Research excites me. And I’m not just talking about interesting and thought-provoking findings that get reported as sound bites in the media. At the risk of outing myself as a total geek, I have to say that the actual research process is just fascinating to me. Designing research studies, crunching numbers, analyzing statistical findings – I love that stuff. And this week, the U.S. Census Bureau, in its report titled “Households and Families: 2010,” offered up the perfect opportunity to play with methods and statistics, particularly as they pertain to the LGBTQ community. This report includes, among other things, information about same-sex couples in the United States. Unfortunately, the procedures used by the U.S. Census Bureau to determine the prevalence of same-sex couples were, in my opinion, pretty weak.
To delve into this in some more detail, I’m reprinting, in its entirety, a brief sidebar included in the report titled “UNMARRIED PARTNER HOUSEHOLDS.” I know this will probably bore you to tears, but it highlights the serious methodological errors the U.S. Census Bureau committed in collecting data on the LGBTQ community. To make this a little easier to digest, I’ll break down each section of this sidebar and offer up some commentary. Here goes:
UNMARRIED PARTNER HOUSEHOLDS
An “unmarried partner household” consists of a householder and a person living in the household who reports that he or she is (1) an unmarried partner of the householder and of the opposite sex; (2) an unmarried partner of the householder and of the same sex; or (3) a spouse of the householder and of the same sex. Procedures for the 2010 Census edited same-sex spouse households as unmarried partner households, and these households appear as such in published Summary File 1 tabulations.
Obviously, checking the “unmarried partner” box doesn’t necessarily mean you’re gay or lesbian, because many straight couples are unmarried. To determine how many people who checked the “unmarried partner” box are part of a same-sex couple, the U.S. Census Bureau went back and determined how they answered the gender item. If both partners checked “male,” or if both partners checked “female,” they were assumed to be part of a same-sex couple. This is cumbersome, though – why not just ask if they’re part of a same-sex couple? Or, to get even more basic, why not include a question that asks about sexual orientation? Not asking about sexual orientation (1) renders single gay and lesbian people totally invisible, and (2) renders bisexuality totally invisible, since many bisexuals are in different-sex marriages.
Problem #1: There are no questions in the U.S. Census that specifically ask about one’s sexual orientation.
It’s important to note that same-sex couples don’t have the option to check “married” on the U.S. Census. In fact, there is no “married” box – the option that exists for heterosexual married couples is “husband and wife.” This isn’t surprising, given that the federal government defines marriage as between a man and a woman. However, the U.S. Census Bureau assumes that all same-sex couples would know to check the “unmarried partner” box, even though that might not be the most accurate descriptor. But who knows – some people in same-sex relationships might check “husband and wife,” or “other relative,” or “housemate or roommate.” So that bring us to . . .
Problem #2: The options available in the U.S. Census do not accurately reflect how same-sex couples define their relationships.
Now let’s bring gender identity into the mix:
During the review of the data, counts of same-sex spouses appeared inflated due to mismarking errors in the gender item on the census forms. Up to 28 percent of the total number of same-sex unmarried partner households may actually be opposite-sex households: 62 percent of reported same-sex spouses were probably marked in error compared with 7 percent of reported same-sex unmarried partners. This report presents data both for same-sex households as shown in Summary File 1 tabulations and for a set of “preferred estimates” that attempts to remove statistically same-sex households that are likely opposite-sex households.
“Mismarking errors”? Essentially, they’re saying that a LOT of people who reported that they are a same-sex “husband and wife” probably marked the gender question incorrectly. It’s entirely possible that some of the reported same-sex spouses weren’t marked in error – even if that couple lives in a state that doesn’t recognize same-sex marriage, they might have chosen “husband and wife” because they consider themselves to be married, and since there wasn’t a “married” box, that left – you guessed it – “husband and wife.” This points back to Problem #2. However, the Census Bureau is correct in that the gender item probaby did yield some errors – but not necessarily because people cluelessly marked the wrong gender. Picture this scenario: If you don’t fit the standard identity of “male” or “female,” and there’s no opportunity to identify yourself as “transgender” or some other non-standard gender identity, then which box do you check? And, depending on the box you check, is your relationship status going to be recorded accurately?
Problem #3: There are no questions in the U.S. Census that specifically ask about gender identity, gender expression, or transgender status.
I realize that the inclusion of “unmarried partner” was a big step for the U.S. Census Bureau, and that this option yielded data that will greatly benefit the lesbian and gay community. However, the 2010 U.S. Census data is woefully inadequate in bringing the LGBTQ community into visible focus. This is a big deal, given that the U.S. Census is probably the largest scale and most ambitious data collection effort in the world, and that the data are used in a wide range of policies, initiatives, and funding decisions. And what’s really sad to me is that the methodological errors in the U.S. Census probably weren’t just dumb mistakes – in fact, I bet every word was very carefully chosen. If anything, we’re seeing the fallout from the political tug-of-war that our country is experiencing regarding LGBTQ issues.
The “Households and Families: 2010” report can be found here: http://www.census.gov/prod/cen2010/briefs/c2010br-14.pdf