I would be interested in seeing the results of an investigation into a similar study that also factors in the importance of what is *behind* the password.
I don't think I'm the only one who puts more effort into choosing a 'good' password for things that are of value. I choose really quite poor passwords for things I really don't care about - eg have no sensitive information behind the login. For things like cash point cards, and other things in front of my actual money, I attempt to use much better passwords.
I think there are many things of little or no value, while just a few of high value. I guess this might skew the numbers somewhat. It's probably quite difficult to factor in this aspect, but it makes me question the conclusions.