Development and validation of a music performance anxiety inventory for gifted adolescent musicians

Anxiety Disorders 19 (2005) 725–751

Development and validation of a music performance anxiety inventory for gifted adolescent musicians Margaret S. Osborne*, Dianna T. Kenny Australian Centre for Applied Research in Music Performance (ACARMP), Sydney Conservatorium of Music C41, The University of Sydney, NSW 2006, Australia Received 24 May 2004; received in revised form 3 August 2004; accepted 13 September 2004

Abstract Music performance anxiety (MPA) is a distressing experience for musicians of all ages, yet the empirical investigation of MPA in adolescents has received little attention to date. No measures specifically targeting MPA in adolescents have been empirically validated. This article presents findings of an initial study into the psychometric properties and validation of the Music Performance Anxiety Inventory for Adolescents (MPAI-A), a new self-report measure of MPA for this group. Data from 381 elite young musicians aged 12– 19 years was used to investigate the factor structure, internal reliability, construct and divergent validity of the MPAI-A. Cronbach’s a for the full measure was .91. Factor analysis identified three factors, which together accounted for 53% of the variance. Construct validity was demonstrated by significant positive relationships with social phobia (measured using the Social Phobia Anxiety Inventory [Beidel, D. C., Turner, S. M., & Morris, T. L. (1995). A new inventory to assess childhood social anxiety and phobia: The Social Phobia and Anxiety Inventory for Children. Psychological Assessment, 7(1), 73–79; Beidel, D. C., Turner, S. M., & Morris, T. L. (1998). Social Phobia and Anxiety Inventory for Children (SPAI-C). North Tonawanda, NY: Multi-Health Systems Inc.]) and trait anxiety (measured using the State Trait Anxiety Inventory [Spielberger, C. D. (1983). State-Trait Anxiety Inventory STAI (Form Y). Palo Alto, CA: Consulting Psychologists Press, Inc.]). The MPAI-A demonstrated convergent validity by a moderate to strong positive correlation with an adult measure of MPA. Discriminant validity was established by a weaker positive relationship with depression, and no relationship with externalizing behavior problems. It is hoped that the MPAI-A, as the first empirically validated measure * Corresponding author. Tel.: +61 2 9351 1386; fax: +61 2 9351 1385. E-mail address: [email protected] (M.S. Osborne). 0887-6185/$ – see front matter # 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.janxdis.2004.09.002

726

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

of adolescent musicians’ performance anxiety, will enhance and promote phenomenological and treatment research in this area. # 2004 Elsevier Inc. All rights reserved. Keywords: Music performance anxiety; Adolescents; Assessment; Psychometric properties

1. Introduction The developmental trajectory and phenomenology of music performance anxiety (MPA) in children and pre-tertiary adolescents has received little attention to date. A comprehensive database search identified only three relevant studies exploring MPA in children and early adolescents (Maroon, 2003; Ryan, 1998, 2003). All three studies are methodologically compromised because they used volunteer students with no history or diagnosis of MPA. However, a clear relationship emerged between MPA and standard anxiety measures, indicating that MPA may constitute a specific manifestation of generalized anxiety problems, and that performance was impaired in highly state anxious students. Most studies on MPA have been undertaken on professional musicians or young adult students studying music in tertiary institutions. These studies indicate that music performance anxiety is highly prevalent and problematic among this occupational group and that performance anxiety appears to be both a public health and occupational health issue for musicians. For example, up to 59% of professional musicians experience performance anxiety severe enough to affect their professional and/or personal lives (Van Kemenade, Van Son, & Van Heesch, 1995), and 70% reported that their playing is adversely affected by it (James, 1997). A recent study has found that opera singers have significantly higher trait anxiety than community samples, and this makes them more vulnerable to the occupational stressors associated with their profession (Kenny, Davis, & Oates, 2004). Evidence also suggests a strong relationship between MPA and social anxiety, one of the most common disorders in the community (Alfano, Beidel, & Turner, 2002; Lewinsohn, Gotlib, Lewinsohn, Steeley, & Allen, 1998). Many anxiety disorders first manifest in childhood and adolescence (Kashani & Orvaschel, 1990), but it is not known whether this is true for MPA. Further, prevalence in adolescents has not been assessed and there are no empirically robust, reliable and valid self-report measures of MPA for adolescent musicians. A number of theories have been developed to explain the onset of MPA in adult musicians. MPA is commonly viewed as a constellation of three interactive yet also partially independent factors: cognitions, autonomic arousal, and overt behavioral responses (Craske & Craig, 1984; Lederman, 1999; Salmon, 1990). There is some dispute amongst researchers regarding the nature of this interaction. Zinn, McCain, and Zinn (2000) argue that performance anxiety is primarily a psychophysiological event where the autonomic nervous system initiates and maintains MPA. Alternatively, Kirchner (2003) maintains the

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

727

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 performer’s constitutional and learned tendency (trait anxiety) to become anxious in response to situations of social stress; the degree of task mastery, and the amount of situational stress. 1.1. Critical review of MPA measures A review of the English-speaking literature using PsycINFO and MUSE databases identified 20 MPA self-report measures published in peer-reviewed journals (excluding Dissertation Abstracts). These are listed (with abbreviations) in Appendix A. All scales were developed for specific research projects with college- and/or adult-aged samples. Most measures assessed MPA across a variety of musical instruments and performance situations, although some were specifically created for pianists (e.g. Piano Performance Anxiety Scale) and string players (e.g. Stage Fright Rating Scale). All scales apart from the Music Performance Anxiety Questionnaire (MPAQ), the Performance Anxiety Selfstatement Scale (PASSS) and the State Emotion Questionnaire (SEQ) assessed MPA as trait anxiety, asking respondents to rate their levels of MPA based on a retrospective evaluation of music performance across various contexts. Levels of MPA occurring at particular points in time, such as music performances in the research context, were assessed using a state anxiety measure [commonly Spielberger’s State-Trait Anxiety Inventory-State subscale (Spielberger, 1983)]. Many of the available scales are adaptations of existing scales, assessing nonmusic performance anxiety. For example, Appel’s Personal Report of Confidence as a Performer (PRCP) was adapted from Paul’s (1966) Personal Report of Confidence as a Speaker; Cox and Kenardy’s (1993) Performance Anxiety Questionnaire (PAQ) was adapted from Schwartz, Davidson, and Goleman (1978) Cognitive–Somatic Anxiety Questionnaire (as cited in Cox & Kenardy, 1993) and the Performance Anxiety Inventory (PAI); the PAI was based on Spielberger’s Test Anxiety Inventory (1980, as cited in Nagle, Himle, & Papsdorf, 1989); and the Achievement Anxiety Test Scale (Alpert & Haber, 1960) was modified by both Sweeney and Horan (1982), and Wolfe (1989) into the Adaptive–Maladaptive Anxiety Scale (AAS–MAS). The most recent scale, the Kenny Music Performance Anxiety Inventory (K-MPAI) (Kenny et al., 2004) was constructed to specifically address each of the components of Barlow’s emotion-based theory of anxiety disorders (Barlow, 2000). Only the K-MPAI, PRCP and PAI assess all three components—cognitive, behavioral and physiological—that are now commonly believed to comprise MPA and other anxiety disorders (Barlow, 2002; Hardy & ¨ hman, 2000; Morris, 2001). Five of these scales Parfitt, 1991; Lang, Davis, & O were reproduced in full in the journals in which they were published to facilitate future research (K-MPAI; Musician’s Questionnaire; PAI; PAQ—Cox & Kenardy, 1993; and PAQ—Wesner, Noyes, & Davis, 1990). Factor analytic

728

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

studies have been reported on the MPAQ, AAS–MAS and Trait Anxiety Scale (TAS). A critical review of the psychometric data presented in the articles was conducted using criteria adapted for self-report measures from McCauley and Swisher’s (1984) review of language and articulation tests. These constitute the minimum criteria that should be considered in evaluating any norm-referenced test (Standards for educational and psychological testing, 1985). They include a description of the standardization population, sample sizes of 100 or more per subgroup, systematic item analysis (including factor analysis), presentation of raw score means and standard deviations for each subgroup, evidence of concurrent and predictive validity, estimates of test–retest reliabilities, test administration details, and qualifications of the administrator or scorer. Table 1 summarizes the criteria met by each MPA measure. The reporting of basic psychometric properties was limited, and in some cases absent. Very few of the tests reported test–retest or internal reliability. Overall, subject numbers in the various research studies were small (20–53). Although Wesner et al. (1990) had the largest sample size (n = 302) and described their sample thoroughly, they provided no results other than percentages of gender, age group and category of musician (i.e. impaired/unimpaired) who endorsed various symptoms of MPA. No single measure satisfied all the criteria, although the KMPAI met most of the criteria and reported the best psychometric properties. In sum, the MPA measures reviewed are idiosyncratic, with inadequate norms and standardization procedures and all have been developed for and used with adult musicians. Table 1 MPA self-report measures meeting each of the nine psychometric criterion

1 2 3 4 5

Criterion

Number of measures (n = 20)

Measures

Description of normative sample Sample size Item analysis

10

MPSS, PAQ-WND, PAQ-CK, PASSS, AAS–MAS, MPAS, TAS, SAS, PMCI PAQ-WND, MPAQ PRCP, PAQ-WND, K-MPAI, PI, MPAQ, PAI, SSQ, AAS–MAS, MPAS, TAS, SAS, PMCI PRCP, PAQ-WND, PAQ-CK, SES, PASSS, K-MPAI, PAI, AATS, PPAS, AD PAQ-WND, PAQ-CK, K-MPAI, PI, SEQ, MPAQ, SSQ, AATS, PPAS, AD, AAS–MAS, MPAS, TAS, SAS, PMCI – – PRCP, MPSS, PAQ-WND, PAQ-CK, SES, PASSS, K-MPAI, MPAQ, PI, SEQ, PAI, SSQ, AATS, PPAS, AD, AAS–MAS, MPAS –

Means and standard deviations Concurrent validity

6 7 8

Predictive validity Test–retest reliability Description of test procedures

9

Description of tester qualifications

2 13 10 16

0 0 18

0

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

729

Currently, there is no self-report measure of MPA for child or pre-tertiary adolescent musicians published in the public domain. Both child and adolescent MPA studies by Ryan (1998) and Maroon (2003) developed research-specific measures, but neither presented any data as to their psychometric properties. In order to address these major deficits in this research domain, we aimed to develop a reliable and valid self-report measure of MPA for the population of young musicians aged 12–19 years. We hope that this scale will facilitate MPA research in this age group through the provision of a brief, reliable and valid instrument that can serve as both a simple screening aid for high school students at risk for developing serious MPA, and as a pre–post intervention measure for treatment outcome studies.

2. Method Five analyses were undertaken to establish the psychometric properties of this scale. Initially, a pilot test was conducted to assess the content of items that were either newly developed or modifications of existing measures. We then administered all scales to a larger sample to determine the factor structure, internal consistency and criterion validity of the Music Performance Anxiety Inventory for Adolescents (MPAI-A). Subsequent concurrent and discriminant validity analyses were undertaken using an additional student sample. Because all studies were conducted using schools within the NSW Department of Education in Australia, a brief summary of the educational system is presented below. 2.1. Australian education setting In Australia, children commence school in kindergarten, aged between 4.5 and 6 years. There are 3 years of elementary school (Kindergarten to Year 2); 4 years of primary school (Years 3–6) and 6 years of secondary school (Years 7–12). Students may exit at Year 10 with their School Certificate or at Year 12 with their Higher School Certificate (HSC). The Conservatorium High School (CHS) is a unique selective government school for gifted young musicians. Successful applicants must have demonstrated musical potential and/or achievement, and have a satisfactory academic record. The Newtown High School of the Performing Arts (NHSPA) is a government school and one of the leading performing and visual arts schools in Australia. Students are considered for a placement to both high schools on the basis of an audition and interview. The McDonald College (MDC) is a private specialist academic and performing arts school for gifted children. Entry to the school is open to all students with a strong performance arts background, as well as students who have a general interest in the performing arts and wish to be educated in a creative environment. All schools teach grades 7–12. Both the MDC and NHSPA have students majoring in dance and theater, as well as music.

730

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

2.2. Scale development The initial item pool was generated by consulting available adult MPA scales. Items were chosen to represent each of the three domains (cognitive, physiological and behavioral) known to be affected in MPA. Care was taken with the wording of the items and all were tested for readability by 12-year-olds. The aim of the pilot study was to assess the readability and comprehension of each of the items on each proposed scale and to examine preliminary psychometric data. 2.2.1. Participants Protocols were completed in class time by 19 Year 8 and 18 Year 10 students (14 boys, 23 girls), with a mean age of 14 years (S.D. = 1.12 years) from the CHS. 2.2.2. Measures Two scales were developed. These were: 1. Demographics: age, gender, language spoken at home, principal and other instruments, length of time studied, time practised each day, desire to be a professional musician, importance of music in the family, age first performed in front of audience, pattern and frequency of performing. 2. Music Performance Anxiety Inventory for Adolescents (MPAI-A): This 18-item scale was designed for use with adolescents to assess the somatic, cognitive and behavioral components of MPA (after Cox & Kenardy, 1993; Nagle et al., 1989). Items such as ‘‘Before I perform, I get butterflies in my stomach’’ (somatic subscale), ‘‘I often worry about my ability to perform’’ (cognitive) and ‘‘I would rather play on my own than in front of other people’’ (behavioral) were answered on a seven-point Likert scale ranging ‘‘0—Not at all’’ to ‘‘6—All of the time’’.

2.2.3. Results 2.2.3.1. Demographic characteristics. Students had been studying for an average of 6.83 years (S.D. = 2.76), and practiced an average of 1.5 h a day (S.D. = 0.74 h). Forty-six percent of students indicated they wanted to be professional musicians, 49% were unsure, and 5% did not want to be a professional musician. Approximately half of the students (53%) performed mostly solo with some group work; 22% indicated half solo-half group performances; 11% performed mostly in ensembles with some solo work; 8% performed only as soloists; and 6% performed only in ensembles. 2.2.3.2. Scale characteristics. Descriptive statistics for the MPAI-A were mean = 50.77, S.D. = 17.14, minimum and maximum scores 8 and 20, respectively. Analysis revealed that improvements could be made to the reliability of the MPAI-A

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

731

with the deletion of some items. Cronbach’s a increased from .88 to .91 by deleting three items ‘‘It is easier to play in front of my family and friends, than in front of strangers’’, ‘‘It is easier to play in front of strangers, than in front of family and friends’’, and ‘‘I usually feel well prepared before I perform’’ (reduced scale given in Appendix B). Because of student confusion in completing the demographic scale questions on practice, this section was modified to request daily practice time on the major instrument only. 2.3. Psychometric properties 2.3.1. Participants A new sample of students (124 boys and 174 girls) from the three performance high schools in the inner Sydney metropolitan area described above agreed to participate in the research. The respective number of participating students at each high school were CHS n = 128, NHSPA n = 117, and MDC n = 53. The mean age of the sample was 14.23 years (S.D. = 1.70 years, range: 11–19 years). 2.3.2. Procedure The study was introduced to students by the head music teacher and music staff, and consent forms were distributed and returned in advance of survey administration. Early in Semester 1, 2002, the scales developed for the study and some standardized psychological assessment measures were distributed to students to complete in class. Students who did not wish to participate continued with their schoolwork for the duration of testing. Teaching staff were not given access to student survey results. Students then participated in a focus-group discussion regarding their experiences of MPA and management techniques. In addition, students completed up to four performance-based protocols throughout the semester, which assessed state anxiety experienced during different music performances. 2.3.3. Measures 1. Demographics: As described in 2.2.2 Scale Development. 2. Music Performance Anxiety Inventory for Adolescents (MPAI-A): As described in 2.2.2 Scale development. 3. State-Trait Anxiety Inventory-Trait (STAI) (Spielberger, 1983): The STAI Trait subscale measures relatively stable individual differences in the tendency to perceive stressful situations as dangerous or threatening. This standardised test was used to validate the MPAI-A and to provide a comparison with standardised norms and other MPA research. This measure has excellent stability with high school students (30 day test–retest interval males = .71, females = .75) and internal consistency (males and females = .90) (Spielberger, 1983). 4. Social Phobia and Anxiety Inventory (SPAI) (Turner, Beidel, & Dancu, 1996), and Social Phobia and Anxiety Inventory for Children (SPAI-C) (Beidel,

732

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

Turner, & Morris, 1998): These scales screen for maladaptive social anxiety in social encounters and settings, assess information on the intensity of distress, and indicate possible social phobia diagnoses. The SPAI-C successfully differentiates children with social phobia from those with other anxiety disorders (Beidel et al., 1998). Adolescents up to and including 14 years completed the SPAI-C, and those 14 years and older completed the SPAI. These measures were included to examine the relationship between MPA and social phobia. Both measures have very high stability and internal consistency (Beidel et al., 1998; Turner et al., 1996). 2.3.4. Results 2.3.4.1. Factor analysis. The factor structure of the MPAI-A was assessed. Prior to factor analysis, we assessed the factorability of the data. An anti-image correlation matrix was used to assess the sampling adequacy of each variable. No item had a measure of sampling accuracy <.5. Bartlett’s test of sphericity was large and significant and the Kaiser–Meyer–Olkin measure was >.6, hence factorability was assumed (Coakes & Steed, 2001). To examine whether the scale was multi- or uni-dimensional, we assessed its factor structure using principal axis factoring with oblimin rotation. Examination of both the number of eigenvalues greater than one and the factor loadings supported a decision to treat the scale as multi-dimensional, although the first factor accounted for most of the variance. Using a best-fit solution, the factor analysis identified three eigenvalues greater than one, which together accounted for 53% of the variance. This three factor solution provided a better fit to the data than either a one or two factor solution (details of factor analyses can be obtained from the authors). Individual items with factor loadings less than .3 were eliminated (Tabachnick & Fidell, 2001). The items and their specific factor loadings are presented in Table 2. The first factor, Somatic and Cognitive Features, accounted for 43% of the variance. The majority of items loading on this factor consisted of those describing the physical manifestations of performance anxiety immediately prior to, and during a performance. Two items related to worry and fear of making mistakes. The second factor, Performance Context, accounted for 6% of the variance, and described the preference performers have for either solo or group contexts and the nature of the audience. The third factor, Performance Evaluation, contained items relating to the evaluation that both the audience and performer may make of a performance, the consequences stemming from these evaluations (particularly when a mistake is made), and difficulty concentrating in front of an audience when performing. This factor accounted for 3% of the variance. 2.3.4.2. Reliability analysis. Cronbach’s a was used to determine the internal consistency of the MPAI-A. Item-total correlations were examined, and items with negative or item-total correlations below .3 were deleted to improve the

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

733

Table 2 MPAI-A factor analysis pattern matrix Item number

Item content

Loadings Factor 1

Somatic and cognitive features (43%) 1 Before I perform, I get butterflies in my stomach 12 Just before I perform, I feel nervous 6 When I perform in front of an audience, my heart beats very fast 4 Before I perform, I tremble or shake 2 I often worry about my ability to perform 9 When I perform in front of an audience I get sweaty hands 15 My muscles feel tense when I perform 5 When I perform in front of an audience, I am afraid of making mistakes

Factor 3

.93 .77 .77 .73 .61 .56 .44 .43

Performance context (6%) 14 I would rather play in a group or ensemble, than on my own 11 I try to avoid playing on my own at a school concert 3 I would rather play on my own, than in front of other people Performance evaluation (3%) 13 I worry that my parents or teacher might not like my performance 8 If I make a mistake during a performance, I usually panic 10 When I finish performing, I usually feel happy with my performance 7 When I perform in front of an audience, I find it hard to concentrate on my music

Factor 2

.38

1.02 .58 .36

.58 .54 .43 .33

.38

reliability of each scale (de Vaus, 1995). The ‘‘a if item deleted’’ column was also examined to decide on further item deletions. The original 18 item MPAI-A scale had high internal reliability (Cronbach’s a = .89). Inspection of the item-total correlations revealed three problematic items (correlations in parentheses). ‘‘It is easier to play in front of strangers, than in front of family and friends’’ ( .07), ‘‘It is easier to play in front of my family and friends, than in front of strangers’’ (.23) and ‘‘I usually feel well prepared before I perform’’ (.24). Deleting these items raised Cronbach’s a to .91. The remaining 15 items also provided the highest Cronbach’s a in the reliability analysis on the MPAI-A in the pilot test. Cronbach’s a values for the three factors determined in Study 2 were calculated as follows: Factor 1 ‘‘Somatic and Cognitve Sensations’’ = .90; Factor 2 ‘‘Performance Context’’ = .77; and Factor 3 ‘‘Performance Evaluation’’ = .69.

734

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

Table 3 MPAI-A descriptive statistics by gender and school grade Gender

Number

Mean

Standard deviation

Male Female

124 174

38.23 46.27

17.93 19.83

3 0

87 84

Grade 7 8 9 10 11 12

54 50 51 55 50 38

39.96 43.83 43.53 47.71 42.05 39.38

20.07 20.20 19.11 20.61 18.03 17.57

0 10 5 0 7 5

84 87 84 82 80 81

298

42.92

19.44

0

87

Total

Minimum

Maximum

Modification indices indicated that deleting items C4 ‘‘When I finish performing, I usually feel happy with my performance’’ (.32) and C5 ‘‘I worry that my parents or teacher might not like my performance’’ (.40) improved the a to .92, with item-total correlations for the reduced scale ranging from .56 to .74. However, considering the factor analysis results, the 15 item MPAI-A scale was retained for all future analyses, as this measure provided a desirable balance between desired psychometric properties and the broadest possible assessment of MPA features. The descriptive statistics for the reduced MPAI-A measure are given in Table 3. Girls scored significantly higher than boys [t(296) = 3.59, P < .001]. An inspection of the means by grade shows a curvilinear trend, with a rise from grade 8 that peaks at grade 10, thereafter declining to pre-grade 8 levels. 2.4. Criterion validity 2.4.1. Procedure In order to determine the construct validity of the MPAI-A, various wellknown, standardized self-report measures of trait and social anxiety were administered. The STAI was used to determine baseline levels of trait anxiety, and depending on the age of students, the SPAI-C or SPAI was used as a measure of social anxiety. Given previous research into both childhood (Maroon, 2003; Ryan, 1998) and adult MPA (Kenny et al., 2004; Lehrer, Goldman, & Strommen, 1990), the MPAIA was hypothesized to demonstrate a moderate to strong positive correlation with trait anxiety. Further, since a diagnosis of social phobia is warranted for those musicians who suffer extreme performance anxiety (Clark & Agras, 1991; Cox & Kenardy, 1993) and hence MPA as a construct is more specifically related to social anxiety than the more general trait anxiety, it was hypothesized that there would be a stronger positive correlation between the MPAI-A and social phobia measures than for trait anxiety.

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

735

Table 4 Means and standard deviations for the STAI, SPAI-C and SPAI by gender (norms in italics) Scale

Boys

Girls

Total sample

Mean

S.D.

Mean

S.D.

Mean

S.D.

STAI Norm data

41.70 40.17

10.05 10.53

42.52 40.97

10.06 10.63

42.18 –

10.05 –

SPAI-C Whole sample Social Phobia present Social Phobia absent

10.12 14.9 23.4 10.2

10.55 9.6 7.6 7.6

10.56 18.8 26.8 15.1

7.86 10.4 9.8 8.4

10.39 16.8 26.1 –

8.99 10.2 8.6 –

SPAI Social Phobia present Social Phobia absent

48.48 – –

24.69 – –

41.87 – –

28.92 – –

44.60 64.7 43.0

27.35 30.6 26.4

2.4.2. Results Table 4 displays the means and standard deviations for males, females and the whole sample, and compares these values against normative data presented in the test manuals. Distributions for each variable were examined and no significant deviations from normality were detected. Norms for the STAI were taken from Spielberger (1983). Normative data for the SPAI-C is quoted in the manual from research by Beidel, Turner, Hamlin, and Morris (2000), on a sample of 249 children both with (n = 63) and without social phobia diagnoses, with a mean age of 10.8 years (S.D. = 1.5 years), ranging from 8 to 14 years. Data for the SPAI is provided in the manual from research by Clark et al. (1994), who have provided the only reliability and validity data on the SPAI for adolescents. Their sample of 223 adolescents ranged from 12 to 18 years (no average age was reported), 39 of which met criteria for DSM-III-R Social Phobia. Normative values are given in italics. There were no significant differences between male and female mean scores on the STAI [t(296) = 0.70, P > .05]. However, girls in this sample scored significantly higher than girls in the norm group [t(173) = 2.04, P < .05], but there was no difference for boys in this sample and the normative group [t(123) = 1.70, P > .05]. The scores on the SPAI-C for this study were much lower than those given by Beidel et al. (2000) for the whole sample. The SPAI data shows the total sample mean is comparable to the mean value reported for adolescents with no psychiatric disorder. Cutoffs for the SPAI-C and use of SPAI with adolescents have not been specified. Considering that both measures were administered in order to capture information across the entire high school age range, various investigations of the data using different screening cutoff scores were undertaken. Using a cutoff of 18 on the SPAI-C, and including the top two categories of difference scores on the SPAI, there is reasonable consistency in the percentage of possible social phobia

736

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

Table 5 Number of students (percentage in parentheses) meeting screening criteria for social phobia Protocol

Unlikely Social Phobia

SPAI-C Cutoff  18 Cutoff  20 Cutoff  15 Cutoff  13 Cutoff  10

115 123 102 94 81

SPAI

Combined SPAI-C and SPAI

(79%) (85%) (70%) (65%) (56%)

37 (28%)

152 (55%)

Possible Social Phobia 30 22 43 51 64

(21%) (15%) (30%) (35%) (44%)

94 (72%) Mild possible = 65 (50%) Possible = 17 (13%) Probable = 12 (9%) 124 (45%)

cases in the sample across both measures (see Table 5). In addition, given that Clark et al. (1994) found lower scores in adolescents compared to adults using the SPAI, rates of possible and probable social phobia were investigated taking into account all three social phobia categories (mild possible, possible and probable social phobia), thereby lowering the cutoff score for possible social phobia diagnosis. This second method of analysis increases the number of screened social phobics, leading to an almost inverse relationship between unlikely and likely presentations of social phobia across the SPAI-C and SPAI. The mean value for this sample on the SPAI-C was significantly lower than the whole norm sample [t(144) = 8.58, P < .01], and the social phobia group [t(144) = 21.03, P < .01]. The mean scores for girls and boys were not significantly different [t(143) = 0.29, P > .05]. Considering that Beidel et al. (2000) found significant differences between boys and girls in the norm samples for the SPAI-C, separate analyses were conduced by gender. There was no difference between boys in this sample and the SPAI-C social phobia-absent group [t(57) = 0.06, P > .05]. However, boys in this study scored significantly lower than both the whole norm sample [t(57) = 3.45, P < .001] and boys diagnosed with social phobia [t(57) = 9.59, P < .01]. The girls scored significantly lower than the social phobia-absent [t(86) = 5.38, P < .01], whole sample [t(86) = 9.77, P < .01] and social phobia-present [t(86) = 19.26, P < .01] norm groups. Similarly, there was no difference between boys and girls on the SPAI [t(129) = 1.37, P > .05], which justified whole group comparisons to the norm data. The total sample mean was significantly lower than the socially phobic diagnostic group [t(130) = 8.41, P < .001], but not significantly different to adolescents without a psychiatric disorder [t(130) = 0.67, P > .05]. The MPAI-A correlated significantly with the STAI (r = .45, P < .01), SPAI-C (r = .42, P < .01) and SPAI (r = .43, P < .01) scales (see Table 6).

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

737

Table 6 Correlations between MPA, trait and social anxiety (r and n values respectively) STAI MPAI-A

.45* 298

SPAI-C

145 SPAI-C *

.43* 131

.47*

.46* 131

145

STAI

SPAI

.42*

–

All correlations significant at P < 0.001.

2.5. Convergent and discriminant validity In light of the unconvincing correlations between the MPAI-A and the STAI, SPAI-C and SPAI, a follow-up study was conducted to clarify the extent to which the MPAI-A is associated with trait anxiety and social phobia. An adult measure of MPA was included to assess concurrent validity. Since the current sample were gifted students, we believed that none would experience any difficulty with understanding an adult version, and indeed we found that no student reported difficulty. We wished to assess the correlation of the MPAI-A with another MPA measure in order to further assess the MPAI-A’s construct validity. The Children’s Depression Inventory (Kovacs, 1992) and the externalizing profile of the Youth Self-Report (Achenbach, 1991) were included to investigate the discriminant validity of the MPAI-A. 2.5.1. Participants Music students, of whom 61% formed part of the criterion validity sample (n = 64; 30 boys and 34 girls) from the NHSPA in Years 8 (Elective and Extension), 9 (Extension), 10 (Elective) and 12 (Music 2) were invited to participate. The mean age of this sample was 14.6 years (S.D. = 1.3, range: 13–17 years). 2.5.2. Procedure Students who were not involved in the previous validity sample were given consent forms, and asked to obtain their parents signature as consent, and return to the liaison music teacher. Surveys were completed in class time, with the music teacher and one of the researchers present. Teaching staff did not have access to individual student results. 2.5.3. Measures 1. Kenny Music Performance Anxiety Inventory (K-MPAI) (Kenny et al., 2004): Because no other child or adolescent MPA measures exist, this scalewas included to test the convergent validity of the MPAI-A. This inventory was developed to assess the emotion-based theory of anxiety proposed by Barlow (2000).

738

2.

3. 4. 5.

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

Items were either specially constructed or selected from other scales to address each of Barlow’s theoretical components, including evocation of anxious propositions (e.g. uncontrollability, unpredictability, negative affect, situational cues); attentional shift (e.g. task or self-evaluative focus, fear of negative evaluation); physiological arousal and memory bias. Questions are answered on a 7-point Likert scale ranging from ‘‘0—Strongly disagree’’ to ‘‘6— Strongly agree’’. The maximum score is 156, with higher scores indicating greater anxiety and psychological distress. This scale demonstrated excellent internal reliability (Cronbach’s a = .94) (Kenny et al., 2004). Youth Self-Report-Externalizing Scale (YSR-Ext) (Achenbach, 1991): The YSR is a self-report form used to assess the feelings and behaviors of 11–18 year olds. Thirty questions included in the externalizing scale of the YSR (comprising Delinquent and Aggressive Behavior factors) were used to assess discriminant validity. These items are answered on a three-point Likert scale of 0—Not true to 2—Very true or often true. The Internalizing scale of the YSR was not included due to administration time constraints. However, the STAITrait and CDI scales were considered adequate in assessing internalizing characteristics such as worry and depression. The externalizing scale has acceptable internal reliability (Cronbach’s a = .89) and 1-week test–retest reliability (r = .81). The scale significantly differentiates the competence and problem scores of non-referred and referred youth presenting for mental health services, and thus demonstrates acceptable validity (Achenbach, 1991). Social Phobia and Anxiety Inventory (SPAI) (Turner et al., 1996), and Social Phobia and Anxiety Inventory for Children (SPAI-C) (Beidel et al., 1998). State-Trait Anxiety Inventory-Trait (STAI) (Spielberger, 1983). Children’s Depression Inventory (CDI) (Kovacs, 1992): The CDI was included to further assess the construct and discriminant validity of the MPAI-A. The CDI is a 27-item self-rated symptom oriented scale measuring symptoms associated with depression such as sleep disturbance, appetite loss, and anhedonia. It is suitable for school-aged adolescents aged 7–17 years. It quantifies a range of depressive symptoms including disturbed mood, hedonic capacity, vegetative functions, self-evaluation, and interpersonal behaviors, providing a total score (CDI-Total), and five subscale scores: Negative Mood (CDI-NM), Interpersonal Problems (CDI-IP), Ineffectiveness (CDI-IF), Anhedonia (CDI-AN), and Negative Self-Esteem (CDI-NS). For each item, the child rates the degree to which each statement describes him or her for the past 2 weeks, using one of three choices, keyed 0 (Absence of symptom), 1 (Mild symptom), or 2 (Definite symptom). Higher scores indicate increasing severity, with scores ranging from 0 to 54. It has been used extensively with normal and clinical children’s populations. The measure effectively distinguishes normal from clinical samples, with improved accuracy obtained when included as part of a multiinstrument battery. It shows good concurrent validity with other measures of depressive symptoms and demonstrates good internal consistency (r = .71–.89) and test–retest reliability (r = .38–.87) (Kovacs, 1992).

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

739

Table 7 MPAI-A descriptive statistics by gender and school grade Gender

Number

Mean

Standard deviation

Male Female

30 34

29.43 38.24

13.28 15.21

7 5

57 72

Grade 8 9 10 12

27 13 17 7

35.04 29.31 35.71 35.57

14.32 17.19 15.36 12.58

5 7 9 20

69 61 72 53

64

34.11

14.90

5

72

Total

Minimum

Maximum

2.5.4. Results Descriptive statistics for the MPAI-A are presented in Table 7. Distributions for each variable were examined and no significant deviations from normality were detected. The reduced 15-item scale again demonstrated acceptable internal reliability for this sample (Cronbach’s a = .88). This sample evidenced much lower scores on the MPAI-A than the criterion validity sample. As in the previous sample, girls scored significantly higher than boys [t(62) = 2.45, P < .05]. Similarly, scores in this sample were lower on the social anxiety measures (see Table 8). The mean value obtained for boys on the SPAI-C was half that obtained in the criterion validity sample (see Table 4), although the gender difference in this Table 8 Means and standard deviations for all scales by gender (norms in italics) Scale

Males Mean

S.D.

Mean

S.D.

Mean

S.D.

STAI SPAI-C SPAI K-MPAI

39.20 4.94 30.79 34.10

9.74 3.45 17.38 19.69

43.56 9.88 40.06 50.74

11.73 10.41 29.90 26.19

41.52 7.41 35.87 42.94

10.98 8.03 25.09 24.65

YSR-externalising

13.93 11.6

8.92 7.0

14.12 10.3

8.05 6.3

14.03 –

8.40 –

10.36 11.36 1.86 2.41 1.03 1.02 2.14 2.35 3.17 3.60 1.36 1.98

10.05 – 1.86 – 1.67 – 2.00 – 3.31 – 2.08 –

13.18 9.74 2.62 2.35 0.79 0.64 2.53 1.66 3.60 3.34 2.97 1.75

9.24 – 2.12 – 1.12 – 1.69 – 2.77 – 2.38 –

11.86 9.18 2.26 2.38 0.90 0.81 2.35 1.98 3.40 3.46 2.22 1.86

9.66 – 2.02

CDITotal Negative Mood Interpersonal Problems Ineffectiveness Anhedonia Negative Self-Esteem

Females

Total sample

1.40 – 1.84 – 3.02 – 2.37 –

740

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

sample was not significant [t(30) = 1.80, P > .05]. The female mean was consistent on this measure across both samples, for both the SPAI-C and SPAI. However, the boys’ scores on the SPAI were highly discrepant between samples. In this sample, on average, boys scored much lower than girls, although the difference was not significant [t(29) = 1.03, P > .05], whereas in the criterion validity sample, boys scored significantly higher overall. Their mean score is almost 18 points higher than the current score. The STAI results are reasonably consistent across both samples for boys and girls, who did not score significantly differently from each other in this sample [t(62) = 1.60, P > .05]. Considering that this convergent and discriminant validity analysis was conducted using NHSPA students only, it was conceivable that these differences could be attributed to sample differences. Therefore, criterion validity sample results for NHSPA only were compared with the NHSPA data to determine whether values in this sample were consistent for NHSPA students. Means and standard deviations for the SPAI-C in the criterion validity sample were as follows: boys = 7.97 (S.D. = 9.96), girls = 9.97 (S.D. = 7.27), total sample = 8.93 (S.D. = 8.73). SPAI means and standard deviations for boys = 49.74 (S.D. = 21.45), girls = 43.95 (S.D. = 21.23), and total sample = 46.70 (S.D. = 21.26). Thus, NHSPA students obtained mean scores comparable to all three high schools combined. Therefore, differences in this sample cannot be attributed to differences in the three high school samples. Boys scored lower than girls on the K-MPAI [t(62) = 2.84, P < .01], and obtained much lower average scores than the girls on the CDI-Negative SelfEsteem subscale [t(62) = 2.86, P < .01]. Norms for the STAI, SPAI-C and SPAI are presented in Table 4. Norms for non-referred youth were used to compare the YSR-Externalizing scale (Achenbach, 1991). Means for the CDI are given for the age group 13–17 as presented in the test manual (Kovacs, 1992). The percentage of this sample that met criteria for possible or probable social phobia (Table 9) was much lower than the percentage of students participating in the criterion validity sample (Table 5). Table 9 Number of students (percentage in parentheses) meeting screening criteria for social phobia Protocol

Unlikely Social Phobia

SPAI-C Cutoff  18 Cutoff  20 Cutoff  15 Cutoff  13 Cutoff  10

30 30 30 27 26

SPAI

14 (45%)

17 (45%) Mild possible = 10 (32%) Possible = 6 (19%) Probable = 1 (3%)

Combined SPAI-C and SPAI

44 (70%)

19 (30%)

(94%) (94%) (94%) (84%) (81%)

Possible Social Phobia 2 2 2 5 6

(6%) (6%) (6%) (16%) (19%)

Table 10 Correlations between MPA, trait, social anxiety, externalising and depression scales (r, n and P values, respectively) K-MPAI MPAI-A

STAI .54 64 <.001 .81 64 <.001

STAI

SPAI-C

SPAI-C

SPAI

.65 32 <.001

.63 31 <.001

.59 32 <.001

.73 31 <.001

.55 32 <.01

.69 31 <.001

YSR-Ext

CDI-Total

CDI-NM

.39 64 <.01

.48 64 <.001

64

64 <.001

.66 64 <.001

.62 64 <.001

.40 64 <.01

.85 64 <.001

.128 64 .313 .44

–

.11 32 .54

SPAI

.08 31 .67

YSR-Ext

CDI-Total

CDI-NM

CDI-IP

CDI-IF

CDI-IF

CDI-AN

CDI-NS

.02

.21

.34

64 .09

64 <.05

.50 64 <.001

.26 64 <.05

.51 64 <.001

.60 64 <.001

.70 64 <.001

.80 64 <.001

.47 64 <.001

.63 64 <.001

.80 64 <.001

.75 64 <.001

.35 32 <.05

.49 32 <.01

.20

32 .84

.27

.35 32 <.05

.36 32 <.05

.57 31 <.01

.50 31 <.01

.15 31 .43

.30 31 .10

.55 31 <.01

.67 31 <.001

.31 64 <.05

.23 64 .07

.23 64 .07

.28 64 <.05

.31 64 <.05

.18 64 .16

.90 64 <.001

.73 64 <.001

.78 64 <.001

.91 64 <.001

.89 64 <.001

.54 64 <.001

.62 64 <.001

.78 64 <.001

.81 64 <.001

.53 64 <.001

.60 64 <.001

.60 64 <.001

.60 64 <.001

.63 64 <.001

.874

.04 32

.71 64 <.001

741

CDI-AN

CDI-IP

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

K-MPAI

.69 64 <.001

742

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

Pearson’s correlations between the MPAI-A and STAI, SPAI-C and SPAI were all slightly higher than obtained in criterion validity sample and all highly significant (see Table 10). As hypothesized, correlations with the social phobia measures were greater than with the STAI. The MPAI-A correlations with the social phobia measures showed that these measures share approximately 40% of their variance. The concurrent validity of the MPAI-A was established by a higher correlation with the K-MPAI than any other measure. Discriminant validity was established in two ways; firstly, by the very low and non-significant correlation between the MPAI-A and the YSR-Externalizing scale, and secondly by the weak correlations between the MPAI-A and the CDI-Total, Negative Mood and Negative Self-Esteem subscales.

3. Discussion The aim of this study was to develop and assess the psychometric properties of a new scale of music performance anxiety for use with child and adolescent musicians. To be considered a reliable and valid measure, a scale should demonstrate good internal consistency, with an a of at least .9 when total test scores are used in placement decisions (Salvia & Ysseldyke, 1988) and good construct, concurrent and discriminant validity. Although validity coefficients in the order of .3–.4 are considered high, and values >.6 are rare (Kaplan & Saccuzzo, 1989) we considered a coefficient greater than or equal to .8 to be a more stringent and appropriate criteria. Inter-correlations among items (the internal consistency) may be used to support the assertion that a scale primarily measures a single construct. Substantial relationships of a scale to other measures that purportedly assess the same construct, and weaknesses of relationships to measures of different constructs, support both the identification of constructs and the differentiation (discrimination) between them (Standards for educational and psychological testing, 1985). The MPAI-A displayed a parsimonious three-factor structure of Somatic and Cognitive Features, Performance Context and Performance Evaluation, which together accounted for 53% of the variance. The majority of the loading (43%) was attributed to the first factor, Somatic and Cognitive Features. Interestingly there are some parallels between these factors and those of the MPAQ, particularly between Somatic and Cognitive Features and the MPAQ Factor 2 High standards/Judgmental attitudes toward performance, and Performance Evaluation with the MPAQ Factor 5: Concern about distraction in self and audience. These results are currently being cross validated using a new sample of young American band musicians in Grades 6 and 7. Reliability analyses demonstrate that the MPAI-A is a psychometrically robust measure, with very good internal consistency (.91). The 15-item MPAI-A achieved this same high level of internal reliability in the scale development and

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

743

reliability samples, although it did reduce slightly to .88 in the concurrent and discriminant validity sample. Construct validity was demonstrated by the MPAI-A correlating most highly with the adult measure of MPA, the K-MPAI, the least with the YSR, and obtaining moderate and highly significant correlations with the STAI, SPAI-C and SPAI. The MPAI-A shared between 18% and 44% of the variance with the two measures of social phobia. These findings are at odds with previous research that compared MPA and social anxiety. Huston (2001) found a correlation in the order of .35 (12% of variance shared) and concluded that MPA and social phobia were relatively unrelated. This difference may be attributable to sample differences. Huston’s sample consisted of an older sample of predominantly professional musicians with much performance experience, where number of years playing an instrument was associated with lower rates of MPA. Slightly weaker yet significant correlations were obtained with the CDITotal and CDI-Negative Self-Esteem scales, which is consistent with the frequent comorbidity between anxiety, particularly social anxiety, and depressive disorders (Essau, Conradt, & Petermann, 1999; Kerr, 2003). Interestingly, this research replicated Ryan’s (1998) correlation between the STAI and Negative Self-esteem in 12-year-old piano students, a finding that is consistent across test and sports performance anxiety research, but equivocal in the MPA field. McCoy (1999) did not find a significant relationship between the two constructs, yet Sinden (1999) found self-esteem to significantly predict MPA. The significant weak to moderate correlations between the MPAI-A and the CDI can be explained by virtue of anxiety and depressive disorders both factoring on the internalizing syndrome of children’s behavioral and emotional problems (Achenbach, 1991). Further, as measures of anxiety and depression are highly intercorrelated in adults, adolescents, researchers have suggested there is a general negative affectivity component that is common to both anxiety and depression (Clark & Watson, 1991; Lonigan, Carey, & Finch, 1994). 3.1. Rates of social phobia The diagnosis of social phobia in this research was impeded not only by the inconclusive screening cutoffs prescribed by the instruments themselves but also by the inconsistent population social phobia morbidity rates found in other research using different assessment tools (Essau et al., 1999; Kashani & Orvashel, 1990; Kerr, 2003). SPAI-C normative results for females across the whole sample suggests that a large proportion of females met criteria for social phobia, which calls into question the validity of using 18 as a suggested cutoff, although the social phobia absent group scored lower than the cutoff overall (Beidel et al., 2000). The authors of the SPAI-C state that the development of separate

744

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

cutoff scores based on gender may be necessary for general populations, if this difference is replicated with other samples. What these figures do emphasise is the somewhat arbitrary nature of the normative data provided for this test. The range of SPAI-C cutoff scores provided in the manual allows users to determine their own cutoff according to the percentage of falsenegatives to false-positives the user is prepared to accept. Unfortunately, the authors have determined these initial normative data using separate research (Beidel, Turner, & Morris, 1995), with a sample of 52 socially anxious and 48 normal control children, without presenting means and standard deviations for the two groups either in the manual or other published literature, thus making normative comparisons difficult. However, for the purposes of screening and determining the potential presence or absence of social phobia, a cutoff score of 18 was used as suggested in the test manual (Beidel et al., 1998). The authors of the SPAI also fail to provide screening cutoff scores for adolescents. The test authors observe from the research by Clark et al. (1994) that the mean score for the socially phobic adolescents on the social phobia and agoraphobia subscales is much lower than that for adults and closer to the figures reported by the socially anxious (but not phobic) college group. They postulate that it may be attributable to adolescents having fewer opportunities to engage in some of the activities listed on the SPAI because they are not as relevant for this age group. Thus, they advise that ‘‘when used with adolescents as a screening device, adjustments to the recommended cutoff need to be made’’ (Turner et al., 1996, p. 32). The APA (1995) reports lifetime presentation rates of social phobia in epidemiological and community studies as between 3 and 13%. Studies specifically assessing the 12–17-year-age group using diagnostic interviews place the social phobia rate lower, between 1.4 and 1.6% (Essau et al., 1999; Kashani & Orvashel, 1990). The prevalence of social anxiety symptoms is much higher, ranging between 46 and 56% for ‘‘fear of social situations’’, and 67% for ‘‘worry what others think of me’’ (Kashani & Orvaschel, 1990). Fear of doing something in front of other people was the most frequent symptom of social phobia, followed by public speaking and/or talking with others (social talk) (Essau et al., 1999). The resonance and similarity of these types of social fears with the context of performing music before an audience is striking. Indeed, studies of social phobia in college and adult musicians with high MPA place diagnosis rates higher, between 27% (Osborne, 1998) and 95% (Clark & Agras, 1991). Further evidence for the arbitrary and inconclusive cutoffs in the SPAI-C and SPAI can be seen in the almost inverse relationships between unlikely and combined likely probabilites of social phobia across the two measures. Investigations using the recommended cutoff score for the SPAI-C in the criterion validity sample found that 79% of 11–14-year-olds assessed using the SPAI-C were unlikely to have social phobia, which reduces dramatically to 28%

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

745

for 15–19-year-olds assessed using the SPAI. Developmentally, this suggests that rates of social phobia increase as people progress through adolescence, which is consistent with previous research indicating social phobia onset in mid-to-late adolescence (APA, 1995; Liebowitz et al., 1985; Turner et al., 1986). However, considering the aforementioned psychometric problems with these tests, this result needs to be treated cautiously. The percentage of this sample who met criteria for possible or probable social phobia in the convergent and discriminant validity sample was much lower than the percentage in the criterion validity sample. These discrepancies can be at least partially explained by the difference in the time of year that the studies were conducted. The criterion validity study was conducted during the middle of Term 1 in the new school year, at the beginning of class assessments and concert preparation. The convergent and discriminant validity study was conducted at the end of the school year, in the week after the last major music and other academic assessments, and in the third last week before the end of school year and summer break. Students were less concerned about academic and social pressures at this time than they had been at the start of the school year, and this may have been reflected in their second group of scores. Consistent with research by Bandura (1991), this may be a case where very small changes in context or task-expectancy may change a person’s appraisal of a situation as anxiety-provoking or not. Although there are inconsistencies in the screening procedures for social phobia using the SPAI-C and SPAI, the various investigations of unlikely-possible social phobia percentages (particularly in Table 5) provide evidence for substantially higher rates of social phobia in this sample of elite adolescent performers than found in general adolescent epidemiological research. Considering that Dadds et al. (1999) found that up to 54% of anxious children can progress to a formal anxiety disorder in adulthood if left untreated, these figures indicate a need for further research into valid and reliable assessments of prevalence, prevention and early intervention of clinical levels of anxiety in this population. 3.2. Gender differences on measures of trait and social anxiety Gender differences were apparent on the three normed measures. Girls demonstrated significantly higher trait anxiety than the norm group but boys’ average scores were not significantly different. Prior research has found that girls consistently present with more anxiety diagnoses and social anxiety fears than boys, often in the ratio 2:1 (Essau et al., 1999; Ford, Goodman, & Meltzer, 2003; Kashani, Orvaschel, Rosenberg, & Reid, 1989; Lewinsohn et al., 1998). The current results are at odds with theses findings. In criterion validity sample girls obtained an equivalent mean score to boys on the SPAI-C, and scored lower (though not significantly) than boys on the SPAI. In the convergent and discriminant validity sample, however, girls scored higher on

746

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

average than boys on both the SPAI-C and SPAI, although the differences were not significant. 3.3. Gender differences in MPA Levels of MPA are generally higher in women than men (Huston, 2001; Steiner, 1998), although Ryan (2003) found a differential pattern of anxious responding across boys and girls. Girls had substantially higher heart rates than boys immediately prior to but not during performance. Boys had significantly more anxious behaviors than girls both prior to, and during a performance. Social/interpersonal anxiety is common for boys and girls as peers become the increasing focus of attention (Kashani et al., 1989). Girls but not boys, become increasingly concerned with age (Kashani et al., 1989). Girls also have significantly greater fears of performing in front of others (Essau et al., 1999). Consistent with these findings, girls obtained significantly higher MPAI-A scores than boys. 3.4. Conclusion This paper presents data on a new instrument for assessing MPA in adolescent musicians. The MPAI-A is a 15-item self-report measure that assesses the somatic, cognitive and behavioral components of anxiety experienced by adolescent musicians. Although further research is needed to confirm its discriminant, external and predictive validity, these initial studies indicate that the MPAI-A is psychometrically robust according to current accepted standards (Salvia & Ysseldyke, 1988). Consistent with previous research, there was a moderate relationship between MPA and social phobia. Similarly, there was a modest positive relationship between high in trait anxiety music performance anxiety. There was a strong situational effect in the self-reporting of subjective experience of anxiety according to the time of school year and related performance demands. Situational factors appear to impact on the degree to which young musicians report performance anxiety, and these deserve further research attention. Additional questions regarding the basis of identified gender differences and the developmental trajectory of performance anxiety from late childhood through adolescence need further investigation to determine if possible protective factors against MPA can be identified. The MPAI-A will be useful to pedagogues, researchers and clinicians in the assessment and/or treatment of MPA in young performers.

Acknowledgment We wish to thank Associate Professor Ross G. Menzies for his advice on the design of the convergent and discriminant validity study.

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

Appendix A

List of MPA measures

Abbreviation Measure

Citation

AAS–MAS

Wolfe (1989)

AATS AD K-MPAI MPAS MPAQ MPSS MQ PAI PASSS PAQ-CK PAQ-WND PI PPAS PRCP SES SEQ SFRS SSQ TAS

747

Adaptive–Maladaptive Anxiety Scale Achievement Anxiety Test Scale Anxiety Differential Kenny Music Performance Anxiety Inventory Music Performance Anxiety Scale Music Performance Anxiety Questionnaire Music Performance Stress Survey Musician’s Questionnaire Performance Anxiety Inventory Performance Anxiety Self-statement Scale Performance Anxiety Questionnaire-CK Performance Anxiety Questionnaire-WND Performance Inventory Piano Performance Anxiety Scale Personal Report of Confidence as a Performer Self-efficacy Scale State Emotion Questionnaire Stage Fright Rating Scale Self-statement Questionnaire Trait Anxiety Scale

Sweeney and Horan (1982) Sweeney and Horan (1982) Kenny (2004) Wolfe (1989) Lehrer et al. (1990) Brodsky, Sloboda, and Waterman (1994) Wills and Cooper (1988) Nagle et al. (1989) Kendrick, Craig, Lawson, and Davidson (1982) Cox and Kenardy (1993) Brodsky et al. (1994); Wesner et al. (1990) Kubzansky and Stewart (1999) Sweeney and Horan (1982) Appel (1976) Craske and Craig (1984) Kubzansky and Stewart (1999) Neftel et al. (1982) Steptoe and Fidler (1987) Wolfe (1990)

748

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

Appendix B

What I think about music and performing (MPAI-A).

Please think about music in general and your major instrument and answer the questions by circling the number, which describes how you feel.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Before I perform, I get butterflies in my stomach I often worry about my ability to perform I would rather play on my own, than in front of other people Before I perform, I tremble or shake When I perform in front of an audience, I am afraid of making mistakes When I perform in front of an audience, my heart beats very fast When I perform in front of an audience, I find it hard to concentrate on my music If I make a mistake during a performance, I usually panic When I perform in front of an audience I get sweaty hands When I finish performing, I usually feel happy with my performance I try to avoid playing on my own at a school concert Just before I perform, I feel nervous I worry that my parents or teacher might not like my performance I would rather play in a group or ensemble, than on my own My muscles feel tense when I perform

Not at all

About half the time

All of the time

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0 0

1 1

2 2

3 3

4 4

5 5

6 6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

0 0

1 1

2 2

3 3

4 4

5 5

6 6

0

1

2

3

4

5

6

0

1

2

3

4

5

6

#

M. S. Osborne and D. T. Kenny (2001) .

References Achenbach, T.M. (1991). Manual for the youth self-report and 1991 profile. Burlington, VT: University of Vermont Department of Psychiatry. Alfano, C. A., Beidel, D. C., & Turner, S. M. (2002). Cognition in childhood anxiety: conceptual, methodological, and developmental issues. Clinical Psychology Review, 22, 1209–1238.

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

749

Alpert, R., & Haber, R. N. (1960). Anxiety in academic achievement situations. Journal of Abnormal Social Psychology., 61, 207–215. American Psychiatric Association. (1995). Diagnostic and statistical manual of mental disorders international version (4th ed). Washington, DC: American Psychological Association. Appel, S.S. (1976). Modifying solo performance anxiety in adult pianists. Journal of Music Therapy, 13(1), 2–16. Bandura, A. (1991). Self-efficacy conception of anxiety. In: R. Schwarzer & R. Wicklund (Eds.), Anxiety and self-focused attention. New York: Harwood Academic Publishers. Barlow, D.H. (2000). Unravelling the mysteries of anxiety and its disorders from the perspective of emotion theory. American Psychologist, 55(11), 1247–1263. Barlow, D.H. (2002). Anxiety and its disorders. The nature and treatment of anxiety and panic. (2nd ed.). New York, NY: The Guilford Press. Beidel, D. C., Turner, S. M., Hamlin, K., & Morris, T. L. (2000). The Social Phobia and Anxiety Inventory for Children (SPAI-C): external and discriminative validity. Behavior Therapy, 31(1), 75–87. Beidel, D. C., Turner, S. M., & Morris, T. L. (1995). A new inventory to assess childhood social anxiety and phobia: the Social Phobia and Anxiety Inventory for Children. Psychological Assessment, 7(1), 73–79. Beidel, D. C., Turner, S. M., & Morris, T. L. (1998). Social Phobia and Anxiety Inventory for Children (SPAI-C). North Tonawanda, NY: Multi-Health Systems Inc. Bracken, B.A. (1987). Limitations of preschool instruments and standards for minimal levels of technical adequacy. Journal of Psychoeducational Assessment, 4, 313–326. Brodsky, W., Sloboda, J. A., & Waterman, M. G. (1994). An exploratory investigation into auditory style as a correlate and predictor of music performance anxiety. Medical Problems of Performing Artists, 9, 101–112. Clark, D. B., & Agras, W. S. (1991). The assessment and treatment of performance anxiety in musicians. American Journal of Psychiatry, 148(5), 598–605. Clark, D. B., Turner, S. M., Beidel, D. C., Donovan, J. E., Kirisci, L., & Jacob, R. G. (1994). Reliability and validity of the Social Phobia and Anxiety Inventory for adolescents. Psychological Assessment, 6(2), 135–140. Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100, 316–336. Coakes, S. J., & Steed, L. G. (2001). SPSS: analysis without anguish: version 10.1 for Windows. Milton, Qld: John Wiley. Cox, B. J., & Kenardy, J. (1993). Performance anxiety, social phobia, and setting effects in instrumental music students. Journal of Anxiety Disorders, 7, 49–60. Craske, M. G., & Craig, K. D. (1984). Musical performance anxiety: the three-systems model and selfefficacy theory. Behavior Research and Therapy, 22(3), 267–280. Dadds, M. R., Holland, D. E., Laurens, K. R., Mullins, M., Barrett, P. M., & Spence, S. H. (1999). Early intervention and prevention of anxiety disorders in children: results at 2-year follow-up. Journal of Consulting and Clinical Psychology, 67(1), 145–150. de Vaus, D.A. (1995). Surveys in social research. (4th ed.). St Leonards, Australia: Allen & Unwin. Essau, C. A., Conradt, J., & Petermann, F. (1999). Frequency and comorobidity of social phobia and social fears in adolescents. Behavior Research and Therapy, 37, 831–843. Ford, T., Goodman, R., & Meltzer, H. (2003). The British Child and Adolescent Mental Health Survey 1999: the prevalence of DSM-IV Disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 42(10), 1203–1211. Hardy, L., & Parfitt, G. (1991). A catastrophe model of anxiety and performance. British Journal of Psychology, 82, 163–178. Huston, J.L. (2001). Familial antecedents of musical performance anxiety: a comparison with social anxiety. Dissertation Abstracts International: Section B: the Sciences & Engineering, 62(1-B), 551.

750

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

James, I. (1997). Federation Internationale des Musiciens 1997 survey of 56 orchestras worldwide. London: British Association for Performing Arts Medicine. Kaplan, J., & Saccuzzo, D. D. (1989). Psychological testing: principles, applications and issues. Pacific Grove, CA: Brooks Cole. Kashani, J. H., & Orvaschel, H. (1990). A community study of anxiety in children and adolescents. American Journal of Psychiatry, 147(3), 313–318. Kashani, J. H., Orvaschel, H., Rosenberg, T. K., & Reid, J. C. (1989). Psychopathology in a community sample of children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 28(5), 701–706. Kendrick, M. J., Craig, K. D., Lawson, D. M., & Davidson, P. O. (1982). Cognitive and behavioural therapy for musical-performance anxiety. Journal of Consulting and Clinical Psychology, 50(3), 353–362. Kenny, D. T., Davis, P. J., & 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(6), 757–777. Kerr, M. (2003, 17 October). Social Anxiety Disorder much more prevalent than previously thought [Internet]. Medscape Medical News. Kirchner, J.M. (2003). A qualitative inquiry into musical performance anxiety. Medical Problems of Performing Artists, 18, 78–82. Kovacs, M. (1992). Children’s Depression Inventory Manual. Toronto, Canada: Multi-Health Systems. Kubzansky, L. D., & Stewart, A. J. (1999). At the intersection of anxiety, gender and performance. Journal of Social and Clinical Psychology, 18(1), 76–97. ¨ hman, A. (2000). Fear and anxiety: animal models and human cognitive Lang, P. J., Davis, M., & O psychophysiology. Journal of Affective Disorders, 61(3), 137–159. Lederman, R.J. (1999). Medical treatment of performance anxiety: a statement in favor. Medical Problems of Performing Artists, 14(3), 117–121. Lehrer, P. M., Goldman, N. S., & Strommen, E. F. (1990). A principal components analysis assessment of performance anxiety among musicians. Medical Problems of Performing Artists, 5(1), 12–18. Lewinsohn, P. M., Gotlib, I. H., Lewinsohn, M., Seeley, J. R., & Allen, N. B. (1998). Gender differences in anxiety disorders and anxiety symptoms in adolescents. Journal of Abnormal Psychology, 107(1), 109–117. Liebowitz, M. R., Gorman, J. M., Fyer, A. J., & Klein, D. F. (1985). Social phobia: review of a neglected anxiety disorder. Archives of General Psychiatry, 42, 729–736. Lonigan, C. J., Carey, M. P., & Finch, A. J. J. (1994). Anxiety and depression in children and adolescents: negative affectivity and the utility of self-reports. Journal of Consulting and Clinical Psychology, 62(5), 1000–1008. Maroon, M.T.J. (2003). Potential contributors to performance anxiety among middle school students performing at solo and ensemble contest. Dissertation Abstracts International, 64(2-A), 437. McCauley, R. J., & Swisher, L. (1984). Psychometric review of language and articulation tests for preschool children. Journal of Speech and Hearing Disorders, 49, 34–42. McCoy, L.H. (1999). Musical performance anxiety among college students: an integrative approach. Dissertation Abstracts International, 60(4-A), 1059. Morris, T.L. (2001). Social phobia. In: M. W. Vasey & M. R. Dadds (Eds.), The developmental psychopathology of anxiety. New York: Oxford University Press. Nagle, J. J., Himle, D. P., & Papsdorf, J. D. (1989). Cognitive-behavioural treatment of musical performance anxiety. Psychology of Music, 17, 12–21. Neftel, K. A., Adler, R. H., Kappeli, L., Rossi, M., Dolder, M., Kaser, H. E. et al. (1982). Stage fright in musicians: a model illustrating the effect of beta blockers. Psychosomatic Medicine, 44(5), 461–469. Nunnally, J.C. (1978). Psychometric theory. New York: McGraw-Hill. Osborne, M. S. (1998). Determining the diagnostic and theoretical adequacy of conceptualizing music performance anxiety as a social phobia. Unpublished Honours thesis, Macquarie University, Sydney.

M.S. Osborne, D.T. Kenny / Anxiety Disorders 19 (2005) 725–751

751

Paul, G. L. (1966). Insight vs. desensitization in psychotherapy: An experiment in anxiety reduction. Stanford: Stanford University Press. Ryan, C. (1998). Exploring musical performance anxiety in children. Medical Problems of Performing Artists, 13, 83–88. Ryan, C. A. (2003). A study of the differential responses of male and female children to musical performance anxiety. Dissertation Abstracts International Section A: Humanities and Social Sciences 63(7-A), 2487. Salmon, P. (1990). A psychological perspective on musical performance anxiety: a review of the literature. Medical Problems of Performing Artists, 5(1), 2–11. Schwartz, G. E., Davidson, R. J., & Goleman, D. J. (1978). Cognitive and somatic processes in anxiety. Psychosomatic Medicine, 40, 321–328. Sinden, L.M. (1999). Music performance anxiety: contributions of perfectionism, coping style, selfefficacy, and self-esteem. Dissertation Abstracts International, 60(3-A), 0590. Spielberger, C.D. (1983). State-Trait Anxiety Inventory STAI (Form Y). Palo Alto, CA: Consulting Psychologists Press, Inc. Standards for educational and psychological testing. (1985). Washington, DC: American Psychological Association. Steiner, A. (1998). The relationship between performance anxiety and narcissism in musicians. Dissertation Abstracts International, 59(4-B), 1869. Steptoe, A., & Fidler, H. (1987). Stage fright in orchestral musicians: a study of cognitive and behavioural strategies in performance anxiety. British Journal of Psychology, 78, 241–249. Salvia, J., & Ysseldyke, J. E. (1988). Assessment in special remedial education. Boston: Houghton Mifflin. Sweeney, G. A., & Horan, J. J. (1982). Separate and combined effects of cue-controlled relaxation and cognitive restructuring in the treatment of music performance anxiety. Journal of Counseling Psychology., 29(5), 486–497. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. (4th ed.). Needham Heights, MA: Pearson. Turner, S. M., Beidel, D. C., & Dancu, C. V. (1996). SPAI: Social Phobia and Anxiety Inventory. North Tonawanda, NY: Multi-Health Systems Inc. Turner, S. M., Beidel, D. C., Dancu, C. V., & Keys, D. J. (1986). Psychopathology of social phobia and comparison to avoidant personality disorder. Journal of Abnormal Psychology, 95, 389–394. 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. Wesner, R. B., Noyes, R., Jr., & Davis, T. L. (1990). The occurrence of performance anxiety among musicians. Journal of Affective Disorders, 18(3), 177–185. Wilson, G. D. (2002). Psychology for performing artists (2nd ed.). London, UK, Whurr. Wills, G., & Cooper, C. L. (1988). Pressure sensitive: popular musicians under stress. London: Sage. Wolfe, M.L. (1989). Correlates of adaptive and maladaptive musical performance anxiety. Medical Problems of Performing Artists, 4(1), 49–56. Zinn, M., McCain, C., & Zinn, M. (2000). Musical performance anxiety and the high-risk model of threat perception. Medical Problems of Performing Artists, 15, 65–71.