When a British science journal published an American study in January showing that emotional problems are more than twice as prevalent for children with same-sex parents than for children with opposite-sex parents, nobody expected the author, or the journal editors, to escape criticism. The “consensus” within the social science establishment is that the kids being raised by same-sex couples are doing fine, and will do even better if these parents are allowed to marry. Any researcher who finds anything different must be wrong, incompetent and homophobic.
What is a little surprising is that scholarly associations would criticise their peer, Paul Sullins, for not doing things which he patently did do in his study, and for benefiting from an allegedly slipshod peer review process that, in fact, is far more rigorous and open than anything their own favourite studies have been subjected to. Such is the quality of briefs submitted by the American Psychological Association and the American Sociological Association to the US Supreme Court last month in support of a ruling in favour of same-sex marriage.
Sulllins, an associate professor of sociology at the Catholic University of America, has replied in detail to these misrepresentations in a brief to SCOTUS jointly submitted with fellow sociologists Loren Marks, of Louisiana State University, and Mark Regnerus, of the University of Texas at Austin, and two organisations – the American College of Paediatricians and Family Watch International.
Regnerus was the first to feel the full wrath of the gay marriage movement with the publication of his landmark New Family Structures Study in 2012 and its finding that the adult children of women who had had same-sex relationships did significantly worse on many measures than the children of married heterosexual couples.
The main scientific pretext for dismissing Regnerus’ study was that he compared the most stable form of family with lesbian or mixed orientation households that happened to be particularly unstable. The APA and ASA has lazily levelled this charge at Sullins as well – ignoring the fact that most studies alleging to show “no differences” between children raised in heterosexual and same-sex families have not accounted for this variable either.
Testing the instability thesis
In fact, Sullins did control for family instability in Emotional Problems among Children with Same-sex Parents: Difference by Definition — in two ways.
First, he used as a proxy for stability whether a household was in its own home or in rented accommodation, since there is abundant research showing that home-owning families are much more relationally stable than renters. At the same time he controlled for the effects of income and education on the type of housing.
Sure enough, home ownership had a strong effect on emotional problems across the board (“children of families in rented quarters are 31 percent more likely to experience emotional problems than children of homeowner families”). But it accounted for only 3 percent of the difference in risk of emotional problems between opposite-sex and same-sex families. Children in stable same-sex families are still about twice as likely to suffer serious emotional problems as are children in stable opposite-sex families.
Second, Sullins compared same-sex families with only opposite-sex step-parent or “blended” families to test the effect of past family breakdown. In other words, he compared his whole sample of same-sex families with the least stable of heterosexual families in his data, conceding as much ground as possible to the thesis that instability is the only or main reason that children in the former seem to be at a disadvantage. Incomprehensibly, says Sullins, the APA calls this a “methodological flaw”.
“Perhaps it has to do with the outcome,” he adds, because the comparison reduced the overall risk of child emotional problems due to same-sex parents by only 13 percent, from a risk of 2.4 to a risk of 2.2. Far from explaining the difference away, the effects of prior divorce or family dissolution accounted for only a small part of the substantially higher risk of emotional problems faced by children with same-sex parents.
Stigma, bullying, and unexpected harm
Ah, but there’s the stigma these children have to contend with, says the “no difference” school. Take away the bullying and you have well-adjusted children. Sullins tested that and found that children with same-sex parents did not experience more bullying than did their counterparts with opposite sex parents – except those in the former group who had ADHD. In other words, the bullying was associated with ADHD rather than the kind of parents these child had.
It is important to note that these findings are based on more data than any previous study — 512 children with same-sex parents drawn from the US National Health Interview Survey. It is one of only eight studies on this subject in the past two decades that have used a random sample large enough to provide any reliable evidence on the question of child wellbeing in same-sex families.
What is more, the findings of the Emotional Harms study have been confirmed by another of Sullins’ studies based on data from the National Longitudinal Study of Adolescent Health (Add Health). As yet unpublished, “The Unexpected Harm of Same-Sex Marriage…”, compared two groups of same-sex parents, one unmarried, often following a prior heterosexual relationship, in which children had lived with their current parents for an average of four years; the other married, in which children had lived with their current parents an average of over 10 years.
Contrary to what is claimed by same-sex marriage advocates, child outcomes were consistently worse in the second group, the one with married parents and longer stability.
In one of its potshots at Sullins the APA has pointed to some coding errors in the NHIS survey that placed some opposite sex couples in with same-sex couples. But, as he points out, such contamination of data would make it more difficult to show differences between the two family types, and lead to an understatement of children’s emotional problems in same-sex households.
Surprisingly, Sullins told MercatorNet, neither the APA nor the ASA briefs “contains a word of substantive critique of the study’s central finding, that child emotional harm is 2-4 times higher with same-sex parents than with opposite-sex parents. There is some critique of the explanatory variables, but no attempt to rebut the findings themselves.”
When all else fails, attack the journal
However, they do resort to a now familiar strategy: attacking the journals in which his work has been published and the peer review process.
For the two studies relevant to this article, the publishers are: the British Journal of Medicine and Medical Research (“Child Attention-Deficit Hyperactivity Disorder (ADHD), and in Same-Sex Parent Families in the United States: Prevalence and Comorbidities,” 2015); and the British Journal of Education, Society and Behavioural Science (“Emotional Problems among Children with Same-sex Parents: Difference by Definition,” 2015).
The APA complains that “none of the journals in which Sullins’s papers were published are indexed in major social science databases.”
Sullins points out that, after the attacks on Mark Regnerus, his editors and reviewers after publication of the NFSS study in a sociological journal, he decided to pursue publication of his studies, based on a large public health survey, in international hard science medical journals, “where the standards of evidence are generally more rigorous, but the imposition of groupthink orthodoxy is much less, than in American social science journals.”
These journals are found in medical indexes. They are also abstracted by all the major scholarly services, making them available to researchers, says Sullins, adding that the publisher of his studies is a member in good standing of the official certifying agency for such journals, the Directory of Open Access Journals.
If only the APA had a review process as rigorous
Both APA and ASA allege that the peer review for Sullins’ articles was substandard. However, he shows that the time involved was not short compared, for example, to the practice of leading medical journals (BMJ and JAMA). Indeed, it was longer than the 13 days it took for one of the iconic “no difference” studies (Jennifer L. Wainwright & Charlotte J Patterson, “Peer relations among adolescents with female same-sex parents”) to be accepted by the journal Developmental Psychology.
Nor did the process lack rigour, as alleged by APA. For Sullins’ central study on child emotional problems, the editor sent it to four instead of the usual two reviewers and appointed two independent editors to approve publication. Approval was given only after two rounds of review and response instead of one. And one reviewer presented Sullins with an extensively revised draft. “This high level of scrutiny is very rare in American social science journals, to say the least,” Sullins notes.
Most tellingly he points out that the peer review history for his article is available online – itself a sign of peer review quality, and one not found among his critics:
“No APA journal practices this level of transparency. It is not possible to examine the peer review history for any articles in the ASA/APA roster of harm denial. Not a single on has dared to publish its peer review history.”
And if more credentials are needed, try this: the publisher of Sullins’ articles was ranked among the top 7 percent of journals worldwide for peer review rigour in a recent independent assessment published by Science, the world’s premier scientific journal. The only American social science publisher involved in the assessment failed the test.
In the end, all the baseless criticisms of Sullins’ studies, publishers and peer review – like those of Regnerus before him — have nothing to do with their scientific rigour, but with his findings, which do not conform to the ideology of harm denial. One can only agree with his conclusion: “Despite mounting evidence to the contrary, the APA and the ASA will doubtless continue to deny that any study has found evidence of harm to children with same-sex parents.”
Two months ago when the Emotional Harms study was published, Michael Cook wrote that Sullins would have to “be ready to go all 15 rounds” in its defence. This round, on any objective assessment, must go to him.
Carolyn Moynihan is deputy editor of MercatorNet.
Paul Sullins research can be viewed on his author page at the Social Science Research Network.