The numbers aren't that important (though I do think you're underestimating). But surely your point that Kinsey's research is 50 years old only supports the argument that the real figure could be higher. 50 years ago people were far less likely to admit to either being gay or having a gay experience than they are now.