Another way to read the report, to be released Tuesday, is that E-Class drivers tend to drive more carefully than Blazer drivers.

I have always felt strongly that this was the case with these kinds of studies. I am glad to see it getting some press. There is a strong self-selection bias at work in any use of real-world data on vehicle safety.

Of course, from the perspective of the insurer, in the aggregate, the difference is immaterial, but on an individual level, it can be unfair to some drivers.