Mark Regnerus Did Not Kill the Dragon

michael and dragon

(Not Mark Regnerus)

**Before you read this: Please read Jim Burroway’s and/or John Corvino’s excellent analyses of Mark Regnerus’ new study in Social Science Research**

We grad students love Karen Kelsky at  She gives hard-nosed advice that we hope will help us get jobs in this ridiculous academic job market.  She says we need to sell our research; and to do this, she says we have to present our research as a story.  Specifically, a hero narrative: our research question is a dragon threatening the village; our task is to tell the villagers 1) why that dragon is a threat (why our research matters) and 2) why we are uniquely suited to slay that dragon.  Why we’re the heroes, in other words.

I bring this up because the hero narrative offers a nice lens for looking at Mark Regnerus’ New Family Structures Study, an overview of which appeared this week in Social Science Research.  The study is being used to suggest that children of gay parents are at greater risk for a number of negative outcomes than their peers raised by straight parents.  Anticipating controversy, Regnerus set up a Q and A on his blog, and wrote this post for

For Regnerus, the dragon is what he calls the “no differences” hypothesis on gay parenting—the idea that parents’ gender has no meaningful difference on their ability to raise a child. The idea, in other words, that all other things being equal, a gay couple can raise a child as well as a straight couple. This notion, he writes, dominates our political debates, our facebook and email conversations, even our legal depositions on the subject of gay marriage.  And Regnerus wants to test it.  Killing the dragon, for Regnerus, means finding out whether that hypothesis holds water.

Of course, other scholars have tested it before, and found it to be sound.  So many in fact, and so conclusively, that the American Psychological Association stated in 2005 that “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.”

But Regnerus points out, with justification, that these studies have all relied on small data samples, or vague census data, or problematic convenience samples.  They may have wounded the dragon, but they haven’t really vanquished it—their work leaves open too many questions.

Regnerus is uniquely suited to slay this dragon, on the other hand, because he has a shiny new sword: three quarters of a million dollars, funding enough to get a random, national sample of adults raised by gay parents.  Compared to this new, expensive weapon, all the would-be heroes that came before only had crude homemade spears.

And all that may be so.  The problem is that unlike Regnerus, those previous scholars actually attacked the dragon.

As I said above, the relevant question in debates about gay marriage or gay parenting is: How does gender affect a couple’s ability to raise healthy children?  All other things being equal, can two men or two women raise a child as well as a male/female pair?

The variable to isolate is gender—not family stability, not income, not sexual behavior.

Because we all know that family instability causes negative outcomes in children.  As does poverty, as does single parenthood.  And we know that adopted children fare worse on standard measures than do children who grew up with their intact biological families.  None of that is under debate.  To say anything relevant at all in the gay marriage debate, we have to make appropriate comparisons of outcomes across comparable groups.  How do children raised from birth by a committed pair of women, for example, fare when compared to children raised in similarly stable heterosexual environments?  And how do children who face divorce and then live with a same-sex stepparent fare next to those who live with opposite-sex stepparents?

As Burroway outlines, Regnerus doesn’t do this.  Instead, he lumps anyone whose parent had a homosexual relationship—of any duration, at any point—into one category.  That category includes children whose parents divorced, children raised by single parents, adopted children, etc.  And then he compares those kids to peers raised in intact biological families.  And, unsurprisingly, he finds that the kids in the latter group do better.  To use Corvino’s phrase, “well, duh.”

Regnerus has defended his methods online by writing that it would have been impossible to maintain a randomized, generalizable sample while dividing his subjects into appropriate comparison groups*.  That’s totally understandable.  But then he shouldn’t position his paper as the next word in the conversation about gay marriage.  And he definitely shouldn’t claim that his paper is better than previous studies at answering the relevant questions in that debate**.  It’s worse.  Way worse.  It’s a more expensive version of the Sarantakos 1996 Australian study, which the APA reviewed and dismissed in its 2005 statement in support of gay parenting, in part because in it “most or all of the parents being raised by gay or lesbian parents, but not the children being raised by heterosexual married parents, had experienced parental divorce.”  It was an invalid comparison then, and it still is.

Basically, Regnerus set out to slay the dragon in a shiny new set of armor, bearing his shiny sword.  Ironically, the sword proved to be a problem—every time Regnerus approached the dragon, the dragon saw it shining in the distance and flew to safety.  But instead of dropping the damn sword and picking up a crude spear, Regnerus gave up and decided to use his fancy sword to beat up a cardboard facsimile of the dragon.

There’s nothing wrong with that, per se.   The dragon is hard to catch.  And Regnerus is careful in his writing not to claim outright that he killed the actual dragon.  So while “Can Mark Regnerus beat up a cardboard dragon?” is not a particularly interesting question, his research does provide the answer, and that’s all that he reports, and, hey, maybe that info will prove useful somewhere down the road.

But his Slate piece, especially, is the equivalent of taking a picture with the severed (cardboard) head of his fake dragon, and then passing that photo around the village pub while saying, “Remember that dragon?  Let’s just say he won’t be bothering this village anymore.”

And we all know that a certain segment of the villagers will be buying Regnerus beers for years to come.

*He writes: “And I’m telling you that it cannot be feasibly accomplished. It is a methodological (practical) impossibility at present, for reasons I describe: they really didn’t exist in numbers that could be amply obtained *randomly*. It may well be a flaw–limitation, I think–but it is unavoidable. We maxxed Knowledge Networks’ ability, and no firm is positioned to do better. It would have cost untold millions of dollars, and still may not generate the number of cases needed for statistical analyses. If randomness wasn’t the key priority, then we could’ve done it. And we’d have had a nonrandom sample that was no better than anything before it.” 

**From the article: “So why did this study come up with such different results than previous work in the field? And why should one study alter so much previous sentiment? Basically, better methods.”  GAH!!!!!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s