From John Ray's shorter notes

November 09, 2005

Spare the rod and spoil the child?

I did myself have a very permissive upbringing and, as a libertarian, I always supported my son in whatever he wanted to do. Though I would not of course have supported him if he had wished to do anything destructive. But Christians are often taught that a more directive style of parenting is appropriate. Barring children from watching TV or certain parts of TV is, for instance, common. My son was and is a "Simpsons" devotee so I was delighted for him to watch as much of all that political incorrectness as he wished.

Now I am not going to pontificate on exactly what is the right child-rearing style. The genetics research tells us that what we do makes very little difference anyway. What I DO want to do is to make it known to parents who favour a stricter approach that attacks on such approaches are very poorly founded. Leftists are of course great advocates of "there is no such thing as right and wrong" and Leftist psychologists have been claiming for decades that a permissive approach founded on such thinking is far and away better for your child's psychological health. Such claims are rubbish and I want to examine just one example of the "research" supporting such claims in order to show what a crock the whole research tradition concerned is. Below is the centrepiece of one such set of findings, from "Self-Reported Narcissism and Perceived Parental Permissiveness and Authoritarianism" by Ramsey, Watson, Biderman and Reeves and reported in the Journal of Genetic Psychology, Volume: 157. Issue: 2 of 1996. Page Numbers: 227ff.

"A negative relationship appeared between the OMNI and the Rosenberg Self-Esteem Scale ( r = -.18, p <.001), and the OMNI correlated directly with the PAQ maternal measure of permissiveness ( r = .18, p < .001), with the PAQ paternal measure of permissiveness ( r = .11, p < .05), but not with the combined ( r = .04, p < .15) measure of permissiveness. All three authoritarianism scales were associated with greater tendencies toward narcissism: PAQ maternal authoritarianism ( r = .11, p < .05), PAQ paternal authoritarianism ( r = .14, p < .01), and the combined index ( r = .16, p < .01). No significant linkages appeared between the OMNI and any measure of authoritativeness ( r s = -.07 to .05, p s < .15). When the data for men and women were examined separately, no significant gender differences appeared in the OMNI correlations with self-esteem or with any parenting style. The coefficient alpha for the OMNI was .66"

So the study purports to show that "authoritarian" parents produce Narcissistic children -- which would indeed be a matter of concern if it were true. But the study is hilariously flawed. The data for it were obtained by handing out a bunch of questionnaires to 370 Introductory Psychology students at the University of Tennessee at Chattanooga! Their report of how their parents treated them is accepted unquestioningly and data derived from a non-sample of Chattanooga students is assumed to tell us something about humanity in general!

The great advantage of such "research" from the point of view of the psychologist is that students are pretty alert to what the psychologist expects and tend to give him the answers he hopes for. But on this occasion even that did not work out. Look at the correlations reported. They average around .15, which is effectively zero. The correlations are "significant" in a statistical sense but that is simply a reflection of the large sample size. Even if we take the correlations seriously, they are telling us that there is only around a 2% overlap between the upbringing and the personality variables -- which means that knowing about the upbringing of the child gives you virtually no power to predict how the child will turn out. What a lot of nonsense it all is! I will back Solomon (Proverbs 13:24; 29:15) as a better guide to child-rearing than that any day.

As a Parthian shot, it might also be noted the the measure of Narcissism (OMNI) used by the researchers showed very poor internal consistency (alpha of .66) and also showed virtually no correlation with another measure of what should have been largely the same concept -- the widely-used Rosenberg Self-Esteem scale. So whether their measure of Narcissism did in fact measure Narcissism seems dubious. Try not to laugh!

Go to John Ray's Main academic menu
Go to Menu of longer writings
Go to John Ray's basic home page
Go to John Ray's pictorial Home Page
Go to Selected pictures from John Ray's blogs