The fundamental constant of sociology?

A friend emailed me a copy of a recent unpublished meta-analysis that doesn’t sit too well with La Griffe du Lion‘s fundamental law of sociology. The authors conducted a follow up to Roth et al.(2001)‘s massive meta-analysis, a meta-analysis which was cited by Rushton and Jensen in defense of their hereditarian position (e.g. Rushton and Jensen, 2006). While finding a 1.1 SD Black-White gap, Roth et al. (2001) noted that they were unable to “assess the influence of time on standardized ethnic group differences” and that they “tried to control for the influence of time by choosing the most recent studies when there was an option.” The authors of this meta-analysis coded for decade, limiting their sample to individuals above the age of 16, and found that as of the 90s the gap in non selected samples was 0.8 SD, down one third from the 70s. They went onto interpret the findings as support for a sociocultural origin of differences and ended their paper on a political note: “Policy makers are warranted to continue to further support and reinforce these developments through equality and affirmative action policies that increase opportunity for minority ethnic groups, and that promote a pro-diversity perspective in organizations.” (For reference, here is a list of meta-analyses of the Black-White gap in industry; herein (table 12.5) is a list of mean d values on an assortment of cognitive assessments.)

(K = 91; N = 1,110,038).

I don’t feel like critically commenting on this study, but I would note that one gets out of a meta-analysis what one puts into it and that it’s clear that the authors of this study did not include some large samples with large d’s that have been fairly invariant across time (e.g., the SAT and GRE). I would also add that the results are consistent with an interpretation of no change (comparing the difference in the 60s to that in the 90s).

This entry was posted in Uncategorized. Bookmark the permalink.

5 Responses to The fundamental constant of sociology?

  1. JL says:

    Is there a list of data sets used in that meta-analysis?

    • Chuck says:

      I’m sure that there’s a list; but I don’t have it. I was basing my claim that “it’s clear” on the Ns divided per decade per racial percentage

  2. cornbob says:

    How did you come to your conclusion? Just curious.

    [If SAT and or GRE data was included the n’s would be higher. Refer to Roth et al 2001 table 1.]

  3. Steve Sailer says:

    With tests like the SAT, there could be a pattern that any narrowing of the white-black gap encourages the powers that be to scrape the bottom of the barrel harder (e.g., by encouraging more students unlikely to thrive in college to take the test for free), which then serves to keep the white-black gap stable.

    • Chuck says:


      So if you compare AFQT scores based on NLSY 79 and 97, you see a closure of about .25 SD. But if you compare NLSY 97 data to AFQTaverages based on tests given under Selective Service (under which the AFQT was given to nearly every able Black and White 18 year old) you see no difference. So was there a closure or not? It’s hard to argue that the Selective Service samples were less representative than the NLSY 79.sample. Fluxes in the Black score of course need not actually be fluxes in the Black score; apparent Black increases/decreases could be due to White decreases/increases. Whenever I see my diversity lottery Berber x-roommate — a great guy but not too bright — I’m reminded of that possibility.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s