Gallup's Gates and Newport summarized the findings.
These results are based on responses to the question, "Do you, personally, identify as lesbian, gay, bisexual, or transgender?" included in 206,186 Gallup Daily tracking interviews conducted between June 1 and Dec. 30, 2012. This is the largest single study of the distribution of the LGBT population in the U.S. on record, and the first time a study has had large enough sample sizes to provide estimates of the LGBT population by state.
As was outlined in the first report of these data in October, measuring sexual orientation and gender identity can be challenging because these concepts involve complex social and cultural patterns. There are a number of ways to measure lesbian, gay, and bisexual orientation, and transgender status. Gallup chose a broad measure of personal identification as LGBT because this grouping of four statuses is commonly used in current American discourse, and as a result has important cultural and political significance. One limitation of this approach is that it is not possible to separately consider differences among lesbians, gay men, bisexuals, or transgender individuals. A second limitation is that this approach measures broad self-identity, and does not measure sexual or other behavior, either past or present.
The number of interviews conducted in each state between June and December is large enough to allow for reasonable estimates of each state's LGBT population. Only eight states had less than 1,000 completed interviews, including the lowest sample size of 613 in Alaska. Gallup also asked the LGBT question of 493 District of Columbia residents. The number of completed interviews conducted in each state is presented in the accompanying table.
The margin of error for each state's estimate varies, depending on the state's sample size. Except for the District of Columbia, all are below ±2 percentage points. A caution comes in interpreting rankings of states with relatively small populations. Although, as noted, the margins of error in general are quite small with all of these estimates, differences in the rankings of the states with the smallest numbers of interviews are more prone to being byproducts of sampling error.
Overall, the results from this analysis of LGBT identity by state may run counter to some stereotypes that portray the LGBT community as heavily grouped in certain states of the union. With the exception of the District of Columbia, the range in percentage LGBT is 3.4 percentage points, from 1.7% in North Dakota to 5.1% in Hawaii.
They note that, broadly speaking, proportions of non-heterosexuals seem to be higher in gay-friendly states than not. This may suggest that the national proportion is actually higher than 3.5%, as one would expect if people in gay-unfriendly states felt a need to be closeted, or this may alternatively indicate the results of migration to gay-friendly states.
I'll note in passing (as I did at the time) a July 2012 poll by Forum Research done for the National Post suggesting that 5% of Canadians identified as GLBT. If Canada is overall more tolerant of GLBT people and identity than the United States, that 5% might indicate an upper level. (Or, it might not, as 10% of under-34s in that poll identified as GLBT.)