The gay-parenting research they hope you never see

They said the kids would be fine. Here's the gay-parenting research that got buried.

A peer-reviewed study reported that children with married same-sex parents had more than four times the odds of reported emotional difficulties as kids with a married mom and dad.

There's a lot of mischief being done very quietly in the weeds of what seems like respectable scientific studies.

Oh, wait. You never heard that? Yeah, I know you never heard that.

The research that produced this finding got buried in the academic literature. And today I'm going to show you how easy it was for that correction to escape public notice — and why it matters for real families like yours.

I suspect the pro-gay establishment hopes you don't see this.

Meet Chris from Ohio

I want you to imagine somebody who really needs reliable answers.

Let's imagine a 20-something young lady. I'll call her Chris. Chris from Ohio.

Chris thinks of herself as bisexual or nonbinary. In her mind, she has a real choice between partnering and co-parenting with a person of the opposite sex or a person of the same sex.

Her decision-making is going to partly depend on what she thinks it would be like to be a mom with another woman versus being a mom with a man.

So let's keep Chris in the back of our minds. And if you're reading this because someone you love is in this situation, or because you've lived it yourself, I want you to know this article is for you.

The study behind the controversy

Now, back to the study. The study reporting that children with married same-sex parents had more than four times the odds of emotional difficulties compared with children of married opposite-sex parents. That doesn't mean Chris' own future children would be four times as likely to have problems, but it's certainly a finding she might want to know about.

This whole research saga started with a paper published in 2016 in the peer-reviewed journal Demography that seemed to show the kids are all right with same-sex couples as parents.

Let's call this Study No. 1.

The researchers expected the children of same-sex married couples would look pretty much the same as the children of opposite-sex married couples. And they found what they were looking for.

This study showed no statistically significant difference in emotional difficulties between kids raised by same-sex couples and kids raised by their married mom and dad. The researchers did report a disadvantage for children with married same-sex parents on a separate behavioral-problems measure, but the emotional-difficulties finding showed no significant difference.

But in 2017, one short year later, the very same researchers analyzed a corrected and expanded version of the same dataset and found the result I already mentioned: Children of same-sex married couples had more than four times the odds of emotional difficulties.

We'll call this Study No. 3.

I'm going to show you exactly how this happened.

What the numbers actually say

Okay, so let's look at some numbers. I'll highlight which figures to look at and what they mean.

For Study No. 1, showing no differences, look at Table 5.

The researchers report what are called odds ratios. You use this method when you're trying to compare how likely something is between two or more groups.

You start with a reference category, and all the other groups are compared to that reference category.

In this case, the reference category is children raised by different-sex married couples. In other words, kids raised by ordinary married moms and dads.

Now, if the odds ratio equals 1, that means the two groups have the same odds of the outcome.

In this case, the researchers wanted to say children raised by married couples are equally likely to have emotional difficulties whether their married parents happen to be the same sex or the opposite sex.

And that's essentially what we see in the first line of Table 5. The estimated odds ratio is close to 1, with no statistically significant difference.

This is the no-differences result.

If, on the other hand, the odds ratio is greater than 1, that means the group has higher odds of emotional difficulties than the reference group.

So let's look at the results the very same set of researchers found using the corrected and expanded data in Study No. 3. That would be Table 7.

Doing the math

Now, stay with me here, because what I'm about to show you is something I suspect the academic establishment hoped you would never notice.

Look at the column labeled emotional difficulties odds ratio. For kids raised by married same-sex couples, the odds ratio is 4.16.

That means the estimated odds of reported emotional difficulties were more than four times as high as for children of opposite-sex married couples.

Odds aren't quite the same as probabilities, and the small sample means we should be cautious about exactly how much weight to put on that figure. But it's a substantial change from the original no-differences result.

So now let's remember Chris from Ohio for a moment.

She may have heard about studies like the first one, saying her potential kids would be fine whether she marries a man or a woman.

She probably didn't hear too much about this correction. You would think she has a right to know just how fragile that no-differences result really was.

So how did the researchers get from no significant difference to more than four times the odds in under a year?

From no differences to twice the odds

Now, here's where the Ruth Institute comes into play. We had a little role to play in the background.

If you guessed that there was a Study No. 2 between Study No. 1 and Study No. 3, you win the prize.

Study No. 3 actually came about after some back-and-forth with a critic.

This critic pointed out that the National Center for Health Statistics had issued an advisory about the dataset. Some of the data had been miscoded.

Specifically, the problem involved the recorded sex of married partners. Some opposite-sex married couples had been incorrectly classified as same-sex couples. Because genuinely same-sex married couples were such a small proportion of the population, even a relatively small number of miscoded records could seriously distort the findings.

The NCHS had advised researchers against using the affected survey years to analyze same-sex married couples. But Study No. 1 had included those years.

So the critic reclassified the data, excluding the problematic years. And his reclassification obliterated the original estimate showing no difference in emotional difficulties, although the replacement estimate remained statistically uncertain.

In fact, the children of same-sex married couples had approximately twice the estimated odds of emotional difficulties. The difference wasn't statistically significant in that small sample, but it was a substantial change from the original estimate.

You can see the results in Table 3 of Study No. 2. The panel excluding the miscoded data shows the odds ratio for same-sex married couples is now about 2 instead of 1.

The author of Study No. 2 corrected what appears to have been an honest mistake. So fair enough.

This is how science is supposed to work: back and forth. Everybody's work can improve from this type of give and take.

But still, let's remember Chris from Ohio, trying to decide what to do with her life. Do you think she heard about this miscoded data?

Probably not.

The biological parenthood problem

Anyway, we still haven't gotten to the statistics showing more than four times the odds of emotional problems.

This critic pointed out another problem in the original study, Study No. 1. The authors used biological parenthood as a control variable.

Now, wait a minute. Let's think about this for a minute.

If the focus of the study is the difference between same-sex parents and opposite-sex parents, you really have to be careful about controlling for biological parenthood.

No child of a same-sex couple will have two biological parents in that couple.

So the variable for biological parenthood may be soaking up variation that is part and parcel of being a child with same-sex parents.

That's the critic's argument. Researchers may reasonably want to separate the effects of biological relatedness, family stability, and parental sex. But depending on the question they're trying to answer, controlling for biological parenthood can also remove an important part of the difference they're studying.

The author of Study No. 2 argued that the biological-parenthood variable should be removed and the analysis redone.

By taking out biological relatedness, the estimated association between living with married same-sex parents and emotional difficulties becomes a whole lot stronger than you originally would have thought.

This is exactly the sort of thing I mean when I talk about tricks with data. There's a lot of mischief being done very quietly in the weeds of what seems like respectable scientific studies. And I suspect the academic establishment is hoping you'll be too lazy to figure that out.

RELATED: The 3 failed promises of gay 'marriage'

Samuel Corum/Getty Images

Out of cite

Now, who was this mystery critic, you might ask? The critic whose name I haven't yet mentioned?

The critic the authors referred to when they wrote, in their response, that they were conducting an analysis in response to commentary by Sullins?

Who is this mysterious Sullins person? It is none other than Father D. Paul Sullins, Ph.D., my colleague at the Ruth Institute.

Father Sullins has been a senior research associate with us since 2018. When he wrote this paper, he was still a sociology professor at the Catholic University of America.

But we'll take a bit of credit, though, because we had the good sense to snag him when he retired.

Anyhow, notice something about the sentence referring to Father Sullins in the authors' revised paper. It does not give the customary parenthetical citation with his name and the publication date of the article they're talking about.

When I first saw this, I thought, "What?"

I scurried back to the references page looking for the full citation to a paper by Sullins. I was shocked to find that his paper, the one to which they're responding, is not even listed in the references.

Well, that is not only bad science, that is bad manners.

How rude.

Anyway, at least Father Paul got them to redo their calculations. But if somebody wanted to put these papers side by side, the way we've done here, it would not be easy to do.

I'll let you decide for yourself whether you think this is an oversight.

Unsettled science

The academic establishment is committed to the no-differences viewpoint.

You can see why I think so in that missing citation. You can see it in the fact that studies showing no differences get plenty of media airtime, while corrections and contrary findings stay quietly hidden in academic journals.

Those are facts readers deserve to know.

And that brings us back to Chris from Ohio.

Perhaps someone like Chris is reading this today. Perhaps someone reading knows somebody who learned from experience that same-sex parenting was not as easy as you had been led to believe.

I want to say to any of you in that situation: I am sorry. This should not have happened to you.

And I want you to know today that you're probably not the only person going through this. The science about same-sex parenting is not nearly as settled as you've been led to believe.

Please share this article. You'll be doing something the academic establishment too often fails to do. You'll be giving Chris from Ohio the information she actually needs and deserves.