Sunday, 26 October 2014

Inbreeding: Two Tribes?


Mohd Fareed has provided the sample sizes for each degree of relatedness, and an explanation for the larger number of first cousin marriages, which is the cultural norm in those communities. I have added in, from his paper with Mohammad Afzal, the degree of relatedness for each group and the mean IQ and standard deviations. The results are intriguing.

                                                           n         Relatedness  IQ       sd

Non-inbred                          225         .0000         96.51 (16.25)

Second cousin                       31          .0156          88.57 (13.99)
First cousin once removed   21          .03125        79.70  (6.51)
First cousin                         105          .0625          68.72   (7.15)
Double first cousin               26          .125             59.52 (14.99)

Mohd says: Now a question may arise in the mind of the public, why aren’t the sample sizes uniform? The answer is: among Muslims,  first cousin marriages constitute about 70% or even more of marriages, as compared to the other inbred categories.

Anyway, let me get stuck into these results, which are the sort of statistics I can understand. First cousins, the most numerous in the sample by far, have half the normal standard deviation, as do first cousins once removed. Why? The first cousin result cannot be a small sample size problem. Why this massive restriction in variation? Second cousins and double first cousins have the normal variation. Why? Variance reduces with degree of inbreeding and then bounces back again for the most inbred. Why? I can think of a possible psychometric hypothesis, suggested to me by an esteemed person of my acquaintance, but I haven’t got round to reading all the textbooks on population genetics kindly recommended to me by distinguished scholars (extremely expensive books, by the way) so could population geneticists please help?

That perplexing problem aside, the IQ for the entire inbred category comes out at IQ 72. The “marry your cousin” meme reduces the average IQ by almost 2 standard deviations. Cruelly, these believers will have been cast into such a state of mental weakness that convincing them of the downside of cousin marriage may take a long time, Bollywood notwithstanding.

If we put together the 225 outbred with the 183 inbred we get an overall IQ of 85.5. Given that Lynn’s mean value for India is about 82, Pakistan about 84 these results are within measurement error, and closer to the Pakistan figures. Lynn’s lists are not complete, and need updating,(he does the job himself, without any funds of the sort that PISA commands)  but they are given below for comparison anyway.

In summary, this sample seems very much like the other samples from India and Pakistan in terms of mental ability. What seems clear is that to get an understanding of the potential underlying ability we need to make a massive allowance for inbreeding. Scores of about IQ 96 should be possible in outbred groups.

To understand the regional social and political behaviours as they are at the moment, we need to understand that there are two tribes in contention: higher status outbreeds of average IQ 96, with considerable freedom from their cousins; and lower status inbreeds of average IQ 72, condemned by chains of obligation to perpetual, uncomprehending cousinhood.

Now, as Mohd Fareed says, Muslim India is not the only part of the world that encourages cousin marriage, but I hope that altruists not yet bound for West Africa could give Fareed and Azfal a call, to see if they can give them a hand with genetic counselling materials, training genetic councillors, and working with local researchers and activists to change the cultural meme of arranged cousin marriages. Better still, if you are a Hollywood actor or director, why not Tweet them now, or if that is too much effort, get your people to send their people a cheque?


Summary data from Lynn.

India     1,339       CPM      88             Gupta & Gupta, 1966

India    1,359        SPM       87             Chopra, 1966

India    5,607       CPM       81              Sinha, 1968

India    1,050       CPM       82             Rao & Reddy, 1968

India  3,536         SPM       84             Majumdar & Nundi, 1971

India      180         SPM       79              Mohanty & Babu, 1983

India      100         SPM       78              Agrawal et al., 1984

India      748         WISCR  79             Afzal, 1988

India      500         CPM       86            Bhogle & Prakash, 1992

India         29         CPM       82           Jyothi et al., 1993

India       569        SPM        82           Raven et al., 1996

India      828        CPM        80           Barnabus et al., 1995

India    8,040       SPM       88            Raven et al., 2000

India        569        SPM      81             Raven et al., 2000

India: median                   82


Pakistan    349     GEFT    84            Alvi et al., 1986

Pakistan    140      SPM     84            Rahman et al., 2002

Pakistan  1,662   SPM      82            Ahmad et al., 2009

Pakistan  2,016   SPM      86            Ahmad et al., 2009

Pakistan: median            84


  1. This is astounding and may account for some of the extreme inequality there.

    Another way to reduce cousin-marriage is to assist with and encourage family planning. Lower birth rate means fewer cousins around to marry.

  2. "Second cousins and double first cousins have the normal variation. Why? Variance reduces with degree of inbreeding and then bounces back again for the most inbred. Why? "

    Sounds like a question for Mr. Razib Khan.

  3. More important than marriage cousins is the average intelligence of couples. Seems more correlation than causality. Richer and traditional families like royal families marry within families OR rednecks. Less tolerated (and less socially smart people) can be marry with biological first or second cousins because they are more psychologically similar.

    Intelligence of sons of cousins decreased specially if parents have lower intelligence too.

    If cousins marriage really decrease intelligence, ''arabs'' today would be retards (well, today they are or tend to be ''quasi-retards'' ;).


  4. Indavl er ligegyldig hvis man er gedehyrde på den Anatolske højslette og gift med tre kusiner, hvor den ene sikkert bliver stenet for utroskab, men i Vesten vil det føre til evig offentlig forsørgelse og lediggang, med kriminalitet til følge.

  5. How you explain relatively higher intelligence of amish people who are strongly endogamic???


  6. Endogamy can reduce iq-intelligence, RELATIVELY SPEAKING (or necessarily not).


  7. Sorry, I think I made many mistakes here.