Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Americans Greatly Overestimate U.S. Gay Population

news.gallup.com

Post reply on HN
Print view View original

41–50 of 109 posts

Re: Americans Greatly Overestimate U.S. Gay Population

#41

Post by tedsanders »

modeless wrote: post #15

These people were asked to choose between 6 options. Every number in every option was higher than the true value. Is it any wonder people overestimate when primed this way? I wonder how many other poll results people rely on are this obviously biased. Edit: I now believe this is incorrect, see below.

You've misinterpreted. The data is presented in six bins. But that doesn't imply the question was multiple choice. If it had been, then there's no way to get a mean of ~25% from the answers.

The question asked was: "Just your best guess, what percent of Americans today would you say are gay or lesbian?"

Re: Americans Greatly Overestimate U.S. Gay Population

#43

Post by DoreenMichele »

trackofalljades wrote: post #3

Is it possible that what you're seeing in this data is less about inaccuracy and more about how comfortably out people are? It says the estimates skew greater for younger people and women responding. Well are younger people and women more likely to know which of their friends are LGBT, because of a host of societal reasons that tend to influence men or older folks to wall off their feelings and communicate less about…

Well are younger people and women more likely to know which of their friends are LGBT

Studies about various social things plus first-hand experience lead me to say "probably."

Re: Americans Greatly Overestimate U.S. Gay Population

#44

Post by cloverich »

tastygreenapple wrote: post #26
tantalor wrote: post #21

Earlier quoted context omitted.

> LGBTQ in the American mind share far outsizes their population What does that even mean? Population is one thing, but can mindshare be measured? You're comparing apples and oranges.

Well, what other groups make up 5% of the population and get corporate logos changed in recognition of that group? About 5% of Americans watched the NBA finals, did brands go out of their way to make NBA fans feel included? Do news shows bring on their NBA correspondent to talk about NBA issues?

The reason for that is NBA fans were not specifically and generally excluded from various societal institutions. The level of public show is an effort not just to recognize inclusion but normalize it. Any attempted change in societal behavior takes concerted effort. For NBA fans (and numerous other small groups) -- there is no change needed.

Re: Americans Greatly Overestimate U.S. Gay Population

#45

Post by abathur »

drawnwren wrote: post #27
modeless wrote: post #15

These people were asked to choose between 6 options. Every number in every option was higher than the true value. Is it any wonder people overestimate when primed this way? I wonder how many other poll results people rely on are this obviously biased. Edit: I now believe this is incorrect, see below.

Eh, it's clear priming didn't have too much of an effect here. The single most selected answer in every year was >25%. Respondents thought the number was too low to begin with.

There are a lot of sources of sloppiness here, though.

The question is sloppy relative to the definitions. They asked "Just your best guess, what percent of Americans today would you say are gay or lesbian?", but the measuring stick is what portion of people identify as gay, lesbian, bisexual, or transgender. I would be a little surprised if people use exactly the same framework to decide if someone else is gay as they use to decide if they want to personally identify as gay.

There's no I-have-no-clue choice.

If you assume some fraction of all respondents are just guessing, and they know they're guessing based on a 100% scale, it would make some sense for the results to skew towards a choice given 75% of the scale.

Young people give estimates that are significantly higher, but a cursory search (not sure about the veracity) suggests that LGBT self-reports also sort very strongly by age (https://williamsinstitute.law.ucla.edu/visualization/lgbt-st...). As far as I can tell, the survey Gallup is using (on identification) doesn't provide an age breakout. I don't want to be too hyperbolic, but the change in the acceptability of identifying as LGBT over the last 70 years is pretty profound, and it seems at least plausible from the data that much of the difference in age groups (and, for that matter, partisan affiliation) could disappear if controlled for different identification rates among people's actual peer groups.

If so, the headline is more like, "Most Americans assume everyone in America is a bit more like their peers than they are."

This feels a little bit like asking one set of people "How many Americans have ever cheated on a test?" and asking another set "Have you ever cheated on a test?" and then publishing an article about how "Americans Greatly Overestimate How Many Americans Have Cheated on a Test"

Re: Americans Greatly Overestimate U.S. Gay Population

#46

Post by narrator »

I don't mind gays or lesbians. I do wonder where all this stuff comes from if it's such a small part of the population though. The transexual population is even smaller. How exactly does a bill become a law in the context of some cause is now the next civil rights crusade? It's not like these people have big enough numbers to get their interests on the agenda as a genuine democratic interest group that a politician would care about strongly.

Re: Americans Greatly Overestimate U.S. Gay Population

#47

Post by djdidf »

trackofalljades wrote: post #3

Is it possible that what you're seeing in this data is less about inaccuracy and more about how comfortably out people are? It says the estimates skew greater for younger people and women responding. Well are younger people and women more likely to know which of their friends are LGBT, because of a host of societal reasons that tend to influence men or older folks to wall off their feelings and communicate less about…

If your an urbanite you have a huge observational bias simply if you accept that rural homosexuals are more likely to migrate to the city.

Reason for the migration include less discrimination, but also higher concentration (making hooking up easier).

Therefore your perceived rate of homosexuality would be grater.

For what it’s worth the self reported rate of 5 percent is the same when I first cared to look it up, about 15 years ago. I would imagine that, if the self reported rate had significant error with it, the decreasing stigma would have changed it

Re: Americans Greatly Overestimate U.S. Gay Population

#48

Post by DoreenMichele »

The LGBTQ community seems to be especially visible online. If you spend a lot of time online, it's possible to have your perception skewed for that reason.

You also have to consider that different people define these things differently. Some people will self identify as bisexual based solely on feeling attracted to other people, having no actual experience. Meanwhile, there are people who feel that you have to actually have (substantial) sexual experience with both genders to qualify for the label.

I've seen advocates say things like "You don't have to give your sexual history to prove you are bi." and I've seen people talk about feeling attracted to members of their own sex and realizing they are actually bi, but don't yet have experience with members of their own sex because of taboos.

So you can potentially have two different people look at the exact same data and draw different conclusions concerning who qualifies for what label.

Different cultures also have different baseline concepts concerning sexuality. Some cultural differences will not be readily captured by the questions used in a survey, especially if they don't take the time to clearly delineate their terms.

Re: Americans Greatly Overestimate U.S. Gay Population

#49

Post by modeless »

tedsanders wrote: post #41
modeless wrote: post #15

These people were asked to choose between 6 options. Every number in every option was higher than the true value. Is it any wonder people overestimate when primed this way? I wonder how many other poll results people rely on are this obviously biased. Edit: I now believe this is incorrect, see below.

You've misinterpreted. The data is presented in six bins. But that doesn't imply the question was multiple choice. If it had been, then there's no way to get a mean of ~25% from the answers. The question asked was: "Just your best guess, what percent of Americans today would you say are gay or lesbian?"

Thanks, I think you're right. But wow what a misleading presentation. They really ought to state the answer format explicitly rather than make you infer it.

Re: Americans Greatly Overestimate U.S. Gay Population

#50

Post by modeless »

tomcatfish wrote: post #38
modeless wrote: post #15

These people were asked to choose between 6 options. Every number in every option was higher than the true value. Is it any wonder people overestimate when primed this way? I wonder how many other poll results people rely on are this obviously biased. Edit: I now believe this is incorrect, see below.

It appears that "Less than 5%" was an option. Where are the six you see?

5 is larger than 4.5, agreeing with my statement that all the numbers are bigger.
Post reply on HN