EFFECTS OF MOOD STATES ON CREATIVITY
AbstractExperimental studies of the effects of emotion on creativity have yielded conflicting results. We present a study in which mood was induced through recall of life events and performance on creativity-related tasks was measured. The results show that the effects of emotions on creativity depend on numerous conditions. In general, a positive mood promoted creative performance. However, the nature of the creative task, the kind of performance score, and the nature of the induced emotion modulated this trend. A post-experimental questionnaire suggested, moreover, that people differ on their conceptions of how emotion may influence task performance.
*Universite Rene Descartes - Paris 5 Laboratoire Cognition et Developpement C.N.R.S., D.M.R. 8605, France Correspondence to : Franck Zenasni Laboratoire Cognition et Developpement C.N.R.S., U.M.R. 8605 71 Avenue Edouard Vaillant 92100 Boulogne-Billancourt E-mail:
[email protected]/
[email protected]
Inspired by the importance of emotions for eminent creators, their mood instability and emotional disorders related to creative work, research conducted during the last fifteen years has examined the effect of induced emotional state on creativity in "average" people. The results diverge. Some studies indicate that a positive mood state improves creativity and a negative mood has no influence. Others show that a positive mood state such as happiness inhibits creativity and a negative mood state facilitates it. Most studies focus on mood valence (the positive--negative nature of mood states) but some indicate that mood intensity has an effect on creativity. Various interpretations have been proposed for the diverse results that are observed.
In order to study the relation between a mood state and creativity, the experimental approach has dominated. In this work, mood is induced; analysis of variance is employed to test mean differences between an experimental condition (induction of a positive or a negative mood state) and a control condition (no induction or induction of a neutral mood state), with creative thinking task performance used as the dependant variable. Isen and colleagues, in a series of studies, examined systematically the effect of mood (or emotion) on creativity (Estrada, Isen, & Young, 1994;Isen, Daubman & Nowicki, 1987;Isen & Williams, 1988). They observed that a positive mood induction consistently fostered creativity. Participants in a positive mood state, compared to those in a control group, increased significantly their performance on the Candle Task (Duncker, 1945) and the Remote Associate Test (RAT,Mednick, 1962). In the Candle Task, participants are given a box of matches, a candle, and a box of thumbtacks. The goal of the task is to attach the candle on the wall in such a way that the wax does not drip on the wall or on the floor when the candle burns. In the RAT, the task consists of finding one word that is associated with three apparently unrelated words. According to Isen, positive mood facilitates access to positive elements present in memory (Isen, 1985; Isen, Shalker, Clark & Karp, 1978;Teasdale & Forgaty, 1979). As positive information in memory is hypothesized to be more richly interconnected than negative information for most people, positive emotional states should promote spreading activation and cue a larger and a more diverse set of cognitive material constituting a "complex cognitive context". This, in turn, may facilitate perception of diverse characteristics and qualities of the objects included in the task and enhance the possibility of combining different elements, and perceiving
new relations (Carnevale & Isen, 1986;Isen & Daubman 1984).To support this interpretation, physiological data are employed to argue that positive effects of mood state on creativity are due to the release of dopamine. Dopamine is believed to facilitate flexible deployment of attention and multiple cognitive perspectives, consequently enhancing cognitive flexibility and creativity (Ashby, Isen & Turken, 1999). Kaufmann (1995), and Kaufmann and Vosburg (1997), criticized, in part, Isen's experiments and disagreed with her interpretations. They used insight tasks similar to the ones used by Isen and found that a positive mood led to decreased creativity,whereas a negative mood led to better performance. Based on differences in the creativity tasks used (Isen's procedure allowed participants to adjust their answers due to feedback during task performance whereas Kaufmann and Vosburg's procedure consisted of a paper-pencil task with no feedback), they suggested two alternative interpretations. First, according to cognitive tuning theory (Schwarz, 1990)a person may have lower criteria of satisficing in a positive mood state than in a neutral or a negative mood state. A positive mood state signals that a person is in a satisfactory position. Consequently the participant may not be motivated to exert extra cognitive effort. Thus, in a creative task, participants may produce fewer ideas because they are more rapidly satisfied by their initial ideas. In contrast, a negative mood state indicates implicitly that there is a problematic situation and that some extra effort is needed in order to return to a neutral situation. Participants may be more optimizing (seeking the best solution) in such situations and consequently more efficient in creative tasks. Second, they noticed that a positive mood increased participants' sensitivity to different cognitive biases (Isen, Means, Patrick & Nowicki, 1982).In both creativity and insight tasks, participants activated responses that did not correspond to the problem. Finally, in 1997,Kaufmann proposed a model which integrates his data with Isen's results (see Kaufmann, 1997;Vosburg & Kaufmann, 1998).According to this model, optimizing versus satisficing criteria for solutions must be distinguished. Optimizing refers to an ideal method, examining all the alternative solutions of a problem and persevering in this work until finding the most effective solution. However, due to the limited capacity of the cognitive system for information processing, individuals often use a strategy of satisficing. Thus, the first solutions considered satisfactory will be accepted. Kaufmann places these two types of orientations at the extremes of a continuum; there are different degrees of satisficing and of optimizing behavior. Participants in a positive mood state will perceive a task as a problem requiring a lax criterion for solution, whereas participants in a negative mood will select answers according to a stricter criterion.
These last interpretations are coherent with Abele's (1992)work. Observing a facilitative effect of negative mood on creativity, she defined emotion as a motivational moderator. An individual in a negative mood state is motivated to find ways that will help him or her to return to a neutral mood state. Thus, flexibility of ideas and fluency are strategies of "mood repair". The induction of positive mood can also favor creativity with emotion acting as a cognitive moderator: individuals in a positive mood are "looser." Consequently they are more intuitive and more flexible. Fluency is greater when the idea produced is congruent with the mood state of the participants. According to Hirt (1999),we may consider that Martin, Ward, Achee and Wyer (1993) have a similar view of the effects of emotion on production. They observed that mood effects for an association task (generate a list of birds) are relative to context. Experiencing a same positive emotion, the quantity of ideas produced is different if participants stopped when they no longer enjoyed the task then when they stopped based on their achievement (mood as input paradigm). Happy participants generated a greater number of responses with an enjoyment-based stop rule whereas sad participants produced more responses with a perfomance-based stop rule. According to the mood as information theory (Schwarz & Clore, 1983), Martin et al. explained that positive mood informs individuals that they feel good about their performance, so they think that the productions are already sufficient and do not continue to produce ideas. Negative moods inform participants that their performance is not sufficient and leads them to greater effort to produce more ideas. Hirt, Levine, Mc Donald, Melton and Martin (1997) confirmed these results and found that individuals in a positive mood were more interested and had more fun with the task compared to individuals in a negative mood. However, they found also that this effect depends on context, and it is observed only for the quantitative aspect of idea production and not for the qualitative characteristics (originality) of the ideas produced. The effects of mood on quantitative measures of performance (such as the quantity of ideas) are distinct from those that underlie qualitative indices of performance (originality).Thus, mood may be linked to creativity in various ways. Some authors have interpreted differently the effects of mood on creativity. Adaman and Blaney (1996)observed that the induction of positive or negative mood shows better results than the induction of a neutral mood for the Torrance Test of Creative Thinking. They emphasized the intensity of the emotion, suggesting that people seek to decrease intense mood states through creative activities. Russ (1993, 1998),in a synthesis of research on creativity and emotion, proposed a theoretical model indicating that per-
sonality-related individual differences could effect emotion and hence creativity, for example openness to internal emotions influences tolerance for ambiguity, which in turn influences creativity.
The literature indicates that there is no consensus on a single effect of emotion on creativity. It is possible that the experimental results and the related interpretations diverge because the nature of the induced emotional state is not identical from one study to another. In the majority of studies, authors induced "negative mood state" and "positive mood state" but they rarely examined the specific quality of the emotion. A negative mood state may correspond, for example, to sadness, disgust, or anger. Isen et al. (1987) through her positive-mood induction with a comic movie may have induced more specifically humor than happiness. Some studies point out that a specific emotion can have a specific effect on cognitive processes. For example, in a study concerning the influence of mood state on social judgement, Bodenhausen et al. (1994a,1994b)found that angry people produced patterns of results comparable with those of happy participants but differed significantly from the pattern shown by sad people. Bodenhausen concluded that all negative affects do not create the same kind of response and introduced the notion of mood-state-specific effects. In a similar way, we may expect different effects of emotion on creativity as a function of their nature, rather than as a function of their valence. Finally, the results may vary because the creativity tasks differed from one experiment to another (divergent thinking task, Remote Associate test, Insight Task). Given that creativity is considered partially domain and task specific, it would not be surprising if the effect of emotional state differed from one study to another.
These differences of variables, procedures and creativity tasks do not enable us to compare closely the existing experiments. However, certain commonalities among the studies allow us to make some general remarks: First, it seems that emotion has an effect on ideational fluency and flexibility as opposed to originality. In this regard, Weisberg (1995)studying the link between manic-depression and creativity observed that a positive
mood state in the composer Robert Schumann was linked to an increase of the quantity of his production, but not the quality. Vosburg (1998)observed a significant link between the valence of the emotional state and the number of ideas produced by the participants in a divergent thinking task, but not between the valence and originality of ideas. Second, there seems to be two categories of effects of mood states on creativity. One of these corresponds to a motivational aspect. Authors speak about mood repair, research of optimal or appropriate responses: individuals are motivated to be creative. The other category corresponds to perturbing or disturbing effects of mood state on creativity. Authors speak, for example, about use of automatic processing and satisficing which decrease the quality or the quantity of ideas produced in a creativity task. Third, it appears that the authors have not taken into account individual differences in mood management, which may explain, in part, the divergence of results. Some traits may modulate the influence of an emotional state on creativity. The main goal of the current study was to examine several variables such as the nature of the emotion, intensity of the emotion, the nature of the creativity task, and interindividual differences that may be relevant to the relationship between emotion and creativity. The experiment was designed to determine the conditions that influence the link between emotion and creativity. We induced mood state and then measured the specific nature of participants' affects. These measures were linked to multiple indices of performance in two creative thinking tasks (verbal and figurative).
Materials Induction of mood. To induce mood, we used the life-event recall technique (Brewer,Doughtie & Lubin, 1980).Participants were asked to recall an event of their life that was either happy, sad or neutral, depending on the experimental condition, and to describe this event in writing. This procedure took
approximately 10 minutes. This technique has been used in previous research studying links between emotion and creativity (Abele, 1992) as well as other studies (Bless, Bohner, Schwarz & Strack, 1990; Brewer, Doughtie & Lubin, 1980) and is considered effective. Some research suggests that the induced mood state tends to last approximately 15 minutes (Abele, 1992;Morris, 1989). Evaluation of mood. Participants evaluated their mood state using four different scales. The first scale measured valence of the emotional state with a 9-point continuum between a positive mood (high score) and a negative one (low score) (Bradley & Lang, 1994).The center of the scale corresponded to a neutral mood. A second 9-point scale measured the intensity of the emotional state (1=low,9=high, Bradley & Lang, 1994).Also we used two ninepoint unipolar scales to measure happiness and sadness respectively. These scales allowed us to consider happiness and sadness as independent dimensions, and they evaluated more specific emotional states than those evaluated by the valence scale. Creativity tasks. Two tasks from the Torrance Tests Creative Thinking (Torrance, 1976)were used. In the "box" task, participants had to generate during 10 minutes as many unusual ideas as possible for using a cardboard box. In the "parallellines" task, participants had to make as many drawings as possible during 10 minutes using sets of two parallel lines. Three scores were calculated for each task: (a) fluency; the number of ideas produced (b) flexibility; the number of different categories of ideas (c) mean originality, in which total originality is divided by fluency (see Mouchiroud & Lubart, 2001).For a sample-specific originality score based on the relative rarity of ideas, two points are given for each idea generated by less than 2% of participants, one point for each idea generated by 2% to 5% of participants, and zero points for each idea generated by more of 5% of participants. Post-experimental questionnaire. A post-experimental questionnaire measured if participants considered that the mood induction was effective and how mood was perceived to influence creativity. We asked participants whether they believed their emotional state had an effect on their task performance and if so to describe the nature of this influence.
Participants Participants were 120 psychology students (mean age = 22.67, SD = 3.21, range 20-39,49 males, 71 females). They were distributed in six groups of 20 participants, with each group corresponding to an emotional state (positive,
negative, neutral) and a creativity task (Box, Parallel lines). Participants received course-related credit for their participation. Procedure Participants were told that the study concerned thinking and problem solving. Participants were randomly assigned to the mood induction conditions. First, each participant recalled either a sad, happy or neutral event for the emotional induction. Then, mood state was evaluated using the four scales described earlier. Next, the creativity task was completed. Finally, participants evaluated again their emotional state and answered the post-experimental questionnaire. Each participant completed only one creativity task after mood induction due to the limited duration of induced mood states and the possible influence of completing a creativity task on subsequent mood. In order to not interfere with the mood induction procedure, we did not evaluate mood before the induction. We were interested primarily in the post-induction emotional state of individuals before they began the creativity task.
Effect of mood induction: analysis for all participants together regardless of their creativity task Mean differences between experimental groups. Analyses of variance indicated that the mood induction was generally effective. We observed significant differences on valence (F(2, 115) = 3.24, P < .05), sadness (f(2, 115) = 4.89 p< .05) and happiness (F(2, 115) = 5.28, P < .05). We did not observe differences for arousal (F(2, 115) = 1.83, n.s.). No gender differences were found. Table 1 Means (and standard deviations) for each mood scalefor the three experimental conditions: Induction of happiness, induction of sadness, induction of a neutral mood Happy condition Neutral condition Sad condition Valence scale Arousal scale Happiness scale Sadness scale
6.92 (1.62) 5.57 (2.17) 6.10 (1.60)
2.55 (1.69)
6.55 (1.58) 5.02 (2.10)
6.15 (1.25) 2.22 (1.76)
5.89 5.86 5.07 3.55
(2.15) (1.57) (1.99) (2.34)
Correlations between scales. Correlations between mood evaluation scales (valence, happiness, and sadness) corresponded to our predictions. The valence scale correlated positively with the happiness scale (r = .89, P C!l
0 -' 0 J:
U
>-
(J) Q.
f-
Z
w
II: II:
::::l
U
Analysis of the post-experimental questionnaire. The questionnaire offered supplementary information about the influence of induced emotion on creativity. In response to the question of whether mood state influenced creativity, participants replied in several different ways: (a) positive mood state had a favorable effect (6 individuals); (b) emotional state had no effect because they put it aside before undertaking the creativity task (12 individuals); (c)negative mood state had a disturbing effect (8 individuals); and finally (d) mood had no influence because the participant's level of creativity was too low in any case (8 individuals). Other participants were in a neutral state or did not specify an influence of mood state. We observed significant differences on fluency (F(3/30) = 3.20/ p
