You should also point out the sample distribution as well. I have known about 20-30 perl programmers that I have worked with and none of them have ever done testing. Nor have any of them ever visited perlmonks.org or use.perl.org or any other organized mailing list. CPAN is considered some type of spam site(maybe I am exaggerating a little).
The point is, that the people who would see and take the Poll as people who are more active in the community therefore more likely to have better habits than someone who 'already knows' how to do programing and therefore does not consult the web for advice. This distribution will royally screw up any potential estimate that you are trying to get from your Poll.