Validation of the Kenny Music Performance Anxiety Inventory (K-MPAI): A cross-cultural confirmation of its factorial structure

Abstract

This research investigated whether music performance anxiety (MPA) can be theoretically understood as a unidimensional construct, and whether the factorial structure is robust across different populations of musicians with different levels of expertise. K-MPAI scores were obtained from 455 Peruvian tertiary music students (mean age = 21.19 years, SD = 3.13, range = 18–40 years) and 368 Australian professional orchestral musicians (mean age = 42.07 years, SD = 10.21, range = 18–68 years). A high order exploratory factor analysis with the Schmid-Leiman solution was performed on the K-MPAI items. Unweighted Least Squares extraction method and optimal implementation of parallel analysis revealed one high order factor and two first order factors for both samples. High Cronbach’s and ordinal alpha levels for items belonging to each first order and high order factor in both samples were also obtained. Structural similarities between the two samples and an invariance analysis signified a comparable structure and conceptual interpretation of K-MPAI scores in both populations. The factorial structure obtained supported a unidimensional interpretation of the construct of MPA. First order level interpretations are also possible and have been demonstrated to be clinically useful.

Keywords

cross-cultural validation factor analysis Kenny’s typology of MPA K-MPAI music performance anxiety

Kenny (2011; Kenny, Arthey, & Abbass, 2014; Kenny & Holmes, 2015) has argued that music performance anxiety (MPA) can no longer be conceptualized as a unidimensional anxiety construct occurring on a continuum of severity from career stress at the low end to stage fright at the high end (Brodsky, 1996). The focus of this unidimensional conceptualization of MPA is on the symptom complex called the General Activation Syndrome, or the fight–flight–fright response, in which the somatic manifestations of anxiety occur in response to an impending musical performance as if it were a physical threat that posed a real and present danger to the musician.

Clinical work with severely performance-anxious musicians alerted the second author (Kenny) to the severity of the underlying psychopathology, which is more severe in the most anxious musicians. This pathology has been conceptualized as an attachment disorder (Kenny, 2011; Kenny et al., 2014; Kenny, Arthey, & Abbass, 2016; Kenny & Holmes, 2015). Such individuals presenting to mental health facilities are likely to be given diagnoses of an anxiety disorder – in particular, social anxiety disorder, other anxiety disorder, panic or panic disorder, and/or depression. Kenny (2011) concluded that while some musicians experienced reality-based focal anxiety that centered on the proximal somatic and cognitive anxiety symptoms typical of most performers when undertaking competitive or demanding public performances, others, while also frequently manifesting the same symptom constellation as those with focal anxiety, also experienced deeper, long-standing psychological distress that pervaded their lives, and which found excruciating expression in the performance context. Hence, Kenny proposed that MPA is better understood as a typology comprising three subtypes of MPA to account for qualitative differences in clinical presentation as well as variations in severity. The three subtypes proposed were: (a) MPA as a focal anxiety, where there is no generalized social anxiety, depression or panic; (b) MPA that co-occurs with other social anxieties; and (c) MPA that co-occurs with panic and depression. There are different levels of severity within each subtype. The theoretical model underpinning this typology is that MPA represents an intersection between an individual’s developmental history, which may be more or less disturbed – mildly, or not at all, in the case of focal anxiety and more severely in the third subtype – and the specific psychosocial conditions of musicianship – talent, achievement of technical mastery, preparedness, performance demands, exposure, competitiveness, and so on. Accordingly, MPA will have some of the general characteristics of other psychological disorders, in particular, the anxiety disorders, which are shared with nonmusicians, and some that are specific to MPA.

To understand these clinical presentations, Kenny applied Barlow’s triple vulnerability model for the development of anxiety disorders to further understand music performance anxiety. Barlow (2000) proposed an emotion-based model of anxiety development that owes much to Lazarus (1991) and Lazarus and Folkman (1987) whose relevance to understanding performance anxiety has been discussed in detail elsewhere (see Kenny, 2009, 2011). Barlow’s model proposes an integrated set of triple vulnerabilities that can account for the development of an anxiety or mood disorder. These are

a generalized biological (heritable) vulnerability;

a generalized psychological vulnerability;

specific life experiences that establish specific psychological vulnerabilities.

The generalized biological vulnerability infers a genetic contribution to the development of particular temperaments that have been labeled at various times “neuroticism,” “negative affect,” or “behavioural inhibition.” The generalized psychological vulnerability is based in early experiences, in particular negative relational experiences that result in a sense that life is unpredictable and uncontrollable and that one does not have the necessary coping resources to manage such experiences. Uncontrollability is strongly associated with negative affect and, subsequently, anxiety and depression (Allen, McHugh, & Barlow, 2008). The first two processes (a biological vulnerability and a generalized psychological vulnerability based on early experiences) may be sufficient conditions for the development of anxious apprehension, and genetic predisposition and sensitizing early life experiences may be sufficient to produce a generalized anxiety or mood (depression) disorder.

Barlow (1988, 2000, 2002) argued that panic attacks, which he called “false alarms,” arise in response to stressful life events (such as music performance) in people who experience high levels of general anxiety. Those at risk of panic attacks have also experienced specific psychological vulnerabilities whereby anxiety comes to be associated with certain internal (somatic sensations or intrusive thoughts) or environmental (social evaluation) stimuli that have become associated with heightened threat or danger. Social evaluation may be accompanied by heightened somatic sensations (false alarms or fear responses in the absence of a real threat or danger) that become associated with a perceived increase in threat or danger, in the case of social evaluation, the fear of others’ disapproval or rejection. Those perceiving most threat are likely to experience the greatest anxiety, and those who are most anxious are more likely to perceive the social evaluative context as more threatening. If individuals have both high negative affect and high anxiety sensitivity, this renders them more at risk of serious anxiety responses such as high music performance anxiety (Kenny, 2011; Kenny & Holmes, 2015; Kenny, Arthey, & Abbass, 2016). High anxiety sensitivity constitutes the first component (i.e., biological vulnerability) and negative affect the second component (i.e., psychological vulnerability) of the three factor model. Performance breakdown or adverse evaluations of performance are examples of the third component in the model.

Notwithstanding the commonalities between social anxiety and panic disorders with severe MPA presentations, the conceptual autonomy of MPA from other anxiety disorders has also been specified (Chang-Arana, 2015b; Kenny, 2009, 2011). For example, those with MPA are more likely than those with social phobia to (a) have higher expectations of themselves (Abbott & Rapee, 2004); (b) have greater fear of their own evaluation of their performance, as opposed to fear of the scrutiny of others in social phobia (Stoeber & Eismann, 2007), although the latter is also present in music performance anxiety; (c) have a higher degree of post-event rumination (Abbott & Rapee, 2004); and (d) a continued commitment to the feared performance situation, as opposed to avoidance of, or escape from the feared situation in social phobia (Powell, 2004); further, (e) the feared task in MPA is cognitively and physically challenging, unlike in social phobia in which the feared tasks are already in the behavioral repertoire (Kenny, 2011).

In response to the clinical findings and the conceptual specificity of MPA, Kenny (2009) developed a music performance anxiety inventory (K-MPAI) that aimed to assess the underlying psychopathology as well as the symptoms of MPA. It was administered to 379 professional orchestral musicians in Australia and to 159 tertiary music students in New Zealand. Principal axis factoring with varimax rotation of the K-MPAI revealed, for the tertiary level music students, 12 underlying factors, which could be subsumed under the following meta-factors:

Early relationship context: [(7) Generational transmission of anxiety; (4) Parental empathy];

Psychological vulnerability: [(1) Depression/Hopelessness (9); Controllability; (11) Trust (12); Pervasive performance anxiety]; and

Proximal performance concerns: [(3) Proximal somatic anxiety; (2) Worry/dread (negative cognitions); (6) Pre- and post-performance rumination; (8) Self/other scrutiny; (10) Opportunity cost; (5) Memory reliability].

Over the last ten years, Peru has experienced an increase in its capacity to offer higher level musical education at two new music faculties in two universities. The formation of two new youth symphony orchestras and one nonprofit organization inspired by the Venezuelan El Sistema (which targets high-risk and low-income children) have created more opportunities for young musicians. Consequently, the risk for non-adaptive anxiety-coping-behaviors are more likely to occur (Chang-Arana, 2015a, 2015b) particularly when there is fierce competition for places in these universities and orchestras.

In Peru, the empirical study of MPA is in its infancy. One of the hindrances to empirical research has been the absence of psychometrically robust psychological instruments to reliably assess MPA. Consequently, Chang-Arana (2015a, 2015b) adapted the K-MPAI for Spanish-speaking populations and estimated its psychometric properties in a sample of 455 Peruvian tertiary music students. Evidence for validity was based on intended test purpose, test content, internal structure and the relationship with other variables (American Educational Research Association, American Psychological Association & National Council on Measurement in Education, 2014).

One of the key findings in Chang-Arana (2015b) was a high order factorial structure for the K-MPAI named “negative affectivity in relation to music performance anxiety” which underlay two first order factors: “music performance anxiety” and “depression,” which corresponded to Kenny’s identification of a somatic symptoms cluster associated with MPA and an underlying psychological vulnerability. These categories were adopted in line with the tripartite model of anxiety and depression (Anderson & Hope, 2008; Brown, Chorpita, & Barlow, 1998; Clark & Watson, 1991). This structural model stated that generalized negative affectivity underlies both anxiety and depression and can account for the high correlation and the difficulty in finding discriminant differences between them. Thus, for the Peruvian sample it is possible to interpret MPA using the K-MPAI as a unidimensional construct comprised of two first order factors that independently measure music performance anxiety and depression.

In this article, we examined, first, whether MPA can be understood theoretically as a unidimensional construct and, if so, the nature and content of that construct; and, second, whether the factor structure is robust across culturally different populations of musicians, and musicians at different levels of expertise.

Methods

Participants

Two samples were used in this study. The first comprised 457 tertiary music students from three Peruvian music institutions; the second comprised 379 professional orchestral musicians from Australia who were each members of one of the eight premier state orchestras in Australia. Two participants from the Peruvian sample (0.44%) and 11 from the Australian sample (2.90%) had a high percentage of missing responses on the questionnaires (around 50%). Since they represented less than 5% for each sample, these cases were deleted from the analysis (Graham, 2009; Schafer & Graham, 2002). Missing data in both samples was replaced with regression imputation (Little & Rubin, 1987).

Sample 1

Eligible participants were undergraduate music students who were majoring in performance and were at least 18 years old. Students studying composition, music education, musicology, and music production were excluded. The final sample (n = 455 participants) was distributed among three music institutions – two private (69.01%) and one public (30.99%). The private institutions presented both contemporary and classical music curricula; the public institution was exclusively classical in focus. Mean age was 21.19 years (SD = 3.13; range = 18–40 years). There were more male (n = 337, 74.1%) than female participants (n = 113, 24.8%); five participants did not specify their sex (1.1%).

Sample 2

Australia has eight full-time professional symphonic and pit orchestras located in each of the capital cities of Australia. Their musicians represent the country’s most elite orchestral musician population. All members of these eight orchestras were invited to participate, of whom 379 agreed, representing a response rate of 72%. The final sample comprised 180 males (48.9%) and 188 females (51.1%). Their mean age was 42.07 years (SD = 10.21, range = 18–68 years). A comparison of the age, sex and instrument group of participants and nonparticipants was undertaken in order to ascertain the possible presence of systematic bias in the sample. Independent t-tests indicated that there were no differences on these dimensions.

Procedure

Sample 1

Authorizations from the music faculties were obtained to enter the classrooms to administer the test protocol. Data collection took place during either the first or last 30 minutes of each class which all music students attended. Prior to distributing the questionnaires, a consent form was presented to the students, read aloud, and signed. Questionnaires were completed successively in the following order: State-Trait Anxiety Inventory (STAI, Spielberger & Díaz-Guerrero, 1970), and Beck Anxiety Inventory (Thornberry, 2011). Participants were reminded about the anonymity of their answers. Some data collection sessions required research assistants who were trained to follow a standard data collection protocol. Data were analyzed with FACTOR version 10 (Lorenzo-Seva & Ferrando, 2006).

Sample 2

Ethical approval for the Australian study was granted by The University of Sydney Human Research Ethics Committee. Each participant received a package of questionnaires to be completed and returned in the pre-paid envelope provided or dropped into a confidential mailbox at each orchestra’s home venue. This was a large cross-sectional study (see, for example, Ackermann, Driscoll, & Kenny, 2012; Ackermann, Kenny, O’Brien, & Driscoll, 2014; Kenny & Ackermann, 2015; Kenny, Driscoll, & Ackermann, 2014, 2016) but only the results of the K-MPAI are of interest in this article.

High order exploratory factor analysis (HOEFA) was performed in both samples using FACTOR version 10 (Lorenzo-Seva & Ferrando, 2006). Following these analyses an invariance analysis was performed using R (version 3.2.5) and packages “semTools” (semTools Contributors, 2016) and “lavaan” (Rosseel, 2012).

Measures

Revised K-MPAI (Kenny, 2009)

This 40-item inventory was developed to assess the emotion-based theory of anxiety proposed by Barlow (2000) as it applies to anxiety in the context of music performance (see Kenny, 2011). This is a revised expanded version of the original 26-item inventory (Kenny, Davis, & Oates, 2004). Questions are answered on a seven-point Likert scale (0 = Strongly disagree to 6 = Strongly agree). Higher scores indicate greater anxiety and psychological distress. Items from this scale demonstrated excellent internal reliability (Cronbach’s alpha = .94; Kenny, 2009) and a robust factor structure with two music populations—professional orchestral musicians and tertiary music students.

Spanish version of the K-MPAI

A Spanish adaptation of the K-MPAI was developed for use with sample 1 by Chang-Arana (2015a, 2015b). All adaptations used blind back-translations until the precise meaning of each item was captured. Content, construct and discriminative validity, and clinical utility of the Spanish and Portuguese adaptations of the K-MPAI (both the 26-item-version and the 40-item-version) have been demonstrated in studies of Brazilian (Barbar, Crippa, & Osório, 2014a, 2014b, 2014c; Barbar, Souza, & Osório, 2015; Rocha, Dias-Neto, & Gattaz, 2011), Spanish (Zarza, Hernandez, López, & Gil, 2016), and Peruvian (Chang-Arana, 2015a, 2015b) musicians. Construct validity and clinical utility for the K-MPAI has also been reported using a German adaptation of the K-MPAI with German musicians (Peschke & von Georg, 2015), and with the English version of K-MPAI with Indian rock musicians undergoing yogic meditation treatment for severe MPA (Meitei & Kumari, 2014).

Results

Validity evidence for the Spanish K-MPAI (Chang-Arana, 2015a, 2015b)

Validity evidence based on test content

Adequate evidence for face and content validity based on test content was reported for the Peruvian sample by subject matter experts who rated item congruence. Of the 40 items, 39 reached a significant consensus (p < .05, consensus range = .88 – 1.00). The exception was item 27 (consensus level = .68): “As a child, I often felt sad.”

Validity evidence based on internal structure

Validity evidence based on the internal structure of the inventory was obtained by both first order and high order exploratory factor analysis.

First order exploratory factor analysis

A principal axis factor analysis with orthogonal-varimax rotation was performed on the 40 items of the K-MPAI. An appropriate Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy and a statistically significant Bartlett’s test of sphericity were obtained, KMO = .91, χ²(496) = 5281.36, p < .001. Four factors were extracted and retained according to the parallel analysis. Factors were named “proximal performance concerns” (20 items, Cronbach’s alpha = .91), “psychological vulnerabilities” (seven items, Cronbach’s alpha = .80), “confidence in memory” (two items, Conbrach’s alpha = .82), and “early parental relationship context” (three items, Cronbach’s alpha = .71).

High order exploratory factor analysis

The factorial structure obtained was only slightly different from the one proposed by Kenny (2011). In order to explore a higher order structure that could articulate the obtained factors, a first HOEFA with Schmid–Leiman solution (SLS, Schmid & Leiman, 1957) was performed.

The HOEFA with Minimum Rank Factor Analysis (MRFA) as the extraction method was performed on the K-MPAI 40 items using an oblique rotation of promin type since the estimation of a higher/second order factor (G) implies a theoretical dependence between the first order factors. A polychoric correlation matrix was factorized because items were ordinal variables and were completed using a polychoric (i.e., Likert) scale (Burga, 2006). An orthogonal correction was performed using SLS which allowed identification of factorial loading contributions from every item with its first and second order factor, KMO = .91, χ²(780) = 6390.8, p < .001. A minimum average partial (MAP) method suggested retaining three first order factors. Items with factorial loadings equal to or higher than .30 were retained since HOEFA lowers factorial loadings (Wolff & Preising, 2005).

The HOEFA was repeated multiple times given that every item deletion resulted in an adjustment to factorial loadings. The procedure was repeated until a stable structure was achieved, i.e., the high order factor did not contain values lower than .30. At this point, both KMO and Bartlett’s sphericity tests presented adequate values, KMO = .93, χ²(435) = 4948.9, p < .001. MAP suggested retaining two first order factors which explained 58.65% of common shared variance. Based on the tripartite model of anxiety and depression (Anderson & Hope, 2008; Brown et al., 1998; Clark & Watson, 1991), these two first order factors were named “music performance anxiety” and “depression”; the G (higher order) factor was named “negative affectivity in relation to music performance.” Items that did not reach a minimum factorial loading of .30 on G and, hence, were eliminated, were:

- Item 2: I find it easy to trust others (G = .190).

- Item 8: I find it difficult to depend on others (G = .129).

- Item 9: My parents were mostly responsive to my needs (G = .181).

- Item 22: Prior to or during a performance, I experience increased heart rate like pounding in my chest (G = .261).

- Item 23: My parents almost always listened to me (G = .203).

- Item 25: After a performance, I worry whether I played well enough (G = .229).

- Item 33: My parents encouraged me to try new things (G = .141).

- Item 35: When performing without music, my memory is reliable (G = .220).

- Item 37: I am confident playing from memory (G = .261).

- Item 40: I remain committed to performing even though it causes me great anxiety (G = .001).

Scores derived from items belonging either to first or second order factors presented adequate reliability levels according to Nunnally and Bernstein (1995). Cronbach’s alpha, ordinal alpha, and standard error of measurement (SEM; calculation based on ordinal alpha levels) are presented in Table 1.

Table 1.

Cronbach’s alpha, ordinal alpha and SEM for high order factor scores and first order scores.

	Cronbach’s α	Ordinal α	SEM
G (30 items)	.92	.97	4.87
F1 (21 items)	.91	.93	6.11
F2 (10 items)	.81	.92	3.01

Note: Item 14 scored both in F1 and F2.

Validity evidence based on relationships with other variables

K-MPAI scores were correlated with both subscales of STAI and BAI. At p < .001 there were statistically significant relationships between K-MPAI scores and the other questionnaires. The correlation between K-MPAI and STAI-Trait scores, r = .70, was higher than the correlation between STAI-State and BAI at r = .53. Moderate effect sizes were obtained between K-MPAI and STAI-State scores and large effects were registered for K-MPAI and STAI-Trait scores and K-MPAI and BAI scores (Ellis, 2010). Statistical power > .80 (Cohen, 1992) was found for all correlations.

Validity evidence based on internal structure for Peruvian and Australian samples

An Unweighted Least Squares (ULS) extraction method with an optimal implementation of parallel analysis (Timmerman & Lorenzo-Seva, 2011) for determining the number of factors was performed on both the Peruvian and Australian samples.

HOEFA for Australian professional orchestral musicians’ sample

A HOEFA was performed on the K-MPAI 40 items using an oblique rotation of promin type and a SLS varimax correction. Sample size was 368. The correlation matrix was factorized using a polychoric correlation. Both KMO and Bartlett’s test of sphericity were adequate to perform a factor analysis, KMO = .93, χ²(780) = 7497, p < .001. Optimal implementation of parallel analysis suggested retaining two first order factors which explained 43.98% of shared variance. The first order factors’ correlation levels with G were: F1 = .72, and F2 = .91. Items with factorial loadings on G that were equal to or higher than .30 were retained. Accordingly, five items were eliminated:

- Item 9: My parents were mostly responsive to my needs (G = .236).

- Item 23: My parents almost always listened to me (G = .219).

- Item 33: My parents encouraged me to try new things (G = .129).

- Item 35: When performing without music, my memory is reliable (G = .286).

- Item 37: I am confident playing from memory (G = .292).

After two additional factor extractions, a stable structure was achieved; the high order factor did not present values lower than .30. KMO level and Bartlett’s test of sphericity were still adequate to perform a factor analysis, KMO = .95, χ²(595) = 6379.9, p < .001. The ULS-PA advised that two first order factors be retained. Together these explained 47.89% of shared variance. The first order factor correlation levels with G were: F1 = .80, and F2 = .93. Items with factorial loadings on G equal to or higher than .30 were retained. Table 2 presents the final solution for the Australian sample.

Table 2.

High order exploratory factor analysis with SLS for the K-MPAI items based on Australian professional musicians.

Item	F1	F2	G
7. Even if I work hard in preparation for a performance, I am likely to make mistakes	.268		.539
10. Prior to, or during a performance, I get feelings akin to panic	.463		.606
11. I never know before a concert whether I will perform well	.292		.603
12. Prior to, or during a performance, I experience dry mouth	.363		.432
14. During a performance I find myself thinking about whether I’ll even get through it	.295		.616
15. Thinking about the evaluation I may get interferes with my performance	.313		.666
16. Prior to, or during a performance, I feel sick or faint or have a churning in my stomach	.501		.609
17. Even in the most stressful performance situations, I am confident that I will perform well	.324		.515
18. I am often concerned about a negative reaction from the audience	.306		.533
20. From early in my music studies, I remember being anxious about performing	.314		.467
21. I worry that one bad performance may ruin my career	.290		.583
22. Prior to, or during a performance, I experience increased heart rate like pounding in my chest	.562		.478
25. After a performance, I worry about whether I played well enough	.458		.551
26. My worry and nervousness about my performance interferes with my focus and concentration	.510		.656
28. I often prepare for a concert with a sense of dread and impending disaster	.307		.699
30. Prior to, or during a performance, I have increased muscle tension	.470		.484
32. After a performance, I replay it in my mind over and over	.427		.463
34. I worry so much before a performance, I cannot sleep	.400		.572
36. Prior to, or during a performance, I experience shaking or trembling or tremor	.521		.539
38. I am concerned about being scrutinized by others	.399		.550
39. I am concerned about my own judgment of how I will perform	.313		.477
40. I remain committed to performing even though it causes me great anxiety	.464		.452
1. I generally feel in control of my life		.205	.622
2. I find it easy to trust others		.244	.446
3. Sometimes I feel depressed without knowing why		.271	.688
4. I often find it difficult to work up the energy to do things		.250	.552
5. Excessive worrying is a characteristic of my family		.142	.493
6. I often feel that life has not much to offer me		.336	.659
8. I find it difficult to depend on others		.246	.533
13. I often feel that I am not worth much as a person		.298	.715
19. Sometimes I feel anxious for no particular reason		.187	.666
24. I give up worthwhile performance opportunities		.131	.403
27. As I child I often felt sad		.247	.553
29. One or both of my parents were overly anxious		.156	.450
31. I often feel that I have nothing to look forward to		.344	.737

Note: N = 368, total item number: 30. Extraction method: Unweighted Least Square (ULS). Method used for estimating advised number of dimensions to retain: Optimal implementation of parallel analysis.

Scores derived from items belonging to either first or second order factors presented adequate reliability levels according to Nunnally and Bernstein (1995). Cronbach’s alpha, ordinal alpha, and SEM (calculation based on ordinal alpha levels) are depicted in Table 3.

Table 3.

Cronbach’s alpha, ordinal alpha and SEM for high order factor scores and first order scores of Australian professional musicians.

	Cronbach’s α	Ordinal α	SEM
G (35 items)	.95	.96	7.50
F1 (22 items)	.94	.95	6.03
F2 (13 items)	.86	.93	3.58

HOEFA for Peruvian tertiary music students’ sample

A HOEFA was performed on the K-MPAI 40 items using an oblique rotation of promin type and a SLS varimax correction. Sample size was 455. The correlation matrix was factorized using a polychoric correlation. Both KMO and Bartlett’s test of sphericity were adequate in order to perform a factor analysis, KMO = .91, χ²(780) = 6390.8, p < .001. Optimal implementation of parallel analysis suggested retaining three first order factors which explained 39.47% of shared variance. First order factors’ correlation levels with G were: F1 = .67, F2 = .91, and F3 = .57. Items with factorial loadings to G equal or higher than .30 were retained. Thus ten items were eliminated:

- Item 2: I find it easy to trust others (G = .187).

- Item 8: I find it difficult to depend on others (G = .131).

-Item 9: My parents were mostly responsive to my needs (G = .173).

- Item 22: Prior to, or during a performance, I experience increased heart rate like pounding in my chest (G = .270).

- Item 23: My parents almost always listened to me (G = .189).

- Item 25: After a performance, I worry whether I played well enough (G = .235).

- Item 33: My parents encouraged me to try new things (G = .142).

- Item 35: When performing without music, my memory is reliable (G = .245).

- Item 37: I am confident playing from memory (G = .282).

- Item 40: I remain committed to performing even though it causes me great anxiety (G = .001).

After one more factor extraction, a stable structure was achieved, that is, the high order factor did not present values lower than .30. KMO level and Bartlett’s test of sphericity were still adequate to perform a factor analysis, KMO = .93, χ²(435) = 4948.9, p < .001. The optimal implementation of parallel analysis advised retention of two first order factors which explained 41.17% of shared variance. Correlations of first order factors with G were: F1 = .76, and F2 = .92. Items with factorial loadings to G equal or higher than .30 were retained (Table 4).

Table 4.

High order exploratory factor analysis with SLS for the K-MPAI items based on Peruvian professional music students.

Item	F1	F2	G
10. Prior to, or during a performance, I get feelings akin to panic	.464		.528
11. I never know before a concert whether I will perform well	.262		.637
12. Prior to, or during a performance, I experience dry mouth	.319		.453
14. During a performance I find myself thinking about whether I’ll even get through it	.189		.607
15. Thinking about the evaluation I may get interferes with my performance	.368		.462
16. Prior to, or during a performance, I feel sick or faint or have a churning in my stomach	.276		.533
17. Even in the most stressful performance situations, I am confident that I will perform well	.165		.422
18. I am often concerned about a negative reaction from the audience	.363		.345
19. Sometimes I feel anxious for no particular reason	.261		.450
20. From early in my music studies, I remember being anxious about performing	.536		.368
21. I worry that one bad performance may ruin my career	.353		.380
24. I give up worthwhile performance opportunities	.191		.479
26. My worry and nervousness about my performance interferes with my focus and concentration	.482		.520
28. I often prepare for a concert with a sense of dread and impending disaster	.268		.674
29. One or both of my parents were overly anxious	.215		.394
30. Prior to, or during a performance, I have increased muscle tension	.406		.505
32. After a performance, I replay it in my mind over and over	.315		.353
34. I worry so much before a performance, I cannot sleep	.453		.476
36. Prior to, or during a performance, I experience shaking or trembling or tremor	.503		.485
38. I am concerned about being scrutinized by others	.505		.486
39. I am concerned about my own judgment of how I will perform	.364		.331
1. I generally feel in control of my life		.214	.489
3. Sometimes I feel depressed without knowing why		.225	.560
4. I often find it difficult to work up the energy to do things		.192	.480
5. Excessive worrying is a characteristic of my family		.172	.436
6. I often feel that life has not much to offer me		.384	.611
7. Even if I work hard in preparation for a performance, I am likely to make mistakes		.156	.493
13. I often feel that I am not worth much as a person		.395	.715
27. As I child I often felt sad		.236	.486
31. I often feel that I have nothing to look forward to		.283	.639

Note: N = 455; total item number: 30. Extraction method: Unweighted Least Square (ULS).Method used for estimating advised number of dimensions to retain: Optimal implementation of parallel analysis.

Just as with the Australian sample, Peruvian scores derived from items belonging either to first or second order factors presented adequate reliability levels according to Nunnally and Bernstein (1995). Cronbach’s alpha, ordinal alpha and SEM (calculation based on ordinal alpha levels) are presented in Table 5.

Table 5.

Cronbach’s alpha, ordinal alpha and SEM for high order factor scores and first order scores of Peruvian professional music students.

	Cronbach’s α	Ordinal α	SEM
G (30 items)	.92	.94	7.43
F1 (21 items)	.91	.92	6.23
F2 (10 items)	.81	.91	2.89

In both factorial structures there were shared items that were eliminated:

- Item 9: My parents were mostly responsive to my needs.

- Item 23: My parents almost always listened to me.

- Item 33: My parents encouraged me to try new things.

- Item 35: When performing without music, my memory is reliable.

- Item 37: I am confident playing from memory.

Invariance analysis

In order to further assess validity based on internal structure and determine whether it was possible to compare the Peruvian and Australian populations, a measure of invariance was computed. Given that in confirmatory models it is not possible to fit an SLS, the Peruvian model was fitted into a bi-factor model variation. The invariance of the bifactorial model was tested with the Peruvian sample. The fit of the model was good (Hu & Bentler, 1999), χ²(375) = 731.92, p < .001, CFI = .989, TLI = .987, RMSEA = .034, SRMR = .045.

Next, different types of invariance models were fitted to the data. Only the first model was accepted (see Table 6).

Table 6.

Fit index of different types of invariance models.

Model	χ²	df	p	CFI	TLI	RMSEA	SRMR
1 Equal form	1476.41	750	< .001	.991	.989	.049	.053
2 Equal loadings	3207.34	810	< .001	.969	.967	.085	.077
3 Equal loadings and thresholds	3032.85	957	< .001	.973	.976	.073	.066
4 Equal loadings, thresholds and residuals	3032.85	957	< .001	.973	.976	.073	.066
5 Equal loadings, thresholds, residuals and means	4622.89	960	< .001	.953	.957	.096	.059

The likelihood ratio test comparing nested models shows that when there are more equal parameters between groups, the model fit is worsened. When comparing model 1 with 2, it worsens. Comparison of models 2 and 3 did not show a change in fit. Models 3 and 4 showed that they were indistinguishable. Finally, comparison of models 4 and 5 showed that the fit worsened (see Table 7).

Table 7.

Test of significant change in chi square fit between invariance models.

Fit	df	χ²	χ² difference	df difference	p
P1	750	1476.4
P2	810	3207.3	77.481	11.148	< .001
P2	810	3207.3
P3	957	3032.8	−18.839	19.386	< .001
P4	957	3032.8
P5	960	4622.9	149.39	1.4414	< .001

Discussion

The HOEFA procedure revealed a nearly identical high order factorial structure in both samples. This provides strong evidence that the back-translation process of the K-MPAI into Spanish was accurate.

For the Peruvian sample the factorial structure obtained in this research was almost identical to the one reported in Chang-Arana (2015a, 2015b), except for item 14, which in the former was grouped under F1 and in the latter belonged to both F1 and F2. Both structures also suggested eliminating the same ten items even though the extraction method for Chang-Arana (2015a, 2015b) was MRFA and for this research was ULS. Additionally, either using MAP test or the optimal implementation of parallel analysis, data indicated that two first order factors be retained, just as the HOEFA with the Australian sample suggested. Taken together, we can conclude that from a HOEFA point of view the K-MPAI factorial structure for the Australian and Peruvian samples is robust.

Elimination of items 35, “When performing without music, my memory is reliable,” and 37, “I am confident playing from memory,” from both samples suggests that the memory-related items were not relevant to this factorial model. One plausible explanation for this lies in the emotion-oriented content of the K-MPAI. Even though memorization of music is an important skill required of professional musicians, it seems that this skill is more related to self-efficacy or technical mastery than the emotional dimension of personality and/or MPA. Similarly, items 9, “My parents were mostly responsive to my needs,” 23, “My parents almost always listened to me,” and 33 “My parents encouraged me to try new things” were excluded from both samples. It is possible that by the time musicians reach the age of 18, their internal working models of early relationships are well established and, having attained the capacity for self-awareness and self-reflection, are more likely to be focused on self-evaluation rather than evaluation of their internalized parental objects, which have, by this stage of development, become integrated into their self-concept (Goodman, 2005).

Factorial structures in both samples pointed towards a high order structure which can be interpreted as unidimensional. Based on the tripartite model of anxiety and depression (Anderson & Hope, 2008; Brown et al., 1998; Clark & Watson, 1991), this high order factor was named “negative affectivity in relation to music performance anxiety” and the two first order factors have been named “music performance anxiety” and “depression” for both samples. Given the clear content of each first order factor and their items’ high level of reliability it is possible to use the inventory in a bi-dimensional way, which has proven useful in the clinical setting.

In a validated German translation of the K-MPAI, and with 130 German musicians comprising professionals, students and amateurs, Peschke and von Georg (2015) reported a similar factorial structure to that found in this study. A three factor solution was obtained as follows: Factor 1: MPA-related symptoms experienced before, during, and after performance; Factor 2: General depression and psychological vulnerability; Factor 3: Early relationship context, including generational transmission of anxiety and parental empathy. Although the factorial method was not specified, the factor structure of the K-MPAI appears robust to different factorial extraction methods. The extracted factors in this study mapped onto performance competencies in the expected direction, with negative correlations on all three factors with a scale measuring conviction of performance competence, and positively with a factor measuring a deficit orientation to performance.

Results reported in this research resonate with two previous studies conducted with the 26-item-version of the K-MPAI (Kenny et al., 2004). For instance, Barbar et al. (2014a) reported an adequate back-translation process and high reliability values in a sample of 230 amateur and professional Brazilian musicians (Cronbach’s α = .82). More recently, Zarza et al. (2016) also conducted a back-translation for a Spanish sample. They performed a combination of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) on a sample of tertiary Spanish music students using equivalent sampling and data collection procedures to Chang-Arana (2015a, 2015b). They reported three main factors named “Helplessness” (Cronbach’s alpha = .79), “Specific Cognitions” (Cronbach’s alpha = .87), and “Early Relationship Context” (Cronbach’s alpha = .57), thereby providing further support for Kenny’s adaptation of Barlow’s anxiety model (2000) for a Spanish musician sample. Nevertheless, Zarza et al. (2016) chose principal components analysis (PCA) as a “factor” extraction method and the criterion of eigenvalues greater than 1 (K1) to retain seven factors in the EFA. This combination has been shown to lead to “potentially serious negative consequences” (Preacher & MacCullum, 2003, p. 13). On the one hand, PCA’s purpose is to reduce data until components that can account for as much variance as possible are identified. The purpose of EFA is not to explain as much variance as possible, but to assist the researcher’s interpretation of “sources of common variation underlying observed data” (Preacher & MacCullum, 2003, p. 21). Therefore, it is incorrect to use PCA when what is intended is factor extraction. On the other hand, the K1 criterion has been proven to underestimate the number of factors to be retained among other flaws, which can be avoided by more robust methods such as parallel analysis (Preacher & MacCullum, 2003). Despite these statistical reservations, the structure reported in Zarza et al. (2016) supported Barlow’s tripartite model (2000) and aligns with the conclusions in the present study. Since different data reduction techniques demonstrated similar theoretical implications, our results provide further support for the consistency of Barlow’s tripartite model of anxiety and depression (Barlow, 2000) and its adaptation to MPA by Kenny (2011).

Rocha et al. (2011) back-translated the revised 40-item K-MPAI (Kenny, 2009) in a Brazilian sample of 218 amateur and professional musicians. High correlations between the K-MPAI and STAI were reported (r = .64) as well as high internal consistency (Cronbach’s alpha = .96) and test–retest reliability measures. These results are consistent with Chang-Arana (2015a, 2015b) and Kenny (2011). Further confirmatory factor analyses and psychometric research using the revised K-MPAI (Kenny, 2009) is warranted. Additional cross-cultural comparisons of the revised K-MPAI could confirm the encouraging results reported in this and previous research that the inventory is applicable to musicians from different cultural contexts and stage of musical development.

The invariance analysis added empirical support to the structure and conceptual interpretation of the high order structure of the Peruvian and Australian samples; that is, the structure and theoretical interpretation is valid for both populations. Nevertheless, since only the first model fit was accepted, it is not possible to directly compare both populations. This result could be explained by the fact that two different kinds of populations were surveyed: music students and professional musicians. Therefore, further research could test our procedure in equivalent groups of musicians. Hence, there is not yet sufficient evidence for assuming that the interpretations and decisions derived from the scores of the K-MPAI are cross-culturally valid.

In sum, the results of this research suggest a strong high-order factorial structure and consistent theoretical interpretation across both Australian and Peruvian populations. The demonstration of a unidimensional higher order factor in this study provides further strong evidence for Kenny’s theory of the etiology of music performance anxiety and further support for the tripartite typology of MPA that she proposed (Kenny, 2011), comprising focal MPA, MPA with social anxiety, and MPA with panic and depression. Future research should include (a) further psychometric analysis of the K-MPAI, focusing on testing the high order factorial analysis in other contexts; (b) further cross-validations of the K-MPAI to support its application cross-culturally; and (c) experimental attempts to further validate Kenny’s clinical typology, as determination of MPA type has significant implications for treatment.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The data collected for the Australian sample was funded by a grant to the second author (Kenny) with colleagues Dr Bronwen Ackermann and A/Professor Tim Driscoll from the Australia Research Council (ARC) (Project ID: LP0989486) and The Australia Council for the Arts (ACA). The researchers also received in-kind support from the eight major symphony and pit orchestras of Australia.

References

Abbott

M. J.

Rapee

R. M.

(2004). Post-event rumination and negative self-appraisal in social phobia before and after treatment. Journal of Abnormal Psychology, 113(1), 136–144. https://dx-doi-org.web.bisu.edu.cn/10.1037/0021-843X.113.1.136

Ackermann

B. J.

Driscoll

Kenny

D. T.

(2012). Musculoskeletal pain and injury in professional orchestral musicians in Australia. Medical Problems of Performing Artists, 27(4), 181–187. Retrieved from https://www.researchgate.net/publication/233941573_Musculoskeletal_Pain_and_Injury_in_Professional_Orchestral_Musicians_in_Australia

Ackermann

B. J.

Kenny

D. T.

O’Brien

Driscoll

T. R.

(2014). Sound Practice: Improving occupational health and safety for professional orchestral musicians in Australia. Frontiers in Psychology, 5, 1–11. doi:10.3389/fpsyg.2014.00973

Allen

L. B.

McHugh

R. K.

Barlow

D. H.

(2008). Emotional disorders: A unified protocol. In Barlow

D. H.

(Ed.), Clinical handbook of psychological disorders: A step-by-step treatment manual (4th ed.; pp. 216–249). New York: The Guildford Press.

American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (2014). Standards for educational and psychological testing. Washington DC: American Educational Research Association.

Anderson

E. R.

Hope

D. A.

(2008). A review of the tripartite model for understanding the link between anxiety and depression in youth. Clinical Psychological Review, 28, 275–287. doi:10.1016/j.epr.2007.05.004

Barbar

A. E.

Crippa

J. A.

Osório

F. L.

(2014a). Kenny Music Performance Anxiety Inventory (KMPAI): Transcultural adaptation for Brazil and study of internal consistency. Journal of Depression and Anxiety, 3, 167. doi:10.4172/2167–1044.1000167

Barbar

A. E.

Crippa

J. A.

Osório

F. L.

(2014b). Performance anxiety in Brazilian musicians: Prevalence and association with psychopathology indicators. Journal of Affective Disorders, 152–154, 381–386. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1016/j.jad.2013.09.041

Barbar

A. E.

Crippa

J. A.

Osório

F. L.

(2014c). Parameters for screening music performance anxiety. Revista Brasileira de Psiquiatria, 36(3), 245–247. Retrieved from https://dx-doi-org-s.web.bisu.edu.cn/10.1590/1516–4446–2013–1335

10.

Barbar

A. E.

Souza

J. A.

Osório

F. L.

(2015). Exploratory factor analysis of Kenny Music Performance Anxiety Inventory (K-MPAI) in a Brazilian musician sample. Archives of Clinical Psychiatry (São Paulo), 42(5), 113–116. Retrieved from https://dx-doi-org-s.web.bisu.edu.cn/10.1590/0101–60830000000060

11.

Barlow

D. H.

(1988). Anxiety and its disorders: The nature and treatment of anxiety and panic. New York: Guilford Press.

12.

Barlow

(2000). Unraveling the mysteries of anxiety and its disorders from the perspective of emotion theory. American Psychologist, 55(11), 1247–1263. doi:10.1037/0003–066X.55.11.1247

13.

Barlow

D. H.

(2002). Anxiety and its disorders. The nature and treatment of anxiety and panic (2nd ed.). New York, NY: The Guilford Press.

14.

Brodsky

(1996). Music performance anxiety reconceptualized: A critique of current research practice and findings. Medical Problems of Performing Artists, 11(3), 88–98.

15.

Brown

T. A.

Chorpita

B. F.

Barlow

D. H.

(1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107(2), 179–192. https://dx-doi-org.web.bisu.edu.cn/10.1037/0021-843X.107.2.179

16.

Burga

(2006). La unidimensionalidad de un instrumento de medición: Perspectiva factorial [The unidimensionality of a measurement instrument: A factorial perspective]. Revista de Psicología de la PUCP, 24(1), 54–80. Retrieved from http://revistas.pucp.edu.pe/index.php/psicologia/article/view/642/629

17.

Chang-Arana

A. M.

(August, 2015a). Adaptation and psychometric properties of the Kenny-Music Performance Anxiety Inventory (K-MPAI). Paper presented at the X Latin- American Regional Conference and III Pan American Regional Conference of Music Education, Lima, Peru. Retrieved from http://congreso.pucp.edu.pe/isme/wp-content/uploads/sites/8/2013/07/Actas-ISME-Per%C3%BA-2015.pdf

18.

Chang-Arana

A. M.

(2015b). Adaptation and psychometric properties of the Kenny-Music Performance Anxiety Inventory (K-MPAI) (Unpublished license thesis). University of Lima: Peru. doi:10.13140/RG.2.2.14697.49763

19.

Clark

L. A.

Watson

(1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100(3), 316–336. doi:10.1037/0021–843X.100.3.316

20.

Cohen

(1992). A power primer. Psychological Bulletin, 112(1), 155–159. doi:10.1037/0033–2909.112.1.155

21.

Ellis

P. D.

(2010). The essential guide to effect sizes: Statistical power, meta-analysis, and the interpretation of research results. Cambridge, UK: Cambridge University Press.

22.

Graham

J. W.

(2009). Missing data analysis: Making it work in the real world. Annual Review of Psychology, 60, 549–576. doi:10.1146/annurev.psych.58.110405.085530

23.

Goodman

(2005). Empirical evidence supporting the conceptual relatedness of object representations and internal working models. Journal of the American Psychoanalytic Association, 53(2), 597–617. doi:10.1177/00030651050530021101

24.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structural analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. doi:10.1080/10705519909540118

25.

Kenny

D. T.

(December, 2009). The factor structure of the revised Kenny Music Performance Anxiety Inventory. Research presented at the International Symposium on Performance Science, Auckland, New Zealand. Retrieved from http://www.legacyweb.rcm.ac.uk/cache/fl0019647.pdf

26.

Kenny

D. T.

(2011). The psychology of music performance anxiety. Oxford: Oxford University Press.

27.

Kenny

D. T.

Ackermann

B. J.

(2015). Performance-related musculoskeletal pain, depression and music performance anxiety in professional orchestral musicians: A population study. Psychology of Music, 43(1), 43–60. doi:10.1177/0305735613493953

28.

Kenny

D. T.

Arthey

Abbass

(2014). Intensive short-term dynamic psychotherapy (ISTDP) for severe music performance anxiety: The assessment, process and outcome of psychotherapy with a professional orchestral musician. Medical Problems of Performing Artists, 29(1), 3–7. Retrieved from https://www.researchgate.net/publication/260950324_Intensive_short-term_dynamic_psychotherapy_for_severe_music_performance_anxiety_Assessment_process_and_outcome_of_psychotherapy_with_a_professional_orchestral_musician

29.

Kenny

D. T.

Davis

Oates

(2004). Music performance anxiety and occupational stress amongst opera chorus artists and their relationship with state and trait anxiety and perfectionism. Anxiety Disorders, 18, 757–777. doi:10.1016/j.janxdis.2003.09.004

30.

Kenny

D. T.

Driscoll

Ackermann

(2014). Psychological well-being in professional orchestral musicians in Australia: A descriptive population study. Psychology of Music, 42(2), 210–232. doi:10.1177/0305735612463950

31.

Kenny

D. T.

Arthey

Abbass

(2016). Identifying attachment ruptures underlying severe music performance anxiety in a professional musician undertaking an assessment and trial therapy of Intensive Short-Term Dynamic Psychotherapy (ISTDP). SpringerPlus, 5(1), 1591.

32.

Kenny

D. T.

Driscoll

Ackermann

(2016). Is playing in the pit really the pits? Pain, strength, music performance anxiety, and workplace satisfaction in professional musicians in stage, pit, and combined stage/pit orchestras. Medical Problems of Performing Artists, 31(1), 1–7. doi:10.21091/mppa.2016.1001

33.

Kenny

D. T.

Holmes

(2015). Exploring the attachment narrative of a professional musician with severe performance anxiety: A case report. Journal of Psychology and Psychotherapy, 5(4), 1–6. doi:10.4172/2161-0487.1000190

34.

Lazarus

R. S.

(1991). Progress on a cognitive-motivational-relational theory of emotion. American Psychologist, 46(8), 819. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1037/0003-066X.46.8.819

35.

Lazarus

R. S.

Folkman

(1987). Transactional theory and research on emotions and coping. European Journal of Personality, 1(3), 141–169. doi:10.1002/per.2410010304

36.

Little

R. J. A.

Rubin

D. B.

(1987). Statistical analysis with missing data. New York: Wiley.

37.

Lorenzo-Seva

Ferrando

P. J.

(2006). FACTOR: A computer program to fit the exploratory factor analysis model. Behavior Research Methods, 38(1), 88–91. doi:10.3758/BF03192753

38.

Meitei

S. T.

Kumari

(2014). Efficacy of cyclic meditation on reducing music performance anxiety in rock musicians. International Journal of Social Science and Humanities Research, 2(3), 126–132. Retrieved from http://www.researchpublish.com.

39.

Nunnally

J. C.

Bernstein

I. H.

(1995). Teoría psicométrica [Psychometric theory] (2nd ed.). México D.F.: McGraw Hill.

40.

Peschke

von Georg

(August, 2015). The competence of performance: Mental aspects of succeeding and failing in musicians. Paper presented at the Ninth Triennial Conference of the European Society for the Cognitive Sciences of Music, Royal Northern College of Music, Manchester, UK.

41.

Powell

D. H.

(2004). Treating individuals with debilitating performance anxiety: An introduction. Journal of Clinical Psychology/In Session, 60(8), 801–808. doi:10.1002/jclp.20038

42.

Preacher

K. J.

MacCallum

R. C.

(2003). Repairing Tom Swift’s electric factor analysis machine. Understanding statistics: Statistical issues in psychology, education, and the social sciences, 2(1), 13–43. https://dx-doi-org.web.bisu.edu.cn/10.1207/S15328031US0201_02

43.

Rocha

Dias-Neto

Gattaz

W. F.

(2011). Music performance anxiety: Translation, adaptation and validation of the Kenny Music Performance Anxiety Inventory (K-MPAI) to the Portuguese language. Archives of Clinical Psychiatry, 38(6), 217–221. Retrieved from https://dx-doi-org.web.bisu.edu.cn/10.1590/S0101-60832011000600001

44.

Rosseel

(2012). lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48(2), 1–36. doi:10.18637/jss.v048.i02

45.

Schafer

J. L.

Graham

J. W.

(2002). Missing data: Our view of the state of the art. Psychological Methods, 7(2), 147–177. doi:10.1037//1082-989X.7.2.147

46.

Schmid

Leiman

J. M.

(1957). The development of hierarchical factor solutions. Psychometrika, 22(1), 53–61. doi:10.1007/BF02289209

47.

semTools Contributors. (2016). semTools: Useful tools for structural equation modeling. R package version 0.4-14. Retrieved from https://CRAN.R-project.org/package=semTools

48.

Spielberger

Ch.

Díaz-Guerrero

(1970). Inventario de Ansiedad: Rasgo-Estado (IDARE) [State-Trait Anxiety Inventory (STAI)]. Florida: University of South Florida.

49.

Stoeber

Eismann

(2007), Perfectionism in young musicians: Relations with motivation, effort, achievement, and distress. Personality and Individual Differences, 43(8), 2182–2192. Retrieved from https://kar.kent.ac.uk/id/eprint/4471

50.

Thornberry

G. L. M.

(2011). El rol de las creencias irracionales en la relación entre los eventos activadores, y las consecuencias emocionales y conductuales en estudiantes de dos universidades privadas de Lima (Tesis de doctorado inédita) [The role of irrational beliefs in the relationship between activating events, and the emotional and behavioural consequences in students from two private universities of Lima (Unpublished doctoral dissertation)]. Lima: San Martín de Porres University.

51.

Timmerman

M. E.

Lorenzo-Seva

(2011). Dimensionality assessment of ordered polytomous items with parallel analysis. Psychological Methods, 16(2), 209–220. doi:10.1037/a0023353

52.

Wolff

H. G.

Preising

(2005). Exploring item and high order factor structure with the Schmid-Leiman solution: Syntax codes for SPSS and SAS. Behavior Research Methods, 37(1), 48–58. doi:10.3758/BF03206397

53.

Zarza

F. J.

Hernández

S. O.

López

O. C.

Gil

B. M.

(2016). Kenny Music Performance Anxiety Inventory: Confirmatory factor analysis of the Spanish version. Psychology of Music, 44(3), 340–352. doi:10.1177/0305735614567932