Yesterday on Public Discourse, I described the controversy that followed the publication of the New Family Structures Study (NFSS), led by University of Texas sociologist Mark Regnerus. During a summer of unusual abuse, Regnerus remained largely silent but with his head unbowed. As autumn arrived, he found himself vindicated as an honest scientist by his university, with continued support from the journal editor who published his research.

In the November 2012 issue of Social Science Research, Regnerus has published a new article: “Parental same-sex relationships, family instability, and subsequent life outcomes for adult children: Answering critics of the new family structures study with additional analyses.” He accepts “arguably the most reasonable criticism” of his original work, the use of the abbreviations “LM” (for lesbian mother) and “GF” (for gay father) to characterize the family situations experienced by his young adult subjects when they were children.

Since the adjectives “lesbian” and “gay” could lead readers to infer something about these parents’ self-identified “orientation” (though in his original article Regnerus clearly dispelled this misapprehension), he now exchanges “LM” for “MLR” (mother who had a lesbian relationship) and “GF” for “FGR” (father who had a gay relationship), so that the adjectives “lesbian” and “gay” now describe the relationships, not the persons. Regnerus also pauses to note the extreme unlikelihood that his categories swept in any “one-night stand” relationships, since the NFSS interviews asked young adults about romantic relationships they would have observed as children.

Regnerus addresses at much greater length the more serious charge that he compared apples to oranges by placing a sample of “MLR” and “FGR” families with high incidence of instability next to his “IBF” cases of intact biological families (married heterosexual couples that stay together and raise their own offspring to maturity). His critics insisted that he should compare intact, long-term stable gay and lesbian couples with his “gold standard” IBF households.

Start your day with Public Discourse

Sign up and get our daily essays sent straight to your inbox.

On this point, Regnerus yields no ground to his critics whatsoever, but instead only strengthens his case that family instability is not a variable to be controlled for so that it falls out of the comparison; rather it is a “pathway” down which MLR and FGR families typically travel as a social reality.

To begin with, Regnerus notes, “if stability is a key asset for households with children, then it is sensible to use intact biological families in any comparative assessment.” But could Regnerus have produced a data set with a higher number of “stably-coupled” gay or lesbian households? He doubts it.

In his original article, he reported that an initially-screened population of 15,000 young adults aged 18-39 yielded a set of 163 who said their mothers had had a same-sex relationship sometime during their childhood. (There were only 73 who said this of their fathers.)

In his new article, Regnerus has re-sorted a dozen of the FGR cases into the MLR category (since in these cases the subjects reported that both parents had had same-sex relationships). Now focusing on his 175 subjects in the MLR category, he finds that fewer than half of them (85) ever lived with both their mother and her same-sex partner during their childhood.

But that low number tapers off dramatically when subjects report the length of the couple-headed period: “31 reported living with their mother’s partner for up to 1 year only. An additional 20 reported this relationship for up to 2 years, five for 3 years, and eight for 4 years.” He later adds that “only 19 spent at least five consecutive years together, and six cases spent 10 or more consecutive years together.”

How many children were raised by two women staying together from the child’s first birthday to his or her eighteenth? Just two. And how many such cases were there in the FGR category—of children raised by two men together for their whole childhood? Zero. This, out of an initial population of 15,000.

I recite these numbers to make a point of my own that fairly leaps off the pages of Regnerus’s work: that family instability is the characteristic experience of those whose parents have same-sex relationships. This is what Regnerus is getting at when he says that critics who want him to treat stability as a “control variable” are actually “controlling for the pathways.” To go on an endless search for a sizable random sample of long-term, stable same-sex couples raising children is to miss the social reality in front of us, namely that they are conspicuously missing from the lives of children whose parents have same-sex relationships.

Doggedly responding to his critics, however, Regnerus divides his MLR cases into two further categories, those in which children never lived with their mother’s same-sex partner (90 cases), and those in which they did for any length of time at all (85 cases), and takes another look at his outcome variables, while also slicing his other categories thinner, of divorce, remarriage, single parenthood, adoption, etc.

Unfortunately for his critics, it makes very little difference. On multiple outcomes, the children of mothers who had lesbian relationships fared poorly, whether those mothers had a partner in the household with their children or not, and these two groups were more like each other than like the intact biological family (IBF) category. As Regnerus notes, “adult children who report a maternal same-sex relationship—regardless of whether their mother ever resided with her same-sex partner—look far more similar to adult children of other types of household than they do to those from stably-intact biological families.”

But shouldn’t Regnerus have asked the parents of his subjects about their self-identified orientation? Maybe he was actually looking at the fallout of “mixed-orientation” relationships that disintegrated, or at the parenting of people who weren’t “really” gay or lesbian. But again his critics are substituting an imagined social ideal for a messy reality.

Regnerus had good reason to ask adult children about their parents’ behavior, not their orientation: because this is what the children would be able to observe and know about, and because sexual attraction and behavior are highly fluid phenomena, despite the myth of a fixed “orientation.”

As he soberly puts it, “there appear to be plenty of failed heterosexual unions in the data,” in which many of the children of mothers who had same-sex relationships “spend their early years with their biological mother and father” before those relationships occur. Regnerus’s findings do not obscure the realities of family and sexuality in our society; they illuminate them.

And of course he was right to interview the children rather than their parents, because the former could more accurately self-report their current life conditions. Yet the children had to be adults at the time of the study, for ethical reasons that forbid this kind of research being conducted with minors and because he wanted to know the “finished product,” as it were, of their upbringing.

So, could it be that Regnerus captured a snapshot of an outdated social phenomenon, given that his study concerned adults who had been raised when same-sex couples rarely raised children (some more than 20 years ago), and did so under more trying circumstances? Would children being raised by persons in same-sex relationships today show a different pattern? “Perhaps,” he says, “but hardly certain.” Multiple studies show that same-sex couples, particularly lesbians, divorce at higher rates where marriage is available to them, and stay together for shorter periods. If so, then again we could expect to find family instability—and the effects thereof—in the life outcomes of children.

As Regnerus concludes, “Perhaps in social reality there really are two ‘gold standards’ of family stability and context for children’s flourishing—a heterosexual stably-coupled household and the same among gay/lesbian households—but no population-based sample analysis is yet able to consistently confirm wide evidence of the latter.” What we can say at this point is that “the probability-based evidence that exists . . . suggests that the biologically-intact two-parent household remains an optimal setting for the long-term flourishing of children.” There is no other type of household of which that can be confidently said.

Further strengthening the case Regnerus has so ably made is a remarkably comprehensive review of what social science knows about the intersection of sexuality, family structure, and childrearing effects, by Professor Walter Schumm of Kansas State University, in the same issue of Social Science Research.

According to Schumm, we know that it makes more sense to regard “the concept of sexual orientation as ‘fluid’ rather than fixed at birth.” And it appears that sexual orientation is subject in the case of children to profound influence depending on family structure.

As Schumm notes, a number of studies “concur in observing significantly higher rates of same-sex behavior or identity among children of same-sex versus heterosexual parents.” (This finding was also evident in the NFSS results reported by Regnerus.) We know from other studies besides the NFSS that long-term stable lesbian and gay couples raising children are extremely rare, or at least that finding them is so difficult that statistical analyses are problematic.

We know that “many children from eventual gay or lesbian families have been born into heterosexual families.” We have reason to believe that “lesbian parents . . . have substantially higher rates of relationship instability than do heterosexual parents,” and that “given the apparent fluidity of sexual orientation in general, but especially for women, it may even be rare for parents to maintain a same-sex orientation for 18 years, much less remaining with the same partner for that time.” We know that “multiple primary caregiver transitions, presumably regardless of the sexual orientation of parents, are stressful for children and increase the risk of poor child outcomes.” Is it any wonder, then, that the New Family Structures Study yielded the results it did?

Overall, Schumm concludes, Regnerus conducted eminently defensible scientific research, making decisions about research design and analysis “within the ball park of what other credible and distinguished researchers have been doing within the past decade.”

We should conclude that where accusations of an ideological axe to grind are concerned, they should not be directed at Regnerus, but at his critics in the academy and his self-appointed inquisitors in the blogosphere. With the latest issue of Social Science Research, Regnerus can consider himself fully vindicated as a scholar.

The controversy over same-sex marriage, and over the place of social science findings in debating the question, will doubtless continue. But Regnerus’s contribution has complicated a set of breezy assumptions too widely held: that children raised in these new family structures suffer no disadvantages whatsoever, and that stable, long-term same-sex-parent families can even be found in significant numbers. In so doing, Regnerus has moved our national conversation on the family forward, in a positive direction, with greater awareness of what is at stake in the public policy choices we make.