SLATE | Mar. 18, 2013 

In a conservative movement seized by extremists, George Will is one of the sane ones. But his recent rejection of social science as having any role to play in the gay marriage debate is wildly off the mark. It's intellectually dishonest, scientifically ignorant, and—I'll say it—anti-gay.

Syndicated columnist George Will speaks in October 2002 in Bakersfield, Calif.

Photo By Christopher Ruppel/Getty Images

Will claims that reasonable people disagree about gay marriage “because so little is known about its consequences.” He quotes a legal brief by conservative scholars affiliated with a famously anti-gay think tank that calls research about gay marriage and parenting “radically inconclusive.” He then warns the Supreme Court—which will hear oral arguments on two gay marriage cases next week—to be wary “about social science that purports to prove propositions … for which there cannot yet be decisive evidence.” In other words, he suggests the value of research on gay marriage is currently zero.

Suppose a group of people claim that redheads can't enter the town square because they'll drive away commerce, badly harming the economy—and then this group gets a law passed barring redheads from public spaces. To reverse the discriminatory law, they then argue, redheads must spend however long it takes to amass definitive proof that entering the town square won't cause harm (which is impossible since you can't conduct research on scenarios you won't permit). When redheads nevertheless begin to produce a growing body of research that points conclusively to the fact that their presence does not harm commerce, the law's defenders consistently reply, “It still might; more research is needed.”

That is the position Will is defending in the gay marriage debate. The very idea that gay equality would cause serious harm is the kind of belief you hold if you're already inclined to think, usually for moral reasons, that homosexuality is harmful, that is, if you're anti-gay. Otherwise, there'd be no better reason to assume gay marriage is harmful than there is from letting redheads go to market.

Although the real basis of most opposition to same-sex marriage is moral, conservatives have found that arguments alleging harm to society are more powerful than ones that simply declare what they don't like to be morally wrong. So they have spent a generation making the case that gay equality would cause all kinds of social disruptions. It's exactly the same tactic that was recently proven profoundly wrong when the Pentagon lifted its ban on openly gay service despite continuous claims that doing so could destroy the military.

Advocates of equality have responded by conducting research on gay families that has consistently shown such fears to be ungrounded. George Will calls this body of work “spurious social science by supporters of same-sex marriage” and complains that it “purports to prove” something that can't yet be proven. But serious scholars don't claim that social science proves gay marriage is OK. It's the other way around: Anti-gay activists dream up harms that gay equality would cause and claim social science proves it—even though they've never supplied such proof. What the most responsible advocates of gay equality claim is that social science helps dismantle conservative claims of harm and strongly suggests the kids are OK. We use the research defensively, out of necessity, because anti-gay advocates make baseless empirical predictions that we're obligated to refute.

Will also gets the scientific method wrong, offering an impossibly high—and wholly incorrect—bar for what constitutes scientific proof, leaving no way for the pro-gay side to ever convince him that the bogus harms conservatives have dreamed up are unlikely to happen. He suggests that the “scientific standard” requires “large random samples with appropriate control samples” and concludes, quoting the conservative legal brief, that “there neither are nor could possibly be any scientifically valid studies from which to predict the effects of a family structure that is so new and so rare.”

But research doesn't have to use random samples to be scientific. The scientific method demands systematic analysis based on empirical observation. While large random samples are the gold standard, that doesn't mean that all other research is unscientific. It just lessens—somewhat—its predictive value. And bear in mind that even the largest research studies on public policy issues only offer predictions, not guarantees, about future outcomes. The difference here is one of degree.

What does the social science actually tell us? Statistics on marriage rates show consistently that they have not dropped in states and countries that allow gay marriage, as examined in Slate. But since gay marriage is so new, so are these results, and no one ever promised civilization would end in a day.

The bulk of the controversy surrounds research on gay parenting. At least 45 scholarly studies have compared children of gay and straight parents and found no disadvantages to the former. It's true that many have small, nonrandom sample sizes—often about 30-50 families. But when you aggregate 45 consecutive studies—even small ones—that all reach the same conclusion, what you end up with is an almost unheard-of scholarly consensus. “Rarely is there as much consensus in any area of social science as in the case of gay parenting,” said Judith Stacey, the New York University sociologist who is one of the deans of gay parenting scholarship.

In any event, not all the studies are small or nonrandom. A 2010 Stanford study used census data, instead of a convenience sampling or self-reporting, to examine 3,500 children of same-sex couples. It compared their school progress to more than 600,000 kids of straight parents and found no differences. Other studies are “longitudinal,” increasing statistical reliability by tracking developments over many years. The longest-running and largest such study, the U.S. National Longitudinal Lesbian Family Study, followed 78 lesbian-headed households for more than 25 years. Led by Nanette Gartrell, a UCLA-based psychiatrist and researcher, the study found no disadvantages (and a few advantages) to children with lesbian moms.

Is there any research showing disadvantages for kids with gay parents? Try as they might, conservative scholars, often funded by anti-gay think tanks, have failed to produce a single study. In the most embarrassing recent effort, University of Texas professor Mark Regnerus was formally reprimanded for publishing a study (which hewrote about for Slate) that he claimed had a massive sampling pool; in reality the number of children of same-sex couples he surveyed was a whopping two. (Last week it was reported that conservative funders bankrolled the study and urged Regnerus to rush it to maximize its influence on the Supreme Court.) Whatever you may say about the limits of the gay parenting studies—and all research has limits—the pro-gay research is currently winning, 45-0.

Absent actual evidence, the religious right has routinely used studies of single-parent and divorced households to allege that any family lacking a father—that is, even if it has two loving moms—is bad for kids. They cite research showing that two parents are better than one, and since all that research has focused on opposite-sex parents, they conflate number with gender and rally around the talking point that kids need “a mother and a father.” Appallingly, George Will has now stooped to this level, citing the allegation of conservative scholars that “research concluded that growing up without fathers had significant negative effects on boys,” even though that research never included households with two gay parents—and this after complaining about “inappropriate invocations of spurious social science” by liberals.

None of this should matter. Even if gay parenting did disadvantage kids, it wouldn't follow that gay marriage should be banned since gay people—like single and divorced people—will have kids no matter what. How could banning gay marriage help those—or any—kids? And if we based straight marriage rights on predicting durability, half the country wouldn't be allowed to wed.

But the research does matter, because conservatives have made it matter, and it may very well influence the Supreme Court. It's true we don't know everything about gay marriage and parenting, but what we do know is important to get right.