PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES

Decreased Latent Inhibition Is Associated With Increased Creative

Achievement in High-Functioning Individuals

Shelley H. Carson Jordan B. Peterson

Harvard University University of Toronto

Daniel M. Higgins
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Harvard University
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Reductions in latent inhibition (LI), the capacity to screen from conscious awareness stimuli previously
experienced as irrelevant, have been generally associated with the tendency towards psychosis. However,
“failure” to screen out previously irrelevant stimuli might also hypothetically contribute to original
thinking, particularly in combination with high IQ. Meta-analysis of two studies, conducted on youthful
high-IQ samples, demonstrated that high lifetime creative achievers had significantly lower LI scores
than low creative achievers (reffect size ⫽ .31, p ⫽ .0003, one-tailed). Eminent creative achievers
(participants under 21 years who reported unusually high scores in a single domain of creative
achievement) were 7 times more likely to have low rather than high LI scores, ␹2(1, N ⫽ 25) ⫽ 10.69,
␾ ⫽ .47, p ⫽ .003.

How do you know but ev’ry Bird that cuts the airy way,
Is an immense world of delight, clos’d by your senses five?
—William Blake, “A Memorable Fancy,” The Prophetic Books
Creative individuals appear characterized in part by the ability
to perceive and describe what remains hidden from the view of
others. Individual variation in latent inhibition (LI), a cognitive
inhibitory mechanism discovered by animal experimentalists in the
late 1950s, may account for the apparent revelation to the creative
mind of what appears “clos’d by the senses five” to others. LI
refers to the varying capacity of the brain to screen from current
attentional focus stimuli previously experienced as irrelevant
(Lubow, 1989). The LI phenomenon appears robust across a
variety of mammalian species, and its biological underpinnings
have been extensively studied (Lubow & Gewirtz, 1995). In hu-
mans, reduced LI has generally been associated with susceptibility
to or actual acute-phase schizophrenia (Baruch, Hemsley, & Gray,
Shelley H. Carson and Daniel M. Higgins, Department of Psychology,
Harvard University; Jordan B. Peterson, Department of Psychology, Uni-
versity of Toronto, Toronto, Ontario, Canada.
We thank Melanie Glickson for her research assistance in this study. We
also thank Robert Sternberg for his helpful comments. This research was
supported by grants from the Harvard University Department of Psychol-
ogy and the University of Toronto Department of Psychology.
Correspondence concerning this article should be addressed to Shelley
H. Carson, Department of Psychology, Harvard University, 33 Kirkland
Street, Cambridge, Massachusetts 02138. E-mail: carson@wjh.harvard.edu
Journal of Personality and Social Psychology, 2003, Vol. 85, No. 3, 499 –506
Copyright 2003 by the American Psychological Association, Inc. 0022-3514/03/$12.00
1988a, 1988b; Lubow, Ingberg-Sachs, Zalstein-Orda, & Gewirtz,
1992). Recent evidence has suggested, however, that reductions in
LI are also associated with the personality trait Openness to
Experience (Peterson & Carson, 2000; Peterson, Smith, & Carson,
2002). Openness, in turn, has been consistently associated with
divergent thinking and trait creativity (McCrae, 1987) and with
creative achievement (King, Walker, & Broyles, 1996). It there-
fore appears possible that reductions in LI may be associated with
increases in human creativity, as suggested by Eysenck (1995).
Many researchers (e.g., Simonton, 1988, 1999) have proposed
that the cognitive processes of individuals capable of creating the
highest achievements in their fields are both qualitatively and
quantitatively different from those of normal thinkers (although
some, like Weisberg, 1993, dispute the “qualitative” distinction). If
qualitative differences do exist, however, one potential source of
difference in the cognitive processes between eminent creative
achievers and other intelligent thinkers may be in the relative
attenuation of LI. Such attenuation could well increase the number
of available mental elements, described by Simonton (1988) as
key, in part, to the process of creative discovery.
Other avenues of research have buttressed this hypothesis.
Dykes and McGhie (1976) demonstrated, for example, that cre-
ative subjects and schizophrenic subjects were better than controls
at identifying items presented on the irrelevant channel of a di-
chotic shadowing task. This finding appears to support Dellas and
Gaier’s (1970) observation that creative individuals tend not to
screen out so-called “irrelevant details.” Furthermore, in a study
conducted by Martindale, Anderson, Moore, and West (1996), a
DOI: 10.1037/0022-3514.85.3.499

499
 500 CARSON, PETERSON, AND HIGGINS
small group of high-creative subjects showed reduced galvanic
skin response habituation rates to auditory stimuli compared with
low-creative subjects. This finding, which has been replicated in
psychosis-prone populations (Raine, Benishay, Lencz, & Scarpa,
1997), may indicate that highly creative people do not precatego-
rize stimuli as irrelevant in the same manner as less creative
individuals. Finally, Eysenck (1995) has noted that originality (the
ability to produce statistically unusual ideas) is conceptually sim-
ilar to the looseness of associations symptomatic of psychosis.
Such looseness is presumably a byproduct of the failure of an
inhibitory filtering mechanism, functioning to limit associations to
those relevant to current task performance. Reduced LI scores are
theoretically associated with relaxation of this inhibitory mecha-
nism (Gray, Feldon, Rawlins, Hemsley, & Smith, 1991).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
If reduced LI represents a predisposing factor common to psy-
chosis and to creativity, what then distinguishes the psychotic from
the poet? Studies that have associated decreased LI with schizo-
phrenia or schizotypy have typically failed to report subject IQ.
However, among individuals with a predisposition to psychosis,
low IQ generally functions as an unfavorable moderating variable
(David, Malmberg, Brandt, Allebeck, & Lewis, 1997; Jones &
Offord, 1975). Furthermore, several investigators (e.g., Claridge,
1997; Eysenck, 1995; Berenbaum & Fujita, 1994) have suggested
that superior intellectual qualities, such as high IQ, may moderate
the expression of a predisposition to psychosis in the highly
creative individual. It seems reasonable to propose that some
psychological phenomena might be pathogenic in the presence of
decreased intelligence, psychometrically defined, but normative or
even abnormally useful in the presence of increased intelligence.
The purpose of the present studies was therefore to (a) deter-
mine if a variety of indicators of individual creativity, particularly
creative achievement, would be associated with reduced LI; (b)
determine if reduced LI in tandem with high IQ would predict
higher general creative achievement scores than either factor in
isolation; and (c) determine if a combination of high IQ and
reduced LI would identify those participants who have made
especially noteworthy contributions to their respective creative
fields.
We limited our two participant samples to a creatively diverse
but high-IQ population to substantively increase the probability of
assessing individuals who were genuinely highly creative, because
a wide body of research (summarized in Eysenck, 1995) has
already identified IQ as a critically important factor in creative
achievement. This body of research has suggested that there may
be an IQ threshold (usually described as an IQ of approximately
120 points) associated with true creative ability (for a review, see
Sternberg & O’Hara, 1999).
We also present a meta-analysis of the combined results of both
of our studies to provide a more accurate estimate of effect size for
LI and, as well, we describe the methods and results of an addi-
tional analysis of eminent creative achievers versus highly intelli-
gent noncreative individuals from the pooled sample.
STUDY 1: LI, CREATIVE ACHIEVEMENT, AND
OTHER MEASURES OF CREATIVITY
One of the major obstacles to creativity research has been the
variety of diverse concepts and definitions associated with the term
creative (Brown, 1989), a descriptor that applies with equal facility
to achievement, mental processes (Torrance, 1968), a personality
trait or combination of traits (Gough, 1979; McCrae, 1987), and
consensually approved novel products (Amabile, 1996). We are
particularly interested in lifetime creative achievement, per se,
because such achievement is beneficial not only to the individual
but to society as well. Furthermore, other measures of creativity,
including creative personality traits and creative mental process
measures (divergent thinking), have been found to be elevated in
certain nonproductive disordered populations as well (e.g.,
Cramond, 1994; Keefe & Magaro, 1980) and thus may not be
unique to highly creative individuals.
However, our initial examination of the LI– creativity associa-
tion used multiple measures of creativity designed to cover a wide
range of concepts, including a lifetime creative achievement mea-
sure (the Creative Achievement Questionnaire [CAQ]; Carson,
Peterson & Higgins, 2003), a mental process measure (divergent
thinking tasks; Torrance, 1968), and a personality measure (the
Creative Personality Scale [CPS]; Gough, 1979).1
For Study 1, we hypothesized that (a) high scorers relative to
low scorers on the CAQ would be characterized by attenuated LI.
We also hypothesized that (b) high scorers on the CPS and the
divergent thinking tasks would demonstrate attenuated LI relative
to low scorers. Finally, we hypothesized that (c) low LI scores
combined with high IQ scores would predict high CAQ scores.
Method
Participants
Eighty-six Harvard undergraduates (33 men, 53 women), with a mean
age of 20.7 years (SD ⫽ 3.3) participated in the study. All were recruited
from sign-up sheets posted on campus. Participants were paid an hourly
rate.
Procedure
Participants were assigned randomly to either the preexposed (n ⫽ 57)
or the nonpreexposed (n ⫽ 29; a ratio of approximately 2:1) LI condition.
During a pretest interview in which demographic information was col-
lected, participants were assessed for outward signs of depression, anxiety,
or boredom. They were also questioned about alcohol and/or caffeine
intake prior to the lab testing. Two participants were rescheduled because
of recent caffeine intake. All participants then completed creativity mea-
sures, IQ tests, personality tests, and the LI task.
Creativity Testing
Creativity testing consisted of three phases: the CAQ, the divergent
thinking tasks, and the CPS.
CAQ. The CAQ is an empirically sound measure of lifetime creative
accomplishment in the fields of art and science. The instrument has
demonstrated test–retest reliability in the range of r ⫽ .85 as well as good
convergent, discriminant, and predictive validity. Participants check off
actual achievements in 10 separate domains of creative accomplishment
(e.g., “My work has won a prize or prizes at a juried art show”). Scores for
individual accomplishments within a domain are weighted according to the
1
An additional measure related to creativity, the personality variable
Openness to Experience, was also assessed for its relationship to LI in this
sample. However, analysis of this relationship has been reported elsewhere
(Peterson & Carson, 2000).
 LATENT INHIBITION AND CREATIVE ACHIEVEMENT
level of achievement attained, as ranked by experts within that domain.
Weighted scores for each domain are summed to provide a total creative
achievement score (Carson et al., 2003).
Divergent thinking tasks. Four validated divergent thinking tasks were
adapted from Torrance (1968). Participants were given 3 min to write down
their responses to each task, such as naming alternate uses for a common
object. Three aspects of divergent thinking were assessed: fluency (number
of responses generated), originality (unusualness of responses, based on the
statistical infrequency of each individual response within the current sam-
ple), and flexibility (number of different categories of response and the
number of category changes). Fluency, originality, and flexibility scores
were z scored and summed to produce a divergent thinking score for each
subject.
CPS. The CPS (Gough, 1979) consists of a set of 30 items from the
Adjective Check List (Gough & Heilbrun, 1965), which has predicted high
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
levels of creativity across multiple studies and diverse samples. Partici-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
pants describe themselves by checking off 18 positively and 12 negatively
scored adjectives.
IQ Testing
Participants completed the Vocabulary and Block Design subtests of the
Wechsler Adult Intelligence Scale—Revised (WAIS-R; Wechsler, 1981).
Raw scores were age scaled, combined to form a composite, and converted
to a full-scale equivalent, using standard guidelines (Brooker & Cyr, 1986).
IQ estimates compiled from this “short form” correlate at r ⫽ .91 with
full-scale WAIS-R IQ scores (Brooker & Cyr, 1986). Because IQ scores
using the short form typically overestimate IQ by 3 points (Brooker & Cyr,
1986), IQ scores were adjusted by subtracting 3 points from the total for
each participant.
LI Task
Participants were seated in a quiet, semidarkened room and told that they
were to participate in two auditory discrimination tasks. They then donned
stereo earphones and adjusted the volume to ensure clarity and compre-
hensibility. Participants in the preexposed condition were then shown a
two-part video version of a well-validated and commonly used auditory LI
task, after Lubow et al. (1992).
In Part 1, the preexposure phase, a list of 30 nonsense syllables (the
masking material) was presented 5 times at a normal rate of speech with no
indication of the termination and start of each repetition. White noise bursts
(the target stimuli) from 3 to 6 s in duration were superimposed ran-
domly 31 times over the course of the recording, at approximately two-
thirds the volume of the nonsense syllables. Participants listened to the
recording through earphones and were asked to determine how many times
they heard a selected nonsense syllable.
In Part 2 (the test phase), the recording of the nonsense syllables and the
white noise bursts was replayed in identical form while yellow disks
arranged in rows on a black scoreboard were revealed individually on the
video screen. Each yellow disk appeared prior to the offset of the target
stimulus. Participants were asked to determine what auditory stimulus
signaled the appearance of the yellow disks. The individual’s score for the
task (trials to rule identification) was determined by the number of yellow
disks visible on the screen (maximum 31) when the correct answer was
given.
Participants in the nonpreexposed condition were shown an identical
videotape, except that the target stimulus was absent from the preexposure
phase of the task.
Traditionally, LI is demonstrated (and the LI test validated) within a
population when participants of that population tested in the preexposed
condition take more trials to learn the association of the target stimulus
with the yellow disks than similar participants tested in the nonpreexposed
condition (Lubow et al., 1992). Thus, lower LI scores in the preexposed
501
condition suggest attenuated LI, whereas higher scores suggest intact or
“normal” LI.
Results
Table 1 lists the zero-order correlations of all creativity mea-
sures with each other as well as with IQ.
Creative Achievement and LI
Participants were divided into low- (n ⫽ 45, M ⫽ 6.7, SD ⫽ 3.9)
and high- (n ⫽ 41, M ⫽ 27.0, SD ⫽ 11.3) creative achievement
groups by CAQ score median split.2 Members of the low-creative
achievement group in the preexposed condition (n ⫽ 28) scored
significantly higher on the LI task than the low-creative achievers
in the nonpreexposed condition (n ⫽ 17), indicating a significant
LI effect, t(43) ⫽ ⫺2.03, p ⫽ .04, d ⫽ .62 (see Figure 1). Members
of the high-creative achievement group in the preexposed condi-
tion (n ⫽ 29), by contrast, did not score significantly higher on the
LI task than their counterparts in the nonpreexposed condition
(n ⫽ 12), indicating an attenuation of LI in the high-creative
achievement group, t(39) ⫽ ⫺.020, p ⫽ .96, d ⫽ .01. An analysis
of variance (ANOVA) indicated significant differences in LI
scores due to the effects of high- and low-creative achievement,
F(1, 82) ⫽ 4.09, p ⫽ .05, ␩2 ⫽ .05. The main effect for condition,
F(1, 82) ⫽ 2.12, p ⫽ .15, ␩2 ⫽ .03, and the interaction effect
between CAQ and condition, F(1, 82) ⫽ 2.04, p ⫽ .16, ␩2 ⫽ .02,
were not significant.
Because LI scores were clearly bimodal rather than normally
distributed (see Figure 2), nonparametric analyses (Mann–Whitney
U) were also conducted and yielded similar results: Low-creative
achievement participants demonstrated intact LI (Z ⫽ ⫺1.86, p ⫽
.06), whereas high-creative achievement participants demonstrated
reduced LI (Z ⫽ ⫺.043, p ⫽ .96).
As hypothesized, the mean LI score of the low-creative achieve-
ment group (n ⫽ 28, M ⫽ 21.8, SD ⫽ 10.4) was significantly
higher than that of the high-creative achievement group (n ⫽ 29,
M ⫽ 14.3, SD ⫽ 8.8) in the preexposed condition, t(55) ⫽ 2.93,
p ⫽ .006, d ⫽ .79 (nonparametric Mann–Whitney U: Z ⫽ ⫺2.54,
p ⫽ .01) (see Figure 3).
LI and Measures of Divergent Thinking and Creative
Personality
A comparison of means indicated that LI scores in the preex-
posed condition were significantly reduced in the high-scoring
groups (determined by median split) of the CPS and the originality
dimension of the divergent thinking tasks as well as in overall
divergent thinking (see Table 2). Nonparametric Mann–Whitney U
tests revealed similar results.
LI and IQ as Factors of Creative Achievement
The mean IQ of the sample was 128.1 points (SD ⫽ 10.3), with
a range of 97 to 148 points. To determine whether LI and IQ would
2
“Median split” is traditionally used in LI research (e.g. Baruch et al.,
1988a, 1988b; Lubow et al., 1992) and thus is included here. However,
median splits arguably lack theoretical meaning. A theoretically driven
division of high- and low-creative achievers is presented in the ANALY-
SIS OF EMINENT CREATIVE ACHIEVERS section.
 502 CARSON, PETERSON, AND HIGGINS

Table 1
Correlations of Creativity and IQ Measures

Measure CAQ CPS Diverg Fluency Orig Flex

CPS .33*
Diverg .47* .29*
Fluency .38* .11 .85*
Orig .46* .32* .86* .59*
Flex .37* .34* .87* .62* .63*
IQ .13 .26* .10 ⫺.03 .19 .08

Note. CAQ ⫽ Creative Achievement Questionnaire; CPS ⫽ Creative

Personality Scale; Diverg ⫽ divergent thinking; Orig ⫽ originality; Flex ⫽
flexibility.
* p ⬍ .05.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

jointly predict creative achievement scores, we regressed LI scores

(preexposed condition) and IQ scores on the creative achievement
scores of the highest and lowest quartiles of CAQ scorers (n ⫽ 30).
Negative LI scores and positive IQ scores jointly predicted 26% of
the variance in creative achievement scores, F(2, 27) ⫽ 4.89, p ⫽
.006, R2 ⫽ .26, with LI scores alone accounting for 18% of the
total CAQ variance, F(1, 28) ⫽ 6.09, p ⫽ .02, R2 ⫽ .18.3 An
additional analysis of the CAQ, IQ, and LI relationship is offered
below in Study 2 and, cumulatively, in the section ANALYSIS OF
EMINENT CREATIVE ACHIEVERS.

STUDY 2: LI AND CREATIVE ACHIEVEMENT

In Study 2, we attempted to replicate our findings within the
Harvard population, concentrating primarily on the relationship of
LI to lifetime creative achievement. Because we had already
established the existence of an LI effect for this population using
a nonpreexposed control group (and because the LI scores of high
creative achievers did not differ significantly from low achievers
in the control group), we decided to concentrate on individual
difference scores within the preexposed condition.
We also wished to control for the chance that participants would
respond to the self-report measure of creative achievement, the
CAQ, in a self-enhancing manner. We therefore administered the
Figure 2. Bimodal distribution of latent inhibition scores in the preex-
posed condition in Studies 1 and 2.

Marlowe–Crowne Social Desirability Scale (MCSD; Crowne &

Marlowe, 1960), a measure of the tendency of responders to tailor
their responses to appear socially acceptable.

Method

Participants
Ninety-six Harvard undergraduates (53 men, 43 women), with a mean
age of 20.1 years (SD ⫽ 1.6) participated in the study. All were registered
in a psychology course and received course credit for participation.

Figure 1. Creative Achievement Questionnaire (CAQ) and latent inhibi-

3
tion (LI) scores in the preexposed and nonpreexposed conditions. Error The regression of IQ and LI (preexposed) on the entire sample was as
bars represent the standard error (positive value only) for the mean LI score follows: F(2, 58) ⫽ 2.84, p ⫽ .07, R2 ⫽ .10, with LI (preexposed)
of each high–low CAQ ⫻ Condition group. accounting for 9% of the variance.
 LATENT INHIBITION AND CREATIVE ACHIEVEMENT
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Figure 3. Latent inhibition (LI) scores (preexposed condition) of high-
and low-creative achievers as measured by the Creative Achievement
Questionnaire (CAQ). Error bars represent the standard error (positive
value only) for the mean LI score (preexposed) of each high–low CAQ
group.
Procedure
All participants completed a computerized version of the CAQ identical
to the paper-and-pencil version used in Study 1 as well as a computerized
version of the MCSD. In subsequent lab sessions, participants were ad-
ministered the LI task in the experimental condition and the IQ tests. Both
the LI and the IQ procedures were identical to those used in Study 1.
Results
Four participants were eliminated from the study because of
MCSD scores more than 2 standard deviations above the mean,
indicating a strong possibility of self-enhancement in responses on
the CAQ.
Creative Achievement and LI
Participants were divided into low-creative achievement (low
CAQ; n ⫽ 47, M ⫽ 4.9, SD ⫽ 3.1) and high-creative achievement
(high CAQ; n ⫽ 45, M ⫽ 22.3, SD ⫽ 9.7) groups by median split,
as in Study 1 (see Figure 3). As predicted, participants in the
high-CAQ group had significantly lower LI scores (M ⫽ 12.8,
SD ⫽ 10.1) than did participants in the low-CAQ group
(M ⫽ 17.6, SD ⫽ 11.4); t(90) ⫽ 2.13, p ⫽ .04, d ⫽ .45 (Mann–
Whitney U: Z ⫽ 2.09, p ⫽ .04).
LI and IQ as Factors of Creative Achievement
Eighty-two of the 96 participants completed the IQ tasks. The
mean IQ of the sample was 124.6 points (SD ⫽ 11.4), ranging
from 100 to 148 points. As in Study 1, we regressed LI scores
(preexposed condition) and IQ scores on the upper and lower
quartiles of CAQ scores (n ⫽ 46). Negative LI scores and positive
IQ scores jointly predicted 20% of the variance in CAQ scores,
F(2, 43) ⫽ 5.33, p ⫽ .006, R2 ⫽ .20, with LI scores alone
503
accounting for 13% of the total CAQ variance, F(1, 43) ⫽ 6.54,
p ⫽ .01, R2 ⫽ .13.4
META-ANALYSIS: LI AND CREATIVE
ACHIEVEMENT
The results of analyses of the relationship between creative
achievement scores and LI scores from Studies 1 and 2 were
combined using meta-analytic techniques (Rosenthal & Rosnow,
1991). Because both studies used samples of similar IQ and
creative achievement, Studies 1 and 2 were weighted equally in the
meta-analysis. When LI scores were compared between high-CAQ
and low-CAQ groups (median split), the combined results of the
two studies yielded highly significant differences with an effect
size of r ⫽ .31 ( p ⫽ .0003, one-tailed; see Table 3).
ANALYSIS OF EMINENT CREATIVE ACHIEVERS
The results of Studies 1 and 2 indicate that participants who
scored higher in overall creative achievement exhibited signifi-
cantly lower LI scores than participants who did not score high in
creative achievement. However, a high score in combined areas of
creative achievement may not necessarily be indicative of the
likelihood of truly eminent achievement in a specified field. One of
the defining characteristics of the truly eminent creative achiever,
however, is the production of a significant accomplishment in his
or her respective field at a young age (Ludwig, 1995). By pooling
the participants from both studies, we hoped to identify enough
eminent creative achievers (those who made a truly substantive
contribution to a single domain of creative endeavor before the age
of 21) to allow for analysis of LI and IQ.
We defined a score of at least 12 points on the CAQ in any one
of the 10 individual domains of measured creative achievement as
constituting a clearly significant contribution to a creative field.
The minimum levels of achievement that qualified as a significant
accomplishment according to this criterion included having a
novel or book of poetry published and sold, having a musical
composition recorded and sold, having a prototype invention pat-
ented and built, having a private showing of original artwork at a
recognized gallery, or winning a scholarship or national prize for
a scientific discovery. We then compared the LI scores of the
identified group of eminent creative achievers with the scores of
the group of individuals who demonstrated minimal creative
achievement.
Method
Participants included pooled subjects from Studies 1 and 2 who met the
following criteria: (a) They were members of the preexposed condition of
the LI task, and (b) they had scored either 12 or more points on any single
domain of the CAQ (eminent creative achievers; n ⫽ 25), or had a total
CAQ score of 3 or below (controls; n ⫽ 23). Eminent creative achievers
included 4 artists, 5 composers, 2 writers, 2 inventors, 3 dramatists, 7
scientists, and 2 choreographers. Mean LI scores of eminent achievers were
then compared with the LI scores of the control subjects.
4
The regression of IQ and LI (preexposed) on the entire sample was as
follows: F(2, 79) ⫽ 2.70, p ⫽ .07, R2 ⫽ .07, with LI (preexposed)
accounting for 5% of the variance.
 504 CARSON, PETERSON, AND HIGGINS

Table 2
Latent Inhibition (LI) Scores of the High Versus Low Groups of Individual Measures of
Creativity (Based on Median Split)

LI score of LI score of
Measure high group low group t df p d Z p

CPS 3.63 47 ⬍ .01 1.06 ⫺3.48 ⬍ .01

M 11.0 20.8
SD 7.0 10.1
Diverg 2.09 52 .04 0.58 ⫺1.78 .08
M 15.6 21.4
SD 9.5 10.9
Orig 2.34 52 .02 0.65 ⫺2.26 .02
M 15.1 21.5
SD 9.7 10.3
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Fluency 1.58 52 .12 0.43 ⫺1.37 .18

This document is copyrighted by the American Psychological Association or one of its allied publishers.

M 16.4 21.1
SD 9.5 11.5
Flex 1.01 52 .32 0.28 ⫺0.88 .38
M 16.7 19.6
SD 9.5 11.4

Note. Z ⫽ Mann–Whitney U test; CPS ⫽ Creative Personality Scale; Diverg ⫽ divergent thinking; Orig ⫽
originality; Flex ⫽ flexibility.

Results mean IQ for the control group was just below the 120 threshold
(M ⫽ 118.3, SD ⫽ 11.9). Twenty-one of the 25 eminent achievers
LI and Eminent Creative Achievers had IQs above the 120 threshold, whereas 11 of the 23 controls had
IQs above the threshold.
The LI scores of the eminent creative achievers (M ⫽ 11.1,
The combined sample of eminent creative achievers and con-
SD ⫽ 7.6) were significantly lower than the LI scores of the
trols was divided into high-IQ (M ⫽ 130.0, SD ⫽ 6.6, n ⫽ 32) and
control group (M ⫽ 19.4, SD ⫽ 10.5), t(46) ⫽ 3.17, p ⫽ .004, d ⫽
moderate-IQ (M ⫽ 111.1, SD ⫽ 7.6, n ⫽ 16) groups on the basis
.93 (Mann–Whitney U: Z ⫽ 2.65, p ⫽ .008). Whereas control
of the proposed threshold score of IQ 120. The combined sample
subjects were equally likely to display either high or low LI scores
was also divided into high-LI (M ⫽ 27.6, SD ⫽ 3.8, n ⫽ 16) and
(on the basis of the manifest division in the LI distribution),
low-LI (M ⫽ 6.6, SD ⫽ 4.7, n ⫽ 32) groups on the basis of the
eminent creative achievers were seven times more likely to have
naturally occurring split in bimodal LI scores in the preexposed
low rather than high LI scores, ␹2(1, N ⫽ 25) ⫽ 10.79, ␾ ⫽ .47,
condition. A 2 ⫻ 2 factorial ANOVA examined the CAQ scores
p ⫽ .001 (see Figure 4).
using high LI/low LI and high IQ/moderate IQ as factors. The
results indicated a significant difference among the creative
LI and IQ as Factors of Eminent Creative Achievement achievement scores of the groups, F(3, 44) ⫽ 8.58, p ⬍ .001, ␩2 ⫽
As mentioned above, a number of researchers have noted a .37, with the low-LI/high-IQ group demonstrating substantially
correlation between creative ability and IQ levels below 120, higher CAQ scores than all other groups, as predicted (see Figure
whereas correlations between creativity and IQs above 120 are
generally less significant. This threshold theory suggests that an IQ
of 120 may be necessary but not sufficient for creative achieve-
ment (for a review, see Sternberg & O’Hara, 1999; for an alter-
native view, see Lubinski, Webb, Morlock, & Benbow, 2001). In
our sample, the mean IQ for the eminent creative achiever group
was above the 120 threshold (M ⫽ 128.6, SD ⫽ 8.3), whereas the

Table 3
Difference in Mean Latent Inhibition Scores of High Versus Low
Creative Achievers: A Meta-Analysis

Study t Effect size p

1 2.93a .400 .0025

2 2.13b .219 .018
Total .314 .0003

Note. All p values are one-tailed. Figure 4. Contingency table of high and low latent inhibition (LI) scores
a
df ⫽ 55. b df ⫽ 90. (preexposed) for eminent creative achievers versus controls.
 LATENT INHIBITION AND CREATIVE ACHIEVEMENT
5). The main effects for both LI, F(1, 44) ⫽ 7.30, p ⫽ .01, ␩2 ⫽
.14, and IQ, F(1, 44) ⫽ 5.44, p ⫽ .02, ␩2 ⫽ .11, were significant,
and the interaction between the two factors approached signifi-
cance, F(1, 44) ⫽ 3.47, p ⫽ .07, ␩2 ⫽ .07.
When we regressed LI (preexposed condition) and IQ scores on
the CAQ scores, negative LI scores and positive IQ scores jointly
predicted almost one third of the variance in creative achievement
scores, F(2, 45) ⫽ 9.55, p ⫽ .0003, R2 ⫽ .30, with LI scores alone
accounting for 19% of the total CAQ variance, F(1, 46) ⫽ 10.53,
p ⫽ .002, R2 ⫽ .19. The LI ⫻ IQ interaction was highly signifi-
cant, F(1, 46) ⫽ 15.81, p ⫽ .008, ␩2 ⫽ .98.
GENERAL DISCUSSION
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
The results of these studies and analyses indicate a substantial
and significant relationship between a variety of indicators of
creativity and reduced LI. In Study 1, significant differences in LI
scores emerged between high- and low-creative participants on
measures of creative achievement, creative personality, and the
originality facet of divergent thinking. The results of Study 2
replicated the relationship between creative achievement and re-
duced LI. A meta-analysis of the two studies indicated a nontrivial
and highly significant relationship between creative achievement
and reduced LI, and an analysis of eminent creative achievers
(relative to noncreative control subjects) suggested a near-
universal reduction in LI in this group.
How do we reconcile the fact that reduced LI scores in humans,
which have previously been associated with psychotic states or
with psychotic proneness, are associated with high levels of cre-
ativity in these studies? The results provide substantial although
circumstantial evidence that high IQ might serve as a beneficial
moderating factor in the expression of LI as either a deficit in
selective attention or a facilitator of creativity. In all of our studies
and analyses, high IQ, when combined with low LI, was associated
with increased creative achievement. These results are particularly
stunning in the analysis of eminent achievers and high-functioning
controls. High IQ clearly appeared to augment the tendency toward
high creative achievement characteristic of low-LI individuals.
Figure 5. Creative Achievement Questionnaire (CAQ) scores of high–
low latent inhibition (LI) and high–moderate IQ groups in pooled eminent
achievers and controls. Error bars represent the standard error (positive
value only) for the mean CAQ score of each high–low LI ⫻ IQ group.
505
These results lend support to the theory that there may be
qualitative (e.g., failure to filter out irrelevant stimuli) as well as
quantitative (e.g., high IQ) differences in the processes underlying
creative versus normal cognition. These results also support the
theory that highly creative individuals and psychotic-prone indi-
viduals may possess neurobiological similarities, perhaps geneti-
cally determined, that present either as psychotic predisposition on
the one hand or as unusual creative potential on the other on the
basis of the presence of moderating cognitive factors such as high
IQ (e.g., Berenbaum & Fujita, 1994; Dykes & McGhie, 1976;
Eysenck, 1995). These moderating factors may allow an individual
to override a “deficit” in early selective attentional processing with
a high-functioning mechanism at a later, more controlled level of
selective processing. The highly creative individual may be priv-
ileged to access a greater inventory of unfiltered stimuli during
early processing, thereby increasing the odds of original recombi-
nant ideation. Thus, a deficit that is generally associated with
pathology may well impart a creative advantage in the presence of
other cognitive strengths such as high IQ.
One of the limitations of the present research was the use of a
single measure of LI. We are currently developing a multimethod
procedure for assessing LI that will hopefully further clarify the
relationship between reduced LI and creativity in future research.
We are also investigating other possible moderating factors (in-
cluding working memory capacity and personality dimensions)
that may have an impact on the creativity–LI relationship.
References
Amabile, T. M. (1996). Creativity in context. Boulder, CO: Westview
Press.
Baruch, I., Hemsley, D. R., & Gray, J. A. (1988a). Differential perfor-
mance of acute and chronic schizophrenics in a latent inhibition task.
Journal of Nervous and Mental Disease, 176, 598 – 606.
Baruch, I., Hemsley, D. R., & Gray, J. A. (1988b). Latent inhibition and
“psychotic proneness” in normal subjects. Personality and Individual
Differences, 9, 777–783.
Berenbaum, H., & Fujita, F. (1994). Schizophrenia and personality: Ex-
ploring the boundaries and connections between vulnerability and out-
come. Journal of Abnormal Psychology, 103, 148 –158.
Brooker, B. H., & Cyr, J. J. (1986). Tables for clinicians to use to convert
WAIS-R Short Forms. Journal of Clinical Psychology, 42, 983.
Brown, R. T. (1989). Creativity: What are we to measure? In J. A. Glover,
R. R. Ronning, & C. R. Reynolds (Eds.), Handbook of creativity (pp.
3–32). New York: Plenum Press.
Carson, S., Peterson, J. B., & Higgins, D. (2003). Validity, reliability, and
factor structure of the Creative Achievement Questionnaire. Manuscript
submitted for publication.
Claridge, G. (Ed.). (1997). Schizotypy: Implications for illness and health.
New York: Oxford University Press.
Cramond, B. (1994). Attention-deficit hyperactivity disorder and creativ-
ity: What is the connection? Journal of Creative Behavior, 28, 193–210.
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability
independent of psychopathology. Journal of Consulting Psychology, 24,
349 –354.
David, A. S., Malmberg, A., Brandt, L., Allebeck, P., & Lewis, G. (1997).
IQ and risk for schizophrenia: A population-based cohort study. Psy-
chological Medicine, 27, 1311–1323.
Dellas, M., & Gaier, E. L. (1970). Identification of creativity: The indi-
vidual. Psychological Bulletin, 73, 55–73.
Dykes, M., & McGhie, A. (1976). A comparative study of attentional
 506 CARSON, PETERSON, AND HIGGINS
strategies of schizophrenic and highly creative normal subjects. British
Journal of Psychiatry, 128, 50 –56.
Eysenck, H. J. (1995). Creativity as a product of intelligence and person-
ality. In D. Saklofske & M. Zeidner (Eds.), International handbook of
personality and intelligence: Perspectives on individual differences (pp.
231–247). New York: Plenum Press.
Gough, H. G. (1979). A creative personality scale for the Adjective Check
List. Journal of Personality and Social Psychology, 37, 1398 –1405.
Gough, H. G., & Heilbrun, A. B. (1965). The Adjective Check List manual.
Palo Alto, CA: Consulting Psychologists Press.
Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D. R., & Smith, A. D.
(1991). The neuropsychology of schizophrenia. Behavioral and Brain
Sciences, 14, 1– 84.
Jones, M. B., & Offord, D. R. (1975). Independent transmission of IQ and
schizophrenia. British Journal of Psychiatry, 126, 185–190.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Keefe, J. A., & Magaro, P. A. (1980). Creativity and schizophrenia: An
This document is copyrighted by the American Psychological Association or one of its allied publishers.
equivalence of cognitive processing. Journal of Abnormal Psychol-
ogy, 89, 390 –398.
King, L., Walker, L., & Broyles, S. (1996). Creativity and the five-factor
model. Journal of Research in Personality, 30, 189 –203.
Lubinski, D., Webb, R. M., Morelock, M. J., & Benbow, C. P. (2001).
Top 1 in 10,000: A 10-year follow-up of the profoundly gifted. Journal
of Applied Psychology, 86, 718 –729.
Lubow, R. E. (1989). Latent inhibition and conditioned attention theory.
Cambridge, England: Cambridge University Press.
Lubow, R. E., & Gewirtz, J. C. (1995). Latent inhibition in humans: Data,
theory, and implications for schizophrenia. Psychological Bulletin, 117,
87–103.
Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C.
(1992). Latent inhibition in low and high “psychotic-prone” normal
subjects. Personality and Individual Differences, 15, 563–572.
Ludwig, A. M. (1995). The price of greatness: Resolving the creativity and
madness controversy. New York: Guilford Press.
Martindale, C., Anderson, K., Moore, K., & West, A. N. (1996). Creativity,
oversensitivity, and the rate of habituation. Personality and Individual
Differences, 20, 423– 427.
McCrae, R. R. (1987). Creativity, divergent thinking, and openness to
experience. Journal of Personality and Social Psychology, 52, 1258 –
1263.
Peterson, J. B., & Carson, S. (2000). Latent inhibition and openness to
experience in a high-achieving student population. Personality and
Individual Differences, 28, 323–332.
Peterson, J. B., Smith, K. W., & Carson, S. (2002). Openness and Extra-
version are associated with reduced latent inhibition: Replication and
commentary. Personality and Individual Differences, 33, 1137–1147.
Raine, A., Benishay, D., Lencz, T., & Scarpa, A. (1997). Abnormal
orienting in schizotypal personality disorder. Schizophrenia Bulletin, 23,
75– 82.
Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research:
Methods and data analysis. New York: McGraw-Hill.
Simonton, D. K. (1988). Scientific genius: A psychology of science. New
York: Cambridge University Press.
Simonton, D. K. (1999). Origins of genius: Darwinian perspectives on
creativity. New York: Oxford University Press.
Sternberg, R. J., & O’Hara, L. A. (1999). Creativity and intelligence. In
R. J. Sternberg (Ed.), Handbook of creativity (pp. 251–272). New York:
Cambridge University Press.
Torrance, E. P. (1968). Examples and rationales of test tasks for assessing
creative abilities. Journal of Creative Behavior, 2, 165–178.
Wechsler, D. (1981). WAIS-R Manual: Wechsler Adult Intelligence Scale—
Revised. San Antonio, TX: Psychological Corporation.
Weisberg, R. W. (1993). Creativity: Beyond the myth of genius. New York:
Freeman.
Received February 6, 2002
Revision received January 16, 2003
Accepted January 21, 2003 䡲