MATE SELECTION
61
that guide mate selection. These criteria differ from person to
person, thus allowing most individuals to find someone whose
criteria they satisfy The criteria may not be entirely conscious
or easily articulated, but they are effective; the hypothesis in-
sists that mating is regulated by lawful principles involving
conscious or unconscious choice. Each individual’s criteria are
determined by his or her own unique demographic, physical,
and especially, psychological characteristics. Recent research
has demonstrated that virtually every physical, demographic,
and psychological characteristic is to an important extent in-
fluenced by genetic factors (Bouchard, Lykken, McGue, Segal,
& Tellegen, 1990). MZ twins share a common genome and are
genetically identical. Moreover, MZ twins reared together have
also shared most of their developmental history. It is becoming
increasingly apparent that merely being reared together does
not ensure that the effective environment of two siblings is the
same; genetically different individuals in the same household
elicit and effect different environments or react differently to
the same experiences (Plomin, Defries, & Loehlin, 1977; Scarr
& McCartney, 1983). But these gene-environment correlations
and interactions, which can yield marked differences between
siblings, merely increase the similarity of MZ twins.
Therefore, it is plausible to suppose that the mating criteria of
MZ cotwins, if such criteria exist, as the idiographic model
requires, ought to be quite similar. If these criteria exist, they
must be somehow determined by a combination of genetic and
environmental factors, including gene-environment correla-
tions and interactions, and because these presumed causal fac-
tors are all shared by MZ twins, their products-including the
elusive, idiosyncratic mating criteria that we seek-should be
very similar in MZ cotwins.
This line of reasoning suggests that a strong test of the idio-
graphic model would be to compare on an adequate variety of
traits the spouses selected by pairs of monozygotic twins. For
each twin individual, we cannot predict what his or her mating
criteria may be nor can we confidently rely on his or her ability
to tell us. We can be confident, however, that they should be
similar (if such idiosyncratic criteria exist) to the criteria that
guided the selection of the cotwin’s mate. We cannot presume
that any individual’s criteria are comprehensive, prespecifying
every spousal trait that we might think to measure. That is, the
idiographic model does not predict that the spouses of MZ
twins will be similar in all respects. However, if the idiographic
model has any meaning at all, it must require that, for each pair
of married MZ twins, there should be some set of features of
the spouses that were criteria1 for these twins, sufficiently selec-
tive to determine a specific choice, and with respect to which,
therefore, this pair of spouses should be very similar.
To help clarify the idiographic model, it may be useful to
imagine how it might be applied to the study of twins’ choices
of cars rather than mates. Different individuals probably rank
order the attributes of automobiles idiosyncratically, but MZ
cotwins are likely to consider the same set of attributes-cost,
power, color, size, handling, and so on-as most important (i.e.,
as criteria!). Although we would not expect such twins to always
select identical vehicles, we would expect their choices to be
very similar in at least certain respects.
The idiographic model predicts that each pair, /, of MZ
spouses should be markedly similar with respect to at least
some small subset, Kf of the N personality and interest traits
measured, where the several traits included in Ki vary unsyste-
matically from pair to pair. If we transform all variables to have
the same mean and standard deviation, sort for each pair the N
absolute between-spouse differences in order of size, and then
average over spousal pairs, the idiographic model predicts that
there will be more very small differences for MZ than for DZ
spouses and more small differences (i.e., more of the N vari-
ables for which the within-pair difference is, say, 0.5 standard
deviations or less) for DZ spousal pairs than for the spouses of
randomly paired persons of the same sex.
Method
The same data set described above provided scores on the 88 vari-
ables of the spouses of 152 PairsofMZtwinsand 117 pairsofDZ twins.
The twins ranged in age from 29 to 55 years; their spouses ranged in
age from 24 to 67 years. The mean ages were 41.0 years and 36.9 years
for husbands and wives, respectively. All scores except the attitude item
responses were age corrected separately by sex and converted to T
score units with M= 50 and SD = 10. (The 14 individual attitude items
could not be meaningfully converted to T scores and were not included
in this analysis, leaving a total of 74 variables.) The spouses of the MZ
and the DZ twin pairs were separately correlated on all variables, with
the results shown in Table 3; all values of.30 or higher are listed in the
table.
To test the idiographic model as described above, the absolute T
score differeιjges on all 74 variables between Spouses A and B were
rank ordered by size and then averaged over the spousal pairs of MZ
and DZ twins separately, to produce Figure 1. Because this method of
analysis is novel, for comparison purposes it was applied also to the
MZ twin pairs themselves (the twins married to the MZ spouse-
spouse pairs) and to a sample of unrelated or “random’* pairs, formed
by randomly reassorting (within sex) the MZ spouse-spouse pairs.
Results
The top curve in Figure 1 shows that, on average, the MZ
twin pairs had identical scores on 5 of the 74 variables (perhaps
a different set of 5 variables for each pair), differed by one T
score unit on 3 variables, and so on. The mean difference for
the MZ twins was 7.39 T score units; the average correlation
implied by this mean difference (D) is (see Plomin & DeFries,
1980) r = 1 - (D∕ 1.13σ)2 = 1 - ([7.39]/11.3)2 = .57, where D is the
mean absolute intrapair difference and <r is standard deviation.
The mean difference for the unrelated pairs was 10.92 T score
units, corresponding to an average correlation of.06. The mean
differences for the MZ and DZ spouse-spouse pairs corre-
sponded to average correlations of. 14 and . 11, respectively
As can be seen in Table 3, which includes the same variables
listed in Table 2 (and all correlations greater than .29), the
spouses of twins resemble each other even less than they resem-
ble their twin spouses, and what similarity exists consists al-
most entirely in the same religiosity-conservatism and educa-
tion-cultural variables on which twins resembled their spouses.
As can be seen in Figure 1, both groups of spouse-spouse pairs
produced mean absolute differences that very nearly coincide
with those of the unrelated pairs and, most importantly, the
numbers of very small spouse-spouse differences are identical
for the spouses of MZ and DZ twins. That is, most pairs even of