New Gallup survey here does it accord with what you expected?
Interestingly gay people apparently do not love North Dakota.
Ah, found it:
Given like that, the differences might come more from sampling problems than from real differences. Do I think there are no differences? No, I’m sure there’s very real differences. But those numbers and their ranking are very, very fuzzy.
How come South Dakota is that high? Especially when it’s compared to neighbouring state like North Dakota, which is at the very bottom. It’s way up there with Massachusetts and California.
Given those error margins (which they don’t detail), the difference of 2.7 points between the two Dakotas is likely to be smaller than the addition of their error margins. I’d need the detailed list of estimated error margins to tell for sure, but they’re also both among the states with less interviews (and thus with a greater error margin as per the article).
I think one problem with this poll is the method of sampling.
To a certain extent, there are still the proverbial “Gay Ghettos” of certain states and cities where you would find openly Gay and Lesbian percentages in certain communities to be easily in the very high double digits - West Hollywood and San Francisco for instance, or Manhattan or Miami etc.
So, sampling the entire state would be like NOT shooting fish in a barrel.
For instance, if you asked how many people of color there are in Illinois, asking the same number of people down state (mostly white, farming communities) as people in Cook County, you would get a far, far smaller percentage of minorities that might not be the true picture.
As a wild example - ask 200 people in Chicago and say you get 48%, but ask 200 people in southern Illinois and get 3%. If Cook County has the vast majority of the populace of Illinois, treating the other part of southern Illinois as “equal” would skew the numbers.
I hope I phrased that correctly.
Connecticut is high mid-pack, about what I’d expect. Then again, it may be low. I have a fair amount of exposure to gay communities and community members for a lifetime straight, and I have never seen so many casually openly gay couples, predominately older ones, anywhere else. This in a relatively rural area, to boot.
I do notice that the insistence of some groups that the number is 10% or higher, nationwide, hasn’t been supported by a valid poll in a long time. (Nothing anti implied or meant, but in the long run reality has to be accommodated in social policy.)
I don’t understand what you’re saying. If you picked 1000 names randomly from an “Illinois phone book”, it would give you a good sample (within whatever margin of error your sample size produced) of the entire state’s population. You would not undersample any specific geographic area by doing so.
From looking at the poll parameters, I think Gallup has done a pretty good job with this poll. Certainly imperfect from some angles, but large enough and, I would assume, broadly enough dispersed, to be meaningful.
Yes, the result for my state surprises me (Montana – 2nd from the bottom of the list). I know quite a few openly gay people in my little town of 2,300 and we just recently had a very public same-sex wedding here.
Part of it could well be margin of error, but I think the bigger reason is that there are a lot of small towns here where people don’t admit to being gay. That doesn’t necessarily mean fewer people are gay – it just means they don’t want to talk about it.
I give them props, too, for reporting the number of interviews for each state.
Looks about right, although I’m questioning the DC result. Maybe it’s just because it’s all urban, but would you even get that high a percentage in SF?
I suspect many of the states on the low end are under-reported due to closetedness.
It’s not surprising Hawaii is so high on the list. They have all those Chinese people.
Let’s try this one more time.
Here is a map of Cubans living in the USA.
Let’s say you were clueless and wanted to know how many Cubans lived in the USA.
As you can see, and probably guess, random sampling in Montana is going to get about zip.
As you can see, and probably guess, random sampling in Florida is going to be large.
As you can see, and probably guess, random sampling in the Miami area is going to be HUGE!
In the grand scheme of things, there really are not that many (in the most recent census in 2000 there were 1,241,685 Cuban Americans living in the USA).
However - asking random people if they are Cuban (or of Cuban heritage) across the entire USA is going to make the numbers really small. And randomly questioning people in the other 49 states, other than Florida, is going to even greatly lower those numbers.
And if you were really clueless, and simply did not ask anyone or very few people in the Miami area, those number would REALLY be off by a huge margin!
My point is, random sampling nationwide/state by state is great for groups of people who do not tend to necessarily group together in separate communities: Christians; men over 6 feet tall; blind people; fat people; people who are left handed. These groups will indeed be found randomly, in every community.
But trying to do a random sampling nationwide on groups who do somewhat tend to flock together in specific geographical areas: Gays; Cubans; Hasidic Jews; professional surfers; snowboarders; farmers; cotton pickers? Those are groups of people you have to find in more specific locations than just randomly asking to find any of them in Tuscon, Arizona.
Short of actually counting numbers on census polls, I just think random sampling of very specific groups has limits to value - especially if geographical influence is not taken into consideration.
I’m curious how the “phone book” part of that affect results.
Phone books don’t contain cellphone numbers. In my experience, the majority of people under 30 years of age no longer have land lines, which means they don’t get called for the poll. Are younger people more likely to publicly declare themselves to be gay than older people? That would certainly influence results. Would that mean that results would differ in areas with poor cellphone service?
In case anyone else is as curious as I was, but wants to be saved the trouble of looking it up, Wikipedia gives 5,217,080 (2011 estimate) as the population of Cook County, and 12,875,255 (2012 est) as the population of Illinois, which would mean Cook County has a little over 40% of the state’s population.
But apparently only 93.7% of people are not gay. High for the survey, but not the top. Lots of North Dakotans are “none-of-the-above”.
My guess is lots of fed govt jobs with no-discrimination clauses. DC also has a high concentration of libraries, archives and museums, and there’s an unusually high percentage of gay folks in those vocations.
I’m surprised at the Maryland percentage and would be interested in knowing the geographic breakdown of the respondents for that state. I know the percent won’t be as high as for DC, but I’d expect it to be at least around the 3.5% mark.
The DC percentage sounds right to me.
Any meaningful poll will include the relevant parameters. That the media, and most readers, swallow whole polls with crap metholodogy is frustrating.
Two of the key sentences in that article are:
and
In other words, the margin of error is almost equal to the difference between the numbers in the various states.