking about it from what I can see.
I personally am significnatly safer than the average driver. This comes solely down to me not drinking alcohol and thus I never drive while drunk. The typical driver also isn't under the influence and thus is significantly better than the average. (I also try to follow other safety practices, but I'm not sure if I'm really better - I'm aware of and pay attention to one thing which makes me better - but what am I not aware of that others are doing?)
Most people are safer than average most of the time. Risk is a heavily bimodal. The issue is that people keep driving even when it's risky. Maybe they have to get to work even though they didn't sleep much, or they need to get home from the bar, or they're road raging, or they don't have someone with them to drive instead.
People driving home from bars is the biggest issue of the ones you mentioned. People drink alcohol at bars, which impairs their judgement, and they then use this impaired judgement to then decide that it's a good idea to get behind the wheel.
They're all big issues. We have an entire set of federal regulations specifically to address fatigued driving for commercial haulers called Hours of Service.
