Mark Regnerus and the Storm over the New Family Structures Study

Seldom has the publication of a dry, factual report in sociology caused such a storm of controversy. In June 2012, the bimonthly peer-reviewed journal Social Science Research published an article by University of Texas sociologist Mark Regnerus titled, “How different are the children of parents who have same-sex relationships? Findings from the New Family Structures Study.” The answer to his title’s question was: quite a bit different, and most of the differences are not good.

Within minutes, it seemed, Professor Regnerus, a gifted and highly productive scholar with two previous books published on related subjects, was denounced as “anti-gay,” attacked personally and professionally, and his thoughtful, measured research conclusions were buried under an avalanche of invective, abuse, and misunderstanding. For the remainder of the summer months, Regnerus withstood an onslaught of criticism, but as the autumn arrived, it became clear that his reputation and the soundness of his research had been vindicated.

What had happened?

The editor of Social Science Research, Professor James D. Wright of the University of Central Florida, had known that Regnerus’s article would spark discussion about family and sexuality among sociologists. As he would later say himself when others complained that he was trying to drive up the readership of the journal, “guilty as charged.” What editor doesn’t want people reading and talking about what he works so hard to produce?

This is why Wright published, alongside Regnerus’s new research, a probing criticism of the inadequacy of nearly all previous research on the question of parenting by people in same-sex relationships, authored by Professor Loren Marks of Louisiana State University (who was not connected with Regnerus’s new research in any way). It’s also why Wright invited critiques to be published, in the same issue, by three experienced scholars in the sociology of the family (Paul Amato, David Eggebeen, and Cynthia Osborne), with rejoinders by Regnerus and Marks. It made for a very interesting exchange.

The June 2012 issue of SSR was a red-hot topic of controversy because Regnerus and Marks overthrew a “consensus” among sociologists on the “no differences” thesis—the view that there are no meaningful differences, in the life outcomes of children, between those raised by heterosexual parents and those raised by gay or lesbian ones.

In its most extreme form—one that is not even supported by the generally low-quality research published before Regnerus’s article—the “no differences” thesis holds that children raised by parents who have same-sex relationships do just as well as, or in some cases even better than, those raised in the intact biological family by their own natural parents who are and remain faithfully married to each other.

The American Psychological Association, despite the cagy wording of its bombshell assertion, was probably happy to invite this unwarranted inference in its 2005 legal brief, published to influence judicial deliberations in same-sex marriage lawsuits. The APA said “the evidence to date suggests that home environments provided by lesbian and gay parents are as likely as those provided by heterosexual parents to support and enable children’s psychosocial growth.” And who didn’t think of the Ozzie-and-Harriet natural family when reading “heterosexual parents” in that sentence?

But as Loren Marks showed, the 59 studies grounding the APA’s statement were all deeply flawed, with sampling and design problems, inadequate statistical rigor, and conclusions about “no differences” that could not be justifiably generalized to the larger population.

And whereas Marks offered only well-founded criticism of previous research, Regnerus offered something new: the first research employing a large, random sample of the young adult population, directly asking them about their childhood experiences and their present state of life, across a range of variables touching on economic and educational success, romantic and sexual experience, substance abuse, experiences with crime and violence, and so forth.

Regnerus and his colleagues in the New Family Structures Study (NFSS), working with the research firm Knowledge Networks, screened more than 15,000 Americans between the ages of 18 and 39, and interviewed nearly 3,000 of them. The subjects’ family experiences were sorted into eight categories, ranging from the intact biological family (with the subject’s parents still together at the time of interview), across various family structures involving divorce, remarriage, adoption, and single-parenting, with two categories for subjects raised by mothers or fathers who had same-sex romantic relationships during their childhood.

The results were dismal for the “no differences” thesis: on 25 out of 40 outcomes variables, the children of mothers who had had lesbian relationships fared poorly compared to the children of intact biological families. And on 11 of the 40 outcomes, the children of fathers who had had gay relationships fared poorly on the same comparison. (For a summary of the study’s findings, see Ana Samuel’s Public Discourse article, “The Kids Aren’t All Right,” and http://www.familystructurestudies.com/.)

Regnerus was cautious in his conclusions: he didn’t label poor outcomes as effects of parents’ sexuality, and noted that “a variety of forces uniquely problematic for child development in lesbian and gay families” could account for the phenomena. But, he concluded, “the empirical claim that no notable differences exist must go.”

The high quality of the New Family Structures Study’s research design, data collection, and findings, and the firmness of Regnerus’s conclusion that the “consensus” in sociology was exploded, only seem to have encouraged interested parties, in the academy and outside it, to attempt to debunk the NFSS. UCLA demographer Gary Gates assembled about 200 scholars to denounce Regnerus’s article, but to little substantive effect.

In the public arena, Regnerus saw his research crudely hashed over at The New Yorker, The New Republic, and the Huffington Post, among other places—and found himself caricatured as strapped to a Catherine wheel on the cover of the Weekly Standard. These are not normal experiences for your average stay-at-home sociology professor. Clearly Regnerus’s political adversaries saw much at stake in the public reception of his research. (For the legal stakes, see my Public Discourse essay, “Supreme Court Take Notice: Two Sociologists Shift the Ground of the Marriage Debate.”)

The two main criticisms of Regnerus’s article, repeated in numerous variations, are these. First, he had used the abbreviations “LM” (for “lesbian mother”) and “GF” (for “gay father”) to describe subjects who knew that their mother or father had a romantic same-sex relationship of any length before the subject turned 18.

The use of “LM” and “GF” was culpably misleading, critics claimed, because the category might include persons who never “identified” as lesbian or gay, and might only have had a “one-night stand” with a same-sex partner. The second criticism, closely related, was that in comparing these young people raised in “LM” and “GF” households, so defined, with those raised in “IBF” households—married heterosexual couples raising their own biological offspring and staying together throughout the subjects’ lives (even beyond their childhood, to the present)—Regnerus was comparing apples to oranges.

In their view, he should have compared children of IBF households with children of long-term, intact, stable same-sex couples who identify as gay or lesbian. Then, they were sure, the differences he found would largely disappear—as they claimed was shown by the previous research Regnerus and Marks had each criticized for their small, unrepresentative samples. What he was really doing, they claimed, was setting stable family situations next to unstable ones—and so stability was the real variable at work. To make it seem that the differences were “about” sexuality was worse than an error, critics claimed: this was culpable distortion of the social phenomena, a twisting of social science in the service of conservative ideology.

A third, more ad hominem criticism was that Regnerus received the majority of his grant funding from the Witherspoon Institute (publisher of Public Discourse), and a minority from the Bradley Foundation—both of them viewed as “conservative” institutions in their educational and philanthropic efforts. But Regnerus declared these facts in his original article, and told his readers that neither Witherspoon nor Bradley had any role in shaping the conduct or the conclusions of his research, which he has made wholly transparent. No one has ever gainsaid this avowal on his part. For my part, I can say that Regnerus had no input on my choice to write this account of the controversy or its content.

In the less responsible precincts of the blogosphere, Regnerus was the target of vicious calumnies along the lines described above, one of which led to the opening of an official “inquiry” by the University of Texas at Austin, where he teaches, to determine whether he had committed “scientific misconduct.”

At UT, the policy in such matters is that the merest squeak from any party alleging misconduct is enough to trigger a preliminary inquiry, which in 60 days must determine whether a full-blown investigation is warranted. The university swung into action, doing everything by the book, at no little inconvenience to Regnerus, but at the end of August the UT “research integrity officer” concluded that no plausible charge of misconduct could be substantiated. The university’s provost accepted that conclusion, and closed the matter without prejudice to Regnerus’s standing as a scholar and teacher.

Meanwhile SSR editor James Wright was under fire for publishing Regnerus’s article; for appearing to rush it to publication; and for placing Marks’s article alongside it. Opting for transparency at some risk to his own reputation, Wright asked a member of SSR’s editorial board to “audit” the process that led to the publication of Regnerus’s article.

The risk was that he chose Darren E. Sherkat, a sociologist at Southern Illinois University whom Regnerus would later describe (without fear of contradiction) as someone “who has long harbored negative sentiment about me.” Sherkat, speaking out of school, confidently told a writer for the Chronicle of Higher Education in July that Regnerus’s study was “bull****” when his audit was still in draft form and neither Regnerus nor Wright had written a response to it.

Sherkat’s audit and several other items of interest have now been published in the November 2012 issue of SSR, in a special 40-page section introduced by Wright. To his credit, when he sticks to the charge he was given, Sherkat finds that the journal’s editor did nothing wrong in publishing either Regnerus’s article or Marks’s.

Wright referred both papers to knowledgeable scholars of the subjects involved, who held varying views on the politics of same-sex unions, and who unanimously recommended their publication. No violations of normal procedure occurred; Sherkat says he “may well have made the same decisions” Wright did, given the reviews; and he dismisses as “ludicrous” any suggestion that the editor was up to anything political.

To his discredit, Sherkat, a sociologist of religion who does not appear to have done any research on family and sexuality issues (but for a single article studying how religion and political affiliation affect views of same-sex marriage), nonetheless appoints himself a final referee of the merits of Regnerus’s research—not a function he was asked to perform—and opines that it should not have been published.

James Wright, correctly, takes Sherkat’s conclusions as an auditor as vindication of his editorial performance, and rightly discounts his colleague’s attempt to set himself up as a post hoc referee with a veto over publishing Regnerus’s scholarship. If he sent the work to knowledgeable reviewers who unanimously said to publish it (and Wright notes that such unanimity is unusual), that seems to be the end of the affair.

But it isn’t. In the latest issue Wright chose to publish two significant new contributions to the discussion begun in June. The real issues with Sherkat and other critics are joined by Regnerus, who returns to the pages of SSR with a vigorous response and a re-analysis of his data, and by Professor Walter Schumm of Kansas State, who contributes an expert review of what we know from social science today about the interwoven variables of sexuality, family stability, and childrearing outcomes.

I’ll say more on these contributions in tomorrow’s essay.