Psychology of Music http://pom.sagepub.com

Performance degradation under pressure in music: an examination of attentional processes Catherine Y. Wan and Gail F. Huon Psychology of Music 2005; 33; 155 DOI: 10.1177/0305735605050649 The online version of this article can be found at: http://pom.sagepub.com/cgi/content/abstract/33/2/155

Published by: http://www.sagepublications.com

On behalf of:

Society for Education, Music and Psychology Research

Additional services and information for Psychology of Music can be found at: Email Alerts: http://pom.sagepub.com/cgi/alerts Subscriptions: http://pom.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.co.uk/journalsPermissions.nav Citations http://pom.sagepub.com/cgi/content/refs/33/2/155

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

155

A RT I C L E

Performance degradation under pressure in music: an examination of attentional processes

Psychology of Music Psychology of Music Copyright ©  Society for Education, Music and Psychology Research vol (): ‒ [- () :; ‒] .⁄ www.sagepublications.com

C A T H E R I N E Y. WA N U N I V E R S I T Y O F M E L B O U R N E , AU S T R A L I A

G A I L F. H U O N U N I V E R S I T Y O F N E W S O U T H WA L E S , AU S T R A L I A

A B S T R A C T An experiment was carried out to investigate the cognitive mechanisms responsible for performance degradation under pressure in music. The experiment was designed to compare the predictions of two theories, distraction and explicit monitoring. Distraction theory (Eysenck, 1992) explains performance degradation as a result of attentional shifts to task-irrelevant information. Explicit monitoring theory (Baumeister, 1984; Masters, 1992) postulates that performance degradation is due to increased attention to step-bystep control of skill processes. A total of 72 novice musicians were given individual lessons on basic note and rhythm reading skills. They were then trained on a keyboard task under one of three conditions (single-task, dual-task, video-monitoring) before being exposed to either a high-pressure or low-pressure post-test. Results showed that pressure led to skill failure in the single-task and dual-task groups, but resulted in improved performance of the video-monitoring group. These results were consistent with explicit monitoring theory. Furthermore, video-monitoring training was found to ameliorate the performance deficits normally caused by high pressure. Finally, the present study found no evidence that trait anxiety influences the effect of pressure on performance. K E Y W O R D S : anxiety, attention, novices, performance, skill failure

Many musicians have experienced unexpected performance degradation, or ‘slips’, when performing under pressure. A memory lapse or an incorrect note can impair the overall quality of a musical performance. In the sporting context, this phenomenon is commonly known as ‘choking under pressure’. Pressure has been defined as the presence of situational incentives to perform well (Hardy et al., 1996). Forms of pressure include the contingency of rewards or punishments on level of performance, the presence of an evaluative audience or competition, and the presence of ego-relevant threat

sempre :

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

156

Psychology of Music 33(2)

(Baumeister and Showers, 1986). Choking has been defined as the occurrence of suboptimal performance under pressure conditions (Baumeister, 1984). To date, no research has directly examined the underlying causes of performance degradation among musicians. Two theoretical frameworks, the distraction and explicit-monitoring paradigms, have guided researchers in other domains. Both theories explain unexpected performance degradation in terms of interference with the performer’s attentional processes. According to distraction theory (Wine, 1971; Eysenck, 1979, 1992), performance degradation is a result of attentional shifts to task-irrelevant information. In music performance, examples of task-irrelevant information include: fear of forgetting the notes when playing from memory, fear of not being able to play a difficult passage, or fear of public failure and subsequent shame. Task-irrelevant information is said to reduce the amount of working memory available for task performance. Thus, breakdowns under pressure are most likely in tasks that rely on continuous attentional control during execution, and are particularly vulnerable to corruption as a result of attentional interference. Support for distraction theory has been found in studies in which highly anxious subjects tend to perform worse than low anxious subjects on tasks that rely on working memory (e.g. Eysenck, 1985; Calvo et al., 1992). These researchers concluded that anxious individuals become preoccupied with task-irrelevant thoughts and worry more about their performance, and consequently have smaller working memory capacity available for task execution. According to distraction theory, anxiety results in performance degradation only in tasks requiring working memory, but not when tasks are well learned and relatively automatic. Darke (1988) tested this idea by explaining the performance of high and low anxious individuals on a reasoning task that contained two components. One part involved the verification of automatic judgments, while the other part required effortful processing. Automatic judgment performance was uncorrelated with trait anxiety measures. In contrast, trait anxiety was predictive of performance on an effortful processing task, with high-anxious subjects performing less well than low-anxious subjects. Given that working memory is required for the first task, but not the second, the findings suggest that anxiety affects performance by reducing working memory capacity. Distraction is certainly implied in self-reports of performance anxiety among musicians. Tobacyk and Downs (1986) and Steptoe and Fidler (1987) have found, for example, that performance degradation was correlated with such thoughts as, ‘I don’t think I will be able to get through to the end without cracking up’ or ‘I’m almost sure to make a dreadful mistake, and that will ruin everything.’ However, retrospective self-reports may not accurately reflect people’s thoughts during actual performances (Ericsson and Simon, 1984).

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure

In contrast to distraction theory, explicit monitoring (or self-focus) theory (Baumeister, 1984; Masters 1992) postulates that pressure raises selfconsciousness and increases attention to the skill processes involved in a performance. The allocation of conscious attention at this step-by-step level may disrupt a well-learned or proceduralized skill. Explicit monitoring theory assumes a procedural model of skill acquisition; that is, expert performance involves the execution of relatively automatic procedures (Fitts and Posner, 1967; Anderson, 1982). Declarative knowledge is necessary at a novice level, but can become an impediment as the skill becomes well learned. Skilled pianists ‘automatically’ move their fingers over the keyboard. When they try to control their movements consciously, however, they often find that they do not know how to move them, and as a result, ‘slips’ in a performance occur. Tests of explicit monitoring theory have been confined largely to the domain of golf putting (e.g. Masters, 1992; Hardy et al., 1996; Lewis and Linder, 1997; Mullen and Hardy, 2000). More recently, Beilock and Carr (2001) provided additional evidence from golf putting and the non-proceduralized skill of alphabet arithmetic. In both tasks, participants were trained to an asymptotic level of performance under one of three conditions. The first, a single-task practice condition, involved normal training conditions. A second condition was labelled ‘distraction’, in which participants learned the task while simultaneously performing a word generation task. Finally, a ‘selfconsciousness’ group practised golf-putting while being filmed. Following training, all three groups were given a high-pressure post-test. Results showed performance degradation in putting, but not in the alphabet arithmetic task. Furthermore, self-consciousness training helped to inoculate putters against choking. Beilock and Carr concluded that performance pressure induces explicit monitoring, which in turn leads to decrements in the execution of a proceduralized skill.

Experiment overview The present study was designed to compare the relevance of distraction and of explicit monitoring theories in the domain of music. To our knowledge, no research has directly examined the processes underlying music performance breakdowns, despite being described as a detrimental outcome of performance anxiety (e.g. Steptoe and Fidler, 1987). To ensure that our participants had identical music performance experience, we recruited individuals with no musical knowledge and taught them basic note and rhythm reading skills. Using a procedure that was analogous to that of Beilock and Carr (2001), participants were trained on a keyboard task under one of three conditions (single-task, dual-task, and videomonitoring) before being exposed to either a high-pressure or low-pressure post-test. While the single-task condition served as a baseline measure of performance, the dual-task and the video-monitoring conditions were designed

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

157

158

Psychology of Music 33(2)

to expose participants to the causal mechanisms of performance degradation as postulated by the two theories. The present study used the methodology employed by Beilock and Carr (2001) because their study has been one of few to directly contrast diverging predictions of the two theories. In the dual-task condition, training took place under a distracting environment (listening to distracting background music while practising the keyboard task). This secondary task is more appropriate than word generation (e.g. Beilock and Carr, 2001) because previous research reported no distraction effect when individuals were asked to simultaneously shadow auditory passages while sight-reading piano music (Allport et al., 1972). However, when both tasks involved musical material, performance on the primary task was impaired (Martin et al., 1988). It is important to acknowledge that the methodology in the present study differs from that of the traditional dual-task paradigm in which performance on a secondary task is assessed. In the present study, pilot testing revealed that the background music was sufficiently distracting. Dual-task training can also be used to determine whether performance at post-test is proceduralized or reliant on working memory. Following Beilock and Carr, this training condition was implemented to safeguard the possibility that novice performance on the keyboard task might not be proceduralized within the relatively short training session. In the video-monitoring condition, participants were trained while being filmed by a video camera. This placed them in an environment that was designed to encourage attention to step-by-step skill execution. Exposure to a video camera has been used to facilitate the focus of correct behaviours or skill processes (e.g. Duval and Wicklund, 1973; Geller and Shaver, 1976; Hass, 1984; Duval and Lalwani, 1999). The overriding purpose of the present study was to explore the attentional mechanisms responsible for performance degradation in music. According to distraction theory, pressure should have no effect on performance in a skill that has become proceduralized, because there should be sufficient working memory resources to cope with task-irrelevant distraction. Specifically, we would expect no significant differences between high-pressure and lowpressure participants across the three training conditions. In contrast, explicit monitoring theory predicts that pressure should impair a proceduralized skill, because attempts to monitor each individual note would disrupt the automaticity of task performance. Thus, we would expect the single-task group to perform worse at post-test when under high pressure. Furthermore, explicit monitoring theory makes differential predictions about the effects of pressure on the other two training conditions. In the dual-task condition, high-pressure participants should exhibit more errors than low-pressure participants at post-test, because they are unaccustomed to explicitly monitor their skill. In the video-monitoring condition, neither high- nor low-pressure participants should be negatively affected by pressure, because these

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure

individuals are trained under conditions that encourage attention to step-bystep execution. Another variable of interest in the present study was trait anxiety. Our specific research question was whether individuals with a high predisposition to anxiety would be more prone to performance breakdowns when exposed to pressure. One of the few studies investigating this issue in music reported that high anxiety, as measured by Report of Confidence as a Performer Scale and a modified version of the General Trait Anxiousness Scale, was associated with poorer quality performance when playing in front of an audience (Craske and Craig, 1984). The present study administered the more widely used StateTrait Anxiety Inventory (STAI; Spielberger et al, 1983). Consistent with Craske and Craig’s study, we expected trait anxiety to correlate positively with performance degradation. In sum, the primary purpose of the present study was to investigate the attentional processes governing performance degradation in music. Specifically, the study sought to examine whether music performance breakdowns are due to task-irrelevant distraction or to the explicit monitoring of skill processes. This issue was investigated by comparing the effects of high or low pressure on performance following one of three different training procedures, namely, single-task, dual-task, and video-monitoring.

Method PARTICIPANTS

Psychology undergraduate students from the University of New South Wales participated in the study. All participants received course credit for their participation. To ensure that the participants had identical music-related experience, all were ‘naïve’ musicians. That is, they had little or no musical knowledge. Eighteen participants were first used as pilots in the development and refining of the keyboard task and distractor manipulations. These participants did not take part in the full experimental procedure. A total of 78 participants were recruited from the actual experiment. Data from six of them could not be used, because they failed to complete the keyboard orientation session. Thus, the resulting number of participants was 72. DESIGN

The experiment had a 3 (single-task vs dual-task vs video-monitoring) × 2 (high-pressure vs low-pressure) × 2 (training vs post-test) mixed design. There were two phases, training and post-test. In the training phase, participants were randomly assigned to one of three training conditions: singletask, dual-task, and video-monitoring. In the post-test phase, half of the participants were tested under high pressure, while the other half were tested under low pressure.

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

159

160

Psychology of Music 33(2) PROCEDURE

All participants were trained and tested individually. They were informed that the purpose of the study was to examine skill acquisition on a keyboard task. After completing the Trait Scale of the STAI (Spielberger et al., 1983), participants were given an orientation to basic keyboard skills. They then completed a training session followed by either a high-pressure post-test or a lowpressure post-test. Keyboard orientation The purpose of the pre-training keyboard orientation was to introduce participants to basic note and rhythmic reading. During this session, all participants received instructions concerning pitch, followed by rhythm training. First, they were instructed on how to identify notes C, D, E, F, G on the staff and the keyboard, followed by preliminary practice in synchronizing single notes to the metronome (m.m. = 40 bpm). The musical segments used in Ash and Holding’s (1990) study were employed to help participants acquire basic intervallic reading skills. Participants performed these sequences with the metronome until no errors were made. They were then introduced to various note durations (crotchets, quavers, and minims) and were required to tap various rhythms to the metronome beat. Finally, participants were given a two-bar musical sequence, comprising all pitch and note durations. The experimenter demonstrated how this sequence should sound. Participants were then required to identify the notes and to tap the rhythm. Again, participants performed this sequence with the metronome until no errors were made. The same set of instructions was used for all participants. Data from the keyboard orientation were not recorded. Training Following the keyboard orientation phase, participants were given a four-bar keyboard task, where they practised under one of the three training conditions: single-task, dual-task, or video-monitoring. Participants were first instructed to identify the notes and to tap the rhythm. The experimenter then demonstrated by playing the melody once. To the same metronome beat, participants were instructed to perform the sequence from start to finish without any repeats. Participants completed a total of 15 performance trials, consisting of three blocks, each of five trials, with each block being separated by a three-minute rest interval. At the end of each trial, the experimenter informed the participants of the type(s) of errors made. The experimenter said ‘note’ if one or more pitch errors occurred, and ‘beat’ if one or more duration errors occurred. Pilot testing revealed that performance on the keyboard task was reaching asymptote after 15 trials. 1. Single-task condition: Participants practised under a ‘normal’ training environment without any experimental manipulation. Upon completion

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure

of the training session, participants completed either a five-trial highpressure or a five-trial low-pressure post-test. 2. Dual-task condition: While performing the practice trials, participants listened to J.S. Bach’s Fugue No. 20 in A minor, BWV 865 over a headphone. This recording was selected because it had a different tempo from the keyboard task. Pilot testing revealed that this piece was rated as the most distracting compared with three other classical piano pieces. Following the training session, the recording was turned off, and participants then took part in a five-trial (high- or low-pressure) post-test identical to that of the single-task group. It is worth noting that when asked, no participant reported familiarity with the distraction piece. 3. Video-monitoring condition: The distinguishing feature of this condition was that participants were filmed by a video camera while practising the keyboard task. The video camera was set up on a tripod directly in front of participants, approximately 60 cm away. Participants were instructed to ‘pay close attention to what you are doing’. They were also told that experienced musicians would review the tape in order to gain a better understanding of how individuals learn a keyboard skill. Following the training session, the video camera was turned off and faced away. Participants then took part in a five-trial (high- or low-pressure) posttest, identical to that for the single-task and the dual-task groups. Post-test Post-test involved an additional five performance trials on the keyboard task. Participants were randomly assigned to one of two conditions: high pressure or low pressure. In the low-pressure condition, participants were not told when the practice trials ended and when the testing session began. The post-test was simply presented as another series of practice trials. In the highpressure condition, however, participants were given a scenario designed to increase pressure. This scenario was a replication of Beilock and Carr (2001)’s methodology, and involved the manipulation of reward contingency and the presence of ego-relevant threat. Specifically, participants were told that if they improved their accuracy, they would receive $5. Participants were also informed that this award was a ‘team effort’ and that they had been randomly paired with another participant, who had already received a ‘good’ rating. Therefore, if the present participant did not perform well, neither participant would receive $5. Following the post-test, all participants in the high-pressure condition were given the monetary reward, regardless of their performance. MATERIALS

The keyboard task During training and post-test, the same four-bar melody was employed. This melody consisted of the notes and rhythm learned in the initial keyboard orientation (see Figure 1).

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

161

162

Psychology of Music 33(2)

FIGURE

1 The keyboard task used in the experiment.

Equipment A computer-monitored MIDI keyboard was used to collect the data. The computer program (Melody Assistant 2.0 from Myriad) allowed the recording and playback of each performance trial, and also provided the metronome beat. The materials used for the training sessions included a video camera for the video-monitoring group, and a personal stereo, headphone, and a CD for the dual-task group. Self-report questionnaire The Trait Scale of the STAI (Spielberger et al., 1983) was used as a measure of participants’ predisposition to anxiety. ERROR SCORING

Performance accuracy was measured by counting the total number of pitch, duration and correction errors within each trial block. Figure 2 illustrates some examples of the error coding system, adapted from Palmer and Drake (1997). An event was labelled as ‘pitch error’ when the performed pitch component(s) differed from the notated pitch information. The term ‘duration error’ was used to refer to errors where the performed note duration differed from the duration notated in the musical sequence. A ‘correction error’ occurred when the participant played a sequence incorrectly, followed by a

Pitch errors: when the performed pitch component(s) differed from the notated pitch information

FIGURE

2 Examples of the error scoring system.

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure

pause, and then a correct restart. Any errors made after the correct restart were coded in the usual manner. Errors arising from the omission or addition of notes, or from mis-struck notes (i.e. adjacent notes being played simultaneously), were all counted as pitch errors. The number of possible errors per bar was set at a maximum of one error per category (pitch, duration, and correction). Errors of a continuous nature (e.g. adding a rapid string of notes in between two correct notes) were coded as a discrete event (i.e. one error), because it was otherwise impossible to consistently quantify such errors. This measure appeared sufficiently sensitive to diagnose differences in performances. Each performance trial was coded twice, once by the experimenter, and a second time by an experienced musician who was blind to the purpose of the study. Both noted each incidence of error type for every bar interval as it occurred. The interrater reliability was 0.96, as measured by the Pearson product–moment correlation. This indicates that there was a high degree of consistency between the two raters. The error scores reported in the Results section represent the ratings made by the experimenter.

Results PERFORMANCE ACCURACY

The error scores were compared in analyses of variance (ANOVA), followed by analyses of simple effects tests, where appropriate. First, the training phase was examined to determine whether practice led to improvement in performance. Performance changes from the final training to the post-test blocks were then analysed. The mean error scores and their standard errors are in Table 1. The effectiveness of the training phase was analysed using a 3 (single-task, dual-task, video-monitoring) × 2 (first training block, final training block) 1 Mean number of errors across training (first and final) and post-test blocks for the three training groups under high and low pressure (standard errors in parentheses)

TA B L E

First training block

Final training block

Post-test block

Single-task low-pressure Single-task high-pressure

9.08 (0.72) 9.58 (0.59)

1.83 (0.37) 1.75 (0.36)

1.17 (0.37) 3.08 (0.55)

Dual-task low-pressure Dual-task high-pressure

11.00 (0.91) 11.08 (0.74)

2.92 (0.45) 3.08 (0.29)

1.33 (0.38) 3.33 (0.57)

9.33 (1.01) 8.67 (0.81)

2.67 (0.36) 2.50 (0.47)

2.00 (0.58) 0.42 (0.23)

Condition

Video-monitoring low-pressure Video-monitoring high-pressure

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

163

164

Psychology of Music 33(2)

ANOVA with mean number of errors as the dependent variable. All groups showed significant improvement in performance from the first to the final training blocks, as evidenced by a main effect of practice (F1,66 = 643.438, p < 0.0001) However, this effect did not interact with group (all Fs < 3.2, p > 0.07). Within the final training block, performance was similar across groups (F < 3.4, p > 0.07). However, at post-test, significant differences emerged between the dual-task and single-task groups (F1,66 = 8.989, p < 0.004) and between the video-monitoring and single-task groups (F1,66 = 15.746, p < 0.0001). To examine the effect of pressure on performance, a 3 (single-task, dualtask, video-monitoring) × 2 (high pressure, low pressure) × 2 (final training block, post-test block) ANOVA was carried out on the mean error scores. As a factor in the analysis, block refers to the comparison between the final training block and the post-test block. There was a significant main effect of block (F1,66 = 8.413, p < 0.05), and a pressure by block interaction effect (F1,66 = 9.275, p < 0.003). This indicates that there was an overall change in performance at post-test when the pressure manipulation was introduced. All three three-way interactions (involving group, pressure, and block) were tested. Two of them reached significance. Figures 3(a) and 3(c) show the results of the first significant three-way interaction, involving the difference between single-task and video-monitoring, pressure, and block (F1,66 = 6.184, p < 0.015). For the single-task group, high-pressure participants performed worse from the final training block to the post-test block, while their low-pressure counterparts improved. For the video-monitoring group, however, both the high- and low-pressure participants improved from final training block to post-test block. A simple effects test showed that the improvement was significantly greater for high- than for low-pressure participants (F1,22 = 5.812, p < 0.03). Figures 3(b) and 3(c) show the results of the second significant threeway interaction, involving the difference between dual-task and videomonitoring, pressure, and block (F1,66 = 17.423, p < 0.001). Under high pressure, performance of the dual-task participants’ performance was worse at post-test compared to the final training block. The reverse was true, however, when under low pressure. This pattern of results differed from that shown by the video-monitoring group, where both the high- and low-pressure participants improved from final training block to post-test block. As mentioned earlier, simple effects showed that high-pressure participants in the video-monitoring group improved more than low-pressure participants. The third three-way interaction, involving the difference between singletask and dual-task, pressure, and block, was not significant (F1,66 = 2.847, p > 0.09). As illustrated in Figures 3(a) and (b), both single-task and dualtask groups made more errors at post-test when under high pressure, but showed improvement in performance when under low pressure. In the

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure

Mean number of errors

(a) Control goup a) Control group 3.50 3.00 2.50

Low -pressure

2.00

High-pressure

1.50 1.00 0.50 0.00 Final training

Post-test Posttest

Block

(b) Dual-task group

b) Dual-task group

Mean number of errors

3.50 3.00 2.50

Low -pressure

2.00

High-pressure

1.50 1.00 0.50 0.00 Final training

Post-test Posttest

Block

Mean number of errors

c) Video-monitoring group (c) Video-monitoring group 3.50 3.00 2.50

Low -pressure

2.00

High-pressure

1.50 1.00 0.50 0.00 Final training

Post-test Posttest

Block

3 Mean number of errors for the three groups (single-task, dual-task, videomonitoring) across the final training block and post-test block.

FIGURE

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

165

166

Psychology of Music 33(2)

high-pressure condition, simple effects comparing performance change from final training block to post-test block revealed that the decline in performance was significant for the single-task group (F1,44 = 8.23, p < 0.01) but not for the dual-task group (F1,44 = 0.256, p > 0.6). Although these analyses were based on the finding of no differences between groups at final training, it is possible that these differences could reach significance if a larger sample had been used. To test whether pressure had an effect beyond any pre-existing performance differences at finaltraining, an analysis of covariance (ANCOVA) was conducted in which posttest was entered as the dependent variable, and final training as the covariate. The analysis revealed that performance at final training did not predict performance at post-test (F2,32 = 1.129, p > 2.9). When the effect of final training is removed, the effect of training group reaches significance (F2,32 = 11.139, p > 0.05). This indicates that group membership, rather than performance at final training, accounts for the performance differences at post-test. TRAIT ANXIETY

Table 2 presents the mean trait anxiety scores for all six conditions. ANOVA yielded no significant differences between the conditions (F1,66 = 0.102, p < 0.05). To determine whether the STAI would predict the tendency to breakdown under pressure, a regression analysis was conducted between trait anxiety scores and the change in performance within the single-taskstressed group. As recommended by Aiken and West (1991), trait anxiety scores were centred prior to analysis. The single-task-stressed subgroup was selected because we were interested in whether trait anxiety predicted performance degradation under normal, as opposed to artificial, training environments. The result was not significant (r = 0.043, F = 1.72, p > 0.8), indicating that trait anxiety scores as measured by the STAI did not predict the direction of performance change when participants in the single-task condition were exposed to high pressure. 2 Mean trait anxiety scores and their standard errors (in parentheses) across the six conditions (n = 12)

TA B L E

Condition

STAI scores

Single-task low-pressure Single-task high-pressure

42.3 (2.59) 42.8 (2.06)

Dual-task low-pressure Dual-task high-pressure

43.6 (1.88) 43.8 (2.62)

Video-monitoring low-pressure Video-monitoring high-pressure

42.8 (2.72) 43.4 (1.95)

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure

Additional regression analyses were performed to assess whether trait anxiety moderated the relationship between training group and performance change among the high-pressure participants. In the first analysis, trait anxiety scores, the difference between single-task and video-monitoring groups, and their interaction, were regressed against performance change from final training block to post-test block. Trait anxiety scores did not significantly predict performance change (β = 0.55, p > 0.7), but the difference between single-task and video-monitoring group did (β = 0.695, p < 0.05). However, the interaction between these two variables was not significant (β = –0.182, p > 0.2). In the second analysis, trait anxiety scores, the difference between dualtask and video-monitoring groups, and their interaction, were entered as predictors of performance change from final training block to post-test block. Again, trait anxiety scores did not significantly predict performance change (β = 0.07, p > 0.7), but the difference between dual-task and videomonitoring group did (β = –0.59, p < 0.05). Furthermore, the interaction between these two variables did not contribute significantly to the prediction of performance change (β =0.22, p > 0.2). Taken together, our findings suggest that trait anxiety did not moderate the relationship between training conditions and the direction of performance change when exposed to pressure.

Discussion The primary aim of this research was to examine the attentional processes underlying performance degradation under pressure in music. The relevance of two theoretical perspectives, distraction and explicit monitoring, was investigated. According to distraction theory (Wine, 1971; Eysenck, 1979, 1992), there should be no difference between high-pressure and lowpressure participants across three training conditions, single-task, dual-task, and video-montoring. In contrast, explicit monitoring theory (Baumeister, 1984; Masters, 1992) would predict that while the single-task and dual-task groups should make more errors at post-test when under high pressure, the performance of the video-monitoring group should not be negatively affected by pressure. Our results were consistent with the predictions of explicit monitoring theory. For the single-task group participants, those in the high-pressure condition performed worse at post-test than those in the low-pressure condition. This suggests that when one attends to a proceduralized skill, pressure to perform well leads to more errors. In the dual-task condition, high-pressure participants also made more errors than low-pressure participants at posttest. In contrast, the performance of video-monitoring participants actually improved when exposed to high pressure. The video-monitoring training was deliberately designed to raise the participants’ self-consciousness during

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

167

168

Psychology of Music 33(2)

practice (Beilock and Carr, 2001). Participants appeared to have been inoculated against performance degradation under pressure by the condition of their training. Explicit monitoring theory postulates that with pressure, the performer focuses attention on the step-by-step processes in an attempt to execute the task correctly (Baumeister, 1984; Masters, 1992). While such explicit focus of attention is necessary in the initial stages of learning (Anderson, 1982), once the skill becomes proceduralized, a musician cannot consciously attend to all notes and finger positions when performing in real time. Attempts to monitor these processes may disrupt the automaticity of task performance. Explicit monitoring theory can account for why participants trained under the video-monitoring condition were immune to the detrimental effects of performance pressure. Training familiarized them with performance under conditions that encourage conscious monitoring of task processes. They were, therefore, less likely to fail under pressure. The present findings are also consistent with previous research supporting explicit monitoring theory. Lewis and Linder (1997) found, for example, that pressure resulted in poorer performances in golf when participants had not been trained under the presence of a video camera. More recently, Beilock and Carr (2001) found evidence of performance degradation in the proceduralized skill of golf putting but not in a declaratively based alphabet arithmetic task, suggesting that proceduralization determines susceptibility to skill failure. Consistent with the present study, they found that video training could ameliorate the performance deficits normally caused by high pressure within the golf-putting task. The level of proceduralization we achieved might be questioned. Proceduralization implies that attention to step-by-step processes is a decreasing function of skill level, so that skilled performances are thought to operate largely outside of working memory (Anderson, 1982). If the participants had continued to rely on working memory for effective execution, there should be no signs of performance degradation across the three training groups, because information on the skill processes would remain declaratively accessible. Moreover, the dual-task group should perform significantly worse than the single-task and video-monitoring groups during training, because the distraction task interferes with the attentional demands of skill execution. An important finding was, therefore, that neither of these predictions was confirmed in the present study. Clearly, the keyboard skill had become sufficiently proceduralized within the short time-frame of training, and therefore required little working memory. An alternative method to test explicit monitoring theory is to manipulate the level of explicit knowledge available during skill acquisition (e.g. Masters, 1992; Hardy et al., 1996; Bright and Freedman, 1998). For example, Liao and Masters (2001) used analogy learning in order to reduce the amount of explicit knowledge in a table tennis task. The function of an analogy is to

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure

ensure that the underlying mechanisms of the skill are inaccessible to consciousness, thus alleviating degradation under pressure. In table tennis, the coach may ask the learner to imagine a right-angled triangle and swing the bat up along its hypotenuse. The essential rules needed to acquire topspin are disguised in the analogy. Consistent with explicit monitoring theory, Liao and Masters found that analogy learners did not suffer performance impairment under pressure. Thus, future research in the musical domain can utilize analogy learning to test explicit monitoring theory. Developing analogies that integrate the technical components required to perform a musical skill is a challenge for future researchers. The secondary aim of this study was to examine the relationship between trait anxiety and performance degradation under pressure. Our rationale was based on Craske and Craig’s (1984) findings that more anxious pianists produced poorer quality performances when playing in front of an audience. Contrary to expectation, however, trait anxiety appeared to have no systematic effect on performance in the present study. Specifically, there was no evidence that a greater predisposition to anxiety increases vulnerability to skill failure under pressure. Furthermore, trait anxiety did not moderate the relationship between training conditions and the direction of performance change under pressure. Together, our findings did not support the proposition that trait anxiety influences the effect of pressure on performance in music. At first glance, our non-significant results seem inconsistent with previous research reporting strong correlations between performance anxiety and the fear of making mistakes (Steptoe and Fidler, 1987). It is likely, however, that while highly anxious performers exhibit greater levels of fear, those reactions do not necessarily manifest into skill failure during musical performances. That is, the anticipation of failure may not be correlated with actual failure when performing under pressure. Given the support for explicit monitoring theory in the present experiment, it is important for future studies to measure the predisposition to engage in step-by-step focus. For example, Masters, Polman, and Hammond (1993) devised a 20-item Reinvestment Scale, administered to participants who subsequently performed golf-putting, squash, and tennis. Compared to low scorers, high scorers on the Reinvestment Scale tend to fail under pressure. According to Masters et al. (1993), high reinvesters are more prone to skill failure because they are more likely to monitor their processes while executing a proceduralized skill. Thus, future research in the musical domain should consider using the Reinvestment Scale, to investigate skill failure under stressful situations. In conclusion, the study reported here provides preliminary findings concerning the attentional processes governing performance degradation in music. Our results provide indications that skill failure in music is due to explicit monitoring and that video-monitoring training may ameliorate the detrimental effects of performance pressure. As one of the reviewers noted,

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

169

170

Psychology of Music 33(2)

many musicians engage in mock performances during practice. That is, they instinctively implement strategies that might reduce skill failure under pressure. More empirical evidence is required, however, before strong recommendations can be made to teachers and expert performers. The next step should be to extend the present study by using a larger sample, more accomplished musicians, and a more realistic performance setting. Given that expert musicians need to execute complex motor and expressive components during any performance, one would expect their skill to have a greater scope for disruption. The role of trait anxiety on the pressure–performance relationship also requires further research. Our results suggest high levels of trait anxiety are not necessarily translated into skill failure during performances. Other measures such as the Reinvestment Scale may be promising candidates for further investigation. Clarifying the relationship between personality and performance degradation may help identify musicians most likely to fail under pressure, and may potentially inform the development of specialized training programmes. AC K N OW L E D G E M E N T S

This research was presented at the 30th Australasian Experimental Psychology Conference, Sydney, on 26 April, 2003. Special thanks are extended to Dr. William von Hippel and the two anonymous reviewers for their helpful comments. The study was partially funded by the School of Psychology, University of New South Wales. REFERENCES

Aiken, L.S. and West, S.G. (1991) Multiple Regression: Testing and Interpreting Interaction. Newbury Park, CA: Sage. Allport, A.D., Antonis, B. and Reynolds, R. (1972) ‘On the Division of Attention: A Disproof of the Single Channel Hypothesis’, Quarterly Journal of Experimental Psychology 24(2): 225–35. Anderson, J.R. (1982) ‘Acquisition of Cognitive Skill’ Psychological Review 89(4): 369–406. Ash, D.W. and Holding, D. (1990) ‘Backward versus Forward Chaining in the Acquisition of a Keyboard Skill’, Human Factors 32(2): 139–46. Baumeister, R.F. (1984) ‘Choking under Pressure: Self-consciousness and Paradoxical Effects of Incentives on Skillful Performance’, Journal of Personality and Social Psychology 46(3): 610–20. Baumeister, R.F. and Showers, C.J. (1986) ‘A Review of Paradoxical Performance Effects: Choking under Pressure in Sports and Mental Tests’, European Journal of Social Psychology 16(4): 361–83. Beilock, S.L. and Carr, T.H. (2001) ‘On the Fragility of Skilled Performances: What Governs Choking under Pressure?’ Journal of Experimental Psychology: General 130(4): 701–25. Bright, J.E.H. and Freedman, O. (1998) ‘Differences between Implicit and Explicit Acquisition of a Complex Skill under Pressure: An Examination of Some Evidence’, British Journal of Psychology 89(2): 249–63. Calvo, M.G., Ramos, P.M. and Estevez, A. (1992) ‘Test Anxiety and Comprehension

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

Wan and Huon: Performance degradation under pressure Efficiency: The Role of Prior Knowledge and Working Memory Deficits’, Anxiety, Stress, & Coping 5(2): 125–38. Craske, M.G. and Craig, K.D. (1984) ‘Musical Performance Anxiety: The Threesystems Model and Self-efficacy Theory’, Behaviour Research and Therapy 22(3): 267–80. Darke, S. (1988) ‘Effects of Anxiety on Inferential Reasoning Task Performance’, Journal of Personality and Social Psychology 55(3): 399–505. Duval, T.D. and Lalwani, N. (1999) ‘Objective Self-awareness and Causal Attributions for Self-standard Discrepancies: Changing Self or Changing Standards of Correctness’, Personality and Social Psychology Bulletin 25(10): 1220–9. Duval, T.D. and Wicklund, R.A. (1973) ‘Effects of Objective Self-awareness on Attribution of Causality’, Journal of Experimental Social Psychology 9(1): 17–31. Ericsson, K.A. and Simon, H.A. (1984) Protocol Analysis: Verbal Reports as Data. Cambridge, MA: MIT Press. Eysenck, M.W. (1979) ‘Anxiety Learning and Memory: A Reconceptualization’, Journal of Research in Personality 13(4): 363–85. Eysenck, M.W. (1985) ‘Anxiety and Cognitive Task Performance’, Personality and Individual Differences 6(5): 579–86. Eysenck, M.W. (1992) Anxiety: The Cognitive Perspective. Hove: Lawrence Erlbaum. Fitts, P.M. and Posner, M.I. (1967) Human Performance. Belmont, CA: Brooks-Cole. Geller, V. and Shaver, P. (1976) ‘Cognitive Consequences of Self-awareness’, Journal of Experimental Social Psychology 12(1): 99–108. Hardy, L., Mullen, R. and Jones, G. (1996) ‘Knowledge and Conscious Control of Motor Actions under Stress’, British Journal of Psychology 87(4): 621–36. Hass, R.G. (1984) ‘Perspective Taking and Self-awareness: Drawing an E on your Forehead’, Journal of Personality & Social Psychology 46(4): 788–98. Lewis, B. and Linder, D. (1997) ‘Thinking about Choking? Attentional Processes and Paradoxical Performance’, Personality and Social Psychology Bulletin 23(9): 937–44. Liao, C-M. and Masters, R.S.W. (2001) ‘Analogy Learning: A Means to Implicit Motor Learning’, Journal of Sports Sciences 19(5): 307–19. Martin, R.C., Wogalter, M.S. and Forlano, J.G. (1988) ‘Reading Comprehension in the Presence of Unattended Speech and Music’, Journal of Memory and Language 27(4): 382–98. Masters, R.S. (1992) ‘Knowledge, Nerves and Know-how: The Role of Explicit versus Implicit Knowledge in the Breakdown of a Complex Motor Skill under Pressure’, British Journal of Psychology 83: 343–58. Masters, R.S.W., Polman, C.J. and Hammond, N.V. (1993) ‘“Reinvestment”: A Dimension of Personality Implicated in Skill Breakdown under Pressure’, Personality and Individual Differences 14(5): 655–66. Mullen, R. and Hardy, L. (2000) ‘State Anxiety and Motor Performance: Testing the Conscious Processing Hypothesis’, Journal of Sports Sciences 18(10): 785–99. Palmer, C. and Drake, C. (1997) ‘Monitoring and Planning Capacities in the Acquisition of Music Performance Skills’, Canadian Journal of Experimental Psychology 51(4): 369–84. Spielberger, C., Gorsuch, R.L., Lushen, R., Vagg, P.R. and Jacobs, G. (1983) State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press. Steptoe, A. and Fidler, H. (1987) ‘Stage Fright in Orchestral Musicians: A Study of Cognitive and Behavioral Strategies in Performance Anxiety’, British Journal of Psychology 78(2): 241–9.

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010

171

172

Psychology of Music 33(2) Tobacyk, J.J. and Downs, A. (1986) ‘Personal Construct Threat and Irrational Beliefs as Cognitive Predictors of Increases in Musical Performance Anxiety’, Journal of Personality and Social Psychology 51(4): 779–82. Wine, J. (1971) ‘Test Anxiety and Direction of Attention’, Psychological Bulletin 76(2): 92–104.

is a Professor at the School of Psychology, University of New South Wales (UNSW). She has a PhD in Psychology (1986; UNSW) and a Master of Music (1996; UNSW). Her research areas include learning and teaching, and topics within the broad area of psychology of music. Address: School of Psychology, University of New South Wales, Sydney, NSW 2052, Australia. [email: [email protected]]

GAIL HUON

C AT H E R I N E Y. WA N is currently a PhD student in Neuropsychology at the University of Melbourne. The present study formed part of her Masters of Psychology degree, which she completed at UNSW in 2002. Her research interests include music cognition, performance anxiety, neural correlates of auditory abilities, and brain plasticity following intensive skill training. Address: Department of Psychology, School of Behavioural Sciences, University of Melbourne, Parkville, Australia, VIC 3010. [email: [email protected]]

Downloaded from http://pom.sagepub.com at University of Melbourne Library on March 17, 2010