The more general form of this poll ("will you play a male or female character?") has come up several times. Each time, with hundreds of responses, the general percentage has been about the same - ~81% male, ~19% female.
Interesting. And given that there isn't much bias, it's probably one of the more scientifically accurate poll statistics. Potential biases include:
1. The poll respondents are all self-selected - In this case, there's no obvious reason why one type of respondent would be more likely to self-select than another.
2. All of the poll respondents are reading the gamesas forums - No obvious reason why gamesas forum readers would respond differently than non-readers though.
Can't think of any others at the moment.