So, you’ve read the headlines today: a brand, spanking new study has found that doing away with traditional marriage laws has reduced suicide rates among gay teenagers by a whopping 14 percent in states that have embraced same-sex marriage! Here are just some of the glowing headlines:

  • PBS Newshour: “Same-sex marriage laws linked to fewer youth suicide attempts, new study says"
  • Washington Post: “Legalizing same-sex marriage was associated with fewer youth suicide attempts, new study finds"
  • The Guardian: “Drop in teenage suicide attempts linked to legalisation of same-sex marriage"
  • USA Today: “Study: Teen suicide attempts fell as same-sex marriage was legalized”

These headlines, which are just a tiny sampling of the blanket media coverage this study has received, obviously suggest that if conservatives would just drop their goldarned old fashioned opposition to same-sex marriage, kids would stop killing themselves.

The study, printed in JAMA Pediatrics, concludes that same-sex marriage policies “would be associated with more than 134,000 fewer adolescents attempting suicide each year.”

There is only one problem: the study doesn’t really prove that.

The study shows a basic correlation between loosening of same-sex marriage law and suicide rate, but it does not show that same-sex marriage policies “reduced adolescent suicide attempts,” as the study’s conclusions state.

In fact, the study itself acknowledges, “our analysis does not allow us to understand the mechanisms through which implementation of same-sex marriage policies reduced adolescent suicide attempts.” In other words, they say that they know that same-sex marriage policies impacted suicide attempts, but they didn’t know how. Well, then.

They also fail to rule out some of the most basic confounds for any sociological study like this: socioeconomic status of the students themselves. Is it possible that the students are committing suicide at a lower rate because their parents are making more money, for example? Is it possible that socioeconomic status also has outsized impact on particular areas? Or how about the social acceptance of gays and lesbians without the legislative question? The study says they couldn’t check either of those confounds out:

We also could not control for unmeasured individual-level characteristics, including socioeconomic status, or for unmeasured state characteristics that may change over time, such as religious affiliation or acceptance of sexual minorities.

Also, the study fails to explain why non-LGBT students would see their suicide rates decline in states that approve same-sex marriage. How exactly does that work?

The research data shows variability in suicide rate in states over time, as well. Why would the suicide rate drop precipitously between 2005 and 2007 among states that only legalized same-sex marriage in 2013 or 2014? The study doesn’t explain.

The study also averages out state data rather than showing a serious trend common to all states. That means that a serious decline in suicide rate in Hawaii could wash out an increase in suicide rate in Delaware. Look at the charts. They’re pretty messy. There’s a reason for that.

The study states, “The analyses on the association between implementation of same-sex marriage policies and adolescent suicide attempts among those identifying as sexual minorities should be interpreted with caution given the limited data availability on sexual orientation.” Yet their conclusions are bold:

We provide evidence that implementation of same-sex marriage policies reduced adolescent suicide attempts. As countries around the world consider enabling or restricting same-sex marriage, we provide evidence that implementing same-sex marriage policies was associated with improved population health. Policymakers should consider the mental health consequences of same-sex marriage policies.

So much for caution.

Here’s the bottom line. This study shows correlation, not causation, and openly ignores basic data that would be necessary in order to rule out confounding factors. It also averages data in order to draw conclusions that truly are state-specific. The study acknowledges its limitations, then draws extraordinarily strong conclusions. And the media runs with it.

This is how cautious science becomes utterly uncautious conventional wisdom. Maybe same-sex marriage lowers suicide rates. Maybe it doesn’t. But this study doesn’t prove it one way or another.