A notable feature of the legal debate over same-sex marriage in the United States in recent years has been the employment of social science research on the family as evidence. In courts of law across the country, advocates of same-sex marriage have advanced the “no differences” thesis—the view that having gay or lesbian parents makes “no difference” in the development of children into well-adjusted adults. For this proposition they have repeatedly relied on the claims of academic sociologists, psychologists, and other researchers. But as Ana Samuel explains in her recent Public Discourse essay, the findings of the New Family Structures Study led by Professor Mark Regnerus at the University of Texas (partly supported by a grant from the Witherspoon Institute, and published by Regnerus in the latest issue of Social Science Research), using data that for the first time meet the highest standards for research in social science, make it impossible to continue making such claims with any real credibility.

How much does this new study matter for the legal debate over same-sex marriage? A very great deal. Same-sex marriage advocates have argued in state and federal courts that traditional marriage laws have no “rational basis,” or that they fail some other more stringent form of “scrutiny” under constitutional provisions guaranteeing due process and the equal protection of the laws. Defenders of marriage have responded that one prominent basis for the conjugal definition of marriage is that the optimal form of family for raising children is the stable union of a married man and woman bringing up their own biological children. The ties of nature, the presence of both parents in the home, the modeling of manhood and womanhood for the next generation—in addition to a host of material benefits produced by stable relationships of this type—have been presented by defenders of marriage as among the reasons for preserving the status quo in law. With decades’ worth of research supporting the “two-parent family ideal” (a phrase used by sociologists Sara McLanahan and Gary Sandefur) as generally better for children than being raised in families experiencing divorce, or headed by merely cohabiting couples, or by single mothers, or in other variations of broken-and-mended families, the argument would seem to have a lot going for it.

Yet a significant number of social scientists have come forward—in the public press, in legislative hearings, and in courts of law—to suggest the unlikely conclusion that the one form of family that equals or approaches the “two-parent family ideal” of a married man and woman is the family headed by a same-sex couple. The American Psychological Association, in an official publication in 2005 intended to influence the legal debate, claimed that “the evidence to date suggests that home environments provided by lesbian and gay parents are as likely as those provided by heterosexual parents to support and enable children’s pyschosocial growth.” But as Professor Loren Marks of Louisiana State University points out in his own separate contribution in the latest issue of Social Science Research, the studies on which these claims have been based are all seriously flawed. Marks demonstrates that they have defects respecting “sample size, sampling strategy, statistical power, and effect sizes”—and in general that they lack methodological rigor.

For years, however, courts of law in significant same-sex marriage cases have been bombarded with the alleged findings of social science that the “no differences” claim is sound. Some of the courts in the early cases (in Massachusetts and Connecticut, for instance) were circumspect, not relying heavily on the “no differences” claim, or skirting the issue by claiming that it is somehow “irrational” for legislatures to permit gays and lesbians to adopt children but not to enter into same-sex marriages. But repetition of the “no differences” claim eventually paid off. In the 2009 decision of the Iowa Supreme Court, Justice Mark Cady accepted the proposition that “sexual orientation and gender have no effect on children raised by same-sex couples, and same-sex couples can raise children as well as opposite-sex couples.” Citing multiple “leading organizations” in the study of child welfare, Justice Cady took the “no differences” claim as established by science. While he accepted that those who believe “dual-gender parenting is the optimal environment for children” are “thoughtful and sincere,” he declared that this view was “largely unsupported by reliable scientific studies.”

Start your day with Public Discourse

Sign up and get our daily essays sent straight to your inbox.

Judge Vaughn Walker of the U.S. district court in San Francisco, in his 2010 decision overturning California’s Proposition 8, was the most enthusiastic promoter of the “no differences” claim, returning to this point repeatedly in his opinion. Relying heavily on testimony by psychologist Michael Lamb, Judge Walker announced as a “finding of fact” that “children raised by gay or lesbian parents are as likely as children raised by heterosexual parents to be healthy, successful and well-adjusted. The research supporting this conclusion is accepted beyond serious debate in the field of developmental psychology.”

Thanks to Loren Marks, we know that there is indeed “serious debate” among scholars about the quality of the research purporting to establish the “no differences” claim. And thanks most importantly to Mark Regnerus, we have the first large, national, probability-sample-based study examining whether this claim withstands close scrutiny with methodological rigor. It does not.

Regnerus’s ground-breaking peer-reviewed article, “How Different Are the Adult Children of Parents Who Have Same-Sex Relationships? Findings from the New Family Structures Study,” compares eight different kinds of family structure, including those in which a child was raised by a parent who had had a same-sex romantic relationship. Regnerus finds that the “intact biological family,” headed by a man and woman married to each other throughout the raising of their own children, is indeed the “most secure environment for child development.” Children raised by a parent who had been involved with someone of the same sex are, to a significant degree, more likely to record, in their own estimation in interviews, a variety of negative outcomes related to their upbringing. Regnerus is careful to make no claims of causality, but the controls in the study are careful ones and the correlations remain strong. What Iowa’s Justice Cady said in 2009 was lacking from the case for “dual-gender parenting”—namely “reliable scientific studies” supporting that case—is lacking no longer. (For a fuller summary of the research and its findings, see Ana Samuel’s article.)

The breakthrough character of this new research can be seen in the responses of three prominent scholars asked to comment on it by the editors of Social Science Research. Paul R. Amato and David J. Eggebeen, both of Penn State, and Cynthia Osborne, of the University of Texas, all praise the New Family Structures Study for advancing our knowledge of child-rearing outcomes from various kinds of families, including those headed by parents who had had same-sex relationships. (Amato was consulted on the NFSS’s initial design, and Osborne is a member of the team led by Regnerus.) Their criticisms of Regnerus’s interpretation of the new data are few and mild, and all of the commentators praise Regnerus for having made no firm claims of his own about causality. Yet all of them openly speculate—with no data for support of their speculations—about the possible causes of the markedly different, often negative outcomes among the children of parents who had been involved with others of the same sex.

In essence, all three commentators want to change the subject. Most notably, two of them would like to halt the use of social science findings in the same-sex marriage debate in courts and legislatures across the country. Given their stated sympathy for the cause of same-sex marriage, and their apparent failure to supply this cautious counsel in the past, this can only be because they know how much the Regnerus study shifts the ground of debate. It is no longer possible to claim that the best social science shows “no differences” between heterosexual and homosexual parenting outcomes. Therefore it must be time for a truce on the use of social science in public and legal debate!

Amato, for instance, writes:

It would be unfortunate if the findings from the Regnerus study were used to undermine the social progress that has been made in recent decades in protecting the rights of gays, lesbians, and their children. . . . [T]he legality of same-sex marriage is a constitutional issue and not one that should be decided on the basis of social science research. . . . Opponents as well as proponents of same-sex marriage have produced a parade of social scientists to provide testimony about the quality and stability of same-sex unions and the adjustment of children from these unions . . . Too much attention has been given to this evidence.

Osborne similarly remarks:

Given the sensitive political and cultural environment surrounding same-sex relationships, preliminary findings of “differences” such as those reported by Regnerus should not be used to support punitive legislation aimed at limiting the family formation and fertility choices of gays and lesbians.

Eggebeen does not call for a halt on using social science research in the marriage debate. But, rather desperately speculating on the causes of the differences Regnerus reports, he seems unable to credit the possibility that anything but extra-familial factors could be at work: “Parents, regardless of sexual orientation, are equally motivated to provide the best care possible for their children.” Even if we were to concede the absolute truth of this surprisingly axiomatic statement, it is not the motivation of parents, but the results of their parenting that matter. Eggebeen’s certainty that family structure cannot itself be any part of the explanation is breathtakingly unscientific.

It might be tempting to say, “It’s a deal.” All individual social scientists, and all scientific professional associations, could declare publicly that they have nothing meaningful to contribute to the debate over same-sex marriage in courts and legislatures and popular referenda. No more testimony, no more legal briefs purporting to represent any “established knowledge” or “consensus of the experts” on the question of parenting outcomes by unconventional new family structures.

Such a truce would strongly support the preservation of the historic norms of marriage. For democratic decision-makers would clearly have a “rational basis” for the status quo—a basis the challengers cannot refute. This was clearly understood by the three dissenters in the 2003 Goodridge decision of the Massachusetts high court—the first state supreme court ruling inventing a constitutional right to same-sex marriage. As Justice Martha Sosman said in her dissent, “the attempts at scientific study of the ramifications of raising children in same-sex couple households are in their infancy and have so far produced inconclusive and conflicting results.” Justice Robert Cordy dwelt on this matter at greater length, concluding:

That the State does not preclude different types of families from raising children does not mean that it must view them all as equally optimal and equally deserving of State endorsement and support. For example, single persons are allowed to adopt children, but the fact that the Legislature permits single-parent adoption does not mean that it has endorsed single parenthood as an optimal setting in which to raise children or views it as the equivalent of being raised by both of one’s biological parents. The same holds true with respect to same-sex couples—the fact that they may adopt children means only that the Legislature has concluded that they may provide an acceptable setting in which to raise children who cannot be raised by both of their biological parents. The Legislature may rationally permit adoption by same-sex couples yet harbor reservations as to whether parenthood by same-sex couples should be affirmatively encouraged to the same extent as parenthood by the heterosexual couple whose union produced the child.

Just so. When the social science is at best highly uncertain, the default position of the judiciary in constitutional cases should be that the burden has been met by traditional marriage laws as having an eminently “rational basis.” The argument would then turn to other traditional canons of legal reasoning: the force of precedent, and the authority of legislatures and other democratic decision-makers to settle basic moral norms for society.

As I say, after the publication of Loren Marks’s critique of past research, and especially of Mark Regnerus’s impressive new research, it could be tempting to seize the implied offer of their rather shell-shocked critics and agree that social scientists may be excused from any further role in the marriage debate.

But then again, why agree to a truce when one is winning an argument? For the plainly rational basis of traditional marriage laws is strongly supported by the studies of Professors Marks and Regnerus. As the Perry case on Prop 8, and related litigation on the Defense of Marriage Act, reach the Supreme Court, the counsels of good social science can be added to the standard norms of constitutionalism to counsel against the willful judicial invention of a right to same-sex marriage.