Regnerus Revisited: American Sociological Association is Wrong about Homosexual Parents and “Stability”
April 24, 2015
One of the key issues regularly raised in the debate over redefining marriage to include same-sex couples has to do with the well-being of children.
Defenders of the historical and natural definition of marriage as the union of one man and one woman argue that children have a right to a relationship with both the man and woman whose union created them, and that research shows children raised by their married, biological mother and father have, on average, better life outcomes than children from other family structures.
Advocates of redefining marriage, on the other hand, argue that children already being raised by same-sex couples would benefit from the legal stability and social affirmation associated with marriage. They also insist that research which has specifically compared children of homosexual parents to children of heterosexual parents shows, "The scholarly consensus is clear and consistent: children of same-sex parents fare just as well as children of different-sex parents."
The latter quote is drawn from an amicus curiae (friend of the court) brief filed in the U.S. Supreme Court by the American Sociological Association (ASA). It's one of dozens filed by advocates on both sides in the case being argued under the name Obergefell v. Hodges. Advocates of redefining marriage are asking the high court to overturn a decision of the U.S. Court of Appeals for the Sixth Circuit, which upheld laws in Michigan, Ohio, Kentucky, and Tennessee that define marriage as the union of one man and one woman.
A separate brief filed in support of the states on behalf of the American College of Pediatricians (ACP), Family Watch International, and several scholars, has thoroughly rebutted most of the arguments in the ASA brief. The ACP brief, using the most recent studies with the best methodology (drawing on large government surveys in Canada and the United States), shows children of same-sex parents are disadvantaged on a number of measures relative to children raised by their married, biological mother and father. (Patrick Fagan of the Marriage and Religion Research Institute -- MARRI -- has summarized the ACP rebuttal of the ASA in an op-ed.)
This revolution in homosexual parenting research began with publication in 2012 of findings from the New Family Structures Study, organized by Dr. Mark Regnerus, a sociologist at the University of Texas (and one of the signers of the ACP brief). Although homosexual activists embarked on a nearly hysterical campaign of vilification of Regnerus, his research was not "discredited," as liberals commonly claim. I have written in detail about the Regnerus study here and here.
While the ASA brief is misleading on the nature and state of the evidence, there is one assertion they make about the Regnerus research which is flatly false -- so plainly inaccurate, in fact, that it makes one question whether the authors of the brief have even looked at the scholarly articles they are critiquing.
The false statements are these:
The study [cited as Regnerus 2012a] stripped away all divorced, single, and stepparent families from the different-sex parent group, leaving only stable, married, different-sex parent families as the comparison. (ASA Brief, p. 23)
Regnerus 2012b [a follow-up report with more detailed analysis of the data] continues to ignore stability as the primary factor in child outcomes. (ASA Brief, p. 26)
In reality, Regnerus 2012a included comparisons for eight different family structures. In addition to what Regnerus calls the "intact, biological family" or "IBF" (the "stable, married, different-sex parent families" referred to by the ASA) and two involving homosexual parents ("lesbian mother" or "LM" and "gay father" or "GF"), his comparison included five other family structures with heterosexual parents which might be considered "unstable," including divorced, stepfamily, single-parent, and adoptive households. Even compared with these "unstable" households, the households headed by homosexual parents did not fare well. For example, Regnerus notes:
Of the 239 possible between-group differences here -- not counting those differences with Group 1 (IBFs) already described earlier -- the young-adult children of lesbian mothers display 57 ... that are [statistically] significant ... and 44 ... that are significant after controls .... The majority of those differences are in suboptimal directions, meaning that LMs display worse outcomes [emphasis added]. (Regnerus 2012a, p. 764)
In his follow-up article ("Regnerus 2012b"), Regnerus broke down the data on households in more detail, listing fifteen different household structures. To make the meaning of the homosexual parent categories more clear, he described them as "fathers who had a gay relationship" or "FGR" and "mothers who had a lesbian relationship" or "MLR." In response to criticism that he did not distinguish between children who had lived in the same household with a homosexual parent and that parent's partner from those who did not, the latter category was broken down to "MLR + partner" and "MLR no partner." (Only two subjects, however, had lived with a homosexual parent and one partner for the entire duration of their childhood from birth to age 18 -- as Regnerus clearly noted.)
Apart from the IBF and (now) three homosexual parent categories, the other eleven categories all involved children of heterosexual parents who had "unstable" household settings -- households where the parents divorced, were never married, one parent died, or in which the child was adopted as an infant (with variations based on subsequent relationships and/or remarriage).
Critics had also claimed that children were more affected by family "instability" than by parental sexual orientation, and those who lived with a partner were presumed to have more stable relationships. Directly contrary to what the ASA brief claimed, Regnerus reported specifically on comparisons between the "MLR + partner" category and other "unstable" categories:
Group 3 (MLRs who lived with their mother's partner) compare less favorably with:
- Group 8 (divorced, lived with mother, no subsequent relationships): 12 differences.
- Group 13 (parents married until one died, no subsequent relationships): 15 differences. (Regnerus 2012b, p. 1376)
Regnerus reports the raw data for all fifteen family structures, and all forty outcome measures. There are thus a total of 440 comparisons between households with children of a lesbian mother and a partner and unstable heterosexual households (11 unstable household categories times 40 outcome measures). He also reports which of the differences reach the level of statistical significance, not only in comparison to the intact biological family, but also in comparison to the "MLR + partner" category.
I did my own analysis of the data in the charts of the Regnerus 2012b article, and found that 84 of these comparisons showed statistically significant differences. Of those, 76 showed worse outcomes for the children of lesbian mothers with a partner; only 8 of those comparisons showed better outcomes for those children. Five of the eight "better" outcomes for "MLR + partner" households were on a single outcome measures -- daughters of lesbian mothers with a partner (but not of those without a partner) have fewer male sexual partners. Having multiple sexual partners of either sex is, rightly, considered an undesirable outcome.
More recent studies using government survey data (which also show negative outcomes for children of homosexual parents) have, arguably, surpassed Regnerus in the quality of the sample studied. However, the Regnerus study remains extremely valuable for the large number of separate outcome measures (forty).
It is also valuable because it does make direct comparisons (contrary to the ASA's claim) between households with homosexual parents and many different unstable household forms headed by heterosexuals. It thus thoroughly debunks the myth that only "instability" explains the harmful outcomes identified for children of homosexuals.