On the Abuse of Social Science


One of the reoccurring talking points of opponents of marriage equality is that social science research demonstrates that children do better with a heterosexual couple than any other type of parenting situation. For an example, see this exchange between CNN’s Brooke Baldwin and hate-group leader Tony Perkins. Note how she fails to challenge Perkins on his oft-repeated assertion that the social science literature backs up his position? Yeah, me too. Unfortunately, this is not just a problem with Baldwin—most journalists don’t challenge him or any other bigots on this point.

As GLAAD’s Aaron McQuade points out, one problem with Perkins’ claim is that he is referring to studies comparing two-parent heterosexual to single-parent homes, which says nothing of children raised by two homosexual parents. And the research that does compare children raised by two straight parents and two gay or lesbian parents demonstrates that there is no discernable difference in how well adjusted the children are.

The 30+ years of literature that has built consensus on this issue was called into question this week. Mark Regnerus, an Associate Professor of Sociology at UT-Austin, published an article on Slate this week discussing the findings of his New Family Structures Study (NFSS). There have been quite a few criticisms and take-downs of this study (see here, here, here, and here), but there are some things I would like to address.

First, the money trail is extremely troublesome to me. As pointed out in some of the articles linked above, funding to the tune of approximately $800,000 was provided for the NFSS from the Witherspoon Institute and the Bradley Foundation, both right-wing organizations with links to the Family Research Council and the National Organization for Marriage. Regnerus states in the data collection section of the academic paper that this funding played “no role at all in the design or conduct of the study, the analyses, the interpretations of the data, or in the preparation of this manuscript.”

That Regnerus’s previous research is lauded by right-wing bigots for its pro-Christian message should not be lost on us. I simply do not think that right-wing organizations are going to throw that much money at research without some level of confidence that the results produced will be what they want to hear. Of course, I have not seen any proof that there’s any wrongdoing on Regnerus’s part, and I’m not making accusations of wrongdoing. I’m just awfully suspicious.

Another part of Regnerus’s article on Slate that bothered the hell out of me was this little gem:

The rapid pace at which the overall academic discourse surrounding gay and lesbian parents’ comparative competence has swung—from the wide acknowledgement of challenges to “no differences” to more capable than mom and pop families—is notable, and frankly a bit suspect. Scientific truths are seldom reversed in a decade. By comparison, studies of adoption—a common method by which many same-sex couples (but even more heterosexual ones) become parents—have repeatedly and consistently revealed important and wide-ranging differences, on average, between adopted children and biological ones. The differences have been so pervasive and consistent that adoption experts now emphasize that “acknowledgement of difference” is critical for both parents and clinicians when working with adopted children and teens. This ought to give social scientists studying gay-parenting outcomes pause—rather than lockstep unanimity. After all, many children of gay and lesbian couples are adopted.

Oy vey! First of all, the academic discourse surrounding gays and lesbians over the last decade is a result of increased social visibility (as evidenced by things like increases in the amount of people who say they know a gay or lesbian person and how Americans hugely over-estimate how much of the population is LGBT). Increased visibility of gay and lesbian families means an increase in academic research that looks at gay and lesbian families. This is not rocket science (it’s social science! teehee). I’d be shocked by the assertion that there’s a gay agenda behind this research (the shift is “a bit suspect”? really?) except that nothing really shocks me anymore. Let’s not fool ourselves: That assertion is a dog whistle. And it’s one of the statements that leads me to question this researcher’s integrity.

Returning to that passage, the idea that scientific truths are seldom reversed in a decade (a) is a stupid thing to say because even if we granted it were rare it wouldn’t mean it’s not possible and (b) assumes that the place we were at before the research on gay and lesbian parenting was a place of scientific truth. The absence of research is not scientific truth! In fact, it’s exactly the opposite of scientific truth.

And the red herring of adoption is just the cherry on this bullshit sundae. You cannot use studies comparing adopted and biological children and conclude that since there is a difference there that there must be a difference for all gay and lesbian parents because “many children of gay and lesbian couples are adopted.” As a social scientist, Regnerus should know better. The studies he is referencing were not designed to answer that sort of question. You simply cannot be intellectually honest and jump from adoption to same-sex parenting like that.

Perhaps the most damning issue with Regnerus’s study is that he utterly fails to adequately define same-sex parenting. Honestly, I don’t know such a glaring theoretical and methodological omission got past peer review. From his article on Slate (emphasis mine):

Instead of relying on small samples, or the challenges of discerning sexual orientation of household residents using census data, my colleagues and I randomly screened over 15,000 Americans aged 18-39 and asked them if their biological mother or father ever had a romantic relationship with a member of the same sex. I realize that one same-sex relationship does not a lesbian make, necessarily. But our research team was less concerned with the complicated politics of sexual identity than with same-sex behavior.

In other words, “I realize that same-sex behavior does not necessarily define a person’s orientation, but I’m going to define it that way anyway!” The problem is that you cannot adequately separate out the “politics of sexual identity” from “same-sex behavior” when talking about sexual orientation. If people don’t identify as gay or lesbian, then you should not count them as gay or lesbian. It’s really that simple.

As someone who researches issues concerning sexual orientation, one of the major theoretical and methodological hurdles I’ve had to jump is how to define my population. According to Randal Sell, operationalizing sexual orientation as a demographic category is extremely problematic, often because “the methods used to measure sexual orientation in these studies do not always correspond with the most common conceptualizations of sexual orientation.” In other words, how the researcher conceptualizes sexual orientation should (as closely as possible) match how the subjects or informants conceptualize sexual orientation.

The wrong way to approach research on sexual orientation is to dismiss the aspects of sexual orientation (“the complicated politics of sexual identity” according to Regnerus) that you don’t like and go with the most simplistic concept that fits your prejudices. Put another way, if Regnerus had provided a robust and theoretically sound definition of sexual orientation for this study, he would not have gotten the results that he did. The results here depend on his shitty definition.

Regnerus is being disingenuous. When you frame your discussion with the question “Does it really make no difference if your parents are straight or gay?” and you propose that shifts in scientific understandings are a result of political posturing, and conclude that a heteronormative couple is best, you do not get to turn around and say “I’m not claiming that sexual orientation is at fault here…” Excuse me, but that’s exactly what you’re saying. The entire point of this article is to challenge the scientific consensus that children of gay and lesbian parents are at least as well adjusted as children of heterosexual parents. And you’re doing it with piss-poor theoretical positioning and equally problematic methodologies.

Regnerus admits in the article that this is all speculation because “the data are not poised to pinpoint causes.” Yes, folks, you read that right. He acknowledges that his research is limited and cannot be used to derive causes while simultaneously talking about his data as if it demonstrates causality. And you know how skeptics feel about correlation and causation!

Image from xkcd:

This is just bad research. That in and of itself isn’t that surprising—there are lots of bad researchers. What’s really troublesome is that his horrible research is being published in a peer-reviewed journal, and right-wing bigots will latch onto it as proof that gay and lesbian parents are bad for children. The abuse of social science research by the Religious Right cannot be allowed to continue. We social scientists need to do a better job of speaking out and educating people about social science research. Education is the key for being able to spot theoretical and methodological issues in research. Without this scientific literacy, people with power and an agenda will continue to attempt to cloak their research in the authority of science and hope that no one has the knowledge to critically analyze it.

And considering the way education is going in this country, they just might succeed.

Featured image from

  • Facebook
  • Twitter
  • Google+
  • Linkedin
  • Pinterest


  1. I’m exceptionally familiar with the academic publishing process in science/engineering, but not with the social sciences at all – if someone publishes something that’s BS in, say, Molecular Systems Biology and multiple people involved in the field write rebuttals, that tends to be enough to get a retraction, or at the very least no one takes it seriously.

    With the social sciences, there’s obviously a much more direct potential for abuse of published research in service of a political agenda – is it likely that this will be retracted, when faced with the very deserving criticism of its ridiculous methods? I’m kind of horrified by the prospect that it wouldn’t be, since of course a conservative sniffing around for “scientific proof” of zir prejudices isn’t going to do anything so… ‘liberal and elitist’ as critically examining shit that agrees with them. 😛

    • It really depends on the specific journal. The journal that the Regnerus article was published in also published 3 commentaries on the paper. They are critiques, but none of them are particularly damning about the paper. In fact, they seem to think the science is sound (one of them even described it as robust).

      I don’t know what it takes for this particular journal to retract an article. I doubt this will ever be retracted. Unfortunately, most of the academic responses to this are admiring the quantitative work and not addressing about the failure in theoretical framing. He does not claim a causal relationship in the paper, but the ways he’s talking about the research publicly do not make this clear. Unfortunately, people without access to the actual paper (or people with a political agenda) will read causation into it because of his popular writing. Which is also part of what makes me think he has an agenda.

  2. Will is right – Regnerus has covered his bases too well in the peer reviewed material. In fact so well that it shows he knows what he’s doing, and to me that makes it seem like he’s being deliberately disingenuous when he misrepresents the implications.

    It seems like he’s purposely created a goldmine of quotable quotes for ‘family’organisations, all the while including a few impotent caveats that he knows are easily ignored as it suits.

    P.S. I just found this site for the first time from It looks great, I didn’t know it existed!

  3. I’m inclined to say that self-identity is not really the appropriate operational definition either. Nobody’s especially worried about the children of gay Mormons in loving heterosexual relationships, ha. The appropriate operational definition is people with two same-sex parents. Unfortunately, the question only asks if either mother or father was at some point (from age 0 to 18) involved in a same-sex romantic relationship. But how long was that relationship, and did the partner act as a parent for any significant amount of time, or at all? Someone on Box Turtle Bulletin found out that 2 people in the entire sample had same-sex parents for most of their childhood. Then they decided to compare this group with only the stable and intact biological families.

    The more I read about this, the more it seems like a huge waste of money. National probability samples don’t grow on trees! And then they asked the wrong questions, did the wrong analysis, and made the wrong comparison.

    • I’m inclined to say that self-identity is not really the appropriate operational definition either. Nobody’s especially worried about the children of gay Mormons in loving heterosexual relationships, ha. The appropriate operational definition is people with two same-sex parents.

      You may have a point about how to properly operationalize a definition for this study. But I think I was unclear in my original post. What I mean is that the way they operationalized “same-sex parent” by asking the child if a parent was involved in same-sex behavior/romance at some point tells us nothing about children of people who identify as gay or lesbian. It tells us only about children of parents who had same-sex relations. And Regnerus is attempting to equivocate “child with a parent who engaged in same-sex romance” with “child of a parent who is gay/lesbian.” The fact that he’s so insouciant in his Slate article about this point is telling.

      Any way you slice it, though, I think we agree that he has failed to adequately define his population in a meaningful way–especially if he wants to turn around and write in the popular press about it.

      The more I read about this, the more it seems like a huge waste of money. National probability samples don’t grow on trees! And then they asked the wrong questions, did the wrong analysis, and made the wrong comparison.

      I agree completely. I guess his private funders are okay with spending $800,000 if it has the potential to further their cause.

      BTW, glad you’re still around. I do enjoy your critical commentary! =)

      • I didn’t rage quit, I just don’t want to waste time with internet arguments unless I enjoy them. Haha, most of my commentary is critical because when I can’t think of a disagreement I tend not to say anything.

Leave a Comment

This div height required for enabling the sticky sidebar