Williams Institute study analyzes characteristics of same-sex couples raising children

Williams Institute demographer Gary Gates begins his new article in National Council of Family Relations by indicating that the gay parents in the hilariously funny Modern Family (okay- the hilarious part is my editorializing, not Gary's analysis) are decidedly not the typical same-sex couple raising children.

The most important conclusion from Gates's review of census data and several other large scale surveys is that large numbers of children of same-sex couples almost certainly are the product of previous heterosexual relationships.  For example, 28% of those who were previously married have children in their home, while the figure is 16% for those who were never married.  When looking at who has a biological or step-child,  23.5% of those who were previously married are in that category;  for those who were never married the figure is 9.5%.  Conversely, couples who do not report a previous marriage are twice as likely as those who do to have an adopted child.

There is other, fascinating, evidence supporting the likelihood that most lesbians raising children have a child from a previous heterosexual relationship.  In the 2009 California Health Survey, which asks respondents to identify their sexual orientation (unlike census data, which can only report numbers of same-sex couples raising children, thus exclusing gay men any lesbians raising children without living with a partner ), 22.4% of heterosexual women reported having a child before age 20, while 37.9% of lesbian and bisexual women reported having a child before age 20.  (Does denial about one's sexual orientation lead to riskier behavior? less likelihood of using birth control?  The data doesn't give us the "why," only room to speculate...)

In this article, Gates repeats information he has provided elsewhere, for example that the greatest percentage of same-sex couples raising children is in the south.  Also, couples with less than a high school education are almost three times as likely to be raising children as couples with a graduate degree.  (This discrepancy does not exist for heterosexual couples).  Furthermore, African-Americans in same-sex couples are 2.4 times more likely than their White counterparts to be raising children.  On the other hand, looking at adopted children only, White same-sex couples are almost twice as likely to have an adopted child when compared with couples where at least one partner is not White, and the couples with adopted children are more likely to have completed higher education.

Nineteen percent of same-sex couples with children have an adopted child, almost double the percentage in 2000.  Yet the percentage of all same-sex couples raising children has decreased.  It looks like lesbians and gay men are less likely to have children in heterosexual relationships now -- hence the more recent decline, perhaps because they are coming out earlier -- and that for all the attention to the "gayby boom," the actual number of children deliberately born or adopted into gay or lesbian families cannot make up the shortfall.

Couples raising one partner's biological child from a prior relationship have legal concerns that can be different from those of couples raising children planned for by the couple together.  For example, if there is another biological parent in the picture at all, that parent would have to consent to a second-parent adoption (where that's possible -- which it isn't in many of the southern states with concentrations of such couples).  And we shouldn't forget that in some parts of the country a heterosexual parent or relative can still challenge a lesbian or gay parent for custody of a child, a circumstance that isn't going away any time soon.

Once again, Gary Gates's data collection and analysis makes a huge contribution to our community and gives us lots to think about.

Let the maps begin! Williams Institute releases first data from Census 2010

Long before I met the incomparable Gary Gates, I admired his work. While at the Urban Institute, Gates co-authored The Gay and Lesbian Atlas, a book with color-coded maps, by state and county, of the numbers of same-sex couples in the entire United States. It is the book that proved the old adage that "we are everywhere." Literally. The data in Atlas came from the 2000 Census. It was quite a revelation that we could learn so much from that source.

Well, Gary Gates is now at the (also incomparable) Williams Institute at UCLA, and today he released the first state maps (again color-coded) with data from Census 2010. Turns out there are 11,259 same-sex couples in Alabama, 27% of whom are raising children. There are 4,248 same-sex couples in Hawaii, 23% of whom are raising children.

Gates has also prepared an explanation of his methodology. The FAQ's are here, with a link to a longer and more technical brief. Next week, Williams will release reports on California, Delaware, Pennsylvania, and Wyoming. Census data by state will be released weekly over the course of the summer.

I've had the pleasure of seeing the glee on Gary Gates's face as data about LGBT people come across his computer screen. How lucky for the rest of us that he can tell us so much about ourselves. Yes, it is very limited. Most especially by the fact that the census counts only same-sex couples, and only couples who live together. There are some data sets that ask directly about sexual orientation, but not the census. Still, no other data set can give us maps like this.

Lucky for me that I will be at UCLA beginning in the middle of July, as the Visiting McDonald/Wright Chair in Law and Faculty Chair of the Williams Institute. That means many chances to see Gary's glee in person!

How same-sex couples fill out the 2010 census form -- is it really asking how we think of our relationships?

If thinking of yourself as married to your partner turned you into husband and husband or wife and wife, then we wouldn't need a marriage equality movement, right? So imagine my puzzlement to find gay organizations instructing us to fill out the 2010 census based on how we think of our relationships. The Williams Institute, to whom I turn for all things demographic about gay men and lesbians, offers this advice, which other groups are repeating:

Same-sex couples who have been legally married or consider themselves to be spouses should identify one person as a "husband or wife".

Other same-sex couples may be more comfortable using the term "unmarried partner". In general, this designation is designed to capture couples who are in a "close personal relationship" and are not legally married or do not think of themselves as spouses. (emphasis mine).

Now I understand the census is an imperfect instrument (very) for counting our relationships. If a couple does not live together, they will not be counted, because the census counts households and the relationships of the people in each household. There is also no option for those who are registered as domestic partners or in civil unions. I applauded when the Williams Institute and others won from the Obama administration the right to be counted as same-sex spouses when they were same-sex spouses.

But now it appears that labeling the person you live with your husband or wife is actually not going to measure the number of same-sex married couples but rather the number of couples who consider themselves spouses, whatever that means -- and I truly do not know what it means.

Gary Gates, demographer par excellence at Williams, explained to me that the census does not ask marital status. In other words, it does not ask you to say whether you are single, married, divorced, etc. He's right. But it does seem to me that asking us to choose "husband/wife" or "unmarried partner" actually is asking us to say if we are married. Admittedly, whether we are married can be contingent. Those couples who have married in a state or country that allows same-sex couples to marry are married in some places. I think they should mark the "husband/wife" box. Those who have not married are "unmarried partners." To me that is not a lesser status; it's just a different one.

The Williams Institute materials are clear that you can identify only one adult in your home as a "husband/wife." What about those who consider themselves married to more than one person? If the line isn't legal recognition, what is the limiting principle?

How about those who have entered a civil union or the kind of domestic partnership that confers virtually all the state-based consequences of marriage? This is a challenge. My partner and I have been registered domestic partners for many years. When I look at the census form I gravitate to "unmarried partner" because we are not married (and we don't plan to marry even though DC now allows it). Yet I admit that some couples who enter this status because it's available where they live may well consider each other husband/wife, and since there is no way to accurately capture their legal relationship then I'm okay with selecting whichever designation fits their own understanding.

But here's another puzzle in the advice from Williams. What does it mean to say that we "do not think of ourselves as spouses"? Either "spouse" has a meaning and you either are or are not, or, well, it has no meaning at all. If my partner and I were to marry I am not sure I would think of her as "my spouse" if that means some traditional notion of marriage. I know I would never call her my "wife." But if we marry, am I not supposed to check the "wife" box for her regardless of how we think of ourselves?

I've tried to think of this from a straight person's point of view. What do an engaged couple living together mark? "Fiance" is not an option, and they may never have thought of each other as "unmarried partners," but they know they are not yet husband/wife. What do they check? Or...how about the couple who think they are "common law" married but they aren't, because their state does not recognize common law marriage (only 10 and the District of Columbia do)? They will check "husband/wife" and it won't be accurate.

Gary Gates tells me that the Census Bureau wants all people who are not sure what to check to select the answer that best reflects their household as they understand it. I could not find that advice anywhere on the Census2010 website. But I did call the census "help line" and said I was in a same-sex registered domestic partnership and did not know which box to check. The person I spoke with said it was my "preference," and if I saw her as a "married partner" I should check "husband/wife" and if I saw her as a "unmarried partner" I should check that.

Gates also says, and I suspect he is right on this, that no amount of education by gay organizations would yield an accurate count of legally married same-sex couples given the constraints of the form itself. So what will gay groups say the census has shown once it's tabulated? Will they qualify the number of claimed "married couples" with the caveat that it is couples who think of themselves as married? I'm guessing there will be comparison of the geographical location, income, etc of those who identify as same-sex unmarried partners and those who identify as same-sex husbands/wives, rather than simply an adding together of the two catgories to tell us about same-sex couples in general. But the categories are unstable and I have trouble imagining what legitimate conclusions could be drawn from the raw data.

And here's another tantalizing nugget from Gates. Apparently the American Community Survey forms (they replaced what were once census "long" forms) ask both marital status and the relationship of the people in the household, and more same-sex couples check "husband/wife" than report being married. He's trying to sort out what that means. Fascinating, isn't it? He's going to have lots more sorting to do over the next several years.