In my previous post I introduced the topic of same-sex marriage as a civil right, only to digress into a discussion of how many gay, lesbian, bisexual, and transgendered (LGBT) people might actually live in Howard County. Now I’ll return to the question of same-sex marriage, starting with an discussion of how many people in Howard County might be candidates for it (or even already married).
As I noted previously, the US Census Bureau does not create or publish direct estimates of the LGBT population. However since 1990 the Census Bureau has surveyed the population to determine the number of unmarried partners (in addition to the data that’s always been collected on married couples), and as part of that survey has collected data on the sex of each partner. This data has in turn been used by others to create estimates of the number of same-sex couples as well as the overall LGBT population both nationally and at a state level.1
Unfortunately these estimates have been fraught with problems (beyond just the garden-variety statistical problems having to do with limited sample sizes). In particular the Census estimates of same-sex unmarried partners were historically too high2, and more recently the Census survey methodology hasn’t properly accounted for same-sex couples who were in fact married (i.e., in Massachusetts or elsewhere). The first problem was not properly corrected for until 2008.3 The second problem is being addressed in the 2010 census, when for the first time the Census Bureau will generate an official estimate of the number of same-sex couples who consider themselves to be spouses (e.g., married or in a civil union or domestic partnership arrangement).4
Thus only when the 2010 census results are released will it be possible to get a reasonably clear picture of how many same-sex couples (married or unmarried) exist in the US, Maryland, and Howard County. In the meantime the best public Census data is from the 2008 American Community Survey, which estimates that 0.5% of all US households consist of same-sex unmarried partners. Because the ACS estimates are based on only a small sample of the population5 (about 1% of all households), and because the number of same-sex couples is so small (less than 1% of all households in almost all jurisdictions), using ACS data below the national level is problematic; in many cases the margins of error on the estimates are comparable to the estimates themselves.
However we can get at least a general sense of the relative numbers of same-sex couples in various local jurisdictions by looking at the ACS estimates. As you might expect, the District of Columbia has lots of same-sex couples (about 1.4% of all households per the ACS). Montgomery County has about half that fraction (about 0.6% of all households), Howard County about a third (0.4%), and Frederick County about a quarter (0.3%). This is reminiscent of the figures on comparative ethnic diversity in Montgomery, Howard, and Frederick that I discussed in an earlier post.
Note however that unlike the situation with ethnic diversity, the proportion of same-sex couples in Howard County is actually below the overall national average. It’s not much below the national average, which makes me think that the total LGBT population of Howard County is closer to the 3% figure than to the sub-1% figure of my range of estimates in the last post. But it does put paid to the idea that Howard County is a bastion of diversity in this particular context.
Finally, note that the above Maryland counties do have higher proportions of same-sex couples than comparable counties in Virginia: Montgomery County has relatively more same-sex couples than Fairfax County (0.6% vs. 0.4%), and Howard County relatively more than Loudoun County (0.4% vs. 0.2%). At first glance this seems to be consistent with the conventional
blue state vs. red state narrative: As a supposedly liberal Democratic-leaning state Maryland is presumably a more hospitable place for same-sex couples than a conservative Republican-leaning state like Virginia (where one of the Ten Commandments seems to be
But wait a minute: Both Maryland and Virginia have about the same number of same-sex couples, at about 0.4% of all households (slightly under the national average). In addition, Arlington County has even more same-sex couples than DC, at 1.5% of all households, and thus seems to be more like the part of DC it once was than part of Virginia.
So what’s going on here? Why isn’t Maryland attracting more same-sex couples than it is? Why can’t Howard County, that supposed haven of diversity and tolerance, attract relatively more same-sex couples than Maryland as a whole, or the nation as a whole? And why hasn’t this supposedly ultra-liberal state enacted a same-sex marriage law, or even a law to allow civil unions? More on that in the next and final post in this series.
1. Since there are no good direct estimates of the LGBT population at the state level, the estimate of the national LGBT population (as discussed in the previous post) is apportioned into each state based on the percentage of all same-sex couples who reside in that state.
For example, since nearly 15 percent of same-sex couples live in California, the estimated size of the GLB population in California is approximately 1.3 million (15 percent of 8.8 million GLB people in the U.S.). See page 4 of Same-sex Couples and the Gay, Lesbian, Bisexual Population: New Estimates from the American Community Survey.↩
2. In many cases mixed-sex married couples responding to the surveys incorrectly entered either their own sex or the sex of their partner, causing them to be incorrectly reclassified as same-sex unmarried partners, under the assumption that a same-sex couple couldn’t possibly be married. (Or, more charitably, they couldn’t be considered married from a Federal perspective, as a result of the Defense of Marriage Act.) This artificially inflated the estimates for same-sex unmarried partners.↩
3. For a very inside baseball look at this issue, see the Census Bureau paper Changes to the American Community Survey between 2007 and 2008 and their Potential Effect on the Estimates of Same-Sex Couple Households.↩
5. See the congressional testimony by Linda Jacobsen of the independent Population Reference Bureau for an overview of problems with the ACS. Note that the ACS replaced the census long form, but has a sample size considerably smaller than the long form, even when combining data from multiple years.↩