Yes definitions are fluid, and yes self-reporting is questionable, but I have no clue why the answer to the question "what percentage of US adults do you guess are gay" is so often so very high. I wonder if the inferred definition is different from the standard - maybe people took it as "how many are open to some same-sex activity" (although that may be a bit low then imo). Since no survey variant of "are you gay" has ever gotten near this level (ro my knowledge - welcome corrections) I wonder why the disconnect. Wishful thinking? Irrational fear magnifying "enemies"? (but then why liberal guesses so much higher than con)? Could underreporting be THAT bad and this is closer to reality? Personally I'm not driven to either maximize or minimize any percentage beyond the silly hyper-low estimates of wishful-thinking fundies, but interested in where the perception of percentages comes from. Anybody more savvy on the history and context of this poll?
http://www.gallup.com/poll/147824/Adults-Estimate-Americans-Gay-Lesbian.aspx