Four (More) Arguments Against the Torrance Tests
John Baer
Rider University
I want to thank Division 10, and especially Kyung Hee Kim, for
making this debate possible. It has been interesting and illuminat-
ing, and I look forward to future Division 10 debates.
I have space here to counter only a few of Kim’s arguments (this
issue), most of which I’ve already addressed (Baer, this issue). In
this short response, I’d like to focus on four ideas: (a) the use of
multiple measures of giftedness, (b) interpretation of Torrance
Tests of Creative Thinking (TTCT) subscale scores, (c) what
polymaths mean for domain specificity theory, and (d) the need for
TTCT proponents to be consistent in their standards.
(a) Like Kim, I strongly endorse the use of multiple measures of
ability, and especially ones that tap creativity— but not ones that
lack validity, which is the problem with TTCT scores. I agree with
the many authors Kim cited (in her Is Using the TTCT Inappro-
priate? section) who have argued for more diverse and inclusive
measures of ability. But it’s important to note that none of these
experts has endorsed the use of the TTCT, and some of them are
actively working on measures of creativity that can be used instead
of the TTCT. This certainly isn’t due to lack of familiarity—no one
would argue that the TTCT lacks a substantial track record—and
one can only conclude that, although all of these authors endorse
and encourage the use of multiple measures of ability, none of
them has found the TTCT a sufficiently valid or useful tool. They
all seem to agree with Simonton’s (2003) summary of the current
state of creativity assessment:
None of these suggested measures can be said to have passed all the
psychometric hurdles required of established ability tests. For in-
stance, scores on separate creativity tests often correlate too highly
with general intelligence (that is, low divergent validity), correlate
very weakly among each other (that is, low convergent validity), and
correlate very weakly with objective indicators of overt creative
behaviors (that is, low predictive validity; p. 216)
(b) Although Torrance strongly discouraged the use of single
scores (as Kim noted) and encouraged looking for patterns among
subscores to provide insight into individual abilities, this is rarely
how the tests are in fact used. Overall scores are commonly used
without reference to subscale scores, and when subscale scores are
used, they are often simply used as proxies for overall creativity.
(See, e.g., Ke ´ri’s recent (2009) Psychological Science report of a
link between psychosis and creativity. Rather than seeking patterns
among subscale scores, each is treated as an independent measure
of overall creativity.) It is possible that nuanced interpretations of
subscale scores might provide insight into an individual’s cogni-
tive functioning, but because this is not how the scores are gen-
erally used, I agree with the advice that Kim reminds us Torrance
frequently gave: Don’t use any TTCT scores (including index
scores) as general measures of creativity. If that advice were
followed, however, it would eliminate most common uses of the
TTCT.
(c) Domain specificity theory predicts the existence of poly-
maths. Contrary to Kim’s argument, their existence supports do-
main specificity, which never claims that creative ability in any
one domain limits creative ability in other domains. It simply says
they are unrelated. Just as someone could have musical talent,
leadership skills, and high-jumping ability, people can also be
creative in multiple domains. Domains specificity theory predicts
polymathy— but it also predicts it will be somewhat rare. Domain
generality (as assumed by the TTCT) would predict polymathy to
be far more common than it actually is (Kaufman, Beghetto, &
Baer, in press a; Kaufman, Beghetto, Baer, & Ivcevic, in press b).
(d) The same evidence used to validate the verbal TTCT inval-
idates the figural TTCT. Kim claimed that Plucker’s (1999) re-
analysis of Torrance’s data shows the TTCT is “the best predictor
for adult creative achievements.” But Plucker’s analysis showed
that only one of the two TTCT tests predicted the self-reported
measures of creative accomplishment used in the validation. The
verbal TTCT did not. If a positive prediction supports the verbal
TTCT, intellectual consistency requires Kim to acknowledge that
the failure of the figural TTCT to predict the same outcomes
invalidates the figural TTCT. I think a more reasonable response is
to question the self-reported measures of achievement used in that
study—I don’t believe they can possibly validate any version of
the TTCT— but if one accepts Plucker’s reanalysis and the self-
report data on which it relies, one must then dismiss the figural
TTCT as an invalid measure. You can’t simply move the goalposts
to fit the outcomes you hope to find.
Sometimes no data is better than bad (invalid) data. That’s why
we would be better off without the Torrance Tests, at least in the
ways they are most commonly used.
References
Baer, J. (this issue). How divergent thinking tests mislead us: Are the
Torrance Tests still relevant in the 21st century? The Division 10 debate.
Psychology of Aesthetics, Creativity, and the Arts.
Kaufman, J. C., Beghetto, R. A., & Baer, J. (in press a). Finding young Paul
Robesons: The search for creative polymaths. In D. D. Preiss & R. J.
Sternberg (Eds.), Innovations in educational psychology: Perspectives
on learning, teaching, and human development. New York, NY:
Springer.
Kaufman, J. C., Beghetto, R. A., Baer, J., & Ivcevic, Z. (in press b).
Creativity polymathy: What Benjamin Franklin can teach your kinder-
gartener. Learning and Individual Differences.
Ke ´ri, S. (2009). Genes for psychosis and creativity: A promoter polymor-
Correspondence concerning this article should be addressed to John
Baer, Rider University, Memorial Hall 102, Lawrenceville, NJ 08648.
E-mail:baer@rider.edu
Psychology of Aesthetics, Creativity, and the Arts © 2011 American Psychological Association
2011, Vol. 5, No. 4, 316 –317 1931-3896/11/$12.00 DOI: 10.1037/a0025211
316