[<< | Prev | Index | Next | >>]

Monday, October 28, 2002

Meta Skepticism

I happened, through a reverse link, to find someone's journal entry from long ago mildly poo-poo'ing a reference I had made elsewhere to this article about body type and temperment--he compared it obliquely to phrenology.

I have a current and growing pet peeve with this trend of what I might call dogmatic skepticism--the rejection of any claimed statistical correlation which happens to sound foofy on the surface. To have that bias is just as non-objective as to believe the foofy things just because they sound good or meet with your untested intuitions. I care a bit more about this spreading affliction because unlike most epistemological memes it tends to infect otherwise highly rational people.

Are you infected? Consider:

I read an article about a study recently which found a correlation between the time of year a baby is born and their odds of having some health problem or another later on in life (I forget the details); the authors hypothesized in this case that it may correlate with the odds of the mother catching the flu during various stages of the baby's development.

Sounds reasonable, no? I mean, from just that we're not going to believe it; but we're not going discard it out of hand either.


There is a standard birth-date cutoff that determines what year a child starts kindergarten. If they're just a little younger, they get held back an entire year, and they are stuck with this offset for the next twelve years and beyond as well. Consider how quickly children are changing from month to month at that age, and then imagine how much of an impact that must (likely) have on matters such as self-confidence, social status, and so on, not to mention the intellectual impact of being given material a whole year behind or ahead of others nearly your age.

Sure, individual variation is huge, but on average one might expect some correlation between date of birth and various personality attributes. Sounds reasonable, no? Again, not something to believe just like that, but not something to discard out of hand either.

Now, those are just two examples, but there is also the impact of Seasonal Affective Disorder and other seasonal correlates to the emotional ambience which surrounds a developing child during their earliest stages of development; and then there is the seasonal swing of pollen and other alergens, and, related to SAD, the amount of sunlight in each day of the year; and there are the older siblings free from school in the summer, nose to the books the other nine months; and so on, and so on.

In short, there are many potential correlates between the time of year a person is born and how they develop, particularly socially and psychologically but even in terms of basic health.

Sounds reasonable, no? Not to discard out of hand.

Now, the last time someone alluded to personality correlations with astrological sign, did you discard it out of hand?

The issue here is that many "rational" people are biased against observing phenomina which they don't have an a-priori rational explanation for. It's a case of throwing out the baby with the bathwater; assuming that because the explanation is irrational or unsupported, that the observation itself must be equally unfounded.

The truth is so often in the middle--the alignment of particular stars probably has no bearing on one's personality but, more coarsely, maybe Virgos are more "intellectual" on average.

If you never look, you'll never know, and a good example of this is Jungian typology. Honestly I don't even know what the historical/psychological rationale is for it, and perhaps that's a good thing--maybe it's allowed me to observe the actual correlations more objectively.

The dogmatic skeptic is so biased against any system of that sort that they fail to observe what is often right under their nose. Let me give an example for my own case of how readily available the evidence is:

My last four girlfriends, and the only one before that who's type I know, were all INT's. Only 2.6% of women in general are INTs. If we assume that the Jungian types bear no correlation to relevant personality attributes--i.e., that our significant others, for instance, should be a random sampling--then the odds of the above outcome are 2.6% to the fourth power (four, not five, because we want the odds of five in any one comparable category), or about one in two million. [This is not entirely correct, because the metric itself--selecting on the first three type letters--was chosen to fit the data; but since the number of comparably measurable metrics which I personally have considerred is fairly small to begin with, this is a relatively slight bias.]

I can make a similar analysis with similarly convincing numbers just based on the gross preponderance of my friends being N types (and mostly NTs) despite Ns as a whole being only 25% of the general population. Or that the women who's personalities I have found most attractive throughout my life have all turned out to be INFPs or marginal INtPs (but alas, never in the right circumstances). Each of these retrospective studies has outcomes completely defying even odds.

In short, even though I'm sampling only from my own personal experience, still the variation is so much less than the correlation that even my small sample size is adequate to prove the validity of the system with high confidence. (Here "validity of the system" does not, obviously, extend to anything and everything that is claimed about it, but rather simply to the statistical import of the metric, whos total implications remain open to further study.)

The body type and temperment issue is similar, though not nearly so obvious and requiring a bit more sampling. In that direction, we are heading toward the more general category of stereotypes, which are also often rejected without study, typically based on the apparent stereotype-breaking reality of individual variability. But there goes the baby with the bathwater again--yes, it is bad to assume someone fits their stereotype even when you have evidence to the contrary; but it is equally naive to never wonder what aggregation over time created the stereotypes in the first place.

To me the final proof that something is amiss with the dogmatic skeptics is their "I'll see it when I believe it" behavior--one day, some study comes out showing a concrete and scientific cause for a correlation, and suddenly the dogmatic skeptic can observe an effect they insisted wasn't there just the day before.

Keep the baby. There are valuable truths threaded throughout many otherwise nonsensical mysticisms--the competitive evolution of memes, if nothing else, helps see to that. Bhudism, for instance, is a brilliant study of human psychology. Approach it as such, and suddenly its wisdom is obvious, and the spiritual, mystical trappings become just superfluous semantic ornimentation. Approach it as a foofy mystic religion, and you'll probably never notice the thousands of years of empirical wisdom behind it.

Alright, 'nuff said. I'm just rambling now.

[<< | Prev | Index | Next | >>]

Simon Funk / simonfunk@gmail.com