Tuesday 22 May 2012

The Gloomy Prospect, and Why Considerations of Interactivity Cannot Rescue Stefan Molyneux's Claims

Let's dwell on the Freedomain Radio action for just a bit longer, shall we? :)

An extremely aggressive forum member by the name of Haplo has quoted a passage from Eric Turkheimer's article "Three Laws of Behavior Genetics and What They Mean", in which Turkheimer argues that some statistical procedures have been badly misapplied by "developmental psychologists, geneticists, and environmentalists", and that the "[s]mall relations (...) found between predictors and outcomes" should not be taken seriously. The quote doesn't provide enough context to tell the reader what applications of statistical procedures Turkheimer was writing about.

He then contrasts the Turkheimer quote with a quote by Steven Pinker, in which Pinker described certain results as being very robust, and the validity of different psychometric tests as being amply validated.

Haplo's aim is to suggest not only that Turkheimer and Pinker disagree with each other, but also that Turkheimer disagrees with the "Three Laws" that he himself formulated at the beginning of that same article, and with his own statement that, as far as the three laws are concerned, "the empirical facts are in and no longer a matter of serious controversy".

This would, of course, be warranted if, in their respective quotes, Pinker and Turkheimer were actually talking about the same thing; but, of course, they aren't. Anybody who wants to take the time to read Turkheimer's article (linked to above) and my earlier blogpost about Steven Pinker's account of the evidence can easily verify this.

It's quicker to just take it from me, of course :) Steven Pinker, in the passage Haplo quoted, was talking about the finding that the shared environment (= "who raised you") is a very small if not negligible predictor of most psychological outcomes by the time the subject is an adult, whereas hereditary factors (= "who you got your genes from") and the nonshared environment (= "everything else") are both substantial factors; these findings result from very straightforward and surprisingly conclusive analyses of variance of the data from a large body of adoption studies and twin studies, and, while Turkheimer, at least in his article in question, was more conservative than Pinker in his description of how small of a factor shared environment is, he certainly does agree with his own "Three Laws". (He has later stated that they shouldn't actually be seen as "laws", but as "the new null hypotheses", and he discouraged any further research aimed at demonstrating the heritability of any outcomes since it wouldn't tell us anything new.)

The Turkheimer quote posted by Haplo, on the other hand, was about the attempts that had been made to discern specific factors within the category "nonshared environment", and attempts that, at the time of his writing, were soon going to be embarked upon in hopes of discerning specific genomic patterns (from the "heridity" category), that would turn out to be strong predictors of psychological outcomes.
He quotes Robert Plomin and Denise Daniels' 1987 article "Why are children in the same family so different from one another?":

One gloomy prospect is that the salient environment might be unsystematic, idiosyncratic, or serendipitous events such as accidents, illnesses, or other traumas . . . . Such capricious events, however, are likely to prove a dead end for research. More interesting heuristically are possible systematic sources of differences between families.

Turkheimer's argument was that, with all the research that had since been conducted to try and discern relevant factors from the nonshared environment, it was time to conclude that the gloomy prospect was true, and that any research of this kind should be discontinued.
Furthermore, he wrote:

Nonshared environmental variability predominates not because of the systematic effects of environmental events that are not shared among siblings, but rather because of the unsystematic effects of all environmental events, compounded by the equally unsystematic processes that expose us to environmental events in the first place (Turkheimer & Gottesman, 1996).

Ok, as a brief digression, here is a passage from Jonathan Haidt's book The Righteous Mind, which refers to another book by his fellow social psychologist Tom Gilovich:

His simple formulation is that when we want to believe something, we ask ourselves, "Can I believe it?" Then (as Kuhn and Perkins found), we search for supporting evidence, and if we find even a single piece of pseudo-evidence, we can stop thinking. We now have permission to believe. We have a justification, in case anyone asks.

In contrast, when we don't want to believe something, we ask ourselves, "Must I believe it?" Then we search for contrary evidence, and if we find a single reason to doubt the claim, we can dismiss it. You only need one key to unlock the handcuffs of must.

I bring this up because I see a lot of people saying "oh, but there's a huge amount of interaction between the factors, and their effects are completely unsystematic, therefore behavioural genetics can be safely dismissed and I can continue believing Stef's empirical claims." And I've seen other arguments that share this basic pattern with the previous one: People point something out about the methodology of adoption or twin studies that is both correct and likely to be unexpected to anyone who is completely new to the field, and then act as if what they have said disproved all the findings from behavioural genetics and confirmed their own position without ever checking what that thing they've pointed out actually implies. Examples: "oh but heritability is only defined within a given population"; "oh but the identical twins used in twin studies both shared the womb of the same highly stressed mother"; "oh but the adoptive siblings used in adoption studies were probably deliberately placed with families that were similar to them"; "oh but the range of the studies is limited".

End digression.

So, can the fact of high interactivity of factors and of low systematicity of their effects rescue Stefan Molyneux's claims? Of course not. His claim is that the effects of the factors pertaining to the shared environment are not only very substantial but highly systematic. Abuse your children and they will be a great deal more likely to develop mental disorders and irrational beliefs and attitudes; parent them peacefully and they will be virtually guaranteed to grow up to be spectacularly functional and rational anarcho-capitalist "philosophers". This is uncomplicatedly contradicted by everything in Turkheimer's article.

Ok, some will say, but doesn't the "interaction" thing still leave some room for believing that, under some circumstances, bad parenting can have the kinds of long-term negative effects that Stef says it has?
Well, theoretically, yes, BUT: given that the "shared environment" does not show up as a significant factor in the variation of adult psychological characteristics, in order for it to be the case that "under circumstance X, bad parenting causes long-term harm", this would have to be evened out by the fact that, "under other circumstances, the same bad parenting causes long-term benefits". Otherwise you're still contradicting the data, since even if bad parenting only negatively affected some substantial segment of the population while leaving everybody else unaffected, that substantial sample would still skew the average, leaving us with a significant amount of variation attributable to the shared environment.

Stefan Molyneux's empirical claims about the long-term effects of parenting are unambiguously wrong, and he should retract them. And he, as well as other members of the Freedomain Radio community, would be very well advised to address their propensity toward self-overestimation. If you are new to a topic, I strongly recommend using phrases like "I don't see how this follows", rather than "This obviously doesn't follow, all those scientists are morons!"

Peace out :)
Seb

3 comments:

  1. I found myself quite distracted by Haplo's aggression in that thread which made it hard to focus on what was actually being said, so thanks very much for laying this out :D
    Am very tired, so will re-read tomorrow.

    Also, it would be WKD to read about your experiences with nootropics on here!

    ReplyDelete
  2. Another technique of the FDR people is to latch on to an alternative theory that seems to rescue their core beliefs.

    Studies show that we are affected almost none by relatively normal parenting! Well, if we cherry pick epigenetics, which has some scientaculicity to it even though we've done nothing more than read and listen to pop-science and worse about it, that ought to enable us to continue with our beliefs.

    You see, science continues to show that soul-like omniscient subconsciouses that can be damaged greatly at any stage of life before you're 18 are quite important in understanding all ills of the world. Science FTW!

    That's what the Blink book was about, after all.

    ReplyDelete
  3. Thx for your comments! :)

    Indeed, Mr. C, the epigenetics thing often comes up as a good example for Jon Haidt's statement that "we only need one key to unlock the handcuffs of _must_".
    Some may find it helpful to notice the analogy between the "Epigenetics, therefore childhood trauma" argument, and the "Gaps in the fossil record, therefore creationism" argument.

    ReplyDelete