The Kids May Not Be All Right, Part 2

(by Dave)

In my last post, I noted that many advocates of same-sex marriage claim that “ we don’t have much data showing that kids raised by same-sex parents are worse off or ‘harmed’.”  I noted that while this might be technically true–given the little evidence there is concerning same-sex parents–the vast majority of social science research done over the last several decades confirms that the best situation for children is living with a mother and a father.

Recently, we saw two new studies released bearing on this topic.  The first, a very large study conducted by researchers from the University of Chicago’s Booth School of Business and the National University of Singapore, looked at 5,000 boys from across the nation dealing with behavioral problems.  The upshot of the study is that boys behave better when they have both a mother and a father, and that this is related to the differences between men and women as parents.

The other study that came out was from the National Longitudinal Lesbian Family Study, purporting to show that lesbian parents are much better than the old-fashioned mother-father combo.  This study and others were highlighted in a LiveScience article that trumpeted “Why Gay Parents May be Better Parents.”  In this article Judith Stacey, a New York University sociologist, claims that a 2010 study of all the studies out there on same-sex parenting shows that kids raised by lesbian couples turn out just as well as kids raised by heterosexual couples: “There’s no doubt whatsoever from the research that children with two lesbian parents are growing up to be just as well-adjusted and successful as children with a male and a female parent.”  She even offered that while there is almost no research on gay men as parents, she suspects they will be “the best parents on average.” But there is a critical problem with her conclusions.

Representative vs. non-representative samples

Stanton Jones, professor of psychology and provost of Wheaton College, notes that “The Achilles’ heel of research into the homosexual condition today is the issue of sample representativeness. To make general characterizations about any population or subpopulation, scientists must know that they have sampled individuals who truly represent the broader group about which they are going to make generalizations.” Jones makes this comment in an article in praise of Evelyn Hooker, a scientist and activist who changed the way homosexuality was looked at, precisely because she was aware of the problem of representative samples of people with same-sex attractions. Before Hooker, scientists treated all homosexuals as “necessarily and inherently pathological”—not surprising considering researchers had simply used convenient groups of people with same-sex attractions in hospitals or prisons.  Hooker found “ordinary people” with same-sex attractions, thus debunking the inherent pathology claim.  But this honesty in debunking claims based on unrepresentative samples is not something that has been carried on rigorously by Hooker’s scientist/activist successors.

In a much longer article, “Same-Sex Science,” Jones makes his own meta-study of what we know about same-sex attractions.  He observes that while Evelyn Hooker, unsurprisingly, proved that not all homosexuals were pathological, studies with representative samples show that people with same-sex attraction—whether they live in what he defines as “homophobic” environments or not—are at a greater risk of mental and emotional health issues.

One of the most exhaustive studies ever conducted, published in 2001 in the American Journal of Public Health and directed by researchers from Harvard Medical School, concludes that “homosexual orientation . . . is associated with a general elevation of risk for anxiety, mood, and substance-use disorders and for suicidal thoughts and plans. Other and more recent studies have found similar correlations, including studies from the Netherlands, one of the most gay-affirming social contexts in the world. Depression and substance abuse are found to be on average 20 to 30 percent more prevalent among homosexual persons. Teens manifesting same-sex attraction report suicidal thoughts and attempts at double to triple the rate of other teens. Similar indicators of diminished physical health emerge in this literature.

Similarly, with regard to the social and psychological aspects of same-sex relationships, Jones finds that claims of same-sex relationships being equivalent to male-female ones are typically based on studies that are themselves plagued, again, by the problem of representative sampling.

Jones writes of a 2007 study by L. A. Peplau and A. W. Fingerhut that has been relied upon in court cases, such as the California one in which Judge Vaughn Walker overturned that state’s Proposition 8, outlawing same-sex marriage. Peplau and Fingerhut, Jones writes, “typically launch into discussions about various characteristics of homosexual couples without ever clearly stating that the studies they cite do not examine representative samples” and “offer only intriguing hints that the studies on which they rely may be unrepresentative and hence potentially biased. They also raise in passing the provocative possibility that homosexual couples may bias their self-reports to look good.”

All of the problems Jones identifies in this earlier study apply to the much heralded National Lesbian Longitudinal Family Study. In an article at the Ruth Institute’s blog, Michael Worley observes that the study only involved people who were one of the 78 couples involved in the original study who both responded to advertisements (showing they wanted to be a public example of lesbian parenting) and conceived via insemination.  The latter condition “eliminate[ed] the vast majority of lesbian couples which have children from previous relationships” and also limited the study to people wealthy enough to afford such procedures.

Folks, what we have here is a very small, unrepresentative sample.

Worley addresses yet another aspect of this study.  He notes that the stability of the relationships even in this highly self-selecting study were not so stable. In 55% of cases the two women raising the child have separated by the time the child turns 17, yet the study reports that “adolescents whose mothers were still together and those whose mothers had separated did not differ on reported [Quality of Life].”  It’s a strange sort of “equality of life” in which the stability of a child’s parents’ relationship has no effect whatsoever on the outcome of the child’s development, isn’t it?

Stanton Jones notes that other studies show that same-sex relationships are much more unstable than heterosexual ones, particularly married relationships. Stability is, he observes, a concern of direct relevance, for instance, to adoption agencies involved in determining fitness for adoptive parenthood.

How does equivalence look in this area? Peplau and Fingerhut cite one study that found that over a five-year period, 7 percent of married heterosexual couples broke up, compared with 14 percent of cohabiting male couples and 16 percent of cohabiting lesbian couples. They also summarize, without mentioning specific numbers, a more representative study from Norway and Sweden, which have sanctioned same-sex partnerships since the 1990s, reporting “that the rate of dissolution within five years of entering a legal union is higher among same-sex partnerships than among heterosexual marriages, with lesbian couples having the highest rates of dissolution.” Their rendering underplays the magnitude of the actual findings, which was that gay male relationships are 50 percent more likely to break up than heterosexual marriages, while lesbian relationships are 167 percent [more than one and a half times] more likely to break up than heterosexual marriages.

Are the kids really all right?  Many of the researchers who are happily quoted by reporters from LiveScience are at pains to say that there is no evidence that anything is different.  One researcher, Michael Rosenfeld of Stanford, is quoted in another article concerning his examination of Census Bureau data that shows a nearly equal school retention rate (getting held back for a grade) for children of same-sex couples and heterosexual couples.  While this may be something, it doesn’t distinguish between statistics for children of cohabiting couples and children in families where Mom and Dad are married.  So, the problem remains. Even Rosenfeld admits that almost all of the studies he sees are done from unrepresentative samples:  “The problem from [a] statistical point of view is that convenient sample studies don’t amount to much.”

Advocates of easier divorce and single-parenting long argued that there was no difference in the outcomes for children whether they were raised by a mother and father, or just a mother.  Yet the long-term research has shown that this is simply not true, as the University of Chicago study on boys—sample size: 5000—has shown yet again.  And yet the push to affirm same-sex parenting goes on, with advocates citing studies with tiny sample sizes (National Lesbian Longitudinal Family Study: 78) that are clearly unrepresentative to convince us that, yes, indeed, the kids are all right.

When we talk about this subject it is important to be clear.  Same-sex couples raising children no doubt care for their children and strive to be good parents. And, there will always be a “success” example of a child of divorce, single-parenting or same-sex parenting to hold up for the world to see. I am truly glad that these individuals have fared well. But in reality, the claim that kids with same-sex parents have the same experience that kids with married moms and dads have, and that their lives are unaffected by their “separated” or fragmented family structure, is just not convincing.


N.B. The original research article on which the two First Things articles are based, including the footnotes to his claims, is found at CACE Hot Topic.  A reference page to the Same Sex Science article also can be downloaded.