I wasn't the one asking for stats, you were 
I don't need stats to be able to see that wearing a puffy coat is incompatible with the safety advice that the harness needs to be snug to the child's body to work correctly.
The straps on the model in the video are not halfway down the dummy's shoulders if you look at the version I shared. They are too loose yes, and the seat is forward facing - it's possibly less of an issue when rear facing as you get a bit more leeway, but if they were that loose then actually I'd expect the child to come out regardless of seat direction.
How would you expect stats on whether anyone has died due to sliding out of a car seat with a coat on to be recorded? We sometimes collect stats about seatbelt use and car seat use (tends to be self report on admission to hospitals) but not about whether the car seat was being used correctly as this is self reported and most people don't necessarily know that they are using their car seat incorrectly.
If a child is ejected or moves further than expected out of a car seat how could an observer state that this was to do with their clothing or the seat installation or merely crash forces or anything else? Their focus is most likely on securing the accident site, rescuing any survivors and preserving dignity. First responders are also usually emergency workers or laypeople - they aren't trained car seat technicians or experts. Also the car seat has potentially moved during the crash, so it's difficult to say whether it was installed incorrectly or whether the forces of the crash have caused it to uninstall. There's no forensic investigation of a crash site to see exactly how injuries were caused in terms of physics moving the people around in a car. That would be of limited use which is probably why it is not done. Also, there were 15 children killed in cars in the UK in 2018 - it's such a small number that each case would be very individual and potentially harmful to a family's privacy. If a particular family wished to share the story of their child's death in order to raise awareness, that would be their choice, but I would imagine this must be extremely painful and difficult to process for a lot of families and opening yourself up to all sorts of ridiculous comments on the internet would not be the way I would want to deal with things if I lost a child.
But this is why we have crash tests - so we can simulate different types of car seat misuse in a lab to see which are more problematic than others. We know that a large proportion of parents misuse car seats because of spot checks, which is why modern car seats have plenty of features designed to reduce misuse - isofix for one, indicators about correct tightness of various features, a couple of seats on the market even have self-tightening harnesses now.
One last thought - if you were going to do a bungee or parachute jump, would you be happy for the safety harness/parachute to be secured over a puffy, slippery coat? Or would you expect the organiser of the jump to insist on lightweight clothing only? That's about the kind of level of force you should be imagining when you think of a severe car accident. I've not done a bunjee/parachute myself, but I've done rope climbing where you wear a harness and they ask for outer clothing to come off so that the harness can get a snug fit.