Why We Sing the Blues: The Relation Between Self-Reflective Rumination, Mood, and Creativity

Emotion
Copyright 2005 by the American Psychological Association
2005, Vol. 5, No. 2, 226 –232
1528-3542/05/$12.00
DOI: 10.1037/1528-3542.5.2.226
Why We Sing the Blues: The Relation Between Self-Reflective
Rumination, Mood, and Creativity
Paul Verhaeghen
Jutta Joormann
Syracuse University
Stanford University
Rodney Khan
Syracuse University
Past research has shown that creative behavior is associated with a higher risk for depression. The authors
hypothesized that a 3rd underlying factor, namely, self-reflective rumination, may explain the connec-
tion. This hypothesis was examined in a sample of 99 undergraduate college students, using path analysis.
The authors found that self-reported past depressive symptomatology was linked to increased self-
reflective rumination. Rumination, in turn, was related to current symptomatology and to self-rated
creative interests and objectively measured creative fluency, originality, and elaboration. No direct link
existed between currently depressed mood and either creative interest or creative behavior. These results
suggest that the association between depression and creativity is solely the result of rumination.
One of the more puzzling findings in the field of creativity
These findings are puzzling. Depression is a debilitating disor-
research is the well-documented link between creative behavior
der, and its defining symptoms include diminished interest in all
and mood disorder. Although the nature and extent of this rela-
activities, loss of energy, indecisiveness, and lack of concentra-
tionship has been debated, a wide range of studies using diverse
tion—not symptoms readily associated with creative behavior. A
methods clearly suggest a connection. For instance, in one of the
direct link between symptomatology and creativity is unlikely.
first studies investigating creative individuals, Andreasen (1987)
First, there is anecdotal evidence concerning creative productivity
compared 30 creative writers and 30 matched controls, including
in individuals with bipolar disorder that suggests that the depres-
first-degree relatives of both groups. The creative writers and their
sive episode decreases productivity rather than increases it. The
first-degree relatives showed significantly higher rates of affective
composer Robert Schumann, for instance, wrote most of his works
disorders, both unipolar and bipolar depression. Indeed, 24 of the
in a state of hypomania and remained silent during episodes of
30 writers were diagnosed with an affective disorder compared
severe depression (Slater & Meyer, 1959). Furthermore, in a more
with only 9 of the 30 participants in the control group. In a similar
recent study on creativity and bipolar diathesis, Shapiro and Weis-
study, Ludwig (1994) included 59 writers and a control group and
berg (1999) found no indication that current depressive symptom-
reported that whereas 59% of the writers were depressed, only 9%
atology was linked to creativity. Additionally, the finding that
of the control group fulfilled diagnostic criteria for depression. In
ingestion of mood stabilizers such as lithium actually aids produc-
his survey of the biographies of 1,004 eminent individuals living in
tivity rather than diminishes it (see studies reviewed in Jamison,
the 20th century, Ludwig (1995) found a lifetime prevalence of
1993, p. 245) suggests that the depressive episode itself is indeed
depression of 50% for people working in the creative arts, com-
debilitating, as expected from its defining symptoms.
pared with 20% of those in the field of enterprise, 24% of scien-
The lack of a reasonable direct causal link between depressed
tists, and 27% of important social figures. Particularly vulnerable
mood and creative behavior makes it likely that a third, underlying
to depression were writers of poetry (77%) and fiction (59%) and
variable is at play. In our reading of the depression literature, we
visual artists (50%). Jamison (1993) concluded from an overview
were struck by the close relationship between self-reflective rumi-
of primary research literature that major depressive illness is 8 –10
native thinking and negative affect, most notably depression. In-
times as prevalent in writers and artists than in the general popu-
deed, in their analysis of the high prevalence rates of mental
lation; additionally, writers and artists are about 10 times more
likely to commit suicide.
disorder in female poets, Kaufman and Baer (2002) recently pro-
posed a link between rumination, depression, and poetry writing.
They argued that the introspection and rumination that character-
izes depression may also be involved in writing poetry and, in
Paul Verhaeghen and Rodney Khan, Department of Psychology and
addition, that rumination and introspection may in fact increase
Center for Health and Behavior, Syracuse University; Jutta Joormann,
instability in poets who are already vulnerable to mental disorder.
Department of Psychology, Stanford University.
Rumination has been defined as “a class of conscious thoughts
Correspondence concerning this article should be addressed to Paul
that revolve around a common instrumental theme and that recur in
Verhaeghen, Department of Psychology, Syracuse University, 430 Hun-
the absence of immediate environmental demands requiring the
tington Hall, Syracuse, NY 13244, or to Jutta Joormann, Department of
Psychology, Jordan Hall, Stanford University, Stanford, CA 94305. E-
thought” (Martin & Tesser, 1996, p. 7). Persistence of the thought
mail: pverhaeg@psych.syr.edu or joormann@psych.stanford.edu
is an additional important characteristic (Martin & Tesser, 1996).
226

---

RUMINATION, MOOD, AND CREATIVITY
227
Self-reflective rumination (e.g., Nolen-Hoeksema, Morrow, &
between depressed symptomatology and creative behavior will be
Fredrickson, 1993; Treynor, Gonzalez, & Nolen-Hoeksema, 2003)
explained by the relation of both to self-reflective rumination.
uses the self, that is, one’s inner feelings, memories, and so forth,
as the recurrent theme. According to Nolen-Hoeksema (1991),
Method
rumination is characterized by being a style of thought rather than
by its negative content. This implies that rumination by itself does
Participants
not necessarily promote depression, because even when healthy
people show this style of thinking, they do not necessarily focus on
Participants were 99 undergraduates from Syracuse University, who
negative affect or on negative personal attributes. However, a large
received either course credit or $10 in cash in return for their participation.
body of literature suggests that a ruminative thinking style in-
In order to maximize the variety of creative behavior in the sample, we
recruited participants from the introductory class in psychology, from the
creases one’s vulnerability to depression and maintains negative
Arts and Sciences Honors program, from selected classes in the Fine Arts
affect when the focus is on negative life events or when the
program, and from one campus writers group.
individual experiences frequent negative mood states. In a large
meta-analysis, Mor and Winquist (2002) found that the effect of
rumination on negative affect is indeed quite large (d
1.08 in
Measures
correlational studies and 0.76 in experimental studies). Mor and
Depression.
Severity of current depressive symptoms was measured
Winquist argued for a reciprocal relationship between mood and
using the Center for Epidemiological Studies Depression Scale (CES-D;
rumination on the grounds of experimental evidence that shows, on
Radloff, 1977). The CES-D is a short self-report scale designed to measure
the one hand, that inducing negative affect leads to increased
depressive symptomatology in the general population. It consists of 20
self-reflective rumination and, on the other hand, that inducing
items assessing depressed mood, feelings of guilt and worthlessness, help-
self-reflective rumination leads to increased negative affect.
lessness and hopelessness, psychomotor retardation, concentration prob-
Rumination could be linked to creativity through a common
lems, appetite disturbance, and sleep problems. For each item, participants
underlying style of thought. In particular in writers and poets, a
indicate on a 4-point scale ranging from 0 (rarely or none of the time [less
than 1 day]) to 3 (most or all of the time [5–7 days]) how frequently they
focus on the self and one’s feelings may be an important part of
have experienced that symptom during the past week. Thus, scores can
creative activity. Interestingly, bipolar patients in a study by Jami-
range from 0 to 60, with a recommended cutoff score of 16 or greater,
son, Gerner, Hammen, and Padesky (1980) reported that their
indicating a significant level of depression. The CES-D has acceptable
depression led to a heightened sensitivity toward their feelings,
reliability and validity (e.g., Gotlib & Cane, 1989). In the present sample,
which they claimed had an impact on their creativity. Other au-
the CES-D had a Spearman–Brown reliability coefficient of .78.
thors have suggested that depression’s facilitative role in the
Past depressive symptoms were measured using a brief checklist of the
creative process is because of increased introspection, resulting in
10 symptoms of depression, as indexed in the Diagnostic and Statistical
a heightened sensitivity to inner content (Richards, 1981). Rich-
Manual of Mental Disorders, 4th edition (American Psychiatric Associa-
ards further suggested that this form of introspection contributes to
tion, 1994). Participants were instructed to indicate by circling yes or no
the content of creative work that later, during periods of improved
next to each symptom whether they experienced this particular symptom
over the past year for a period of 2 weeks or longer. In the present sample,
functioning, emerges as a product of the creative process. Another
this measure had a Spearman–Brown reliability coefficient of .78.
possible cause for the link between rumination and creativity may
Creative interests.
We designed a questionnaire that listed 20 artistic/
lie deeper still. Recent research has suggested that the style of
creative activities (painting, drawing, computer graphics, photography,
thinking that is evident in rumination is the consequence of an
writing prose, dancing, acting, film/video, playing an instrument, etc.). For
underlying deficit in executive control processes, namely, cogni-
each activity, participants indicated how many hours per week they en-
tive inhibition (Hertel, 1997; Joormann, 2004, in press). Deficits in
gaged in this behavior (we labeled this variable time). They also indicated
the ability to screen out goal-irrelevant or previously irrelevant
on a 5-point scale ranging from 1 (just fooling around) to 5 ( potential
information may, on the one hand, lead to rumination but may, on
career) how serious they were about this activity (we labeled this serious-
the other hand, also contribute to original thinking. In line with this
ness). Participants could add activities if they wished; few did so. We
suggestion, Carson, Peterson, and Higgins (2003) recently dem-
calculated the average seriousness score and the total number of hours
engaged in creative activities as indicators of creative interests.
onstrated that eminent creative achievers were seven times more
Creative behavior.
The Abbreviated Torrance Test for Adults (ATTA;
likely to have low rather than high scores on a latent-inhibition
Goff & Torrance, 2002) consists of three “activities,” each taking 3 min. In
task.
the first activity, the participants imagine what would happen if “you could
The present study explored the hypothesis that depression and
walk on air or fly” and what problems this may create. They list as many
creative behavior may be linked through the third underlying
of these problems as they can. In the next two activities, simple abstract
factor of self-reflective rumination. More specifically, we hypoth-
shapes are provided on a page, and the participants are invited to use these
esize that self-reflective rumination will be correlated with both
“to make some pictures” that are “unusual” and “interesting” and to give
depressed symptomatology and creative behavior. One of the
each picture a title. The ATTA is scored on three scales. Fluency is the
major challenges of research on creativity has been the measure-
number of distinct answers generated. Originality is the number of re-
ment of this construct, which equally applies to achievement,
sponses that do not appear on the list of common answers provided by the
test manual. Elaboration is the number of details contained within the
mental process, a personality trait, or the creation of a novel
answers; the test manual provides strict scoring criteria for these.
product. No consensus measure has emerged. We therefore used
Three items from the Purdue Creativity Test (PCT; Lawshe & Harris,
multiple measures of creativity, including mental-process mea-
1960) were also included. Each item consists of an abstract line drawing,
sures (viz. the dimensions of creativity: Fluency, Originality, and
and the participant is invited to write down as many answers to the question
Elaboration), and measures of interest and seriousness about cre-
“What is this?” as they can in 2 min. From these answers, a fluency score
ative activities. We predict that a substantial part of the correlation
was calculated by tallying the total number of responses.

---

228
VERHAEGHEN, JOORMANN, AND KHAN
To test whether the three assumed scales (i.e., Fluency, Originality, and
Procedure
Elaboration) were indeed present in the data, a confirmatory factor analysis
Participants were tested in small groups (the largest group included 20
was conducted using LISREL (see Figure 1); in the model, we allowed for
individuals) in quiet rooms. Each session lasted about 40 min. Each
error covariance for measures obtained from the same item or activity. This
participant received a printed booklet with all the tests and questionnaires.
model fit the data well, 2(38, N
99)
50.51, root-mean-square error of
Because of their timed nature, the creativity tests were administered first,
approximation (RMSEA)
.057.
starting with the ATTA, then the PCT. After the creativity testing was
Self-reflective rumination.
Participants filled out the 41-item Rumina-
completed, participants worked through the questionnaires at their own
tive Responses Scale (RRS; Nolen-Hoeksema & Morrow, 1991; this scale
pace. The questionnaires were administered in the following order: creative
includes the subscales Rumination, Distance, Problem Solving, and Dan-
interests questionnaire, RRS, CES-D, and past depressive symptoms. De-
gerous Activities). This scale asks participants to indicate on a 4-point scale
pressive symptoms were assessed last because items dealing with dyspho-
(1
almost never, 2
sometimes, 3
often, or 4
almost always) how
ria may aversely affect the participants’ mood, and we wanted to prevent
often they “think or do” each of the items when they “feel down, sad, or
contamination of the other measures by these potential mood shifts.
depressed.” Thus, the scale measures responses to dysphoric mood that are
focused on the self, on symptoms, or on possible consequences and causes
Data Analysis
of moods. In addition, behavioral responses to sad moods are assessed.
Previous studies have shown good test–retest reliability (Nolen-Hoeksema,
Path analysis in LISREL was used. Because the sample is relatively
Parker, & Larson, 1994) and acceptable convergent and predicitive validity
small compared with the number of variables estimated if a full model were
(Nolen-Hoeksema & Morrow, 1991; Nolen-Hoeksema et al., 1994). Re-
fitted (i.e., a model with a measurement portion and a path-analytic model
cently, Treynor et al. (2003) have argued that many of the Rumination
on latent factors), we decided to conduct the path analysis on unit-weighted
items on the RRS are directly depression-related. They therefore advocate
composites (i.e., sum scores for CES-D, for the past depression question-
using only a 5-item subset of nondepression-related items from the RRS as
naire, for the number of hours of creative activities, and for the Reflec-
a relatively pure and depression-free measure of rumination-as-reflection
tiveness questionnaire, the average for seriousness, and unit-weighted
(hereafter called Reflectiveness). We used this scale for our analysis.
average z scores for the three scales of creativity—Fluency, Originality,
Typical items are “Write down what you are thinking about and analyze it”
and Elaboration). Because the measurement units are arbitrary, we report
and “Go away by yourself and think about why you feel this way.” In the
the coefficients from the completely standardized solution (these can be
present sample, the Reflectiveness scale had a Spearman–Brown reliability
read as analogous to beta weights in a regression analysis). Alpha level for
coefficient of .86.
statistical testing was set at p
.05.
Figure 1.
Results from a confirmatory factor analysis (LISREL) for the three factors of creative behavior
assumed present in the Abbreviated Torrance Test for Adults and the Purdue Creativity Test. 2(38, N
99)
50.51, p
.08, RMSEA (root-mean-square error of approximation)
.057.

---

RUMINATION, MOOD, AND CREATIVITY
229
Results
Table 2
Means and Standard Deviations of the Variables Included in the
Table 1 contains the correlation matrix; descriptive data are
Path Analysis
provided in Table 2. The data analysis was done in two progressive
steps. In the first step, we mounted a path model among the five
Variable
M
SD
creativity variables—seriousness of creative activities, time in-
vested in creative activities, Fluency, Originality, and Elaboration.
Current depression (CES-D)
16.07
9.23
Past depression (checklist)
4.16
2.69
In the second step, we added Reflectiveness and the two depres-
Reflectiveness (from RRS)
2.43
0.75
sion variables to the creativity cluster. This two-step approach was
Seriousness about creative activities
0.99
0.63
taken to minimize the risks associated with data-driven model
Hours/weeks spent on creative activities
33.96
37.63
fitting on a large number of variables, notably, the risk of getting
Elaboration
0.02
0.87
Originality
0.01
0.65
trapped in a local minimum.
Fluency
0.01
0.66
The baseline creativity model assumed that creative interests
(seriousness and time) would influence each of the three aspects of
Note. N
99. CES-D
Center for Epidemiological Studies Depression
creative behavior independently. This starting model is in accor-
Scale; RRS
Ruminative Responses Scale; Elaboration, Originality, and
dance with the literature that suggests that internal motivation and
Fluency are composite z scores.
experience are driving forces for creative behavior (e.g., Amabile,
1996; Ericcson, Krampe, & Tesch-Ro¨mer, 1993). In this model,
99)
18.44, RMSEA
.016. It was further improved by freeing
seriousness directly influenced time invested, which in turn di-
the path between Reflectiveness and Fluency, 2(17, N
99)
rectly influenced Fluency, Originality, and Elaboration, with no
14.61, RMSEA
.000. This final model is depicted in Figure 2.
paths specified between the latter three variables. This model did
We should point out that this model is mathematically equivalent
not fit the data well, 2(6, N
99)
52.03, RMSEA
.281. The
to a model in which the path between past depression and Reflec-
modification indices provided by LISREL were used to guide
tiveness is reversed, that is, a model in which Reflectiveness is the
further fitting efforts. The first additional path that was freed was
exogenous variable, exerting a direct influence on depression and
the path between Fluency and Originality. This resulted in the
on the creativity cluster, which have no direct link between them.
nonsignificance of the path between hours and Originality. After
deleting this path, model fit was, 2(6, N
99)
34.53, RMSEA
Discussion
.221. The next path freed was between seriousness and Fluency.
After deleting the now nonsignificant path between hours and
Previous research, including studies of creative individuals and
Fluency, fit statistics were,
2(6, N
99)
22.64, RMSEA
historiometric analysis, has shown a link between creative behav-
.169. The next path freed was between Fluency and Elaboration,
ior and depression. Our hypothesis was that a common underlying
2(5, N
99)
12.11, RMSEA
.121. In the final model, the
psychological characteristic, namely, a tendency for self-reflective
path between Elaboration and Originality was freed. This model
rumination, may be the source of this correlation. Specifically, we
was considered final because no further significant improvement
claimed that self-reflection independently (a) increases the risk for
could be made; it fit the data well,
2(4, N
99)
5.02,
depression and (b) spurs interest in and ability for creative behav-
RMSEA
.051.
ior. Our results lend full support to this hypothesis.
In the second step, Reflectiveness and the two depression vari-
Our final path model shows that the data can be accounted for
ables (past depression and current depression) were added to the
without any direct link between currently depressed mood and
model. As stated in the introduction, present research suggests that
either creative interest (as measured by an activities checklist) or
the relation between self-reflectiveness and depression is recipro-
creative behavior (as measured by scores on three dimensions of
cal. Therefore, we set up the model with past depression as the
creativity obtained from two standard creativity tests). Rather, the
exogenous variable, influencing current depression and Reflective-
correlation between depressed mood and creative behavior appears
ness, and Reflectiveness as influencing both current depression
to be because of the link of both of these variables with self-
and the creativity cluster through seriousness only (seriousness
reflective rumination. In our final model (see Figure 2), past
was chosen because this was the exogenous variable for the
depression does explain some of the variance in creative behavior,
creativity model). This model fit the data very well,
2(18, N
but only through its link with self-reflection. We should point out
Table 1
Correlation Matrix for the Variables Included in the Path Analysis
Variable
1
2
3
4
5
6
7
8
1. Reflectiveness
—
2. Current depression
.36
—
3. Past depression
.28
.50
—
4. Fluency
.29
.18
.15
—
5. Originality
.09
.06
.05
.50
—
6. Elaboration
.22
.08
.19
.37
.01
—
7. Seriousness
.35
.27
.18
.33
.00
.35
—
8. Time
.15
.17
.08
.16
.00
.27
.48
—

---

230
VERHAEGHEN, JOORMANN, AND KHAN
Figure 2.
Final best-fitting path model (LISREL) to explain the relationships between the depression variables,
self-reflective rumination, creative interests, and creative behavior. 2(17, N
99)
14.61, p
.62, RMSEA
(root-mean-square error of approximation)
.000.
that this link, although significant, explains only 8% of the vari-
creative fluency (i.e., the number of ideas generated within a
ance. The large majority of the variance in self-reflective rumina-
certain time period). In other words, individuals who are more
tive behavior remains unexplained by self-reported past depres-
likely to examine their lives are also more likely to actively search
sion. Additionally, this link can be removed from the model at no
for a creative outlet, maybe as a venue to express and/or share their
cost to model fit by assuming that self-reflective rumination is the
feelings. They then spend more time in creative activities, and this
driving force behind both past depression and current depression.
serves to make them more fluent in the creative realm. In the
Whichever of these models one prefers, it is clear that the connec-
ability pathway, self-reflectiveness has a direct influence on cre-
tion between depression and creativity appears to be spurious, and
ative fluency. No direct link exists between self-reflectiveness and
because of the reliance of both on self-reflective attention.
the two other creative ability variables, Originality or Elaboration,
These results are in line with research on psychopathology and
but Reflectiveness does influence both through creative fluency.
creativity that largely suggests that although depressive disorders
The pathway between Fluency and Originality is in agreement
and bipolar disorders are linked to creativity, it is not the depres-
with predictions derived from variation-selection models of cre-
sive symptomatology, that is, the anhedonia and the negative mood
ative behavior (e.g., Campbell, 1960; Simonton, 1997). Such mod-
state, that is responsible for this link (Shapiro & Weisberg, 1999).
els posit that the number of novel useful creative ideas (i.e.,
Indeed, studies that have examined the influence of mood states on
Originality) will be a function of the total number of ideas gener-
creativity in normal populations find that it is positive affect that
ated (i.e., Fluency). Simply put, if a person is able to generate a
enhances creative problem solving and increases the unusualness
large number of ideas, then she or he will also be likely to generate
of word associations (e.g., Isen, Daubman, & Nowicki, 1987).
a large number of original ideas. The available evidence in the
Rather, as has been suggested by Richards (1981), our results
field of professional creativity indeed suggests that the most pro-
suggest that heightened self-reflection and rumination as major
ductive individuals also deliver the highest number of works
characteristics of depression may contribute to the content of
considered truly creative (e.g., Albert, 1975; Simonton, 1994). The
creative work, which during periods of improved functioning
link between Fluency and Elaboration can be explained through
emerges as a product of the creative process.
variation-selection models (e.g., Simonton, 1997), which posit that
How exactly does self-reflective rumination influence creative
the rate at which ideas are elaborated is directly proportional to the
behavior? Kaufman and Baer (2002) proposed that this relation-
number of ideas that have been generated. That is, individuals who
ship may be because of the fact that individuals who suffer from
are fluent in spawning new ideas will also be likely to work on
depression are more likely to ruminate and introspect and that
those ideas.
some of these individuals may turn this rumination into poetry.
In summary, then, we propose that self-reflective rumination
Alternatively, if writers are already vulnerable to depression, then
prepares individuals to generate a larger number of ideas. This
rumination and introspection may increase their instability. Our
enhanced fluency, in turn, allows for the emergence of more
model suggests two pathways, one through motivation, the other
creative ideas and for increased elaboration. Interestingly, re-
through ability. In the motivational pathway, self-reflection pro-
searchers have recently suggested that a combination of stop rules
motes seriousness about creative endeavors. Seriousness, in turn,
and current mood state may explain perseveration and termination
influences both the number of hours one spends on creative activ-
in analyzing the causes and consequences of problems (Davey,
ities (which in turn leads to enhanced elaboration of ideas) and
Startup, Zara, MacDonald, & Field, 2003; Startup & Davey, 2001).

---

RUMINATION, MOOD, AND CREATIVITY
231
Particularly, Watkins and Mason (2002) hypothesized that rumi-
it is likely that the model captures true psychological mechanisms
nators use a default “as many as can” stop rule and take their
rather than, for instance, the influence of selection mechanisms
negative mood state as information that they are not satisfied with
that certainly exist within the creative professions (e.g., one may
the number of items generated. Consequently, this mood-as-
argue that successful artists and writers are “allowed” to exhibit
information hypothesis may provide an explanation as to why a
pathological behavior because of their relatively marginal status in
tendency to ruminate in depressed participants results in enhanced
society and their self-employment). It is, however, also possible
fluency. The outlined direct connection between self-reflectiveness
that age— or, more likely, continuing experience with the creative
and Fluency and between Fluency and Originality, however, may
life—moderates the relationship between self-focused attention
also point to another underlying common mechanism. As de-
and creativity. For instance, increasing levels of instrumental skill
scribed in the introduction, researchers have suggested that rumi-
in one’s creative endeavors (e.g., Ericcson et al., 1993) or an
nation is because of failures of cognitive inhibitory mechanisms
age-related increase in the skill of emotion regulation (e.g.,
that limit the contents of consciousness to those relevant to a
Carstensen, Fung, & Charles, 2003) may attenuate or overshadow
current goal or to a current task performance (Hertel, 1997; Joor-
the relationships that are strongly present in our sample. In addi-
mann, in press). Similarly, Fluency and Originality have been
tion, if the rumination– creativity relationship is the result of an
linked to dysfunctions in cognitive inhibition. Carson et al. (2003)
underlying lack of inhibitory control, as proposed above, then
recently suggested that attenuated cognitive inhibition may in-
age-related differences in executive control processes (Hasher &
crease the number of available mental elements, described by
Zacks, 1988) may result in significant age-related changes of this
Simonton (e.g., 1994, 1997) as central to creative thought. As
association.
outlined in the introduction, Carson et al. (2003) found a near-
A third limitation is that the present sample size was too small
universal reduction of latent inhibition in a group of eminent
to take different types of creativity into account. The data on
creative achievers. Thus, cognitive inhibitory dysfunctions may
depression and creativity suggest that the effects should operate
lead to an increased access to a greater inventory of unfiltered
mainly in writers and artists and less so in entertainers and actors.
stimuli during early processing. Depending on the thought content
A fourth limitation is that the present sample size did not allow
that preoccupies the individual at a given moment, this process
for the estimation of a full LISREL model, that is, a path model
may increase the odds for original recombinant ideation or in-
including latent variables. Rather, we had to obtain confirmation of
crease the odds for recurrent, negative self-focused thinking.
our measurement model first; for our path model, unit-weighted
Clearly, future studies are needed to clarify possible underlying
composites were used to estimate scores for each construct. This
mechanisms of the self-reflection–rumination and creativity
procedure leaves the path coefficients less free of measurement
association.
error than a full model would.
One negative path between creative behaviors was found,
Our findings have two broad implications. The first implication
namely, between Elaboration and Originality. This suggests that
is that the cliche´ that the artist must suffer is not true. Artists and
thinking original thoughts may come at the expense of pursuing
writers often subscribe to this view (as cited in Jamison, 1993; the
those thoughts to practical fruition. Interestingly, within research
painter Edvard Munch felt that taking his sufferings away would
on rumination and inhibition, it has been shown that reduced
“destroy [his] art,” adding, “I want to keep these sufferings,” p.
inhibition and increased rumination increase the number of off-
241) and may therefore underutilize therapeutic possibilities. The
goal, task-irrelevant thoughts, and hamper the problem-solving
second implication is that self-reflectiveness could well explain
process. For instance, in a recent study, Davis and Nolen-
other aspects of creativity, more specifically, the established link
Hoeksema (2000) showed that dysphoric ruminators cannot inhibit
between creative behavior and asocial personality characteristics,
off-goal thoughts and fail to maintain productive lines of reason-
notably, introversion (Feist, 1998) and Eysenck’s P factor (essen-
ing. Consequently, ruminators tend to become trapped in nonpro-
tially nonconformist tendency; Eysenck, 1995). Both of these
ductive perseveration on negative moods and events. This negative
personality factors may enhance the self-reflective tendencies in
path is offset by a positive path between Fluency and Elaboration,
the individual, and this may be part of the reason why they are
summing up to a near-zero correlation between Originality and
correlated with creativity.
Elaboration.
So why do we sing the blues? Our study suggests that we
We should point out four limitations of the present study. First,
actually do not—it is simply the case that people who are more
the study was not meant to explain all of the variance in creative
likely to have the blues are the ones who are more likely to sing (or
behavior. Real-life creativity is obviously embedded in a much
paint or write). Both—the blues and the singing—appear to be
more complex context than the one sketched here (a context that
rooted in a maybe all-too-closely examined life.
includes variables such as personality, Feist, 1998, 1999; intelli-
gence, Sternberg & O’Hara, 1999; sociohistorical circumstances,
References
Simonton, 1984, 1994; deliberate practice, Ericcson et al., 1993;
and social support systems, Fleming & Hollinger, 1994).
Albert, R. S. (1975). Toward a behavioral definition of a genius. American
A second limitation is that the population we used was a sample
Psychologist, 30, 140 –151.
Amabile, T. M. (1996). Creativity in context. Boulder, CO: Westview
of college students, albeit skewed to yield a larger number of
Press.
creative individuals than we typically would expect. Obviously,
American Psychiatric Association. (1994). Diagnostic and statistical man-
the present study needs to be replicated using a sample that
ual of mental disorders (4th ed.). Washington, DC: Author.
includes professional artists and writers. This limitation may,
Andreasen, N. C. (1987). Creativity and mental illness: Prevalence rates in
however, also be a strength: If these pathways exist in a sample of
writers and their first-degree relatives. American Journal of Psychiatry,
young individuals who are not yet professionals in their fields, then
144, 1288 –1292.

---

232
VERHAEGHEN, JOORMANN, AND KHAN
Campbell, D. T. (1960). Blind variation and selective retention in creative
Kaufman, J. C., & Baer, J. (2002). I bask in dreams of suicide: Mental
thought as in other knowledge processes. Psychological Review, 67,
illness, poetry, and women. Review of General Psychology, 6, 271–286.
380 – 400.
Lawshe, C. H., & Harris, D. H. (1960). Purdue Creativity Test. West
Carson, S. H., Peterson, J. B., & Higgins, D. M. (2003). Decreased latent
Lafayette, IN: Purdue Research Foundation.
inhibition is associated with increased creative achievement in high-
Ludwig, A. M. (1994). Mental illness and creative activity in women
functioning individuals. Journal of Personality and Social Psychology,
writers. American Journal of Psychiatry, 151, 1650 –1656.
85, 499 –506.
Ludwig, A. M. (1995). The price of greatness: Resolving the creativity and
Carstensen, L. L., Fung, H., & Charles, S. (2003). Socioemotional selec-
madness controversy. New York: Guilford Press.
tivity theory and the regulation of emotion in the second half of life.
Martin, L. L., & Tesser, A. (1996). Some ruminative thoughts. In R. Wyer
Motivation and Emotion, 27, 103–123.
Jr. (Ed.), Ruminative thoughts: Vol. 11. Advances in social cognition
Davey, G. C. L., Startup, H. M., Zara, A., MacDonald, C. B., & Field, A. P.
(pp. 1– 47). Hillsdale, NJ: Erlbaum.
(2003). The perseveration of checking thoughts and mood-as-input hy-
Mor, N., & Winquist, J. (2002). Self-focused attention and negative affect:
pothesis. Journal of Behavior Therapy and Experimental Psychiatry, 34,
A meta-analysis. Psychological Bulletin, 128, 638 – 662.
141–160.
Nolen-Hoeksema, S. (1991). Responses to depression and their effects on
Davis, R. N., & Nolen-Hoeksema, S. (2000). Cognitive inflexibility among
the duration of depressive episodes. Journal of Abnormal Psychology,
ruminators and nonruminators. Cognitive Therapy and Research, 24,
100, 569 –582.
699 –711.
Nolen-Hoeksema, S., & Morrow, J. (1991). A prospective study of depres-
Ericcson, K. A., Krampe, R. T., & Tesch-Ro¨mer, C. (1993). The role of
sion and posttraumatic stress symptoms after a natural disaster: The
deliberate practice in the acquisition of expert performance. Psycholog-
1989 Loma Prieta earthquake. Journal of Personality and Social Psy-
ical Review, 100, 363– 406.
chology, 61, 115–121.
Eysenck, H. J. (1995). Creativity as a product of intelligence and person-
Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L. (1993). Response
ality. In D. Saklofske & M. Zeidner (Eds.), International handbook of
styles and the duration of episodes of depressed mood. Journal of
personality and intelligence: Perspectives on individual differences (pp.
Abnormal Psychology, 102, 20 –28.
231–247). New York: Plenum Press.
Nolen-Hoeksema, S., Parker, L. E., & Larson, J. (1994). Ruminative
Feist, G. J. (1998). A meta-analysis of personality in scientific and artistic
coping with depressed mood following loss. Journal of Personality and
creativity. Personality and Social Psychological Review, 2, 290 –309.
Social Psychology, 67, 92–104.
Feist, G. J. (1999). The influence of personality on artistic and scientific
Radloff, L. S. (1977). The CES-D Scale: A new self-report depression
creativity. In R. J. Sternberg (Ed.), Handbook of creativity (pp. 273–
scale for research in the general population. Applied Psychological
296). New York: Cambridge University Press.
Measurement, 1, 385– 401.
Fleming, E. S., & Hollinger, C. L. (1994). Project CHOICE: A longitudinal
Richards, R. L. (1981). Relationships between creativity and psychopa-
study of the career development of gifted and talented young women. In
thology: An evaluation and interpretation of the evidence. Genetic
R. Subotnik & K. D. Arnold (Eds.), Beyond Terman: Contemporary
Psychology Monographs, 103, 261–324.
longitudinal studies of giftedness and talent (pp. 316 –348). Norwood,
Shapiro, P. J., & Weisberg, R. W. (1999). Creativity and bipolar diathesis:
NJ: Ablex Publishing.
Common behavioural and cognitive components. Cognition & Emotion,
Goff, K., & Torrance, E. P. (2002). Abbreviated Torrance Test for Adults
13, 741–762.
manual. Bensenville, IL: Scholastic Testing Service.
Simonton, D. K. (1984). Genius, creativity, and leadership: Historiometric
Gotlib, I. H., & Cane, D. B. (1989). Self-report assessment of depression
inquiries. Cambridge, MA: Harvard University Press.
and anxiety. In P. C. Kendall & D. Watson (Eds.), Anxiety and depres-
Simonton, D. K. (1994). Greatness: Who makes history and why. New
sion: Distinctive and overlapping features (pp. 131–169). San Diego,
York: Guilford Press.
CA: Academic Press.
Simonton, D. K. (1997). Creative productivity: A predictive and explana-
Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension, and
aging: A review and a new view. In G. H. Bower (Ed.), The psychology
tory model of career trajectories and landmarks. Psychological Review,
of learning and motivation (Vol. 22, pp. 193–225). San Diego, CA:
104, 66 – 89.
Academic Press.
Slater, E., & Meyer, A. (1959). Contributions to a pathography of the
Hertel, P. T. (1997). On the contributions of deficient cognitive control to
musicians: 1. Robert Schumann. Confinia Psychatrica, 2, 65–94.
memory impairments in depression. Cognition & Emotion, 11, 569 –584.
Startup, H. M., & Davey, G. C. L. (2001). Mood-as-input and catastrophic
Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). Positive affect
worrying. Journal of Abnormal Psychology, 110, 83–96.
facilitates creative problem solving. Journal of Personality and Social
Sternberg, R. J., & O’Hara, L. A. (1999). Creativity and intelligence. In
Psychology, 52, 1122–1131.
R. J. Sternberg (Ed.), Handbook of creativity (pp. 251–272). New York:
Jamison, K. R. (1993). Touched with fire. New York: Free Press.
Cambridge University Press.
Jamison, K. R., Gerner, R. H., Hammen, C., & Padesky, C. (1980). Clouds
Treynor, W., Gonzalez, R., & Nolen-Hoeksema, S. (2003). Rumination
and silver linings: Positive experiences associated with primary affective
reconsidered: A psychometric analysis. Cognitive Therapy and Re-
disorders. American Journal of Psychiatry, 137, 198 –202.
search, 27, 247–259.
Joormann, J. (2004). Attentional bias in dysphoria: The role of inhibitory
Watkins, E., & Mason, A. (2002). Mood as input and rumination. Person-
processes. Cognition & Emotion, 18, 125–147.
ality and Individual Differences, 32, 577–587.
Joormann, J. (in press). Inhibition, rumination, and mood regulation in
depression. In R. W. Engle, G. Sedek, U. von Hecker, & D. N. McIntosh
(Eds.), Cognitive limitations in aging and psychopathology: Attention,
Received June 25, 2004
working memory, and executive functions. New York: Cambridge Uni-
Revision received November 3, 2004
versity Press.
Accepted November 15, 2004

---