Perfectionistic self-presentation and emotional experiences in music students: A three-wave longitudinal study

Research has shown that perfectionism predicts emotional experiences among amateur, professional and adolescent musicians. In examining these relationships, previous research has measured trait perfectionism and employed cross-sectional designs. The current study builds on existing research by examining whether perfectionistic self-presentation (as opposed to trait perfectionism) predicts negative and positive emotional experiences in music students over time. One hundred and forty-three music students (M age 18.92 years, SD = 2.96) enrolled in music-related degree programs completed measures of perfectionistic self-presentation (perfectionistic self-promotion, non-display of imperfection, and non-disclosure of imperfection) and emotional experiences (positive and negative feelings) at the start, middle and the end of the academic year. Path analysis revealed that perfectionistic self-promotion at the start of the year predicted lower positive feelings in the middle of the year, and non-disclosure of imperfection in the middle of the year predicted lower positive feelings at the end of the year. In addition, negative feelings in the middle of the year also predicted higher non-disclosure of imperfection at the end of the year. The findings suggest that the desire to present oneself perfectly and avoid disclosure of imperfections may contribute to less positive emotional experiences among music students.

mixed and/or negative experiences (e.g., Burt & Mills, 2006). These experiences extend to the emotions music students have and whether they come to develop positive or negative feelings towards the study of music. In the current study, we sought to better understand the factors that might contribute to the emotional experiences of music students. We did so by examining whether positive and negative feelings towards studying music were predicted by the perfectionistic self-presentational styles exhibited by music students across the first year of university.

Multidimensional perfectionism and musicians
Perfectionism is a personality characteristic broadly defined as a combination of a commitment to exceedingly high standards and a preoccupation with harsh self-critical evaluation (Frost, Marten, Lahart, & Rosenblate, 1990). It is typically considered to be a trait in that it reflects consistency in thoughts, feelings, and emotions evident across contexts and time (McAdams & Pals, 2006). There are multiple models and measures that have been used to examine perfectionism. These models and measures often differ in terms of their content and place varying degrees of emphasis on personal and interpersonal dimensions. However, in line with the broad definition of perfectionism, researchers typically differentiate between dimensions of perfectionism that encapsulate striving towards very high personal standards or flawlessness (referred to as perfectionistic strivings) and dimensions of perfectionism that encapsulate self-evaluative concerns, doubts, and perceived pressures from others (referred to as perfectionistic concerns; Stoeber & Otto, 2006).
Research examining these two dimensions of perfectionism attests to their importance in a number of contexts (e.g., sport and education ;Hill, Mallinson-Howard, & Jowett, 2018;Speirs Neumeister, 2007). This research has typically found perfectionistic concerns to be associated with maladaptive correlates, processes and outcomes (e.g., neuroticism, avoidant coping and burnout). By contrast, research has typically found perfectionistic strivings to be more complex and associated with a mix of adaptive and maladaptive correlates, processes, and outcomes (e.g., conscientiousness, problem-focused coping, and better performance versus self-criticism, worry, and anxiety). There is also evidence that some dimensions of perfectionistic strivings may make people vulnerable to motivation, performance, and well-being issues under some circumstances (e.g., Curran & Hill, 2018). Overall, then, perfectionistic concerns and perfectionistic strivings are distinct and both need to be taken into account when considering the likely consequences of perfectionism.
A small number of studies have examined perfectionism among musicians including professional musicians, amateur musicians, and talented adolescent musicians (Kenny, Davis, & Oates, 2004;Kobori, Yoshie, Kudo, & Ohtsuki, 2011;Stoeber & Eismann, 2007). The findings of these studies are generally consistent with research in other contexts. Specifically, research has found that perfectionistic strivings can be highly motivating and have some desirable achievement-related benefits (e.g., hours spent practicing and awards received associated with music; Stoeber & Eismann, 2007;Kobori, Yoshie, Kudo, & Ohtsuki, 2011). By contrast, perfectionistic concerns have no such benefits and are instead associated with more negative emotional experiences such as performance anxiety (Stoeber & Eismann, 2007). Qualitative research has also corroborated these findings with evidence that professional musicians (along with other elite performers) considered their perfectionism to be both central to their success and a source of significant problems in their professional and personal lives (Hill, Witcher, Gotwals, & Leyland, 2015).

Limitations of previous research
While studies are beginning to emerge that suggest perfectionism is important to the experiences of musicians, there is considerable scope for additional research. Two particularly notable areas that need to be addressed are that (a) research to date has focused exclusively on trait perfectionism and (b) previous studies have adopted cross-sectional designs.
In terms of the first limitation of existing research, although examination of trait perfectionism is most common, perfectionism is thought to manifest in a number of other ways. Hewitt et al. (2003) have argued, for example, that perfectionism is also evident in the manner in which individuals seek to present themselves to others. Perfectionistic self-presentation is a separate, expressive, and distinctly interpersonal aspect of perfectionism. It is an attempt to create and maintain an image of perfection in public settings. There are three facets of perfectionistic self-presentation: perfectionistic self-promotion (seeking to demonstrate one's perfection), non-display of imperfection (minimizing the public display of mistakes, flaws, and shortcomings), and non-disclosure of imperfection (minimizing admission of mistakes, flaws, and shortcomings). The first facet is thought to be approach-oriented and the other two facets are thought to be avoidance-oriented (i.e., one is underpinned by motivation to demonstrate competence and the other two by motivation to avoid demonstrating incompetence). In differentiating between perfectionistic self-presentation and dimensions of perfectionism like perfectionistic standards and strivings, Hewitt et al. consider perfectionistic self-presentation to provide "expressive" aspects of perfectionism, rather than "content-related" aspects (i.e., it is concerned with whether an individual seeks to project a perfect image to others, rather than whether someone pursues perfection).
Research examining perfectionistic self-presentation has provided a number of noteworthy findings. In particular, unlike for trait perfectionism, there is much less ambiguity regarding its implications as facets of perfectionistic self-presentation are almost always associated with maladaptive correlates, processes, and outcomes. This includes negative emotional experiences (e.g., negative affect and anxiety; Hewitt et al., 2003) as well as more severe pathological experiences (e.g., depression and suicide ideation; Flett, Besser & Hewitt, 2014;Roxborough et al., 2012). In addition, facets of perfectionistic self-presentation have been found to predict a range of outcomes after taking trait perfectionism into account (anxiety, depression, self-esteem; Hewitt et al., 2003). Finally, when considered independently (i.e., controlling for the relationship between the facets), the two avoidance-based facets (non-display and disclosure of imperfection) tend to be the most problematic (see Hewitt et al., 2003). Therefore, perfectionistic self-presentation is an important dimension of perfectionism that warrants examination alongside, and in addition to, trait perfectionism.
In terms of the second limitation of existing research, the weaknesses of cross-sectional designs are well documented. Cross-sectional designs do not allow inference of causality between variables as there is no temporal component in the design (i.e., all variables are measured at the same time point). In addition, as these designs provide only a static "snapshot" of the relationship, they offer no means of assessing whether the magnitude or direction of the relationships change over time or whether variables act on one another to varying degrees over time (i.e., whether reciprocal effects exist). Such reciprocal effects have begun to receive attention in perfectionism research with some evidence emerging of how perfectionism and its various outcomes may often influence each other (e.g., Nordin-Bates, Hill, Cummings, Aujla, & Redding, 2014). Longitudinal designs are required to examine reciprocal relationships and, although such designs do not have sufficient control to rule out the influence of other variables, they also provide a necessary further step towards establishing causal relationships.
To our knowledge, only one study has examined the relationship between perfectionistic selfpresentation and emotional experiences longitudinally (in the form of a broader concept of well-being). Specifically, Mackinnon and Sherry (2012) examined whether overall perfectionistic self-presentation (a combination of all three facets) mediated the relationship between perfectionistic concerns and well-being in undergraduate students over three time points. They found support for the proposed mediation and, importantly for the current study, they also found a negative relationship between overall perfectionistic self-presentation and well-being over time. In the current study we do not focus on trait perfectionism or mediation, but extend the model proposed by Mackinnon and Sherry by (1) examining the unique relationships of the three facets of perfectionistic self-presentation (rather than overall perfectionistic self-presentation) with emotional experiences, (2) examining both negative and positive emotional experiences over time (not just positive emotional experiences), and (3) examining possible reciprocal relationships between facets of perfectionistic self-presentation and emotional experiences over time (not just unidirectional relationships).

Present study
The purpose of the current study was to examine the relationships between facets of perfectionistic self-presentation and positive and negative emotional experiences in music students over time. Based on the above reasoning and research, it was hypothesized that facets of perfectionistic self-presentation would predict decreases in positive feelings and increases in negative feelings. In regards to reciprocal relationships, no hypotheses were offered as this element of the study was considered exploratory.

Participants and procedures
Participants were 143 (75 males, 68 females) students enrolled in the first year of music-related programs at three universities in the UK (age M = 18.92, SD 2.96, range 18-51). The musicrelated programs were similar in that they all had a broad musical curriculum incorporating elements of Western music history, theory and analysis, ethnomusicology and music psychology alongside performance and composition. Entry onto the programs was also based on academic qualifications rather than performance skill. All participants played one or more musical instruments. The most common instruments were voice, piano, and guitar. The average number of hours they reported practicing was 8.08 h per week (SD = 2.96 h). Participants completed a multi-section questionnaire that contained measures of perfectionistic self-presentation and emotional experiences at the beginning, middle, and end of the academic year (weeks 3, 10, and 15 of a 22-week academic year or weeks 1, 8, and 20 calendar months; all data collection +/-2 weeks). Of the 143 students in the study, 44 completed the questionnaire on one occasion, 21 on two occasions, and 78 on all three occasions. Institutional ethical approval was gained prior to conducting the research. Participants were recruited on a voluntary basis in taught sessions on their degree programs. All participants provided informed written consent.

Measures
Perfectionistic self-presentation. The Perfectionistic Self-presentation Scale developed by Hewitt et al. (2003) was used to measure a perfectionistic self-presentational style. The scale includes 27 items that measure the three facets of perfectionistic self-presentation: perfectionistic self-promotion (10 items; "I strive to look perfect to others"), non-display of imperfection (10 items; "I hate to make errors in public"), and non-disclosure of imperfection (7 items; "Admitting failure to others is the worst possible thing"). Responses are scored on a seven-point Likert scale (1 = strongly disagree to 7 = strongly agree). Hewitt et al. (2003) have provided evidence of the validity and reliability of the scale.
Emotional experiences. The Scale of Positive and Negative Experience (SPANE) developed by Diener et al. (2010) was used to measure feelings of well-being and ill-being. The scale includes 12 items that measure positive emotions (6 items; "Happy") and negative emotions (6 items; "Sad"). Respondents are asked to think about what they have been doing and experiencing during the last four weeks. Here they were instructed to think about their experiences on their university program. Responses are scored on a five-point Likert scale (1 = very rarely or never to 5 = very often or always). Two scores are derived as these experiences are considered partially independent (Diener et al., 2010). Diener et al. (2010) have provided evidence of the validity and reliability of the scale.

Analyses
The hypothesized model was examined using AMOS (24.0; Arbuckle, 2014). Full Information Maximum Likelihood (FIML) estimation was used to assess the model. FIML is an excellent means of estimation when data includes missing data (e.g., Enders & Bandalos, 2001). In the analyses, all variables were included as manifest variables. As in Mackinnon and Sherry (2012), both within-trait, cross-wave correlated error (e.g., perfectionistic self-promotion at time one error correlated with perfectionistic self-promotion at time two error) and same-trait, within-wave correlated error (e.g., perfectionistic self-promotion at time one error correlated with perfectionistic self-promotion at time two error) were included in the model. These correlated errors account for violations of the independence assumption within longitudinal designs (same-trait, within-wave correlated error) and common unmodelled explanatory factors (within-trait, cross-wave correlated errors; Cole & Maxwell, 2003). Fit of the proposed model was assessed using conventional criteria with adequate fit indicated when χ 2 p <. 05, χ 2 /df < 3, Comparative Fit Index (CFI) >. 90, Tucker-Lewis Index (TLI) > .90, and Root Mean Square Error of Approximation (RMSEA) < .10.

Preliminary analysis
The data were screened for univariate and multivariate outliers (see Tabachnick & Fidell, 2007). Standardized z-scores larger than 3.29 (p < .001, two-tailed) were used as criteria for univariate outliers and Mahalanobis distance χ 2 (15) = 37.70, p <.001, two-tailed, was used as criterion for multivariate outliers. This led to the removal of three participants. The remaining data (n = 140) were considered to be approximately univariate normal with three instances of non-normality: positive emotions time 2 (zskew = -2.22), negative emotions time 1 and 2 (zskew = 2.02 and 2.37). After transformation, these variables were normally distributed (all +/-SQRT transformations). Transformed variables and original variables were almost perfectly correlated. Transformed variables were used for bivariate correlations and path analysis. Finally, internal reliability analysis (Cronbach's alpha) was performed on each scale. All instruments demonstrated sufficient internal consistency (α ⩾ .70 for scales with 10 items or more and α ⩾ .60 for scales with five items or more; Loewenthal, 2001). Cronbach's alphas are displayed in Table 1.  Note. * p < .05, ** p < .01, two-tailed. Cronbach's α is reported on the diagonals.

Descriptive analyses and bivariate correlations
The descriptive statistics and bivariate correlations are displayed in Table 1. Participants scored low-to-moderate levels of perfectionistic self-promotion. Scores were highest for non-display of imperfection. Participants also reported moderate-to-high positive feelings and low-to moderate low-to-moderate negative feelings. Mean scores were similar across all three time points.
Bivariate correlations revealed statistically significant positive correlations between the facets of the perfectionistic self-presentation and statistically significant negative correlations between positive and negative feelings at all time points. In addition, at time 1, non-display and non-disclosure of imperfections had significant positive correlations with negative feelings. At time 2, non-display of imperfections had a significant positive correlation with negative feelings (though the size of the correlation for non-disclosure of imperfections was almost identical). At time 3, non-display and non-disclosure of imperfections had significant positive correlations with negative feelings. Additionally, all facets of perfectionistic self-presentation had a significant negative correlation with positive feelings.

Discussion
The purpose of the current study was to examine the relationships between facets of perfectionistic self-presentation and positive and negative emotional experiences in music students over time. It was hypothesized that facets of perfectionistic self-presentation would predict decreases in positive feelings and increases in negative feelings. Reciprocal relationships were also examined but this element of the study was considered exploratory.

Perfectionistic self-presentation on emotions over time
In support of the hypotheses, perfectionistic self-promotion at the start of the academic year predicted decreases in positive feelings in the middle of the academic year and non-disclosure of

Paths β
Time 1 to Time 2 Perfectionistic self-promotion to positive feelings −.27* Non-display of imperfection to positive feelings .20 Non-disclosure of imperfection to positive feelings −.14 Perfectionistic self-promotion to negative feelings .06 Non-display of imperfection to negative feelings .06 Non-disclosure of imperfection to negative feelings .12 Positive feelings to perfectionistic self-promotion .03 Negative feelings to perfectionistic self-promotion −.05 Positive feelings to Non-display of imperfection .03 Negative feelings to Non-display of imperfection .01 Positive feelings to Non-disclosure of imperfection .01 Negative feelings to Non-disclosure of imperfection −.06 Time 2 to Time 3 Perfectionistic self-promotion to positive feelings −.01 Non-display of imperfection to positive feelings .09 Non-disclosure of imperfection to positive feelings −.26* Perfectionistic self-promotion to negative feelings −.07 Non-display of imperfection to negative feelings −.01 Non-disclosure of imperfection to negative feelings .03 Positive feelings to perfectionistic self-promotion −.06 Negative feelings to perfectionistic self-promotion .11 Positive feelings to Non-display of imperfection −.05 Negative feelings to Non-display of imperfection .17 Positive feelings to Non-disclosure of imperfection −.03 Negative feelings to Non-disclosure of imperfection .21* Note. * p < .05, ** p < .01, two-tailed.
imperfection in the middle of the academic year predicted decreases in positive feelings at the end of the year. These particular findings are consistent with those of Mackinnon and Sherry (2012) who found total perfectionistic self-representation predicted decreases in total wellbeing over time (a composite of positive affect, negative affect, and life satisfaction). However, our findings also build on their work by indicating that in order to better understand the relationship between perfectionistic self-presentation and emotional experiences, distinguishing between its three facets may be required. In this regard, our findings allude to a more complex pattern of relationships whereby different facets of perfectionistic self-presentation are important at different times. Here, actively proclaiming perfection initially detracted from the development of positive feelings, later when students became more accustomed to the setting it was the more defensive concealment of shortcomings that was problematic.
Another noteworthy finding was that facets of perfectionistic self-presentation predicted changes in positive feelings but not changes in negative feelings. This is something that is potentially lost when examining total well-being. Why perfectionistic self-presentation was important to positive feelings but not negative feelings is not clear, especially when previous research has found facets of perfectionistic self-presentation to predict negative affect (e.g., Hewitt et al., 2003). However, it is important to note that it is not uncommon for individuals to express a mix of positive and negative emotions and the absence of positive emotions does not necessitate the presence of negative emotions or vice versa (Larsen, McGraw, & Cacippo, 2001). Therefore, instances when factors influence one but not the other are possible. Lower positive or negative emotions might, for example, be considered to reflect indifference ("I feel neither enthusiastic nor apprehensive about my study"). If so, here we may have identified a scenario in which facets of perfectionistic self-promotion are not sufficient to arouse increases in negative feelings but nonetheless detract from the development of more positive feelings.
In terms of reciprocal effects, negative feelings in the middle of the academic year predicted decreases in non-disclosure of imperfection at the end of the academic year. This is an especially novel finding in that it is the first instance, to our knowledge, in which a reciprocal effect involving perfectionistic self-presentation has been observed. In terms of possible explanations for this finding, it may be that negative feelings exacerbate interpersonal sensitivity and threat so to create a greater sense that deficiencies should be hidden and not shared with others. It is also possible that increasing negative feelings reinforce the low self-regard thought to underpin the need to hide deficiencies from others (see Hewitt et al., 2003). Regardless, in combination with other relationships in the model, the model provides evidence of a possible downward spiral of feelings and facets of perfectionistic self-presentation acting on each other in an undesirable manner over time.

Limitations and future research
The findings must be considered alongside the study's limitations. The current study examined the relationship between perfectionistic self-promotion and emotional experiences in a specific context (studying music at university). Future research may wish to explore the degree to which these findings extend to other settings and samples (e.g., conservatoire students). In the meantime, caution is required in regards to generalizing the findings beyond the current context.
The study also included a large amount of dropout across the three time points. Any systematic difference between students who completed the study and those who dropped out will influence the generalizability of the findings. For example, it is possible that the music students who did not complete some of the later questionnaires were not present in classes when the questionnaires were distributed. This could be for any number of reasons but might include factors relevant to the current study such as more negative emotional experiences on the program. As such, our findings may only apply to students who are more likely to attend and complete the first year of the program.
The modest sample size means smaller effect sizes were not statistically significant and, indeed, the ability to detect smaller effects decreased across time due to dropout. This is evident in that one of the relationships is notable in regards to its size but was not statistically significant (non-display of imperfection to positive feelings). This relationship is consistent with the findings that other facets of perfectionistic predict changes in positive feelings but not negative feelings. Employing strategies in future research to help retain participants across time points (e.g., participant incentives or targeted follow-up of non-completers) will help address this issue as well as help ensure a more representative sample.
Finally, the lack of control of other variables, a common problem in non-experimental research, means that unmeasured variables may account for the observed relationships. To address this issue, future research might include a wider array of variables and covariates. Based on research examining perfectionism, variables worth considering include trait perfectionism and perfectionistic cognitions (ruminative thoughts about the need to be perfect; Flett, Hewitt, Blankstein, & Gray, 1998).

Conclusion
The study examined whether perfectionistic self-presentation predicted changes in positive and negative emotional experiences in music students over time. It was found that the desire to present oneself perfectly and avoid disclosure of imperfections was related to decreases in positive feelings while studying music. No facets of perfectionistic self-presentation were associated with changes in negative feelings. As such, facets of perfectionistic self-presentation may influence the experience of students on music-related degrees primarily by decreasing positive feelings but not necessarily by affecting negative feelings. Attesting to the importance of examining these relationships over time, just as facets of perfectionistic self-presentation may act on positive feelings, negative feelings were found to increase a desire to avoid disclosure of imperfections during the academic year.

Authors' note
The data collection associated with this project was undertaken while the lead author was at the University of Leeds.