Some non-effects of adoption

By Razib Khan | July 21, 2011 1:49 am

I mentioned offhand to Trey of Genomes Are Us that I’d look around for the effect of adoptive environments on criminality (what with the recent concern about studying the genetics of criminal predisposition). Luckily I have Steve Hsu in my RSS, as he posted something of interest just yesterday, pointing me to this paper, The environments of adopted and non-adopted youth: evidence on range restriction from the Sibling Interaction and Behavior Study (SIBS):

Previous reviews of the literature have suggested that shared environmental effects may be underestimated in adoption studies because adopted individuals are exposed to a restricted range of family environments. A sample of 409 adoptive and 208 non-adoptive families from the Sibling Interaction and Behavior Study (SIBS) was used to identify the environmental dimensions on which adoptive families show greatest restriction and to determine the effect of this restriction on estimates of the adoptive sibling correlation. Relative to non-adoptive families, adoptive families experienced a 41% reduction of variance in parent disinhibitory psychopathology and an 18% reduction of variance in socioeconomic status (SES). There was limited evidence for range restriction in exposure to bad peer models, parent depression, or family climate. However, restriction in range in parent disinhibitory psychopathology and family SES had no effect on adoptive-sibling correlations for delinquency, drug use, and IQ. These data support the use of adoption studies to obtain direct estimates of the importance of shared environmental effects on psychological development.

The technical issue here is that there’s long been the assumption that because adoptive families are pre-screened they’re a selection-biased environment which won’t exhibit the same environmental variance as the general population. This matters potentially for behavior genetics studies which use adopted children, as it might underestimate shared environmental effect (a robust behavior genetic finding that puzzles many is that most of the environmental effect is non-shared). I’m not too focused on this specific issue, but rather want to pass along two charts which might interest readers:


CATEGORIZED UNDER: Uncategorized
  • Charles Nydorf

    Studies based on quantities not ultimately definable in terms of length, mass, and time are not to be trusted no matter how sophisticated the statistical methodology.

  • http://washparkprophet.blogspot.com ohwilleke

    “a robust behavior genetic finding that puzzles many is that most of the environmental effect is non-shared”

    This is absolutely the standard way of stating this result. But, giving the profoundly insignificant impact of “shared environment” on outcomes, which demonstrably account for a huge part of the environmental influences that people receive as children, the term “non-shared environment” may be inaccurate and may be better described as “non-shared factors”.

    The case that “environmental influences” using the term “environment” in the same sense we mean when we are talking about “shared environment” are really a big component of the “non-shared factors” is increasingly a questionable inference (and it is strictly an inference, I’m not aware of any studies showing any particular non-shared environmental factor producing a non-shared environmental variation directly). Until someone can point to a direct link between some non-shared environmental exposure and outcomes, the standard terminology is really questionable.

    Put another way, to the extent that a trait is plastic and influenced by environmental exposures, it should demonstrate ratios of shared environnment and non-shared environment impacts that are more or less proportionate to the proportion of sibling environmental exposures that are shared and are not shared. If identical twins spend every moment together and non-shared environment accounts for 90% of the difference between them, then those differences aren’t really due to environmental exposures in the conventional sense.

    Other potential non-shared factors include random epigenetic variation and genetic differences between siblings in the proportion of genes actually inherited from each parent.

  • http://infoproc.blogspot.com steve hsu

    @ohwilleke: you are completely correct. The non-shared stuff is mysterious and “factors” is better than “environment” since we don’t know what it is. The terminology reflects the history of the subject: these researchers fully expected to find big shared environmental effects. (As did I!) Current educational and welfare policies are still based on this (now falsified) expectation. Sandra Scarr (one of the Minnesota twins researchers) writes frankly about her initial expectations, how they sat on the data for a while because they simply couldn’t believe the results, etc. Someone who is skeptical about these results, thinking they are produced by researchers with a pro-genetic cause bias, should look into the history of the subject. It’s rare, but some social scientists can actually learn from data and modify their priors. Note, I’m not holding my breath for Chicago school economists ;-)

  • AG

    Data here is quite persuasive. But most people in the world only believe opinions without data.

    I like steve’s topic `Without data, you are just a person with opinion’

  • http://www.personalgenomics.us Trey

    Thanks. I wish it was not behind a subscription wall (35 bucks!) so I could read the whole thing.

    I would like to read it because the point of the paper, looking at range restriction, is very pertinent to that second graph:

    “Relative to non-adoptive families, adoptive families experienced a 41% reduction of variance in parent disinhibitory psychopathology and an 18% reduction of variance in socioeconomic status (SES).”

    I’m not sure how they are correcting for range restriction (which they do indeed find) in SES in adoptive families. Maybe I’m not quite understanding that second graph and how they are compensating for the restricted range of SES.

  • http://washparkprophet.blogspot.com ohwilleke

    @Charles Nydorf

    One of the things that makes this study persuasive despite the admittedly potentially iffy questionaire based measurement instruments is that the results on several completely different measurement instruments are robust and that this is also a good confirmation of other literature that is out there that uses measurement instruments different from these. If anything, the data is so consistent that I worry that the statistical noise might even be suspiciously low.

    The classic example of data that is too good to be true is Mendel’s original pea paper that I recall analyzing in an advanced college statistics class. The Chi square deviation due to random variation from the “expected value” due to genetics in that paper is so aburdly low and impossible to replicate that some of the data was almost certainly faked for publication even though the principle is actually correct. But, the didn’t have nonparametric statistics back then so it looked like it made the finding stronger rather than weaker. See, e.g., http://www.stat.cmu.edu/tr/tr675/tr675.ps

  • Clark

    Studies based on quantities not ultimately definable in terms of length, mass, and time are not to be trusted no matter how sophisticated the statistical methodology.

    Must be a bear for you to do E&M.

  • Charles Nydorf

    @ohwilleke
    As I understand it Mendel lived before the days when people reported results literally. I think he used numbers to indicate the underlying trends in his data.

    @Clarke
    I’m no expert but as far as I know electromagnetic units are reducible to length, mass and time.

  • Darkseid

    i understand the general concept, as Kevin Mitchell wrote about this a year ago, but i’m not grasping how the graphs highlight this. greatly appreciated if anyone can quickly explain that, thanks.

  • ackbark

    I want to ask whether there is anything to suggest that adoptive children may be unconsciously resistant to environmental effects in the adoptive household,

    whether or not there is evidence of a distinction in this between those who always knew they were adopted and those who only found out when they were adults?

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com

ADVERTISEMENT

See More

ADVERTISEMENT

RSS Razib’s Pinboard

Edifying books

Collapse bottom bar
+

Login to your Account

X
E-mail address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »