Thursday, February 1, 2007

Experiments in Punditry


I recently appeared as a talking head on an hour-long cable news show concerning a bill introduced in California that would ban spanking. I was asked if I would make some points as to the limitations of the bill, which I agreed to, so long as I could clarify that I was anti-spanking and that there are many alternatives to corporal punishment that are more effective and less harmful.

But then what problem do I have with a bill banning spanking? Well, basically I believe that an ounce of prevention is worth a pound of cure. If we as a society really care about children, we should teach children how to be good parents – by being good parents ourselves, and by developing programs that teach pregnant women and their partners and families about the challenges of raising children, and how to deal with all the stresses inherent in being a parent. To intervene AFTER the spank has occurred, as this law does is pointless – for whatever harm is inherent in spanking (if any) is already done, when the entire harm could have been avoided in the first place if we only took the steps of educating people before the harm occurs. Also, there are many people who are unfit to be parents, and they need to be discouraged from fostering children until they are ready to be assume what amounts to one of the most difficult social roles imaginable.

We could go on for hours about semantic issues related to the meaning of a spank versus abuse, ontological issues related to what is discipline and punishment, teleological issues related to what is the aim of good or bad parenting, epistemological issues related to how we learn to practice parenthood, and so on. That is an issue for another blog.

My main concern has to do with the use and abuse of research to put forth an agenda and the nature of scientific objectivity. The other guests with whom I was discussing these issues, while I respected their thoughts and viewpoints, all had some agenda. For instance, one was a member of the Family Research Council, a Christian conservative think tank. The other owned a company that runs programs that teaches parenting skills. What struck me was the extent to which these individuals were emotionally invested in the issue. I never really thought much about spanking, because there seems to be larger issues at hand – pardon the pun – such as child physical and sexual abuse. But I respect that in certain cultural contexts, spanking is a normative practice, and even if in some cultures it may not be prescriptive, to apply a universal standard would be to ignore the fact that parental discipline occupies a complex cultural space in which various factors – cultural, social, educational, economic, sociological, and so on – intersect. The inner workings of the family is not a sphere in which a “one size fits all” policy seems appropriate.

I have always taken issue with extremes. “Anything goes” and “Nothing goes” seem to be fruitless approaches towards thinking through most problems. But I believe that the vast majority of people, the vast majority of the time, shy away from thinking through an issue but rather feel through it. Steven Corbert was onto something when he popularized the term truthiness. Most people’s convictions are based upon how they feel, not how they think, and often what we feel we should do, and what we really should do, are in stark contrast. Too many people trust their guts, and unless you are a gourmand, you probably should question what your stomach tells you.

So it was kind of fun to be on television. They had a make-up lady do my face and hair - it was the first time I had to wear makeup since that time in graduate school when I went to a party in drag. There were lights and cameras and satellite hook-ups and all sorts of interesting things. I went into this with one thought in my mind: “Imagine the hypocrisy of a state that practices the ultimate form of corporal punishment – the death penalty – trying to outlaw spanking.” It didn’t set well with the liberal side of me. But neither does spanking children. I do regret one thing that I said…that one alternative to spanking is distracting children – that children have short attention spans and that if they want to touch a hot stove, to distract them with something else, like the television.”

I want to officially say here that I think that spanking is far better than forcing a child to watch the garbage that counts as television these days. I had a nightmare last night of parents following the sage advice of Professor Experimentaholic and putting their children in front of Bill O’Reilly's No Spin Zone. Now there should be a law against THAT. I'd be the first in line to vote for that.

At another level, I guess I am being a bit hypocritical because I say this after having appeared on television myself. Oh well.

I don’t think I would ever make a very good pundit. Pundits tell you what to think, not how to think. Pundits tell you what is right from wrong, not how to determine for oneself what is right and wrong. Any bit of knowledge or wisdom that I could tell you in 30 seconds or less is probably not knowledge or wisdom worth having.

But what I didn’t get to talk about is the nature of the research on corporal punishment, the vast majority of which is fundamentally flawed. The main flaw comes in the form of mistaking correlation for causation. The vast majority of these studies do something like this: Get a group of parents and ask them whether they spank their child, or how frequently, or how hard. Then they measure behavioral problems in the two groups of children, do some multiple regression with dummy variables, and find that children of parents who spank them exhibit more behavioral problems. Then the wise researcher then goes and says that spanking causes behavioral problems. The problem with this approach should be apparent immediately to anyone since Aristotle’s Metaphysics. With this quasi-experimental design, you can not say that the spanking causes behavioral problems because it is equally possible that children with behavioral problems are more likely to get spanked. To do the study right, you need to take a group of new parents, and divide them into two groups – one who spank their children, and one that is prohibited from spanking, and measure the outcome. However, quite fortunately, I doubt there is an institutional board out there that would be willing to allow researchers to tell a group of women who wouldn’t spank their children to do so, and with vigor!

So we are left with some pretty questionable causal links. (There are a number of other problems I won’t go into, such as the fact that parents who spank their kids also are more likely to abuse them, which is not to say that spanking is a gateway practice into abuse, which is again another causal misinterpretation). And I don’t mind studies with poor causal links: they lead you down the path to more careful empirical investigations. My only fear is that given this fact, people – people with Ph.D.s and official-sounding titles at important-sounding institutions who really should know better than to pull this kind of nonsense - seem willing to make sweeping interpretations and inferences based on this paucity of available data. And the only rationale I can come to is that they were convinced of interpretations of the data long before they ever looked at any data of any kind. It doesn’t matter what the data says when you approach it with the zealotry of a pundit. You’ll find what you want in it – it will be like a Rorschach test. And you will always find what you want in your data when you have a vested financial or spiritual interest in the outcome. It’s not even worth citing a study, or citing research, or even pretending that you are even doing science at all.

I caution the reader, as I caution my research methodology class, that warning bells should start ringing in your head when any talking head says, “Research shows that…” or “Studies have determined that…” or variants thereof. What research? Which studies? Are there studies that show the opposite effect? Who did them? Who funded them? How were they collected? Who were the participants? Who were the interviewers? What scales were used? The devil is in the details, and I don’t want to start basing social policy or law or putting people in jail or funding some guy’s company that teaches parenting skills based upon poorly constructed and controlled quasi-experiments. Or worse still, evaluations of former participants of such programs. Or the fact that the president of this company has several lovely children who he never laid a hand upon and who have grown up to become well-adjusted productive members of society. This isn’t evidence of anything. But to parade this as evidence seems to me to be about as dubious as the kind of claims made by alternative resistance exercise machines, diet pills that will make you lose 10 pounds a week, and dead Nigerian kings whose heirs have to remove $10,000,000 from the country.

If I had my thirty seconds of sage advice, it would be this: I think one needs to have an open mind, but not so open that your brains fall out. Likewise, it is useful to have convictions, but you shouldn’t let those convictions lead you blindly down the road to conclusions. And it is easy to be a talking head, but a head not attached to a solid body of empirical data and carefully designed research should have its microphone turned off.