What does social science prove about gay marriage?

David French says that liberals are so committed to gay marriage that they'd be in favor even if it demonstrably harmed children of gay marriage:

"There could exist definitive social science that homosexual families produce — on average — worse outcomes for their children than heterosexual families, and the fervor of the gay-marriage advocates would be undimmed. After all (and like no-fault divorce), the case for gay marriage has never been about the welfare of children, but instead, the fulfillment of adults.  "

At risk of saying, "I know you are but what am I?": Does anybody really think that the mass of social conservatives would drop their opposition to gay marriage even if definitive proof existed that children did better in gay families? I think the mass of opposition to gay marriage is rooted in religious beliefs—people believe it to be morally wrong—and field research probably isn't going to persuade them otherwise. The emphasis on "the welfare of children" is the fighting ground mainly because it offers a secular reason to oppose gay marriage—though advocates undoubtedly believe it to be true, because they believe gay marriage is morally wrong. But if the child welfare argument weren't available to them, they'd find another objection. We all have our predispositions, but contra French, liberals aren't any more committed to theirs than conservatives are.