It doesn't say what you think it says

I hate crap like this:

Study Shows Women More Likely to Cause Traffic Accidents

It's not just the headline, by the way, it's the article itself:

A University of Michigan study of 6.5 million car crashes will undoubtedly be the source of many tense discussions around the kitchen table if not Vegas comedy riffs, finding that an inordinate number of accidents happen when both drivers are women.

The problem?  This:

Using the General Estimate System data from a nationally representative sample of police-reported crashes, the researchers expected to find that male-to-male crashes would account for 36.2% of accidents, female-to-female would make up 15.8% and male-to-female would make up 48% of crashes.

Instead, they found female-to-female accidents made up 20.5% of all crashes, much higher than expected. Male-to-male crashes were lower than expected, at 31.9%, and male-to-female crashes were 47.6%.

Now, perhaps you can excuse the use of "inordinate."  But claiming that it's any sort of evidence women are worse drivers than men is stretching it.  It would be at least as accurate to say that "Men don't cause quite as many wrecks as we thought they did." 

Male-to-male crashes are still a much larger category than female-to-female, after all, and there's no mention made of how many of the male-to-female crashes are caused by the fairer sex.

Then again, if it was worded in a less controversial fashion, I wouldn't be writing about it, so there is that.