Anxiety, Stress, and Coping, September 2005; 18(3): 183 /208

A Systematic Review of Treatments for Music Performance Anxiety

DIANNA T. KENNY
Australian Centre for Applied Research in Music Performance, Sydney Conservatorium of Music, The University of Sydney, Australia
(Received 15 March 2004; revised 18 October 2004; accepted 28 March 2005)

Abstract A systematic review of all available treatment studies for music performance anxiety was undertaken. Interventions were categorised into psychological treatments and pharmacological treatments. Psychological treatments included behavioral, cognitive, cognitive-behavioral, combined treatments and other therapies. Issues such as the number of studies in each treatment modality, their sample sizes and the methodological quality of most of the studies reviewed precluded firm conclusions about the effectiveness of any of the treatments assessed for music performance anxiety. The field is in urgent need of larger scale, methodologically rigorous studies to assist the large minority of musicians who suffer from performance impairing music performance anxiety.

Keywords: Music performance anxiety, behavioral therapies, cognitive therapies, drug therapies, alternative therapies

A Systematic Review of Treatments for Music Performance Anxiety Several international reviews of music performance anxiety (MPA) among professional orchestral musicians indicate that MPA is widespread and problematic (Steptoe & Fidler, 1987). For example, the International Conference of Symphony and Opera Musicians National US survey (Lockwood, 1989), distributed to 48 orchestras (2,212 respondents) reported that 24% of musicians frequently suffered stage fright, defined in this study as the most severe form of MPA, 13% experienced acute anxiety and 17% experienced depression. A Dutch study (van Kemenade, van Son, & van Heesch, 1995) reported that 59% of musicians in symphony orchestras reported performance anxiety severe enough to impair their professional and/or personal functioning. James (1998), in a survey of 56 orchestras, found that 70% of musicians reported that they experienced anxiety severe enough to interfere with their performance, with 16% experiencing this level of anxiety more than once a week. A recent study indicated that MPA is not limited to orchestral musicians, showing that opera chorus artists are also prone to high levels of performance anxiety (Kenny, Davis & Oates, 2004). It is not difficult to imagine that most performers,

Correspondence: Professor Dianna T Kenny, Director, Australian Centre for Applied Research in Music Performance, Sydney Conservatorium of Music, The University of Sydney, NSW, Australia. E-mail: D.Kenny@fhs.usyd.edu.au ISSN 1061-5806 print/ISSN 1477-2205 online # 2005 Taylor & Francis DOI: 10.1080/10615800500167258

184

D. T. Kenny

by the very nature of their profession, would be affected by the general stresses related to having to perform under conditions of high adrenalin flow, anxiety, fatigue, social pressure, and financial insecurity (Lehrer, Goldman, & Strommen, 1990, p. 48). Sternbach (1995) described the working conditions of professional musicians as generating a total stress quotient that far exceeds that observed in other professions. However, since not all performers suffer the same degree of MPA, or indeed report the same levels of occupational stress, individual differences in a range of psychological characteristics are likely to account for variations in the degree to which musicians experience symptoms. A large number of treatment modalities (e.g., behavioral, cognitive, pharmacological and complementary) has been developed for music performance anxiety (MPA). However, a review of this literature indicates that the field is still in its infancy with respect to the conceptual and theoretical formulations of the nature of MPA and its empirical investigation. The terminology surrounding the concept is also problematic. For example, MPA is sometimes referred to as stage fright and the terms are used interchangeably, while others view stage fright as an extreme form of MPA. Recently, a number of integrative theories have been proposed and these have the potential to direct future research into the aetiology and treatment of the condition. Barlows (2000) model of anxiety may have heuristic value in understanding performance anxiety in general and MPA in particular. This model proposes an integrated set of triple vulnerabilities thatcan accountfor the development of ananxietyor mood disorder.These are a: i. generalized biological (heritable) vulnerability; ii. generalized psychological vulnerability based on early experiences in developing a sense of control over salient events, and a iii. more specific psychological vulnerability whereby anxiety comes to be associated with certain environmental stimuli through learning processes such as respondent or vicarious conditioning. Barlow argues that genetic predisposition and sensitizing early life experiences may be sufficient to produce a generalized anxiety or mood (depression) disorder. The third set of vulnerabilities appears necessary to produce focal or specific anxiety disorders such as panic disorder or specific phobias. For example, social evaluation may be accompanied by heightened somatic sensations that become associated with a perceived increase in threat or danger. In the case of young performers who are high in trait anxiety (the expression of the generalized biological vulnerability), who come from home environments in which expectations for excellence are high but support for achieving excellence is low (generalized psychological vulnerability), exposure to early and frequent evaluations and self-evaluations of their performances in a competitive environment (specific psychological vulnerability) may be sufficient to trigger the physiological, behavioural and cognitive responses characteristic of music performance anxiety. Anxiety may be triggered by conscious, rational concerns or by cues that trigger, unconsciously, earlier anxiety producing experiences or somatic sensations. Once triggered, the person shifts into a self-evaluative attention state, in which self-evaluation of perceived inadequate capabilities to deal with the threat, in this case, the imminent performance, is prominent. Attention typically narrows to a focus on catastrophic cognitive self-statements that disrupt concentration and performance. In this respect, MPA may share commonalities with social anxiety and its extreme form appears phenomenologically similar to social phobia. One could argue that the conditions under which one performs, that is, the degree of social evaluative threat perceived by the performer, is the defining feature of social phobia. Those perceiving most

Treatment review of music performance anxiety

185

threat are likely to experience the greatest anxiety, and those who are most anxious are more likely to perceive performance conditions as more threatening. In the music research literature, MPA is commonly viewed as a constellation of three interactive yet partially independent factors: cognitions, autonomic arousal, and overt behavioral responses (Craske & Craig, 1984; Lederman, 1999; Salmon, 1990). There is empirical support for this three-factor model of music performance anxiety. Craske and Craig (1984) demonstrated greater response synchrony among high trait, compared to low trait anxious performers, particularly when performing in a stressful situation involving evaluation of their performances by expert judges. Low anxious individuals experienced elevated heart rate but not cognitive or behavioral symptoms in the performance condition when compared to a warm-up baseline condition in contrast to high anxious individuals who experienced elevations in all three areas. High levels of self-reported performance anxiety were also related to lower levels of confidence. Researchers are in dispute regarding the nature of this interaction. Zinn, McCain and Zinn (2000) argue that performance anxiety is primarily a psychophysiological event in which the autonomic nervous system initiates and maintains MPA. Alternatively, Kirchner (2003) maintains the symptomatic aspects of MPA are activated by the perception of threat by the performer, and not the autonomic nervous system. Wilson (2002) describes threat perception as an interaction of three variables that play important roles in the experience of distressing anxiety: the performers constitutional and learned tendency to become anxious in response to situations of social stress (trait anxiety); the degree of task mastery, and the degree of situational stress, where high anxiety is more likely to be experienced in situations that are socially or environmentally instructive. Theorizing in the area of test anxiety and academic competence is instructive demanding. Three similar processes are considered relevant to performances in examinations: cognitiveattentional processes (e.g. worry, task-irrelevant thinking, negative self-preoccupation), cognitive skills (study habits) and self-efficacy or the exercise of human agency (how a person influences his/her thoughts, behaviours, goals, and outcomes). In a simultaneous test of these three theories, Smith, Amkoff, and Wright (1990) found that cognitiveattentional processes accounted for most of the variance in both performance on tests and test anxiety, but that both cognitive skills and self-efficacy measures added additional unique variance. These findings suggest that multi-modal interventions are needed to address the multiple difficulties experienced by test anxious individuals. These formulations are directly relevant to our understanding and treatment of music performance anxiety. The following review of treatment for MPA reflects the differing theoretical arguments regarding the etiology of MPA, with some focusing on behavioral change, some on cognitive change, others on reduction of physiological symptoms through the use of pharmacotherapy, and some on idiosyncratic formulations. Assessment of their relative effectiveness may simultaneously cast light on theoretical issues, as well as provide practitioners and researchers with a basis for intervening more effectively in assisting musicians who suffer this condition. The aim of this paper, therefore, is to provide a systematic review of treatments for music performance anxiety, to identify the problems with this research and to make recommendations for future research. Method The aim of the search was to provide the most comprehensive list possible of both published and unpublished primary studies in the area of MPA interventions as the precision of the

186

D. T. Kenny

estimate of effectiveness depends on the information obtained, and the care taken to avoid publication bias. Because this constitutes the first systematic review of treatments for music performance anxiety, broad inclusion criteria were adopted. To identify as many relevant studies as possible, several databases were searched. The search used gateways Ovid and Science Direct to access data bases Journals@Ovid Full Text, MEDLINE (1966 to October 2002), CINAHL (1982 to October 2002), ERIC (1966- October 2002), PsychINFO (1967-October 2002) and EMBASE (1966-October 2002). A search of Dissertation Abstracts International was also conducted, with the last search for relevant dissertations made in October 2003. The search history followed the protocol recommended by the Cochrane Library. The search included treatment studies for music performance anxiety in student or professional musicians. Studies were included if they had a control group, although a couple of preand post-test designs were included because they represented the only available studies in a particular treatment modality. Articles were limited to English. Most studies and reviews were published after 1985, except for the drug studies, which were produced from the late 1960s to the mid 1970s. There have been no recent pharmacological studies on MPA due to the greater ethical constraints on research that have been implemented in the past decade. Keywords for the search strategy in this review were music, musicians and performance anxiety. Alternative keywords such as stage fright were also included. A second set of keywords defined study methodology: randomized controlled trial; double blind method; therapy; treatment; comparative studies; and evaluation studies. These were combined with keywords music performance anxiety. A third set of keywords was used to search for specific treatment types. Non-drug studies were searched under cognitive therapy, cognitive restructuring, self-instruction, cognitive behaviour/behavior therapy, behaviour/behavior therapy, behaviour/behaviour rehearsal, systematic desensitization, stress inoculation, (deep) muscle relaxation, breathing, meditation, biofeedback, Alexander Technique (see p. 19 for definition), music therapy and hypnotherapy. These keywords were also cross-referenced with music performance anxiety. The same search strategy was repeated for each database. For drug studies, the keywords were beta-blocker [Beta blockers block the effect of adrenaline (the hormone norepinephrine) on the bodys beta receptors. This slows down the nerve impulses that travel through the heart. As a result, the resting heart rate is lower, the heart does not have to work as hard and requires less blood and oxygen] and SSRI [ie drugs belonging to the group known as selective serotonin reuptake inhibitors (SSRI) are antidepressant medications, examples of which are sertraline (Zoloft) and fluoxetine (Prozac )]. Dissertation Abstracts International (DAI) produced 101 dissertations, of which 13 met inclusionary criteria and were retrieved. Many of the dissertations did not appear on other electronic lists and these proved to be a rich source of data. All retrieved articles were hand searched for other relevant papers that may have been missed in the search strategy. An additional nine articles were located using this method. Additional references were retrieved by hand searching. This search yielded 125 articles. Abstracts were assessed for relevance to the search criteria. Key review studies were identified, with a preference for critical and systematic reviews. Brodsky (1996) and Nube (1991) were most useful. Treatment studies were divided into drug and non-drug studies, and these were further subdivided according to the specific form of therapy used. These included drug type and dose for the drug studies, and hypnotherapy, Alexander technique,

Treatment review of music performance anxiety

187

behavioral interventions, cognitive interventions, cognitive-behavioral interventions, other combinations of therapies, and other therapies for the non-drug treatment studies. None of the studies were true randomised controlled trials (although the majority randomly assigned subjects to groups) and very few were directly comparable in terms of methodology, subjects, and outcome measures, so a meta-analysis was not possible. However, some of the studies included sufficient data to make it possible to calculate effect sizes. Cohens d is a measure of the distance between the means (of experimental and control groups). Contrary to the p value, effect size measures are not sensitive to sample size and this is an advantage in assessing the effectiveness of treatment in studies with small numbers. According to Cohen (1977), d /0.20, 0.50, and 0.80 constitute small, medium, and large effects respectively. Effect sizes are provided in Tables I /VI. The formula d /M1-M2/spooled was used to calculate Cohens d in each instance. Where a repeated measures design was employed, post (or follow-up) means and standard deviations of the experimental and control groups were used to calculate Cohens d. Where a repeated measures design was employed, but no control group was involved or control group data were unavailable, pre and post (or followup) means and standard deviations of the experimental group were used to calculate Cohens d. Where a simple two group design was employed, the means and standard deviations of the experimental and control groups were used to calculate Cohens d. Where insufficient data were provided or where no significant differences were found, effect sizes were not calculated. These calculations are supplemented with a comprehensive narrative review of all available outcomes. Results Behavioral Interventions Six studies (see Table I) assessed the therapeutic effect of behavioral treatments on MPA, two of which used samples specifically selected because they were high in MPA. Four of the six studies provided sufficient detail for some effect sizes to be calculated. The interventions assessed included systematic desensitization, progressive muscle relaxation, awareness and breathing and behavioural rehearsal. Kendrick, Craig, Lawson, and Davidson (1982) compared behavior rehearsal and cognitive-behavioral treatments for MPA. The behavior rehearsal group did not show improvements in state anxiety or subjective stress, but they did show significant pre- to post-treatment improvements in performance quality and self-statements about performance anxiety (as measured by the Performance Anxiety Self-Statement Scale: PASS). The behavior rehearsal group also showed greater pre- to post-treatment improvements in the visual signs of anxiety than controls, although the cognitive-behavioral group showed an even greater improvement than the behavior rehearsal group on this outcome measure. Kendricks study suggests that behavior rehearsal may be an effective form of treatment for MPA for some outcome measures (STAI, subjective stress scale) (Spielberger, 1983) but not others (self-efficacy and visual signs of anxiety) for which a CBT intervention was superior. Sweeney and Horan (1982) (see Table IV) found that the behavioral technique of cue-controlled relaxation led to improvements in anxiety, MPA, heart rate, and performance quality in students suffering from MPA, but that CBT was not significantly more effective for these outcome measures than this simple behavioral treatment. Richard (1992) (see Table V) failed to find a therapeutic effect with cue-controlled relaxation but this

188

D. T. Kenny

failure may have been due to insufficient statistical power, given the very small sample size in his study. Reitman (1997) examined the therapeutic effect of two systematic desensitization procedures, and found that the treatment groups did not differ significantly to a waiting list control group on any of the anxiety, heart rate, or performance quality outcome measures employed. However, Reitmans study is weakened by a very small sample size (18 subjects across 3 groups), and thus the lack of significance reported may reflect insufficient statistical power. Four other studies (three of which are dissertations) assessed behavioral treatments for MPA on music students. Grishman (1989) and Mansberger (1988) used standard muscle relaxation techniques, Wardle (1969) compared insight/relaxation and systematic desensitisation techniques, and Deen (1999) used awareness and breathing techniques. These studies indicated pre- to post-treatment improvements on self-report measures of performance anxiety (Deen, 1999; Grishman, 1989; Mansberger, 1988) and heart rate (Grishman, 1989; Wardle, 1969), but not performance quality (Deen, 1999; Mansberger, 1988; Wardle, 1969). Again, demand characteristics may have confounded these results due to the transparent nature of the research and the reliance upon self-report measures of MPA. Further, Mansberger (1988) obtained no pre-intervention measures. In summary, behavioral treatments do appear to be at least minimally effective in the treatment of MPA, although the heterogeneity of the treatment approaches employed makes it difficult to isolate consistent evidence for the superiority of any one type of behavioral intervention. Cognitive Interventions Two studies (see Tables II and IV) assessed the therapeutic effect of cognitive techniques alone on MPA. A dissertation by Patston (1996) reported a comparison of cognitive (e.g. positive self-talk, etc.) and physiological strategies in the treatment of MPA. No significant improvements on vocal and visual manifestations of performance anxiety were found for either treatment or control groups. However, the sample consisted of only 17 opera students who were not specifically selected on the basis of their MPA severity, and the intervention was conducted by the author, a singer and teacher, who had no training in psychology. Consequently, this lack of significance may reflect a floor effect in that subjects may not have been sufficiently anxious pre-treatment for a reduction in anxiety to be detected, a lack of statistical power, or problems with the integrity of the intervention. Further, standardised outcome measures were not employed, and the author does not make it clear how subjects were assigned to groups. In a methodologically superior study, Sweeney and Horan (1982) (Table IV) found that cognitive restructuring techniques may be helpful in treating music students suffering from MPA. They found that a treatment group showed significantly greater pre- to posttreatment improvements on MPA, anxiety, performance quality, and heart rate than controls. All effect sizes were large for these outcomes. While this study gives tentative support cognitive strategies in the treatment of MPA, further studies are clearly needed. A dissertation by Roland (1993) (Table III) suggested that self-instruction alone may be useful in reducing MPA, although his failure to include a control group is a major methodological weakness of his study.

Table I. Summary of Behavioral Interventions for Music Performance Anxiety

Author Year Country Subject No Subject type Study design Treatment Control Outcome measures Results Effect size (Cohens d) Conclusions

Deen

1999 (PhD dissertation)

USA

39

Music students

*2 (Groups)  /2 (Time) repeated measures design *Randomly assigned to groups.

Awareness/ breathing exercise (6 weeks practice with a tape)

No tape

Grishman

1989 (PhD dissertation)

USA

41 (51% female)

Advanced music students and professionals

*2 (Groups)  /2 (Time) repeated measures design *Randomly assigned to groups within instrument family.

Modified jacobsonian progressive muscular relaxation (6  /2 hr sessions over 3 wks)

Waiting list

Kendrick et al.

1982

Canada

53 (91% female)

Student pianists with MPA

*3 (Groups)  /3 (Time) repeated measures design *Randomly assigned to groups.

Behavior rehearsal (3 weekly sessions of 1.5 /2 hrs each / homework)

1. CBT (self-instruction, attentionfocusing techniques) (3 weekly sessions of 1.5 /2 hrs / each homework) 2. Waiting list

Performance anxiety inventory (PAI) Performance quality rating scale *PAI measured pre-treatment and post jury examination. Performance anxiety response Q (PARQ) STAI-T STAI-S Symptom questionnaire Music performance anxiety questionnaire (MPAQ) HR *Measured pre and post-treatment STAI-S Subjective Stress Scale Expectation of Personal Efficacy Personal Efficacy scale for Musicians Performance Scale for Musicians Performance Anxiety Self-Statement Scale (PASS) Visual signs of anxiety Performance quality *Measured pre and post-treatment FU.

pB /.005 .56 NS NS

Treatment group, but not controls, showed significant decrease in performance anxiety (PAI).

NS

NS

NS pB /.05 pB /.05
/.05 pB

NS .64 .72 .74

Treatment group, relative to controls, showed improvements on STAI-S, symptoms, MPAQ, and baseline HR.

Treatment review of music performance anxiety

pB /.05 NS NS
/.05 B

.85 NS NS .62 /1.21 Both treatment groups showed more improvement than controls on PASS and performance quality. CBT group showed more improvement than behavior rehearsal group and controls on Expectations of Personal Efficacy. CBT showed greater improvement than behavioral group who showed more improvement than controls on visual signs of anxiety.

/.05 B

.25 /.83

/.05 B /.05 B

1.24 /1.99 1.20 /1.51

189

190

Table I (Continued ) Author Year Country Subject No Subject type Study design Treatment Control Outcome measures Results Effect size (Cohens d) 1.03 Conclusions

D. T. Kenny

Mansberger

USA 1988 (M. Music Dissertation)

19 (63% female)

Music students

*Simple 2 group comparison *Did not specifically state whether or not group assignment was random

Deep muscle No contact relaxation (3  / 90 min weekly sessions)

Reitman

1997 (PhD dissertation)

USA

18 (72% female)

Music students with MPA

*3 (Groups)  /2 (Time) repeated measures design *Randomly assigned to groups.

Wardle

1969

USA

30

Music students (Brass players)

/2 *3 (Groups)  (Time) repeated measures design

1. Verbal coping systematic desensitization 2. Music assisted coping desensitization /75 min (8  weekly group sessions) 1. Systematic desensitisation /40min (7  sessions) 2. Insightrelaxation (7  / 40 min sessions)

Waiting list

Self-assessed state anxiety scale (SASAS) Level of perceived self-efficacy Strength of perceived selfefficacy Performance quality *All measures taken after treatment, before juried performance only sEMG HR STAI Performance Anxiety Response Questionnaire (PARQ)

pB /.05 pB /.05
/.05 pB

1.03

Treatment group showed higher level and strength of self-efficacy and lower SASAS scores than controls.

1.24

NS

NS

NS NS NS NS

NS NS NS NS

Orientation session *Randomly assigned to groups

Heart rate (HR) Walkins-Famham Performance Scale Behavioral observation Performance ratings by judges *Measured pre and post-treatment

pB /.05 NS

Insufficient date provided NS

/.05 pB

NS

Treatment groups showed reduction in HR from pre- to post-test, controls showed an increase. HR reduction greater Insufficient in insight group than date desensitization provided group. Treatment NS groups showed reduction in instrumental behaviors indicating anxiety compared to controls.

Treatment review of music performance anxiety Cognitive-behavioral Interventions

191

Three studies (see Table III) assessed the therapeutic effect of cognitive-behavioral strategies on MPA. Harris (1987), Roland (1993), and Kendrick et al. (1982) all reported that standard CBT techniques were effective in the treatment of MPA in students specifically selected for study because of the severity of their MPA. Two of the three studies (Kendrick et al., 1982 & Roland, 1993) reported moderate to strong effect sizes for improvement in performance quality (Harris did not assess performance quality). Kendrick et al. (1982) found that CBT was superior to behavioral rehearsal in terms of improvements in expectations of personal efficacy and visual signs of anxiety. Harris (1987) and Roland (1993) reported that CBT led to reductions in state anxiety as measured by the STAI, although Kendrick et al. (1982) failed to find a significant difference between treatment and control groups on this measure. The evidence for improvements in MPA following CBT is quite consistent, although further studies with larger samples are needed to confirm this evidence. While CBT may be superior to drug therapy in treating MPA (see Clark & Agras, 1991), currently there is little evidence to suggest that it is superior to either standard behavioral or cognitive techniques alone. Combined Interventions A number of studies (see Table IV) have examined the effect of combining treatment approaches. Only one of the four studies (Clark & Agras, 1991) provided sufficient data to permit the calculation of effect sizes. Clark and Agras (1991) found that cognitivebehavioral therapy was superior to drug therapy with buspirone in the treatment of MPA, and also that improvement in performance quality, for which a strong effect size was found (1.64), was greater in a CBT/placebo group than in a placebo group alone. [It should be noted that since buspirone has failed in studies of social phobia to separate from placebo, it is not appropriate to describe the CBT/buspirone as a combined treatment. Saying that CBT /buspirone in this context is not conceptually different from saying that CBT / placebo]. Finally, Sweeney and Horan (1982) found that behavioral, cognitive, and cognitive-behavioral treatments were all effective in treating students suffering from MPA, when compared to a control group, but that no significant differences were apparent between the three types of treatment. This was the case for improvements in general anxiety, MPA, and performance quality. The only difference between treatments was with respect to heart rate, where the behavioral and cognitive treatments alone led to greater preto post-treatment improvements than the combined CBT treatment. Brodsky and Sloboda (1997) assessed counseling, counseling/relaxation, and counseling/relaxation/vibro-tactile sensations in the treatment of MPA. This study had no control group, was not conducted with a sample of MPA sufferers, and did not report the statistics required to clearly determine the nature of any group differences. Other studies have assessed the combined effect of behavioral, or cognitivebehavioral, and biofeedback techniques on MPA. Niemann, Pratt, and Maughan (1993) found that students with MPA showed significantly greater pre- to post-treatment anxiety reduction than controls when treated with behavioral and biofeedback techniques, a finding consistent with that of Nagel, Himle, and Papsdorf (1989), who employed a combined CBT/biofeedback treatment. Sweeney-Burton (1997) was unsuccessful in reducing anxiety and improving musical performance following a similar behavioral/biofeedback intervention, but this study was conducted with music students who had not specifically

192 D. T. Kenny

Table II. Summary of Cognitive Interventions for Music Performance Anxiety Author Year Country Subject No Subject type Study design Treatment Control Outcome measures Results Effect size (Cohens d) NS Conclusions

Patson 1996 Australia 17 (71% female) (M.Music dissertation)

Opera students

*3 (Groups) /2 (Time) repeated measures design *Did not specifically say if group assignment was random.

Cognitive strategies (e.g. positive self-talk etc.)

1. Physiological strategies (8 / .5 hr sessions over 9 days) 2. Control group (8 /.5 hr walks over 9 days)

Vocal and visual NS manifestations of performance anxiety *Measured pre and posttreatment.

No pre- or post-test significant differences for any group.

Table III. Summary of Cognitive-behavioral Interventions for Music Performance Anxiety

Author Year Country Subject No Subject type Study design Treatment Control Outcome measures Results Effect size (Cohens d Lowest; highest) 1.25; 1.65 .40 NS NS Conclusions

Harris

1987

USA

17 (47% female)

Students with MPA

/3 *2 (Groups)  (Time) repeated measures design *Did not specifically say if assignment to groups was random.

6 /2 hr weekly sessions involving cognitive self-Instruction, relaxation training, imagery & behavioral rehearsal

Waiting list

PAI STAI-S STAI-T Test Anxiety Inventory (TAI) Anxiety rating from teacher *Measured pre and posttreatment and at 3 mth follow-up

pB /.001 pB /.05 NS NS pB /.01

Kendrick 1982 et al.

Canada

53 (91% female)

Students pianists with MPA

*3 (Groups)  /3 (Time) repeated measures design *Randomly assigned to groups.

CBT (selfinstruction, attentionfocusing techniques) (3 weekly sessions of 1.5-2 hr each / homework)

1. Behavior rehearsal-3 weekly sessions of 1.5-2 hrs each / homework 2. Waiting list

STAI Subjective Stress Scale Expectations of Personal Efficacy Scale for Musicians Performance Anxiety Self-Statement Scale (PASS) Visual signs of anxiety Performance quality STAI-S Musical Performance Anxiety SelfStatement Scale (MPASS) Performance quality *Measured at pre-test recital (1 wk before treatment),

NS NS B /.05

B /.05

/.05 B /.05 B

Roland

Australia 25 (88% 1993 female) (PhD Dissertation)

Student pianists

*3 (Groups)  /3 (Time) repeated measures design *Randomly assigned to groups

1. Self instruction (SI) 2. Progressive muscle relaxation(R) 3. Combination treatment (6  /1 hr weekly sessions; 2hr sessions for combin.)

None Performance quality Measured at pre-test racital (1 wk before treatment), post-test recital (1 wk after treatment), and 6 wk follow-up.

/.05 pB pB /.05

/.05 pB

Treatment group, but not controls, showed lower PAI and STAI-S, at post-test than pretest. Treatment group showed pre to post improvement in teacher anxiety ratings (not 1.31 reported for controls) Improvements in PAI and teacher rating, but not STAI-S, were maintained at follow-up in treatment group. NS Both treatment groups showed more improvement NS than controls on PASS and performance quality. CBT .62; 1.21 group showed more improvement than behavior rehearsal group and controls on .25;.83 Expectations of Personal Efficacy. CBT showed greater improvement than 1.24;1.99 behavioral group who showed more 1.20;1.51 improvement than controls on visual signs of anxiety. 1.37;1.72 Combination group showed 1.13;1.64 for pre-post and pre-FU combination improvement on MPASS, group; .77 SI showed pre-post for SI group improvement only. R .46;1.01 showed pre-post and pre-FU improvement on performance quality, not significant for other groups.

Treatment review of music performance anxiety 193

194

Table III (Continued ) Author Year Country Subject No Subject type Study design Treatment Control Outcome measures Results Effect size (Cohens d Lowest; highest) Conclusions

D. T. Kenny

33 (88% female)

Music students with MPA

*3 (Groups)  /3 (Time) repeated measures design *Randomly assigned to groups.

post-test recital (1 wk after treatment), and 6 wk follow-up. STAI-S Waiting list 1. Standard CBT 2. Modified CBT (included preperformance routine & visual rehearsal etc.) (4  /2 hr weekly sessions)

B /.05

MPASS

/.05 B

Self-efficacy scale Performance quality HR *Measured at pre-test recital (1 wk before treatment), post-test recital (1 wk after treatment), and 5 wk follow-up.

/.001 B

NS
/.05 B

Insufficient data provided for combined treatment groups Insufficient data provided for combined treatment groups Insufficient data provided for combined treatment groups NS Insufficient data provided for combined treatment groups

Standard and modified CBT groups didnt differ on any outcome variables. when treatment groups were combined they were superior to controls on STAI, MPASS, SE, but not PQ, at post and FU, Both treatment groups differed to controls on anticipatory HR (controls lower surprisingly); standard CBT group also differed to controls on HR during performance (controls lower surprisingly); no diffs between treatment groups.

Treatment review of music performance anxiety

195

been selected for their high MPA and thus the lack of significance may be due to a floor effect. In summary, there is little evidence to suggest that combined treatment approaches enhance improvements in MPA over and above those offered by single treatments, but further research in this area is needed.

Other Interventions Meditation. Although over 2,000 studies on meditation were identified in PsycInfo and Medline databases, only one study (Chang, 2001) examined the effects of meditation on music performance anxiety. Meditation was defined as a self-regulatory practice designed to train attention in order to bring mental processes under greater voluntary control (Walsh, 1995, p. 388). It focuses on the use of attention, thought or concentration as a means to achieve understanding or insight (Goleman & Schwartz, 1976). In this study, 19 music students (mostly pianists) were randomly assigned to a meditation or control condition. Pre and post performance anxiety measures indicated very modest support for the role of meditation in reducing performance anxiety. Interestingly, there were no significant differences between the groups on measures of cognitive interference (mind wandering, intrusive thoughts) that the meditation intervention specifically addressed. Once again, the small sample size may have obscured potentially significant effects and a larger study with a more intensive treatment intervention is needed to address this issue. Biofeedback. McKinney (1984) assessed the use of peripheral skin temperature training using biofeedback on self-reported MPA and quality of musical performance in 32 male wind instrumentalists. He found that peripheral skin temperature training had no effect on peripheral skin temperature or anxiety levels assessed by the STAI-S. There was some evidence that biofeedback may be useful in improving performance quality (effect size / .83). This study was not conducted with subjects specifically selected because of their high level of MPA and effects may have been more pronounced with subjects suffering higher performance anxiety. Like Richard (1992), see below), the main effect of time (that is, repetition of the anxiety provoking performance at post-test) was the most salient result in this study, indicating that familiarity with the requirements of the task and practice effects were more effective in reducing performance anxiety. Music therapy. Montello (1989) and Montello, Coons, and Kantor (1990), assessed the effect of a 12-week music therapy intervention on freelance musicians suffering from MPA in two separate studies. The intervention consisted of musical improvisation, three musical performances in front of an audience, awareness techniques and verbal processing of their anxiety responses. The outcome measure in the first study was self-ratings of confidence about performance ability. In the second study, a replication of the first, an attentional control group and additional outcome measures were included. These were the Narcissistic Personality Inventory (NPI), and judgments of raters regarding improvements in the musicality of subjects performances as a function of treatment. The 10 experimental subjects became significantly more confident (effect size /2.46) as performers as measured by the Personal Report of Confidence as a Performer Scale (PRCP) (Appel, 1976) and less anxious as measured by the Spielberger State/Trait Anxiety Inventory (STAI) than the 10 waiting-list control subjects after music therapy intervention. In Study II, experimental subjects became significantly less self-involved, less stressed and more musical after group music therapy intervention as compared to attentional control subjects (all effect sizes /2).

196

Table IV. Summary of Combined Interventions for Music Performance Anxiety

Author Year Country Subject No 54 (50% females approx.) Subject type Professional symphony orchestra musicians Study design Treatment Control Outcome measures Results Effect size (Cohens) NS Conclusions

D. T. Kenny

Brodsky & Sloboda

1997

Israel/ UK

*3(Groups)  /3 (Time) repeated measures design *Subjects matched before being randomly assigned to one of three interventions

1. Counselling None 2. Counselling / relaxation / listening to music 3. Counselling / relaxation / music and musicgenerated vibration sensations

General Health Questionnaire (GHQ-28) STAI

NS

B /.005 NS

Derogatis Stress Profile (DSP) POMS

Insufficient data provided NS

Clark & Agras

1991

USA

29

Musicians (including full-time prof) with social phobia (with regard to performance situations).

*4 (Groups)  /2 (Time) repeated measures design *Randomly assigned to groups. *Double blind study.

Placebo 1. CBT /Placebo medication (P) / 2. CBT Buspirone 3. Buspirone /5 weekly (CBT  meetings) [Busp  / 5mg tab, up to 12/day, for 6 weeks]

Maslach Burnout Inventory Appraisal of Music Performers Stress (AMPS) Music Performance Stress Survey (MPSS) *Measured pre and post-treatment, and 2 month FU. Subjective anxiety during musical performance Subjective anxiety during speech Heart rate during musical performance Heart rate during speech Self Statement Questionnaire Personal report of confidence as performer Musical performance *all measured pre and post treatment (Personal report of confidence also 1mth follow-up)

/.0005 Insufficient B data provided NS NS /.05 B

NS

Insufficient data provided NS

Few differences between treatments. Together the treatments led to pre to post improvements in anxiety, depression, anger, fatigue, confusion, psychological stress, work overload, work relations, performers stress. *Stats poorly done in this study.

pB /.05
/.05 pB

NS NS
/.05 pB /.05 pB

.29 (CBPT /P v Placebo) / .62 (CBT P v Placebo) NS NS .40;1.05 .86;2.36

pB /.05

1.64

CBT groups showed greater pre to post fall in music & speech anxiety than non CBT groups. Placebo groups showed a greater pre to post fall in Self-Statement Questionnaire than buspirone groups. CBT groups showed greater pre to post fall in performer confidence at 1 mth follow-up than non CBT groups. CBT /placebo group showed greater post to FU improvement on performer confidence than each

Table IV (Continued ) Author Year Country Subject No Subject type Study design Treatment Control Outcome measures Results Effect size (Cohens) Conclusions

Nagel et al.

1989

USA

20 (60% female)

Music students with MPA

*2 (Groups)  /2 (Time) repeated measures design *Randomly assigned to groups.

Progressive muscle Waiting list relaxation, cognitive therapy, biofeedback training (6 weekly /6 group sessions weekly biofeedback sessions)

Niemann et al.

1993

USA

18 (78% female)

Music students with MPA

/2 *2 (Groups)  (Time) repeated measures design *Assignment to groups based on study schedule of students.

6 Waiting list /35 min biofeedback sessions /6  /1 hr group meetings (training in coping strategies; muscle relaxation, breathing awareness, imagery) & coinciding practice of the strategies with sedative music during individual biofeedback training.

Performance Anxiety Inventory (PAI) STAI-T STAI-S Test Anxiety Inventory Autonomic Perception Questionnaire (APQ) Rational Behavior Inventory (RBI) *Measured pre and post-treatment STAI-S and T

/.001 B /.05 B NS NS NS

1.51 0.40 NS NS NS NS

of the other 3 groups. CBT /placebo group showed more improvement over time in musical performance than placebo alone group. CBT /placebo group showed improvement from pre to post on the self-statement questionnaire compared to buspirone alone group. Significant greater pre to post reduction in performance anxiety and trait anxiety in treatment than control group.

Treatment review of music performance anxiety

NS

pB /.01 pB /.01

Insufficient data provided

Facilitating/ debilitating anxiety scales *Measured pre and post-treatment

Significant pre to post improvements were found on anxiety scales.

197

198

Table IV (Continued ) Author Year Country Subject No 49 (49% female) Subject type Music students with MPA Study design Treatment Control Outcome measures Results Effect size (Cohens) NS 1.27;1.62 .88;1.000 .24;.86 1.58;2.000 1.56;1.75 Conclusions

D. T. Kenny

Sweeney & Horan

1982

USA

*5 (Groups)  /2 (Time) repeated measures design *Matched by pretest scores then randomly assigned to groups.

1. Cue-controlled relaxation (CCR) 2. Cognitive restructuring (CR) 3. CCR /CR (All had 6 weekly sessions of 60 min each)

1. Standard treatmen t(ST) control (Musical Analysis Training) 2. Waiting list

Achievement Anxiety Test Scale (AATS) Musical Performance Competence (MPC) Behavioral Index of Anxiety (BIA) Pulse rate Piano Performance Anxiety Scale (PPAS) Anxiety Differential (AD) *Measured pre and post-treatment STAI-T and S but results reported only for S. Adjudicator Performance Rating Form *Measured pre and post-treatment.

NS B /.05
/.05 B /.05 B B /.05 /.05 B

SweeneyBurton

1997 (D.Ed. USA Dissertation)

30 (43% female)

Music students

*2 (Groups)  /2 (Time) repeated measures design *Did not specifically say if group assignment was random.

Diaphragmatic No training breathing, progressive muscle relaxation, autogenic training, biofeedback.

NS

NS

NS

NS

CCR and CR alone, but not combined, led to lower pulse rate than ST control; no diff between treatments. All 3 treatments were superior to ST control (and did not differ from each other) on PPAS, AD, MPC (except CR), and BIA (except CCR). No significant diffs between treatment and control groups with regard to pre-post changes in TAI or music performance.

Treatment review of music performance anxiety

199

The holistic group music therapy intervention was recommended as a way to reduce performance anxiety by helping musicians to: (1) become more aware of the underlying dynamics of performance anxiety; (2) experience unconditional acceptance and support in a safe group environment; (3) bond with their music-selves; (4) transform anxiety through creativity (reparation); and (5) bond with others in the spirit of musical community (p. 4). Despite the small sample size, there were many methodological strengths in this study, including the choice of subjects with severe MPA. This approach appears to warrant further consideration as a treatment for music performance anxiety in experienced, professional musicians. Ericksonian resource retrieval. This technique refers to the use of unconscious mechanisms within the individuals personal history to adapt to a current life challenge. Resources are defined as automated patterns of feeling, perceiving and behaving (Lankton & Lankton, 1983, p. 121). Resource retrieval is a process oriented intervention that focuses on assisting the person to access existing strengths rather than teaching him/her new skills. The technique involves identification of current life challenge, identification of relevant personal resource, cognitive and emotional re-experiencing and practicing of resource, and experience of the current challenge in the context of the cognitions and emotions related to the resource. Richard (1992) assigned 21 volunteer music students to one of three conditions: Ericksonian resource retrieval, cue-controlled relaxation and a wait list control. He found that Ericksonian resource retrieval reduced MPA at about the same rate as that of cue-controlled relaxation. However, repeated measures analysis found that all three groups improved over time on measures of anxiety and confidence as a performer. Treatment and control groups did not differ in pre- to post-treatment improvements on self-reported MPA, performance quality, or performer confidence. Small subject numbers and attrition from the treatment conditions make the results difficult to interpret, and a larger replication with better compliance is needed to fully assess the potential of this technique in reducing MPA and in improving jury performances. Hypnotherapy. Only one study (Stanton, 1994) has assessed the therapeutic effect of hypnotherapy on music performance anxiety (MPA). In this study, a group of music students suffering from MPA were given two 50-minute sessions of hypnotherapy, while a control group of students discussed their performance and anxiety with their lecturer for a similar length of time. The Performance Anxiety Inventory (PAI) was administered pre and post-treatment, and at 6 month follow-up. A significant pre- to post-treatment reduction in MPA (as measured by the PAI) was found for the treatment group, but not the control group, and a further significant reduction was found at 6 month follow-up. This study was weakened by the fact that the criteria for subject selection were somewhat subjective, with students selected by their lecturer for inclusion if they appeared to be prone to MPA. Stanton also suggested that some experimental subjects may have discussed the study with control subjects both during and after the treatment period, giving rise to a potential confound. In summary, Stantons findings suggest that hypnotherapy may be useful in the treatment of MPA, but further methodologically superior studies are required. Alexander Technique. Alexander Technique is an educational process in which the student learns a set of skills that result in lessening of the areas of tension in the body, so that movement becomes easier and less effortful. The aim is to cultivate a more natural alignment of head, neck and spine that has associated with it qualities of balance, strength

200

D. T. Kenny

and coordination. The method aims to teach conscious and voluntary control over posture and movement and to undo involuntary muscle tension. The Alexander Technique is a method for eliminating unwanted muscular patterns or habits that interfere with smooth performance. For a performer, the technique is a method for using kinaesthetic cues, the sensations of tension, effort, weight, and position in space, in order to organize ones field of awareness in a systematic way. Only one study has assessed the therapeutic effect of the Alexander Technique on MPA. In this study (Valentine, Fitzgerald, Gorton, Hudson, and Symonds, 1995), one group of music students were given 15 lessons in the Alexander Technique, while a control group received no lessons in the Alexander Technique. The outcome measures were heart rate, musical quality, technical quality, the Nowlis Mood Adjective Checklist (NMAC), and the Music Performance Anxiety Self-Statement Scale (MPASS). Measures were taken at audition, in class prior to treatment, in class post-treatment, and at final recital. Controls showed a greater increase in heart rate variance from audition to recital than the treatment group. The treatment group showed pre- to post-treatment improvement in musical and technical quality, but the effect sizes were weak (.25 and .13 respectively), and an increase in scores on the active and warm-hearted scales of the NMAC, whilst controls showed the opposite pattern of results. The treatment group also showed a greater pre- to posttreatment decrease in anxiety than controls, and an increase in positive attitude to performance (as measured by the MPASS). These findings provide very weak evidence that the Alexander Technique may be effective in improving the quality of performance and mental state of the performer (insufficient data to calculate effect size), and may help to modulate increased variability of heart rate under stress. However, a number of methodological weaknesses need to be highlighted. Subjects were not specifically selected for their high levels of MPA, and thus the extent to which the results can be generalized to genuine MPA sufferers is unclear. Further, the sample size was inadequate, and the design lends itself to confounding by demand characteristics. Data were insufficient to calculate effect sizes for three of the five outcome measures. Given these shortcomings and the scarcity of studies that have investigated the efficacy of the Alexander Technique in treating MPA, any conclusions must at this stage be tentative. In summary, there are too few studies available that provide any evidence for the effectiveness of these treatment modalities from which to draw conclusions. Better and larger studies are needed before any of these treatments can be confidently recommended to anxious musicians. Drug Interventions Lehrer (1987) and Nube (1991) have published comprehensive reviews of the impact of beta-blockers on music performance anxiety and a brief overview of other drugs, such as anxiolytics and antidepressants has been provided by Sataloff, Rosen, and Levy (2000). Accordingly, only a brief summary of drug interventions will be provided in this paper and the interested reader is referred to these earlier papers for a more comprehensive review. Beta-blockers have become increasingly popular among performers in recent years. For example, Lockwood (1989), in a survey of 2,122 orchestral musicians, found that 27% used propranolol to manage their anxiety prior to a performance; 19% of this group used the drug on a daily basis. Performers prefer beta-adrenoceptor blocking agents over anxiolytic drugs (egdiazepam) because of their reduced impact on central functions such as mental alertness and cognitive function. Beta blockers appear to be most effective for those musicians who report primarily somatic manifestations of their anxiety (e.g. palpitations,

Table V. Summary of Other Interventions for Music Performance Anxiety

Author Year Country Subject No 20 (75% female) Subject type Music students in piano, violin and voice Study design Treatment Control Outcome measures Results Effect size (Cohens d) .41 NS NS Conclusions

Chang

2001

USA

*Post-test only control group design *Randomly assigned to groups. 2 (Groups)  /2 (Time) repeated measures design *Did not specifically state whether group assignment was random. *2 (Groups)  /2 (Time) repeated measures design *Randomly assigned to groups. *3 (Groups) /2 (Time)  repeated measured design *Randomly assigned to groups.

Meditation (8 weekly 1.25 hr sessions)

1. Meditation 2. No treatment Control

McKinney 1984

USA

32 male

Music students in wind instruments

Biofeedback training (2  /15 min sessions/wk for 5 weeks)

No training control

Performance Anxiety Inventory (PAI) Cognitive Interference Questionnaire (CIQ) Concert PerformanceHeart rate and blood pressure Peripheral skin temperature (PST) STAI-S Watkins-Famham Performance Scale (WFPS) Performance rating by three judges (PRJ)

pB /.05 NS NS

No significant diffs between treatment and control groups with regard to pre-post changes in STAI or music performance. Biofeedback group showed greater pre to post increase in WFPS than control group. No significant differences between biofeedback and control groups with regard to pre-post changes in PST, STAI-S or PRJ. Therapy group showed greater increase in PRCP than both control groups. Therapy group, compared to attentional controls, became more musical, less stressed, and less self-involved during performance from pre to post. Therapy group moved significantly towards normalcy on the NPI (from pre to post) compared to attentional controls but not waiting list controls.

NS

NS

NS pB /.05 NS

NS .83

NS

Treatment review of music performance anxiety

Montello et al.

1990

USA

Exp1: 17

Freelance musicians with MPA

Group music therapy (12 weekly 1.5 hr sessions)

Waiting list

Personal Report of Confidence as a Performer (PRCP) STAI-T *Measured pre and post-treatment.

pB /.01
/.05 pB

2.61 0.67

Exp2: 24 (53% female)

Freelance musicians with MPA

Group music therapy (12 weekly 1.5 hr sessions)

1. Waiting list 2. Attentional control (this group met weekly with therapist to complete psych tests and discuss musical topics)

Richard

1992 (D.Ed. Dissertation)

USA

21

Music students with MPA

*3(Groups)  /2(Time) repeated measures design *Randomly

1. Eriksonian Resource Retrieval (4 /45 min  weekly sessions)

Waiting list

PRCP STAI-T Narcissistic Personality Inventory (NPI) Performance musicality Performance stress symptoms Performance self-involvement *Measured pre and post-treatment Musical Performance Stress Inventory (MPSI) STAI-S Personal Report of Confidence as a

pB /.001 2.46 NS NS 0.16 (cited pB /.05 in article) /.001 2.29 pB

/.001 2.10 pB /.001 2.47 pB

NS

NS

NS NS

NS NS

Significant pre to post improvement on STAI-S and PRCP for Eriksonian group. No significant group  /time interactions.

201

202

Table V (Continued ) Author Year Country Subject No Subject type Study design Treatment Control Outcome measures Results Effect size (Cohens d) NS Conclusions

D. T. Kenny

assigned to groups.

Stanton

1994

Australia 40

Music students with MPA

Valentine et al

1995

UK

25(84% female)

Music students

*2(Groups)  /3(Time) repeated measures design *Paired on basis of PAI scores then randomly assigned to groups within each pair. *2 (Groups) Alexander  /4 (Time) Technique repeated (15 lessons) measures design *Randomly assigned to groups

2. Cuecontrolled relaxation (6  /45 min weekly sessions) Hypnotherapy (2  /50min sessions 1 week apart)

Performer (PRCP) Cognitive strategies assessment Performance competence Discussion with lecturer (matched for session length) Performance Anxiety Inventory (PAI) *Measured pre, post and 6 months later

NS

NS

NS

pB /.01

1.26 (Pre v Post) 1.98 (Post v FU)

Significant reduction in PAI following treatment, but not control. Further significant reduction in treatment group at 6 month follow-up.

No lessons in AT

Heart rate Musical quality Technical quality Nowlis Mood Adjective Checklist Music Performance Anxiety Self-Statement Scale *Most measures taken at audition, in class prior to treatment, in class following treatment, and at final recital.

/.05 pB pB /.05 /.05 pB pB /.05 /.06 pB

Insufficient data provided 0.25 0.13 Insufficient data provided Insufficient data provided

Controls showed greater increase in heart rate variance from audition to recital than treatment group. Treatment group showed pre to post class improvement in musical and technical quality, controls showed opposite. Treatment group showed pre to post increase in active and warm-hearted scales of Nowlis, controls showed decline. Treatment group showed greater pre to post decrease in anxiety scale of Nowlis than controls. Treatment group showed increase in positive attitude to performance (MPASS) from pre to post, controls showed decline.

Treatment review of music performance anxiety

203

hyperventilation, tremor, trembling lips, sweating palms etc) (Gates et al., 1985; James & Savage, 1984) and less effective for those experiencing more cognitive or psychological effects, such as low self-esteem, social phobias, or generalised free floating anxiety (Lehrer, 1987). Nube (1991) identified nine studies examining the effects of various beta-blockers (Atenolol, Metopolol, Nadolol, Oxprenolol, Propranolol, Pindolol) on MPA. Two additional studies were identified (Berens & Ostrosky, 1988; Lide n & Gottfries, 1974) bringing to 11 the total number available for review. Generally, the sample size in these studies is small (n /8-30 subjects) and most used subjects who had not been formally diagnosed with music performance anxiety. Three used student musicians, two used professionals and four used a combination of students and professionals. Drug type and dosages varied as did time of administration and outcome measures, making meta-analysis impossible. However, in seven of the nine studies reported by Nube (1991) and in both additional studies, symptoms associated with sympathetic overactivity was significantly improved by medication. The findings regarding the effects of beta blockers on other outcome measures were less conclusive. For example, James, Griffith, Pearson, and Newbury (1977) and James, Burgoyne, and Savage (1983) reported a positive impact of medication on intonation in string players while Pearson and Simpson (1978), (as cited in Nube , 1991) found no difference in intonation between medicated and unmediated musicians. Similarly, results were different for different drugs / both propanolol and oxprenolol reportedly reduced anxiety and bow shake while nadolol had no such effect. For those studies that used subjective as well as physiological measures to assess outcome, most reported no effect on measures such as perception of performance (Brantigan et al., 1982; Gates et al., 1985; James et al., 1983), state or trait anxiety (Brantigan et al., 1982; Gates et al., 1985), stage fright rating scale (Neftel et al., 1982), or judges ratings of quality of performance (James et al., 1983). Berens and Ostrosky (1988) assessed the impact of beta-adrenergic blocking agents prior to musical performance on 150 musicians and singers. This study concluded that the drug decreased tachycardia and improved the quality of the performance, while drawing attention to the potential difficulties with drug withdrawal and unwanted side effects. Symptoms that have been reported in at least 10% of users include bradycardia, hypotension, cold extremities, gastrointestinal upset, sleep disturbance and muscle fatigue. Discussion This systematic review of treatment for music performance anxiety indicates that there is considerable scope for the further development and evaluation of appropriate interventions. The number of well-conducted studies that meet current standards for the conduct of randomized controlled trials is very small, and this review was therefore forced to rely on studies of uncertain methodology, unpublished dissertations, small sample size, potentially resulting in lack of power to detect significant treatment effects, and design flaws such as lack of a control group or inappropriate experimental designs (e.g., no pre-test scores). The use of heterogenous volunteer subjects, who were not selected on the basis of a prescribed level of music performance anxiety, and who had not reported problematic music performance anxiety may have obscured significant treatment effects due to floor effects. Some of the studies combined student and professional musicians, a practice not supported by evidence that the nature and causes of anxiety in novice and mature performers may be quite different.

204

D. T. Kenny

A rigorous definition of MPA is needed to advance treatment. However, defining MPA as a social anxiety (social phobia) using criteria set out in DSM-IV-TR (APA, 2000) as the inclusion criteria may be too restrictive, particularly if the musician presenting for treatment experiences MPA as a focal anxiety (ie does not meet other criteria for social anxiety). Few of the intervention studies reviewed acknowledged that performers need a certain amount of arousal or anxiety to maximise their performance. For example, increased anxiety may facilitate performance, especially for performers with high task mastery, and be associated with better adjudicator ratings (Steptoe & Fidler, 1987). Experienced performers may also need more anxiety in order to achieve peak performance (Steptoe, 1989). Wolfe (1989) distinguished between the potentially facilitative aspects of anxiety, such as arousal and intensity, and the negative aspects such as apprehension and nervousness that impair performance. Similar distinctions have been made in the test anxiety literature. For example, Raffety, Smith & Ptacek (1997) found that debilitating trait anxiety was associated with lower examination scores, less effective coping and higher levels of tension, worry and distraction than those with facilitative trait anxiety. Further, MPA is task and context specific, becoming apparent only with difficult pieces performed under stressful conditions, such as examinations, auditions and major performances as opposed to technically less demanding pieces played under neutral or reassuring conditions (Wilson, 2002). These are issues that must be addressed by health professionals planning interventions for music performance anxiety. Such treatments need to promote sufficient relaxation to counteract the negative symptoms of excess sympathetic nervous system activity, while maintaining sufficient arousal and concentration needed for an optimal musical performance (Brotons, 1994). Much more work needs to done on the role of cognition in MPA and its effect on performance quality. Research on the role of cognition in test anxiety is instructive here, although findings are complex and to some extent counter-intuitive. Early research indicated a simple relationship between cognition, test anxiety and test performance. High test anxious subjects engaged in negative cognitions that impaired task performance while low test anxious subjects engaged in problem solving cognitions that enhanced performance (Wine, 1980). Subsequent research reported that high test anxious children engaged in more negative and positive coping cognitions than low test anxious children and that positive coping was not task facilitating (Zata & Chassin, 1985). Several researchers have since concluded that the absence of negative thinking may be the more salient factor than the presence of positive thinking and that this relationship held across low, moderate and high anxious children. Further, coping cognitions of high anxious (but not low and moderate anxious) children interfered with task performance. Only positive self-evaluations showed a positive relationship with task performance (Prins, 1994). The most likely explanation for these findings is that on-task and coping cognitions distract attention from the task or in feet remind children that they are in an unpleasant anxiety-arousing situation. These findings could form the basis for the development of hypotheses to be tested in music performance settings. None of the studies could be pooled in a meta-analysis primarily because too few provided sufficient data to calculate effect sizes, use of diverse subject groups and treatments, duration and intensity of treatment, and use of disparate outcome measures. Further, many of the studies reported in this review constituted the only study of its kind for the treatment genre (hypnotherapy, biofeedback, meditation, Alexander Technique, Ericksonian resource retrieval, and music therapy). Standard treatments for anxiety do not seem to be directly transportable to treatment for music performance anxiety, although it is premature to make such an assertion in view of

Treatment review of music performance anxiety

205

the lack of methodologically sound research with musicians. With respect to behavioral interventions, effect was somewhat dependent on the outcome measure, with some techniques showing improvement on self-report measures but not performance quality (Deen, 1999; Mansberger, 1988; Wardle, 1969), while others showed positive effect on both measures (Sweeney & Horan, 1982). Interventions leading to an improvement in performance quality are most desirable, since they will have a self-reinforcing, confidenceenhancing effect on future performances, obviating the need for further treatment. Behavior rehearsal, cognitive restructuring, combined self-instruction and progressive muscle relaxation and combined self-instruction and attentional training all had significant positive effects on performance quality. Systematic desensitization, deep muscle relaxation, visual rehearsal and awareness and breathing interventions alone did not improve performance quality. There is some evidence that biofeedback combined with a CBT intervention can reduce anxiety but the finding that biofeedback alone did not produce significant post treatment effects suggests that the effective component was CBT. With respect to the other treatments reviewed, even fewer conclusions can be drawn because in each case, results are based on one methodologically compromised study. Tentative support is demonstrated for Alexander Technique although more methodologically robust studies are needed to confirm its potential effectiveness. Similar conclusions may be drawn for hypnotherapy and meditation. Neither biofeedback training to increase peripheral blood flow to the fingers nor Ericksonian Resource Retrieval produced the desired effects of anxiety reduction during performance or improved performance quality. The music therapy model proposed and tested by Montello appears to have real promise in the management of MPA. Although subject numbers were small, the study was methodologically well constructed and a replication study confirming and extending the findings of the first study lends support to the conclusions drawn. The central musical focus of this intervention and the secondary focus on psychological processes may be the underlying factor contributing to the positive outcomes of this intervention. Once again, further research is needed before such an approach can be recommended. This therapy was designed for mature freelance musicians and included music improvisation as part of the treatment. This approach may not be suitable for younger, less experienced musicians as the expectation to improvise may itself cause additional anxiety in young performers with little or no experience in musical improvisation. Of great interest is the finding that a combined cognitive-behavioral intervention and placebo medication resulted in better performance quality ratings and self-report than combined cognitive-behavioral treatment with buspirone (Clark & Agras, 1991). Perhaps the expectation that an external agent (placebo) will control the symptoms of anxiety is better than the use of the actual agent (active drug), since there are no physiological side effects related to placebos. However, as pointed out earlier, buspirone has not been shown to be effective in the treatment of social phobia, so its use for MPA is perhaps equivalent to placebo. In the current research climate with its more stringent requirements related to participant safety and informed consent, drug trials of the type conducted in the 1970s and 1980s are no longer possible, or desirable. Future Directions There are numerous possibilities for future research in the area of MPA including greater attention to the conceptual and theoretical issues underpinning the concept, the need for properly controlled studies that compare treatment effectiveness for MPA across two or more treatment modalities, and which investigate the effectiveness of combined treatments.

206

D. T. Kenny

A small number of comparative studies have been conducted, but they have been weakened by numerous methodological flaws discussed earlier. The publication of well designed and controlled comparative treatment studies is essential if the field is to arrive at a position in which best practice in treating MPA can be determined. Such studies need a sound theoretical basis, drawing on developments in related areas of test and sport anxiety, should use a consistent definition of MPA, appropriate and robust samples, and standardized outcome measures in relation to self-reported MPA, functional impairment, and performance quality. In addition, random assignment to groups, the inclusion of appropriate control groups and control variables, and the use of pre/post repeated measures designs should be common to all future treatment studies in this area. To determine the optimally effective parameters of any treatment approach, future research studies will need to systematically vary the intensity and duration of treatment, with the ultimate aim to develop an effective short-term treatment approach that can be tailored to different sub-groups of MPA sufferers. Given that the aetiology of MPA may differ for students and professional musicians, effective treatment approaches may need to be tailored to the needs of each sub-group. Further, the inclusion of follow-ups in future treatment studies is essential, with maintenance of improvements over time being one of the crucial determinants of the effectiveness of any anxiety-related treatment approach. Researchers will also need to investigate the value of relapse prevention strategies in maintaining improvements in both anxiety levels and performance quality. In addition to further research on treatment effectiveness, there is an urgent need for further work on the assessment of MPA. This systematic review of the literature on treatment approaches to MPA highlights the heterogeneous approach to assessment in the MPA literature, and the subsequent difficulties in comparing the results of studies with substantially different methodologies and outcome measures. It is important that future studies avoid over-reliance on self-report measures of MPA, and that standardized measures of MPA, functional impairment, and performance quality be developed and employed consistently across treatment studies. Until these fundamental assessment and measurement issues are resolved, it will be difficult to develop a progressive treatment literature on MPA. In conclusion, 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. For significant progress to be made, future research will require a clear definition of MPA, consistency and strength in methodology, and the development of robust and appropriate outcome measures. Only then will firm conclusions about the effectiveness of various treatment approaches for MPA be possible. This comprehensive review can form the basis for the development of a more robust science of MPA and its treatment.

References
Appel, S. S. (1976). Modifying solo performance anxiety in adult pianists. Journal of Music Therapy , 13 , 2 /16. American Psychiatric Association (2000). Diagnostic and Statistical Manual-4th Edition-Text Revision (DSM IVTR). American Psychiatric Association (APA). Barlow, H. (2000). Unravelling the mysteries of anxiety and its disorders from the perspective of emotion theory. American Psychologist, November , 1247 , 1263. Berens, P. L., & Ostrosky, J. D. (1988). Use of beta-blocking agents in musical performance induced anxiety. Drug Intelligence and Clinical Pharmacy , 22 , 148 /149. Brantigan, C. O., Brantigan, T. A., & Joseph, N. (1982). Effect of beta blockade and beta stimulation on stage fright. The American Journal of Medicine , 72 , 88 /94.

Treatment review of music performance anxiety

207

Brodsky, W., & Sloboda, J. A. (1997). Clinical trial of a music generated vibrotactile therapeutic environment for musicians: main effects and outcome differences between therapy subgroups. Journal of Music Therapy , 34 , 2 / 32. Brodsky, W. (1996). Music performance anxiety reconceptualised: A critique of current research practices and ndings. Medical Problems of Performing Artists , II , 88 /98. Brotons, M. (1994). Effects of performing conditions on music performance anxiety and performance quality. Journal of Music Therapy , 31 , 63 /81. Chang, J. C-W. (2001). Effect of meditation on music performance anxiety. Doctor of Education Dissertation, Teachers College, Columbia University. Dissertation Abstracts International, AAT 3014754. Cohen, J. (1977). Statistical power analysis for the behavioral sciences (rev. ed). New York: Academic Press. Clark, D. B., & Agras, W. S. (1991). The assessment and treatment of performance anxiety in musicians. American Journal of Psychiatry , 148 , 598 /605. Craske, M, & Craig, K. (1984). Musical-performance anxiety: The three-systems model and self-efcacy theory. Behavior Research and Therapy , 22 , 267 /280. Deen, D. R. (1999). Awareness and breathing; Keys to the moderation of musical performance anxiety . Unpublished doctoral dissertation, University of Kentucky. Gates, G. A., Saegert, J., Wilson, M, Johnson, L., Shepherd, A, & Hearne, E. (1985). Effect of beta blockade on singing performance. Annals of Otolaryngology, Rhinology and Laryngology , 94 , 570 /574. Goleman, D. J., & Schwartz, G. E. (1976). Meditation as an intervention in stress reactivity. Journal of Consulting & Clinical Psychology , 44 , 456 /465. Grishman, A. (1989). Musicians performance anxiety: The effectiveness of modied progressive muscle relaxation in reducing physiological, cognitive, and behavioral symptoms of anxiety. Dissertation Abstracts International , 50 (6-B), 2622. Harris, S. R. (1987). Brief cognitive-behavioral group counselling for musical performance anxiety. Journal of The International Society for the Study of Tension in Performance , 4 , 3 /10. James, I. (1998). Western orchestral musicians are highly stressed. Resonance: International Music Council [France] , 26 , 19 /20. James, L., Grifth, D. N., Pearson, R. M., & Newbury, P. (1977). Effect of Oxprenolol on stagefright in musicians. The Lancet , 2 , 952 /954. James, I. M., Burgoyne, W., & Savage, I. T. (1983). Effect of pindolol on stress-related disturbances of musical performance: Preliminary communication. Journal of the Royal Society of Medicine , 76 , 194 /196. James, I., & Savage, I. (1984). Benecial effect of nadolol on anxiety-induced disturbances of performance in musicians: a comparison with diazepam and placebo. American Heart Journal , 108 , 1150 /1155. Kendrick, M. J., Craig, K. D., Lawson, D. M., & Davidson, P. O. (1982). Cognitive and Behavioral Therapy for Musical-Performance Anxiety. Journal of Consulting & Clinical Psychology , 50 , 353 /62. Kenny, D.T., Davis, P. & Oates, J. (2004). Music performance anxiety and occupational stress amongst opera chorus artists and their relationship with state and trait anxiety and perfectionism. Journal of Anxiety Disorders, 18 , 757 /777. Lankton, S.R., & Lankton, C. H. (1983). The answer within: The clinical framework of Ericksonian resource retrieval . New York: W.W. Norton. Lederman, R. J. (1999). Medical treatment of performance anxiety: a statement in favour. Medical Problems of Performing Artists , 14 , 117 /121. Lehrer, P. M. (1987). A review of the approaches to the management of the management of tension and stage fright in music performance. Journal of Research in Music Education , 35 , 143 /53. Lehrer, P., Goldman, N., & Strommen, E. (1990). A principal components assessment of performance anxiety among musicians. Medical Problems of Performing Artists , 5 , 12 /18. Lide n, S., & Gottfries, C.-G. (1974). Beta-blocking agents in the treatment of catecholamine-induced symptoms in musicians. The Lancet , 2 , 529. Lockwood, A. H. (1989). Medical problems of musicians. New England Journal of Medicine , 320 , 221 /227. Mansberger, N. B. (1988). The effects of performance anxiety management training on musicians self-efcacy, state anxiety and musical performance quality. Unpublished doctoral dissertation, Western Michigan University. McKinney, H. V. (1984). The effects of thermal biofeedback training on musical performance and performance anxiety. Unpublished doctoral dissertation, University of Northern Colorado. Montello, L. (1989). Utilizing music therapy as a mode of treatment for performance stress of professional musicians. Dissertation Abstracts International: Section A: Humanities and Social Sciences (I-A) , 50/10. Montello, L., Coons, E. E., & Kantor, J. (1990). The use of group music therapy as a treatment for musical performance stress. Medical Problems of Performing Artists , 5 , 49 /57.

208

D. T. Kenny

Nagel, J. J., Himle, D. P., & Papsdorf, J. D. (1989). Cognitive-behavioral treatment of musical performance anxiety. Psychology of Music , 17 , 12 /21. Neftel, K., Adler, R., Ka ppeli, L., Rossi, M., Dolder, M., & Ka ser, H. et al., (1982). Stage fright in musicians: a model illustrating the effect of beta blockers. Psychosomatic Medicine , 44 , 461 /469. Niemann, B. K., Pratt, R. R., & Maughan, M. L. (1993). Biofeedback training, selected coping strategies, and music relaxation interventions to reduce debilitative musical performance anxiety. International Journal of Arts Medicine , 2 , 7 /15. Nube , J. (1991). Beta-blockers: Effects on performing musicians. Medical Problems of Performing Artists , 6 , 61 /68. Patston, T. (1996). Performance anxiety in opera singers . Unpublished masters thesis, Sydney Conservatorium of Music, University of Sydney, Australia. Prins, J. (1994). Cognition in tests-anxious children: the role of on-task and coping cognition reconsidered. Journal of Consulting and Clinical Psychology , 62 , 404 /409. Raffety, B. D., Smith, R. E., & Ptacek, J. T. (1997). Facilitation and debilitation trait anxiety, situational anxiety and coping with an anticipated stressor: A process analysis. Journal of Personal and Social Psychology , 72 , 892 / 906. Reitman, A. D. (1997). The effects of music-assisted coping systematic desensitization on music performance anxiety: A three systems model approach, Unpublished doctoral dissertation, Temple University. Richard, J. J., Jr. (1992). The effects of Ericksonian resource retrieval on musical performance anxiety. Dissertation Abstracts International: Section B: the Sciences & Engineering , 55 (2-B), 604. Roland, D. J. (1993). The development and evaluation of a modied cognitive-behavioral treatment for musical performance anxiety. Dissertation Abstracts International: Section B: the Sciences & Engineering , 55 (5-B), 2016. Salmon, P. (1990). A psychological perspective on musical performance anxiety: a review of the literature. Medical Problems of Performing Artists , 5 , 2 /11. Sataloff, R., Rosen, D. C., & Levy, S. (2000). Performance anxiety: What singing teachers should know. Journal of Singing , 56 , 33 /40. Smith, R. J., Amkoff, D. B., & Wright, T. I. (1990). Test anxiety and academic competence: A comparison of alternative models. Journal of Counseling Psychology , 37 (3), 313 /321. Stanton, H. E. (1994). Reduction of performance anxiety in music students. Australian Psychologist , 29 , 124 /127. Steptoe, A. (1989). Stress, coping, and stage fright in professional musicians. Psychology of Music , 17 , 3 /11. Steptoe, A., & Fidler, H. (1987). Stage fright in orchestral musicians: A study of cognitive and behavioral strategies in performance anxiety. British Journal of Psychology , 78 , 241 /249. Spielberger, C. (1983). Manual for the State-Trait Anxiety Inventory . Palo Alto, CA: Consulting Psychologists. Sternbach, D. J. (1995). Musicians: A neglected working population in crisis. In S. L Sauter, & L. R. Murphy (Eds.), Organizational risk factors for job stress (pp. 283 /302). Washington, DC: American Psychological Association. Sweeney-Burton, C. (1997). Effects of self-relaxation techniques training on performance anxiety and on performance quality in a music performance condition. Unpublished doctoral dissertation, University of North Carolina, Greensboro. Sweeney, G. A., & Horan, J. J. (1982). Separate and combined effects of cue-controlled relaxation and cognitive restructuring in the treatment of musical performance anxiety. Journal of Counselling Psychology , 29 , 486 /497. Valentine, E., Fitzgerald, D., Gorton, T., Hudson, J., & Symonds, E. (1995). The effect of lessons in the Alexander technique on music performance in high and low stress situations. Psychology of Music , 23 , 129 /141. van Kemenade, J. F. L. M., van Son, M. J. M., & van Heesch, N. C. A. (1995). Performance anxiety among professional musicians in symphonic orchestras: A self-report study. Psychological Reports , 77 , 555 /562. Wardle, A. (1969). Behavior modication by reciprocal inhibition of instrumental music performance anxiety. Research in music behavior: modifying music behavior in the classroom. In C. K. Madsen, C. H. Madsen & R. D. Greer (Eds.) (1975), Research in Music Behavior (pp. 191 /205). New York: Teachers College Press. Walsh, R. (1995). Phenomenological mapping: A method for describing and comparing states of consciousness. Journal of Transpersonal Psychology , 27 , 25 /56. Wilson, G. D. (2002). Psychology for performing artists (2nd ed). London, UK: Whurr. Wine, J. D. (1980). Cognitive-attentional theory of test anxiety. In I. G. Sarason (Ed.), Test anxiety: Theory, research and applications (pp. 349 /385). Hillsdale, NJ: Erlbaum. Wolfe, M. L. (1989). Correlates of adaptive and maladaptive musical performance anxiety. Medical Problems of Performing Artists , 4 , 49 /56. Zatz, S., & Chassin, L. (1985). Cognition of test anxious children under naturalistic test-taking conditions. Journal of Consulting and Clinical Psychology , 53 , 393 /401. Zinn, M. L., McCain, C, & Zinn, M. A. (2000). Music Performance Anxiety and the High Risk Model of Threat Perception. Medical Problems of Performing Artists , 15 (2), 65 /71.