Looking at the data it’s pretty apparent that none of it is statistically significant. It’s all useless.
For example, in death rates, the CX-9 2wd is a top scorer with 0 deaths but the CX-9 4wd is middle of the pack with twenty deaths. The 4wd version of the CX-9 isn’t a deathtrap nor the 2wd version super reinforced. The sample size is just too small. The 2wd versus 4wd explorer have the same unexplained discrepancy.
I’d assume it’s because the 4wd models are bought by people living in places with snow/ice, therefore more likely to have a weather related crash.
I think that unlikely since for the explorer the 2wd and 4wd splits are reversed. And it makes sense since we aren’t talking very many deaths. One or two bad accidents could swing numbers.
And actually winter weather seems to have little relationship to road deaths. Massachusetts and Minnesota appear to have the lowest death rate per mile driven while the Deep South is the worst. Which makes sense, you drive slower in denser areas (Northeast/big Midwestern cities) and help is closer by and often of higher quality when things go wrong. Also I’d guess all other things equal the more affluent area will have fewer deaths since people drive newer cars which are on the whole safer