02 July 2012

Sloppy science reporting on spanking

Just for the sake of disclosure, I'm going to make my stance on spanking children clear at the outset: I'm against it in all forms and all circumstances. There's no evidence that it's effective, and plenty of evidence that it could be harmful. Spanking is fine among consenting adults, and that's pretty much it. Perhaps the only reason it persists is because of dogmatism that holds the Biblical adage "spare the rod, spoil the child" above actual scientific research. The American Academy of Pediatrics is abundantly clear on spanking:
The American Academy of Pediatrics strongly opposes striking a child for any reason. If a spanking is spontaneous, parents should later explain calmly why they did it, the specific behavior that provoked it, and how angry they felt. They also might apologize to their child for their loss of control. This usually helps the youngster to understand and accept the spanking, and it models for the child how to remediate a wrong. [1]
Not that science has ever deterred Bible-thumpers ("the AAP is a bunch o' liberals!" I can hear them crying, stupidly).

BUT....

An article posted on Yahoo! today shows some sloppy science reporting that's reporting on some sloppy science. The headline is sufficiently alarmist: Spanking Linked to Mental Illness, Says Study.

Wait a second though. What to they mean, "linked to"? Has a study established a causal relationship here, or is it merely correlational? Let's find out!
Researchers examined data from more than 34,000 adults and found that being spanked significantly increased the risk of developing mental health issues as adults.
Ah. So, it's an epidemiological study. That means the data is correlational, not causal. Such studies can be highly misleading – they can have hundreds, even thousands of correlations. But we have to guard against the fallacy of cum hoc, ergo propter hoc – correlation does not prove causality. The article continues:
According to their results, corporal punishment is associated with mood disorders, including depression and anxiety, as well as personality disorders and alcohol and drug abuse. They estimate that as much as 7 percent of adult mental illness may be attributable to childhood physical punishment, including slapping, shoving, grabbing, and hitting. The study reports that spanking ups the risk of major depression by 41 percent, alcohol and drug abuse by 59 percent, and mania by 93 percent, among other findings.
Since the data is correlational, it's possible that spanking doesn't actually cause or even directly contribute to any form of mental illness. It could be that people who are raised by parents who themselves suffer from mental illness and/or addiction are more likely to be spanked, but that the actual mental illness is the result of genetics and/or other environmental factors.

The article is also a little fudgy on what, exactly, is meant by "spanking". Spanking, like many facets of child discipline, exists on a spectrum of severity and frequency that is thick with gray.
"We're not talking about just a tap on the bum," study author Tracie Afifi, PhD, of the University of Manitoba in Winnipeg, explained in a statement. "We were looking at people who used physical punishment as a regular means to discipline their children." However, the analysis excluded individuals who reported more severe maltreatment such as physical abuse, sexual abuse, emotional abuse, physical neglect, emotional neglect, or exposure to intimate partner violence.
So this becomes critical: even if a causal link could be established between spanking and mental illness, where along the spectrum of severity does spanking's impact cross from "trivial" to "problematic"? And note that the study's authors aren't solely referring to spanking, but also to "slapping, shoving, grabbing, and hitting". That just makes the article's headline, and the researchers' conclusions, even more dubious.

Epidemiological studies have their place, but we have to be careful not to draw direct conclusions from them. If anything, this study is little more than a stepping stone that says we ought to investigate the relationship between mental illness and spanking – if there is one – a bit more closely.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.