In criterion-related validity studies, the framework of Brunswik Symmetry
(Wittmann & Suß, 1999) is useful for considering the roles of Gf and Gc, or
maximal performance and typical behaviors. This framework states that
maximal validity is obtained when there is both a direct correspondence be-
tween the predictor space and the criterion space, and the breadth of the pre-
dictors and the breadth of the criterion are matched. That is, broad predic-
tors are most appropriate for broad criteria, and narrow predictors are most
appropriate for narrow criteria, as long as the correspondence is matched. If
the wrong narrow predictor is used to predict a narrow criterion, it will per-
form more poorly than a broad predictor. Thus, when it comes to validating
intelligence measures, a high level of Brunswik Symmetry is obtained when
the intelligence measure has ample representation of both maximal perform-
ance and typical behavior, because primary school success is predicated on
both maximal performance (e.g., aptitude tests) and typical behaviors (e.g.,
cumulative grades on homework, in-class assignments, and end of term
achievement assessments).
Such omnibus intelligence tests, however, are less well suited for predicting
post-secondary academic performance (e.g., the Stanford–Binet or Wechsler
tests have poorer predictive validities for college and university performance,
compared to primary school performance, even after taking into account the
restriction-of-range of talent and explicit prior selection at post-secondary in-
stitutions), because the criteria for academic success are dominated by typical
behavior measures (e.g., term papers and final examinations) and less de-
pendent on maximal performance. Performance in graduate school and be-
yond is better predicted by measures of domain knowledge (which falls under
the wide category of Gc and typical behavior) than it is predicted by abstract
reasoning or other general aptitude measures (e.g., see Willingham, 1974).
One of the popular criticisms of the traditional IQ-type tests is that they do
not well predict occupational performance (Anastasi, 1982). From the per-
spective provided above, it seems clear that one reason for such findings is
that there is a lack of Brunswik Symmetry, stemming from the inclusion of
maximal performance measures when they have relatively less impact on oc-
cupational performance than measures of typical intellectual behaviors.
AFFECT AND TYPICAL–MAXIMAL PERFORMANCE
Where modern intelligence assessments have increasingly focused on maxi-
mal performance (especially in terms of those who advocate using only Ra-
ven-type tests for assessing intellect), measures of affect (personality) have fo-
cused on typical behaviors. Indeed, one could argue that the underlying
theme of the trait–situation controversy—where researchers argued about
whether traits or situations had dominant influences on behavior (Mischel,
1968; Rushton, Brainerd, & Pressley, 1983) was largely predicted on a form
122 ACKERMAN AND KANFER