Just a quick musing…not only is sample size an issue, but I wonder if maybe there isn’t some sample bias. For instance, an internet poll on the CNN website about a political issue is likely to draw response only from CNN viewers who are more likely to have strong political ideas and a general awareness of the intricacies of the issues at hand. Likewise, there may be a disproportionate number of people following this series of threads who have themselves been raped or know someone who has been, and that was why they took an active interest in this topic.
I realize you added a similiar caveat to your OP, I just wanted to reiterate and expand upon it. I see some value in taking such a poll, but only in the limited context of discovering roughly how often sexual assault has been part of the lives of the women on this board. That could of itself be useful, but not necessarily be even close to the figure in the general population.
BTW, I thought we had cleared up the fact that not all the people doubting the high rape statistics were being assholes, just some of them in light of other extrinsic insensitive behaviours they were exhibiting in the context of personal attacks. The skepticism of statistics was not the jerkish behaviour, it was the way said skepticism was presented and the subsequent act of giving irishgirl the 3rd degree.