56 PRETEND GENESMoreover, the MISTRA sample included DZAs, and the investigators
have acknowledged that comparison of MZA with DZA correlations
would be an impor tant control. But they have seemingly refused to pub-
lish such comparisons. In addition, they have restricted access to data by
critical investigators like Jay Joseph, which constitutes a breach of a
crucial condition of scientifi c research. Could it be that correlations for
DZAs are unexpectedly high— perhaps even as high as those for MZAs—
suggesting other imperfections of sampling or that there really is no dif-
ference? Indeed, in the partial reports published to date, with incomplete
numbers, the latter does appear to be the case.^19
What is remarkable is that these studies have been published, and
their conclusions reiterated, in highly esteemed journals, in full knowl-
edge of the profound message thereby conveyed to psychologists, teach-
ers, parents, a sensationalizing mass media, members of the general
public, and policy makers. Th at the situation has not changed much
since then is shown by Jay Joseph’s recent analyses.^20 Th ere seems little
doubt that studies of “separated” twins are universally— and prob ably
intrinsically— fl awed.
Comparing Identical with Nonidentical Twins
Th e alternative method of study is to compare the average resemblance
between MZ and DZ pairs reared together. If the logic of the in de pen dent
/additive model is correct, then on average the resemblance of DZ
twins should be half that of MZ twins. Th at is, they should correlate
around 0.5 compared with 1.0 for MZ twins. Th e extent to which that dif-
ference is found is a mea sure of heritability of the trait. So the latter is es-
timated by doubling the diff erence between the MZ and DZ correlations
or by using other model- fi tting statistical maneuvers based on the same
assumptions.
Most of the conclusions about the heritability of human cognitive po-
tential have come from this “classical twin model.” It is this approach that
Robert Plomin and colleagues have adopted in their massive Twins Early
Development Study started in 1994 and involving more than twelve thou-
sand pairs. However, this approach is also seriously fl awed, making the
results diffi cult to interpret.
This content downloaded from 139.184.14.159 on Tue, 17 Oct 2017 13:51:26 UTChttp://www.ebook3000.com