by Peter Sprigg
August 3, 2012
Recently, I posted a piece responding to last weeks Weekly Standard cover story describing the attacks upon sociologist Mark Regnerus. His recent article in the journal Social Science Research showed that children of homosexual parents (that is, young adults who reported that a parent had a same-sex romantic relationship while they were growing up) suffer disadvantages relative to children in a variety of other family situations.
The author of the Weekly Standard story, Andrew Ferguson, criticized a sentence containing FRCs summary of the studys findingsin part, because it failed to include awkward clarifications like that in parentheses above. In my earlier post, I argued that Ferguson had made mountains out of molehills by criticizing a single sentence for not containing all the qualifications and explanations that were contained in the 3,000-word paper which the sentence introduced (somethingFerguson never mentioned).
However, in his even longer (4,000 words) article,Fergusonhimself made mountains of error with misleading statements that may have seemed like molehills to him.
Here are some statements by Andrew Ferguson of The Weekly Standard that are far more questionable than the FRC statement he takes issue with:
1) For more than a decade now the unchallenged view among social scientists has been that there is no difference between children brought upI mean parentedby lesbian and gay couples and those brought up in households where Ma and Pa are married.
This is not an accurate description even of what the pro-homosexual research has asserted. For example, the statement by the American Psychological Association (APA) that was challenged by scholar Loren Marks (in the same issue with Regnerus’ article) said, Not a single study has found children of lesbian or gay parents to be disadvantaged in any significant respect relative to children of heterosexual parents.
What is the difference between this APA statement and what Ferguson said? The APA made no assertion about children raised by coupleseither homosexual or heterosexual. These studies have focused almost exclusively on the sexual orientation of the parentsnot the family structure. A large number of them compared children raised by lesbian single mothers with children raised by heterosexual single mothers.
Virtually none of the pro-homosexual parenting studies have compared children brought up by lesbian and gay couples and those brought up in households where Ma and Pa are married. (I am aware of only one previous study that limited its comparison to coupleshomosexual, cohabiting heterosexual, and married heterosexual. Published by researcher Sotirios Sarantakos in the journal Children Australia, it found that children of married heterosexual couples did the best, and the children of homosexual couples did the worst, on a majority of the outcomes measured. Marks discusses this study in his article.)
In the debates over the definition of marriage, FRC has focused our primary arguments not on the generic superiority of heterosexual parents over homosexual parents (although the Regnerus study offers support for that as well), but on the fact that children do best when raised by their own, married mother and father. There is no question that children raised by heterosexual single parents or divorced parents are also at a disadvantage relative to those raised by their married mom and dad. That is a principal finding of the Mapping America series produced by FRCs Marriage and Religion Research Institute (MARRI).
2) Ferguson says:
It was also clear that the nature of gay parenting has changed quite a bit from the 70s, 80s, and 90s, when these young adults were children. Typically, Regnerus said, they were born from heterosexual unions that went bust; nowadays the children of homosexual couples are often plannedbrought into a family through adoption, in vitro fertilization, or surrogate motherhood.
Ferguson cites no scientific source, either pro-homosexual or pro-family, for the claim that nowadays the children of homosexual couples are often planned. My suspicion is that this is merely one of the pro-homosexual talking points (which Ferguson relies on too much in his article). Over the last forty years such planned homosexual families have probably gone from virtually non-existent to merely minuscule in number; there is no basis for saying such families are formed often. Chances are that even today, the vast majority of children being raised by gay parents (whether that is defined by self-identification or by behavior) were conceived in the context of a previous heterosexual relationship.
3) Ferguson says,
The differences between the GFs [children of gay fathers] and the IBFs [children of intact biological families] were smaller and less significantthere was no difference, for example, in reports of childhood sex abuse.
This illustrates a pet peeve of mine about how the media reports on statistical significance. While findings that are statistically significant are certainly more compelling (because we can have a higher level of confidence that they accurately represent the real world), that does not mean we should simply dismiss any findings that do not rise to the mathematical standard of statistical significance. In the case of the gay fathers in the Regnerus study, the main reason why fewer of the comparisons involving them were statistically significant is because the sample size was smallernot because the reported differences in outcomes were small.
It is flatly false, for example, to say there was no difference … in reports of childhood sex abuse between IBFs and GFs. In fact, only 2% of the children of married parents reported they were ever touched sexually by a parent or other adult, while 6% of the children of gay fathers said thisthree times as many. In addition, only 8% of the children of married parents said they were ever forced to have sex against their will (not just in childhood), vs. 25% of the children of gay fathers—again, three times as many. However, because of the small sample sizes (both of GFs and of those who experienced sexual abuse), these findings were not statistically significant. That hardly constitutes proof (or even evidence) that there was no difference. The evidence (but, in the absence of statistical significance, not proof) points in exactly the opposite direction.
4) Ferguson paraphrases Gary Gates, a pro-homosexual demographer, as saying that we cant tell from Regneruss data what role homosexualityas opposed to divorce, welfare, single-parenthoodplayed in the bad outcomes.
Unfortunately, Ferguson fails to even mention that Regnerus did include an analysis of his data using controls for several other factors that might influence the outcomes. Specifically, he included controls for respondent’s age, race/ethnicity, gender, mother’s education, and perceived family-of-origin income. If a finding is statistically significant with controls, it means that any apparent differences based on having a homosexual parent cannot be accounted for based on any of the factors for which he controlled.
Even more importantly, he also controlled for having been bullied and for respondents’ current state of residence … according to how expansive or restrictive its laws are concerning gay marriage and the legal rights of same-sex couples. These controls seriously undermine the pro-homosexual talking point (repeated by Ferguson in his critique of FRC’s statement), that the instability, and hence the bad outcomes, could be largely traced to trauma caused by the antihomosexual prejudice of an earlier time. A finding that is significant even after the controls would mean, for example, that even among those who were never bullied, and even in states where same-sex marriage is already legal, children do better when raised by their own mother and father.
Ferguson ignores the fact that Regnerus used controls; but perhaps even more glaringly, he ignores the fact that Regnerus compared the children of homosexual parents not just with parents of married mothers and fathers, but with five other heterosexual family structures as wellchildren adopted by strangers, children whose parents divorced late (that is, after they turned 18), children who lived in step-families, and children of single parents. This part of the study has not gotten as much attention (and it will be the focus of my next FRC Issue Brief on the Regnerus study).
On these comparisons, Regnerus reports that children of lesbian mothers display 57 [differences] that are [statistically significant] … . The majority of these differences are in suboptimal directions, meaning that LMs [children of lesbian mothers] display worse outcomes. Regnerus has dramatically understated his findings here. I went through his tables item by item, and found that the LMs were suboptimal on all 57 statistically significant comparisons.
Critics complain that Regnerus failed to compare stable gay households with stable heterosexual households (more on that later). But Regnerus did compare unstable gay households with unstable heterosexual householdsand the gay households were worse on every statistically significant comparison he reported (and most of the ones that were not statistically significant). This suggests that the problem with the homosexual parents was not merely their instabilitythe problem was their homosexuality. Virtually none of the media reports on the study have addressed this.
5) Ferguson says,
[Researcher Loren] Marks sums it up: In response … to any question regarding the long-term, adult outcomes of lesbian and gay parenting we have almost no empirical basis for responding.
And now, with the publication of Regneruss study … we still dont.
This is a ridiculous statementessentially discounting entirely the many ways in which Regnerus study is superior to any previous gay parenting study ever conducted. Marks was referring specifically to the 59 pro-homosexual parenting studies cited by the American Psychological Association in a 2005 report. Using a large, representative, population-based sample (which almost none of the APA studies did), Regnerus made numerous findings that are statistically significant, and almost all of them showed children of homosexual parents at a disadvantage. While it does not settle every possible question about homosexual parenting, it is false to suggest that the Regnerus study has not provided an empirical basis for responding to such questions. (And again, Regnerus data contains more information than what he published in this initial articleif other researchers want to use it.)
In saying we still don’t, Ferguson is accepting the effort of pro-homosexual activists to change the subject, and insist that the only meaningful findings would be ones that compare intact homosexual parents with intact biological families. At least he notes what Regnerus also points outthat the number of such [intact homosexual] parents in the general population is infinitesimal right now.
This is all somewhat comical, actually. Pro-homosexual activists who for years have said we have all the information we need (to conclude that children of homosexuals suffer no disadvantages) are so desperate to discredit the Regnerus study that they are suddenly insisting that we don’t really know anything about the subject yet.
In reality, thanks to Regnerus New Family Structures Study, we now know far more about the young adult children of parents who had same-sex relationships than we did from all the previous gay parenting studies combined.