Scotland may ban spanking. Decades of research suggests that’s a good idea.

Last week, a member of the Scottish Parliament introduced legislation that would make spanking a crime on par with assault. And the Scottish National Party, which controls the Scottish Parliament, has pledged to make it law sometime early in 2018, Scotland’s national newspaper reports.

When that happens, Scotland will join a growing list of counties, states, and municipalities that have made hitting children illegal. Since Sweden became the first country to ban corporeal punishment in 1979, 51 other countries have adopted similar laws. (The United States is not among them.)

Current law in Scotland (and in the UK at large) allows for “reasonable chastisement” (yes, it’s vague) when it comes to disciplining kids. This new legislation would make Scotland the first UK country to turn that gray area into a fine black line.

“Giving children equal protection against assault will send a clear message to all of us about how we treat each other and underpin Scotland’s efforts to reduce violence,” John Finnie, the member of the Scottish Parliament who introduced the law, told reporters.

Opponents say that spanking — generally defined as hitting a child with an open hand on the behind — leads to a slippery slope of abuse and can cause emotional and behavioral problems down the line. The question of whether mild spanking definitely causes problems is hard to answer. (You can’t exactly run an experiment that involves hitting children.)

Here’s the state of the science. There’s not a lot of evidence that spanking provides a benefit to kids. And there’s some (controversial and contested) evidence that it can actually lead to harm.

In a 2016 meta-analysis in the Journal of Family Psychology, psychologists Elizabeth Gershoff and Andrew Grogan-Kaylor sifted through 75 studies on spanking, excluding research that analyzed other forms of physical punishment and abuse, such as hitting children with objects. The studies employed either cross-sectional (i.e., snapshot) or longitudinal (over the span of several years) designs. In all, the resulting data pool contained information on 161,000 children. It found “no evidence that spanking is associated with improved child behavior.”

The analysis found evidence that spanking is associated with troubling outcomes — like increased aggression, increased antisocial behavior, and mental health problems later in life. According to Gershoff, spanking “makes it clear to the children that you can hit somebody if you have power.” And that lesson sticks with a kid, making them act out later in life.

There are a lot of big caveats to mention here.

The size of these negative effects is small, the study notes, and there’s no proof that spanking specifically caused these behaviors later in life. So it’s hard to tell if spanking leaves a lasting scar.

And the evidence in the broader literature is mixed. A 2013 meta-review on corporal punishment in kids found “trivial” correlations with negative outcomes. But a 2017 analysis in the journal Child Abuse & Neglect found in a correlational study that spanking was linked to similar adverse effects as childhood emotional abuse.

The research is not perfectly clear because behaviors like spanking are some of the hardest for social scientists to study and draw concrete conclusions from.

For one, researchers can’t conduct spanking experiments where they randomly assign children to be hit. They also have to lean on questionnaires, where parents or children have to remember how punishments were administered.

There’s also the case that “bad” children are just spanked more, and are also generally more aggressive and antisocial throughout their lives. And even though this meta-analysis singled out studies only about spanking, it’s still possible the parents in these studies were also using harsher forms of punishment (like hitting children with objects) and didn’t mention it to the researchers.

Overall, when studying real-world behaviors like spanking, it’s very hard, if not impossible, to truly isolate the variable and study its absolute effect on behavior.

“Gershoff and Grogan-Kaylor’s most recent meta-analysis relies on correlational evidence that would be considered woefully inadequate in any other scientific field,” the American College of Pediatricians wrote in an scathing editorial bashing the meta-review. They said that “studies would be rejected if they were being used to halt a medical procedure, such as chemotherapy for combating cancer.” To reiterate: You just can’t study spanking like you could a drug.

The editorial noted only four of the studies limited the definition of spanking to “two open-handed swats to the buttocks for child defiance.” The other studies left the term “spanking” a bit more open to interpretation: They could mean smacks anywhere on the body. Then again, parents who hit their children might not be mindful of these differing definitions, or may use them interchangeably.

Bob Larzelere, the Oklahoma State University family science researcher who co-authored the College of Pediatricians editorial, says in an email that there are a few studies that show something called “conditional spanking,” where children only get spanked if they ignore another mild discipline (like a time out) may be effective in getting children to cooperate. Though the research on this disciplinary technique is unclear as to whether mild “conditional” spanking leads to a slippery slope of increasing violence, or whether it has long-term consequences for a child’s well-being.

At the very least, the 2016 meta-study makes a case that spanking doesn’t lead to any long-term measurable good. Only one study of the 75 studies found an effect linking spanking to a positive life outcome. “If — in the real world — spanking was good for kids, some of these studies should have found that,” Gershoff told Vox in 2016.

In 2016, I spoke with Gershoff in depth about her research. And I asked her why we assume spanking could be a helpful disciplinary strategy. “One is parents think it works,” she said. “And they think it works because it gets an immediate reaction out of the child. Immediately the child cries. The parent goes “Aha! they understood that I am mad.” That’s gratifying to the parent, so the parent is rewarded by getting this reaction out of the child.”

The other reason: Parents themselves were spanked themselves as children, and are continuing the practice. But Gershoff didn’t think parents today should necessarily feel guilty for spanking their kids.

“We know now that children need to be in carseats and seat belts,” she said. “But those of us who grew up in the 1970s were in cars that didn’t even have seat belts. Do I think my parents were bad parents for not putting me in a seat belt? No, because no one understood how important seat belts were.”

Even if the effect of spanking is null — not positive or negative — why should (however mildly) violent actions be condoned? The world over, children are routinely the subjects of violent abuse; the United Nations finds domestic violence is a leading cause of children’s deaths worldwide. And many argue that spanking is the first step on a continuum of violence. If a hand slap is good, why isn’t a rod better? It’s a dangerous slippery slope.