- Top of page
The study longitudinally tracked the relationship among challenge/skill balance, flow, and performance anxiety in 27 student musicians over the course of a semester as they worked toward a recital of a piece of music. Using hierarchical linear modeling, the balance between the challenge of a passage of music and the perceived skills necessary to play that music was found to be significantly and consistently correlated with optimal experience. Results of moderated multiple regression indicated that skill level moderated the relationship between challenge, flow, and performance anxiety. Results also indicated that flow and performance anxiety were antithetical experiences, such that when flow was highest, performance anxiety was lowest and vice versa. These findings are discussed in terms of the application of flow theory to understanding performance, and the practical implications for reducing task-specific anxiety.
- Top of page
Means, standard deviations, and intercorrelations across individuals for all study variables are presented in Table 1. The distributional qualities of the measures of perceived challenge and perceived skill necessary to capably play the passage of music were assessed. The overall mean of the composite sum of the two perceived challenge items was 13.98 (SD= 3.99, median = 14.43) and that of the two perceived skills items was 10.81 (SD= 3.36, median = 11.22). Consequently, the challenge and skills inherent in performing the passages of music was generally perceived by all participants to be above average and to be negatively skewed in the moderate to high level.
Table 1. Means (M), Standard Deviations (SD), and Intercorrelations across Individuals for all Study Variables (N= 27)
| ||M||SD|| 1 || 2 || 3 || 4 |
|1. Flow||30.02||5.99|| || || || |
|2. Performance Anxiety||4.66||1.77||−.57**|| || || |
|3. Challenge/Skill Balance||6.22||2.79||−.62**||.52**|| || |
|4. Perceived Challenge||13.98||3.99||.59**||−.53**||−.85**|| |
|5. Perceived Skill||10.81||3.36||.03||−.02||−.27||.12|
Before proceeding with the tests of the hypotheses, we analyzed the systematic within- and between-individual variance in momentary flow and performance anxiety. The results of the null models that were used to test these statistical assumptions are presented in Table 2. The results indicated that there was substantial within- and between-individual variance for both momentary flow and anxiety (Flow: σ2= 30.95, τ00= 5.14; Anxiety: σ2= 4.47, τ00= 0.26). Chi-square tests indicated that the between-individual variance was significant for both flow and anxiety (Flow: τ00= 5.14, χ2(26) = 204.07, p < .01; Anxiety: τ00= 0.26, χ2(26) = 80.96, p < .01). The intraclass correlation (ICC=τ00 /(σ2+τ00)) for the flow measure was .14, indicating that between-individual variance accounts for 14 per cent of the total variance in flow. This would suggest that 86 per cent of the overall variance (both systematic and error) is attributable to within-individual variation. For anxiety, 5 per cent of the total variance was accounted for by between-individual variance, indicating that 95 per cent of the variance was due to within-individual variation in momentary performance anxiety. These results suggest that both flow and anxiety vary considerably from situation to situation.
Table 2. Parameter Estimates and Variance Components of the Null and Serial Dependent Models
| Model equations ||γ00||γ10|| σ 2 || τ 00 || τ 11 |
| Null models a |
|Perf. Anxietyij=β0j+rij||4.66||–||4.47||0.26**|| |
| Serial Dependent Models (Random-coefficient regression) b |
|L1: Flowij=β0j+β1j (Flowij.t-1) +rij|
|L1: Perf. Anxietyij=β0j+β1j (Perf. Anxietyij.t−1) +rij|
We also tested whether individuals' flow and anxiety ratings were randomly distributed or serially dependent (Ilies & Judge, 2002; see Table 2). Lagged flow was found to be a significant predictor of momentary flow (Flow: γ10= .37, p < .01), and anxiety was also serially dependent on its lagged effect (Anxiety: γ10= .32, p < .01). All subsequent models controlled for lagged effects.
In order to test if there was a relationship between challenge/skill balance and the experience of flow while performing a passage of music (Hypothesis 1), we ran a random coefficient regression model controlling for the lagged effects of flow. The results of this analysis are presented in Table 3. The regression coefficient for challenge/skill balance was significantly different from zero (β21=−0.60, t(1028) =−18.30, p < .01). The direction of the regression coefficient indicates that lower discrepancy or greater balance between perceived challenges and skills is associated with higher levels of flow. Challenge/skill balance accounted for 35 percent of the within-individual variance in flow, after controlling for the lagged effects of flow. In sum, results supported Hypothesis 1 by indicating that challenge/skill balance was significantly related to the experience of flow.
Table 3. Results for the Random-Coefficient Regression Models for Flow and Performance Anxiety Controlling for Lagged Effects
| Model equations || β 1j || β 2j ||γ00||γ10||γ20|| σ 2 |
| Random-coefficient regression models a |
|L1: Flowij=β0j+β1j (Flowij.t−1) +β2j (Challenge/Skill Balanceij) +rij|
|L1: Perf. Anxietyij=β0j+β1j (Perf. Anxietyij.t-1) +β2j (Challenge/Skill Balanceij) +rij|
A similar random regression model was used to assess the relationship between challenge/skill balance and performance anxiety (Hypothesis 2a). The results of the hierarchical linear model that was tested are presented in Table 3. T-tests indicated that the level-1 regression coefficient was significantly different from zero (β21= 0.26, t(1028) = 22.15, p < .01). This indicates that the larger the discrepancy between perceived challenges and the skills necessary to perform a piece of music the greater the experience of performance anxiety. Forty-seven per cent of the within-individual variance in performance anxiety was accounted for by challenge/skill balance, after controlling for lagged anxiety. This would appear to support Hypothesis 2a.
To test Hypothesis 2b and to better understand the relationship between challenge/skill balance and the experiences of flow and performance anxiety, we used moderator multiple regression analyses on the level-1 variables (Aiken & West, 1993). The results of the moderated random-coefficient regression analyses are presented in Table 4. The interaction effect between perceived challenge and perceived skill was found to be significant for flow (γ40= 1.35, t = 8.48, p < .001) and performance anxiety (γ40=−0.75, t = 12.77, p < .001) after controlling for autocorrelational effects. The effect sizes of the challenge × skill interaction effect were significant for both flow (R2= .07) and anxiety (R2= .15).
Table 4. Results for the Moderator Random-Coefficient Regression Analyses
| Model parameters a || γ 00 || γ 10 || γ 20 || γ 30 || γ 40 |
| Random-coefficient regression model a |
|L1: Flow =γ00+γ10 (Flowt−1) +γ20 (Challenge) +γ30 (Skill) +γ40 (Challenge × Skill) + r|
|L1: Perf. Anxiety =γ00+γ10 (Perf. Anxietyt−1) +γ20 (Challenge) +γ30 (Skill) +γ40 (Challenge × Skill) + r|
Using asymptotic variances and covariances, we plotted the simple slopes in order to understand the nature of the interaction effects (Bauer & Curran, 2005) (see Figures 2 and 3). The perceived challenge of the task was not related to flow among those musicians who perceived their skills to be low. For low-skilled performers, flow was uniformly high regardless of the challenge of the piece of music they were playing (Figure 2). However, as skill level increased, the relationship between challenge and skill was stronger and more positive. Among moderate to highly skilled musicians, flow levels were highest when challenge was high, and lowest when the piece of music being played was not challenging (Figure 2). These results would again appear to support Hypothesis 1. The converse was found for performance anxiety as the outcome variable. Low-skilled performers uniformly experienced low levels of anxiety and this was independent of the degree of challenge in the music being played (Figure 3). On the other hand, among moderate to highly skilled musicians, the amount of anxiety experienced increased as the challenge of the task decreased (Figure 3). Anxiety was highest when playing not particularly challenging music, and lowest in the high challenge situation, that is when flow was at its highest level. Consequently, analysis of the data using moderator multiple regression rather than difference scores yields results that are not supportive of Hypothesis 2a. However, our data do suggest that flow and anxiety are incompatible states, in that among moderately to highly skilled musicians anxiety was lowest when flow was highest and vice versa. Therefore Hypothesis 2b was supported.
- Top of page
The current research set out to test the relationship between an important precondition to flow, the balance between action opportunities and action capabilities, and the subjective experience of flow. Recently this relationship has been studied in participants engaging in relatively simple videogame tasks in laboratory settings (Engeser & Rheinberg, 2008; Keller & Bless, 2008; Keller & Blomann, 2008). We wanted to ascertain whether the relationship between challenge/skill balance and optimal experience generalised to a more complex, meaningful, and dynamic task that consisted of moderate to high levels of both perceived challenge and skill.
Using the absolute difference between skill and challenge measures, we found that both flow and performance anxiety were associated with the balance between the perceived skills necessary to perform a task and the perceived challenges inherent in the task. Flow was more likely to occur for tasks where challenges and skills were balanced, whereas performance anxiety was associated with tasks where the challenge did not match the skills of the performer. These results confirm Csikszentmihalyi's theory that the balance between challenge and perceived skills is an essential precursor to flow, even in the performance of a complex task such as playing a passage of music.
However, it must be noted that there has been some controversy about the use of difference scores (Edwards, 2001; Peter, Churchill, & Brown, 1993). One of the main problems is that difference scores are typically less reliable than their component variables (Peter et al., 1993). We felt justified in using difference scores for two reasons. First, the reliability coefficients for challenge (α= .93) and skill (α= .95) were satisfactory and the attenuation in reliability for the difference term was minimal. Second, the correlation between the components was small (r=−.001). A second issue with difference scores is the possible restriction in variance in the difference variable (Peter et al., 1993). This did not seem to be a problem in the current study. Further, such a restriction in variance would have truncated the correlations between challenge/skill balance and both flow and performance anxiety, resulting in a more conservative test of the hypotheses. However, using absolute difference scores confounds the experiences of boredom and anxiety proposed by Csikszentmihalyi's channel theory (1975). Consequently the procedure we used may have caused artificial dichotomies in our data and a loss of fidelity. To avoid these kinds of problems we further analyzed our data using interactive regression analysis (Edwards, 2001) to determine whether the relationship between flow and challenge was moderated by the perceived skill level of the person performing the task.
The results of moderated multiple regression provided a more detailed analysis of the relationship between challenge and flow that is both consistent and contrary to flow theory. The literature on optimal experience suggests that challenges and skills have to be at relatively high levels to experience flow (Massimini & Carli, 1988). Our results found that there was no relationship between the challenge of the task and flow for low-skilled performers, whereas there was a significant positive relationship when skill levels were perceived to be moderate to high. Among moderate to highly skilled performers, flow levels were highest when the piece of music being played was challenging. A similar finding was found with momentary performance anxiety as the outcome variable. Specifically, when skill levels were perceived to be low, there was no relationship between the perceived challenge of the music and task-specific anxiety. In this circumstance anxiety was uniformly low. However, for performers who perceived their musical skills to be moderate to high, the perceived challenge of the task was significantly and negatively related to performance anxiety. Skilled performers experienced the highest level of task-specific anxiety when playing relatively easy pieces of music. This contradicts the eight-channel theory of challenge/skill balance (Csikszentmihalyi, 1988, 1990; Massimini & Carli, 1988) that stipulates that when the challenge of a task is low and the skills required to perform the task are moderate to high, then the resulting state should be one of either boredom or relaxation (Massimini & Carli, 1988) rather than anxiety. Anxiety is proposed to be experienced when the challenge of the task is perceived to be high and skills low (Csikszentmihalyi, 1975). We also found partial support for the notion that flow and performance anxiety are mutually exclusive experiences. When flow was highest, anxiety was lowest, and vice versa. However, this relationship was moderated by the perceived skill level of the person performing the task. Low-skilled performers experienced relatively high levels of flow and low levels of anxiety regardless of the difficulty of the task, unlike the moderate- to high-skilled performers.
These findings are consistent with recent theory and research on the physiological and cognitive anatomy of flow. For example, Dietrich (2006; Dietrich & Stoll, 2010) has proposed that flow is a state of “transient hypofrontality” in that there is a temporary suspension of some higher cognitive functions thereby disabling their interference with the more implicit and automatic cognitive processes that are characteristic of flow. This would explain why participants experienced more anxiety when performing less challenging tasks. It is in such tasks that analytic cognitive processes have more opportunity to interfere with the focused attention of flow and facilitate a more disintegrated or distracted attention that is characteristic of anxiety. This explanation is also consistent with distraction theory, which proposes that when an individual's attention is diverted from the execution of the task by task-irrelevant thoughts, performance is impaired and there is a greater likelihood of anxiety (Beilock & Carr, 2001; Lewis & Linder, 1997).
The finding that flow and task-specific anxiety are incompatible states is corroborated by recent research on the physiology of flow (see Ullén, de Manzano, Theorell, & Harmat, 2010). In a study of professional pianists, flow was shown to correlate with decreased heart period, increased cardiac output, increased respiratory rate and depth, an activation of facial muscles associated with the expression of positive emotions, and a deactivation of muscles implied in negative emotional expression (de Manzano, Theorell, Harmat, & Ullén, 2010). Furthermore, there is evidence that suggests that flow is associated with the activation of parasympathetic systems that have been shown to counteract the catabolic sympathetic activity that is associated with stress (Grape, Sandgren, Hansson, Ericson, & Theorell, 2003).
Some research has identified situational characteristics associated with flow, such as the amount of autonomy and skill variety inherent in the task (Demerouti, 2006; Fullagar & Kelloway, 2009). However no research has investigated the task-specific characteristics that may inhibit flow and induce its opposite state, antiflow. Our research indicated that task-specific anxiety and flow are incompatible states. Specifically, when performance anxiety was highest, flow was lowest, and vice versa. We would argue that flow and anxiety are not antipodal states (in that they are not the opposite ends of the same continuum), but that they are antithetical (in that they are negatively related). Antipodal constructs suggest that the absence of one construct indicates the presence of the other construct, and that both can be measured using the same scale. Our findings indicate that flow and performance anxiety can exist simultaneously, but that the presence of one minimises the magnitude of the other. Also, we would not suggest that flow and performance anxiety be measured using the same scale. As mentioned above, an antithetical state to flow has been briefly referred to in the literature and termed “antiflow” (Allison & Duncan, 1988). Antiflow has been described as a demotivational state characterised by tedium and a lack of autonomy and control (Sorrentino et al., 2001). However, very little empirical research has addressed the nature of antiflow. Our research would suggest that task-specific anxiety may be an important component of the antiflow state that suppresses optimal experience.
Our multilevel analyses of the variance components of flow and performance anxiety indicate that most of the variation in both constructs is due to situational characteristics, even taking into account that some of this variance could be attributable to error. This suggests that both constructs demonstrate state-like rather than trait-like properties, and further supports evidence that flow is a situational or momentary state, rather than a trait (Fullagar & Kelloway, 2009). This has important theoretical implications. The clarification of key concepts and distinguishing between traits and states is important in the construction of theory in psychology (Mischel, 1969). Luthans (2002) has emphasised that a crucial criterion for the inclusion of constructs in positive organisational scholarship should be that they are manageable and capable of being effectively changed to improve performance. Our research adds further evidence that both flow and performance anxiety are responsive to changes in situational contingencies, specifically the degree of balance between the challenge of the task and the skills necessary to meet that challenge.
We do not wish to suggest that there are no trait aspects to either task-specific anxiety or flow. One model that has been used to understand music performance anxiety is Barlow's theory of emotion (2000). This model conceives performance anxiety as consisting of the interplay between inherited biological predispositions and learned environmental contingencies. Our methodology focused predominantly on assessing momentary anxiety and flow, and even though our findings support that of Fullagar and Kelloway (2009) in documenting state-like aspects of flow, there is evidence of trait components to optimal experience. Several studies have indicated that self-motivated individuals who have a high need for achievement exhibit a greater predisposition to experience flow, and are more likely to seek out situations of challenge/skill balance than those who are lower on these characteristics (Asakawa, 2004; Eisenberger et al., 2005). Our research did show that the relationships between flow and challenge, and between flow and task-specific anxiety, were moderated by perceived level of skill. It is conceivable that skill levels are at least partially determined by such dispositional characteristics as intrinsic motivation and need for achievement.
Our findings have several important practical implications. First, the finding that both flow and performance anxiety are predominantly state constructs suggests that there are situational factors that can be managed to facilitate the experience of flow and to reduce performance anxiety. The treatments for performance anxiety fall into two broad categories (Kenny, 2005). First, there are various combinations of cognitive and behavioral approaches (e.g. behavioral rehearsal, systematic desensitisation, stress inoculation, relaxation, meditation, and biofeedback). Second, there are psychopharmacological or drug interventions that reduce the effects of adrenaline (e.g. the use of beta-blockers, selective serotonin reuptake inhibitors, and anti-depressants). However, there are few well-conducted studies that have assessed the effectiveness of these treatments. Kenny (2005), in a review of treatments for music performance anxiety, concludes that “the literature on treatment approaches for MPA is fragmented, inconsistent, and methodologically weak. These limitations make it difficult to reach any firm conclusions about the effectiveness of the various treatment approaches reviewed” (p. 206). Our research suggests that generating a flow state would appear to be one effective way to reduce performance anxiety.
Among moderately to highly skilled performers, we found that the highest levels of task-specific anxiety occurred when the task was perceived to be relatively easy. Our results further indicated that if the performer was anxious, it was much less likely for him/her to experience flow. Conversely, when performers experienced high levels of flow, their anxiety was the lowest. Apparently, being “in flow” enables performers to focus on the task, enjoy their performance, and feel less anxious. This suggests that one way to prevent performance anxiety is to encourage flow. For more skilled individuals, it is even more important that the perceived challenge of the task match their perceived skill level in order to facilitate flow and reduce performance anxiety.
Several factors have been identified as facilitators of flow. Csikszentmihalyi et al. (2005) specifically outline three conditions that are important for encouraging optimal experience or flow. These include having clear goals, the provision of specific feedback concerning task performance, and having the skills necessary to perform the task. We see these three conditions as being dynamically related. Our research suggests that optimal levels of flow are achieved through two interrelated processes. First, individuals should, when possible, choose moderately challenging tasks that are commensurate with their ability. Second, a practice regime should be established that has clear goals and specific feedback, enabling the individual to develop the skills necessary to perform the task. In addition to these factors, Bakker (2005), drawing upon emotional contagion theory, found that flow spilled over from music teachers to their students. Consequently, another way of increasing flow and decreasing performance anxiety is for teachers to model the state both in their own performances and in their instruction.
Our findings have broader implications for the world of work in general. We have argued that both flow and work-related anxiety are predominantly determined by situational factors. Specifically, our results would suggest that organisations may be able to increase the amount of flow experienced at work, and simultaneously decrease work-related anxiety, by increasing the level of challenge in the task so that it is commensurate with the skill level of the worker. However, this effect may only be significant for workers with moderate to high skill levels. Such manipulations of the challenge of the task may not be effective for increasing flow and decreasing work-related anxiety in workers with low skill levels.
Our findings may also help us better understand how subjective worker experiences are related to objective performance outcomes. The factors described by Csikszentmihalyi et al. (2005) as facilitators of flow are remarkably similar to the prescriptions of goal setting theory. Clear goals that are matched to ability levels and accompanied by task feedback have been found to be related to increased performance in a wide range of settings (Locke & Latham, 2002). It is possible that following the prescriptions of goal setting encourages the development of flow, and thereby contributes to reduced stress and anxiety, as well as improved performance. If so, the experience of flow might help explain the positive effects associated with goal setting, as well as suggesting positive emotional outcomes that have not traditionally been studied in goal setting research.
Finally, our results have broader practical implications for individuals who wish to achieve optimal experience. Faced with a discrepancy between challenges and skills, the individual can respond in two ways (Csikszentmihalyi, 1990). First, he or she can increase the perception of challenge in the task. Csikszentmihalyi (1990) points out that this is relatively easy once the individual is aware that challenges exist. The other response is to increase one's skills through practice or training. Which of these strategies is used is dependent on the situation. Increasing perception of the task challenge is more likely when the individual believes that they are over-skilled, while increasing skills through practice is more relevant when the individual believes that they are under-skilled to perform the task. The goal is to achieve a balance between perceived skill and perceived challenge. It is at this point that the individual will become absorbed in the task and experience flow. However, simply balancing challenges and skills is not sufficient to produce optimal experience. Challenges have to be at a level that is the above average for the individual so that there is the opportunity for skill development, as demonstrated by the low levels of flow, and high levels of anxiety, in our data when high skill was matched with low challenge. Task-specific anxiety can be effectively reduced by ensuring a balance between perceived, moderate, task challenge and the skill level required to perform the task, thereby inducing flow.
There are several limitations to the current study that suggest directions for future research. First, we studied a very select sample of music students. It would be interesting to see if the antithetical relationship between flow and performance anxiety generalised to other disciplines, including and beyond the performing arts and athletic activities. Even though MPA appears to be independent of ability and experience (Kenny & Osborne, 2006), future research should investigate whether the mutually exclusive relationship between flow and performance anxiety maintains itself with experience. Second, one of the limitations of the Experience Sampling Method is that it necessitates the use of short scales that are minimally disruptive and require minimal effort. Our research used short composite scales to assess flow, perceived challenge, skill level, and performance anxiety. Such scales have yet to have their psychometric properties confirmed. With flow in particular, there is some debate as to whether it should be assessed as a multi-dimensional or unidimensional construct (Marsh & Jackson, 1999). Although our research utilised a specific composite measure, there are other operationalisations of flow (e.g. Bakker's work-related flow inventory (2008)). Future research would do well to use more extensive, multi-faceted operationalisations, as well as alternative definitions of flow to establish the general validity of the effects found in the current research. Finally, it could be argued that there was not much variation in the musical tasks that were studied, and that any within-subject variance could be due to variations in dispositional anxiety rather than fluctuations in momentary performance characteristics. However, our participants played several instruments and performed a wide variety of musical compositions that were each divided into several musical passages of varying difficulty. We are therefore confident that the variances in challenge/skill balance, momentary flow, and performance anxiety were sufficiently large.
There is growing evidence that the experience of flow is significantly and positively related to performance (Demerouti, 2006; Eisenberger et al., 2005). Flow has also been associated with positive mood (Fullagar & Kelloway, 2009). Fredrickson (1998, 2001) has made compelling theoretical arguments based on empirical evidence that positive emotions broaden individual resources and thought–action repertories such that individuals experiencing such emotions function at an optimal level. This is corroborated by research that has correlated psychological well-being with employee performance and turnover (Wright, Cropanzano, & Bonett, 2007). Our research did not assess students' performance in their final recital. It would be interesting to ascertain whether the relationship between flow and performance is robust, and transfers to a wider range of tasks, including musical performance.
In conclusion, this study aimed to understand the relationship between an important precondition of flow, namely challenge/skill balance, and the experience of flow while performing an applied, complex, task. Although this relationship has been established for simple tasks, such as playing videogames, we found that it translated into a more complex and dynamic musical task where participants were motivated to perform well. Specifically, we found that the experience of flow can be promoted by ensuring a balance between the challenges inherent in a task and the perceived skills necessary to perform that task, particularly among individuals who already perceive their task-related skills to be moderate to high. Our study suggests that understanding the processes that generate the flow state provides both a theoretical and practical framework for reducing performance anxiety.