MATE SELECTION
59
have traditionalism scores in this range, from 28.25 to 7 1.75.
But some 97% of all eligible women score in this range on
traditionalism, so this one-dimensional criterion does not help
much to narrow the search.
Our seeker knows also that his potential wife should have an
attractiveness score in the same range. But about .972 = 94% of
all women have both scores in this range and are thus potential
mates for him. If we knew of 10 mutually uncorrelated variables
on which spouses resembled each other to the extent of r = SO,
then we could target his search on just the .971° = 74% of all
women (or of all Norwegian Lutheran women) who fall within
the lo-dimensional hypercube in which he can expect to find
his future bride. But 10 such orthogonal homogamy variables
have not been identified, and there probably are not more than
the equivalent of 4 or 5, even taking into account the host of
correlated characteristics on which spouses are weakly similar.
It is doubtful that even the most complete database of assorta-
tive mating coefficients could serve to focus our seeker’s search
on fewer than about .975 = 86% of all women in his field of
eligibles. Even if our seeker himself had scores of 70 on all 5
variables, so that his search area was in a less populated region,
he would still be left to choose among some 52% of all eligible
women.
In quantifying the extent to which spousal correlations
narrow the field of eligibles, it is seductive but erroneous to
interpret these correlations as predictive of some criterion, as in
multiple regression. Merely four mutually orthogonal predic-
tors that each correlate .50 with the criterion would account for
100% of criterion variance, whereas five orthogonal predictors
that each correlate .90 with the same criterion is mathemati-
cally impossible. Yet, by way of comparison, it is easy to find
five mutually orthogonal variables on which MZ twins show an
average within-pair correlation of .90 (e.g., age, height, finger-
print ridge count, IQ, and electroencephalogram alpha fre-
quency). By the same analysis used above, an MZ twin seeking
his or her cotwin on the basis of just these five variables will
narrow the search area to less than 1% so that, from a group of
100 candidate cotwins, a specific and singular pairing is likely
to occur. For our lonely bachelor, however, in his neighbor-
hood, at the singles bar, at the church social, at school, or at
work, the entire literature on spousal similarity would elimi-
nate not more than half the young women whom he encounters
and will therefore leave him still unmated.
Just deserts, or equity, model. Another model of mate selec-
tion, promulgated by Goffman (1952) and Blau (1968), suggests
that each of us assesses our ownmate value and then seeks in a
search area populated by prospective mates to whom we attri-
bute mate values similar to that which we attribute to ourselves.
In other words, this model suggests that we seek our “just de-
serts,” scorning potential mates whose value we assess much
lower than our own and not aspiring to mates with mate values
much higher than we could offer in return. This model has
considerable plausibility, so much indeed that one feels it must
be true, at least at the extremes; the ordinary person who be-
comes fixated on a movie star is considered to be odd. Begin-
ning with Walster, Aronson, Abrahams, and Rottman (1966), a
series of studies, reviewed by Berscheid and Walster (1978, pp.
182- 19 1), led these authors to conclude that “People do seem to
end up choosing partners of approximately their own ‘social
worth: ” (p. 190).
The work of Buss(1984,1989)and others demonstrates two
important facts relating to the equity model of mate selection:
With remarkable consistency across diverse cultures, (a) people
express similar preferences and, within sex, similar rank order-
ing of preference in their descriptions of an ideal mate; and (b)
men rank women highest who are young and beautiful,
whereas women pay more attention to status and earning power
in evaluating men. Thus, there appears to be cross-cultural gen-
erality combined with consistent sex-specific differences in the
formulae that people use in computing mate value.
On the other hand, the mate-value assessments that influ-
ence mate selection must be the subjective assessments of the
two individuals involved, and these may not correspond to the
assessments that others would make. Feingold (1992), in a re-
cent meta-analysis, found that self- and other-ratings of physi-
cal attractiveness share less than 5% of common variance. Mur-
stein (1976) asked 98 young married couples to rate, on a 5-
point scale, the physical attractiveness of their spouse and of
themselves; eight judges also rated, on the same scale, the attrac-
tiveness of each spouse from photographs. The judges rated
2 1% to 24% of the spouses as above average in attractiveness,
whereas 67% of the wives and 85% of the husbands rated their
spouses above average (39% and 43%, respectively, rated them-
selves above average, pp. Й9- 150). Because physical attractive-
ness is an important component of mate value, especially for
men (see above), these data suggest the possibility that mate
value may be to some extent a consequence, rather than a pre-
dictor, of mate selection, that the causal sequence is “I want her;
she’s beautiful” rather than “She’s beautiful; I want her.”
As a theory of mate selection, we suggest that the just deserts
model can be regarded as a facet of the similarity model; just as
we tend to confine our search area to people similar to our-
selves in IQ, physical attractiveness, traditionalism, and the
like, so too do we tend to acquire mates not too different from
ourselves in mate value, as we make that subjective assessment.
It is probable that our ratings of mate value will be correlated
with, and predictable from, our estimates of the candidates’
scores (or of their similarity to ourselves) on these same dimen-
sions. Even if we assume that our assessment of a candidate’s
mate value is wholly independent of our assessment of physical
attractiveness, traditional values, IQ, and the like, there is no
reason for supposing that the spousal correlation for mate
value, as it was assessed by both partners before their selection
of each other, is any higher than, say, the spousal correlation for
good looks, that is to say, about .50. Therefore, it seems unlikely
that incorporating the just deserts model into the similarity
model would substantially narrow the search area, and our hy-
pothetical seeker would still be left to choose among at least
half of all the potential mates who meet his eligibility require-
ments. In other words, it appears that the models in question
are capable, at best, of identifying(some of) those whom he will
not select but are not capable of determining-or even strongly
focusing-his affirmative choice. Both models succeed in di-
recting him into the appropriate line at the mating cafeteria, the
line that features the dishes he can afford. that he is accus-
tomed to, and finds appropriate (let us call it the kosher line),