Friday, December 18, 2009

Pitiful research on psych. abuse

You would think that if the psychological profession is training people to make judgment on what is or is not actionable abuse, then they would have some scientific research to back it up. But they don't.

Here is an academic paper (pdf) on psychological abuse of kids.

First the authors look for a definition. They reject a definition in terms of harm to a child because:
It poses a serious problem for research, however, because children may suffer immediate psychological pain from verbal attacks by parents but not display any lasting social or psychological problems.
So they settle on this definition:
Psychological aggression is a communication intended to cause the child to experience psychological pain. The communicative act may be active or passive or verbal or nonverbal.
In other words, they choose a definition that is so broad that it includes all sorts of harmless actions.

The first requirement of a scientific definition is to show that it can be applied consistently. But when they looked at studies using this definition, they found incidence rates varying from 25% to 94%.

So even the experts doing the academic studies cannot agree on what this definition means.

After a bunch of meaningless charts and tables, the authors start getting opinionated:
An alternative hypothesis is based on the theory that any act of psychological aggression against a child, regardless of whether the aggression is a purely expressive emotional outburst or is a means of correcting or controlling misbehavior, puts the child at increased risk for mental disabilities. Consequently, the rate of social and psychological problems will start to increase from the beginning of the psychological aggression distribution. If this hypothesis is supported, it means that any psychological aggression is abusive. If so, the results of this research would mean that close to 100% of American parents, by their own report, psychologically abuse their children, some starting in infancy.

In the absence of research showing that each increase in the amount of psychological aggression is associated with an increased probability of child behavior problems, there may not be many who regard any psychological aggression as abuse.
They are admitting that there is no evidence that psychological aggression is harmful, but they want to hypothesize that every single act of psychological aggression contributes to mental disabilities.

The hypothesis is absurd, and they have no evidence for it.
Regardless of whether occasional instances of psychological aggression damage the child, any act of psychological aggression against a child is an unacceptable mode of human relationships and provides a poor role model for the child. Parents should never use psychological aggression as a means to correct or control misbehavior.
Got that? No parent should ever say or do anything that ever causes a child psychological pain.

The anti-spanking folks are goofy enough, but this paper goes way beyond that. Their rule would seem to prohibit all forms of punishment and discipline. They would even forbid silence, if a flattering comment could be made.

Here is the sleight-of-hand they use to pretend that science supports their nutty ideas. First they claim that there are studies showing a correlation:
Similarly, psychological aggression by parents has been shown to be associated with higher rates of delinquency and psychological problems.
Then they argue that correlation implies causality:
Thus, avoiding discipline techniques that involve psychological as well as physical aggression increases the probability of the child being well behaved and well adjusted, rather than resulting in ‘‘kids running wild.’’
But this is a logical fallacy. There is a correlation between delinquency and punishment in kids, in the sense that there is some overlap between the more delinquent kids and the more punished ones. Is the delinquency causing the punishment or the punishment causing the delinquency? The authors seem to be leaping to the conclusion that the punishment is causing the delinquency. Correlations does not imply causality.

This type of fallacy is common in child-rearing research. Here is how it can go wrong. Suppose American white, black, and asian families are being studied. The black and asian parents spank the most, while the asian kids are the best behaved and the black kids are the worst behaved. If the study sample includes just white and black families, then spanking and bad behavior will be correlated. The black families will have more of both. So if the study authors fallaciously assume that correlation implies causality, then they will conclude that spanking causes bad behavior.

But now suppose that the study includes only white and asian families, with no blacks. Then it will look like spanking is correlated with good behavior, and the study will reach the opposite conclusion.

Because of problems like this, studies on parenting practices tend to be worthless. Furthermore, they tend to be done by over-opinionated childless do-gooders who just want to confirm their prejudices about how everybody else should be rearing their kids.

1 comment:

Anonymous said...

psychology and sociology use tools and reasoning in the scientific method that barely, if at all, work in areas as difficult as clinical trials for drugs. The tools and methods are simply not up to the task of dealing with the complexities of human behavior. They work well enough, if used correctly, on the lab bench. To extrapolate those techniques to a field where we really know nothing that can be quantified is not only absurd but dangerous. On top of all that, from my empirical experience and that of a lot of people I've spoken with over the years it seems most of the practioners in those fields get into them because they themselves need the help...... But you're right, it is indeed social engineering. The only things that have come out of psychology that can be reduced to actual practice are brainwashing (including advertising and marketing) and interrogation methods. Thanks for a great post, George.