High Striving, High Costs? A Qualitative Examination of Perfectionism in High-Level Dance

ABSTRACT

This study provides the first in-depth investigation of how perfectionism is experienced in high-level dance. Seventy-seven students (M age = 15.52, SD = 2.30) completed a perfectionism questionnaire. Next, dancers with the highest and lowest levels of perfectionistic strivings (PS) and perfectionistic concerns (PC) were recruited for interview (N = 8), as representatives of the four subtypes of perfectionism in the 2 × 2 model. Distinct profiles emerged for the four quadrants of the 2 × 2 model in relation to four key themes: self-regulation, achievement goals, views on mistakes, and role of others. Dancers with high PS displayed the highest levels of self-regulation while dancers with low PC appeared most task-oriented. Dancers with high PC reported holding less favorable views on mistakes and placed greater emphasis on the opinions of others. Importantly, PS appeared to be both helpful and hurtful. Findings are discussed in relation to theory intermixed with practical recommendations.

KEYWORDS:

• Perfectionism
• Motivation
• Achievement striving
• Dance science
• Qualitative research

Introduction

Striving for ever higher levels of performance is an inherent part of the performing arts (Hall and Hill 2012). Researchers have therefore studied numerous constructs related to achievement striving, trying to elucidate what might constitute healthy or unhealthy striving. One construct that has yielded growing attention in recent years is perfectionism.

Classic descriptions of perfectionism include setting very high and exacting standards, and various forms of concern or doubt regarding actions and achievements (e.g. Frost et al. 1990). Our understanding of perfectionism has since been furthered by dividing it into perfectionistic strivings (PS) and perfectionistic concerns (PC) (Stoeber and Otto 2006). PS comprise the goal-directed aspect of perfectionism such as setting high goals, wanting to be perfect, holding oneself to high standards, and working very hard. PC are more varied; concern over mistakes is key, but PC may also comprise negative reactions to imperfection, doubts about actions, need for approval, and/or perceptions that others expect one to be perfect (see Stoeber and Madigan 2016 for a literature review).

Using the PS-PC distinction, research has shown that PC are associated with less favorable outcomes than PS (Hill et al. Forthcoming). In this review, Andrew Hill et al. (Forthcoming) re-analyzed 1772 effect sizes from 63 previously published studies into the correlates and effects of perfectionism in sport, dance, and exercise. It was concluded that PC confer no benefits and should be considered harmful to performers, given their association to problems with motivation, emotion, and performance. PS, by contrast, were characterized as complex and ambiguous in their associations with these same outcomes. It was noted, however, that the vast majority of studies underpinning these conclusions are from sport, and so it should not be assumed that findings are mirrored in dance.

While not as abundant as in sport, research into perfectionism in dance has also grown in recent years. What emerges from these studies is that, just like for athletes, PC are clearly detrimental to dancers, with correlates ranging from anxiety and debilitative imagery (Nordin-Bates et al. 2011) to complex problems such as disordered eating (Goodwin et al. 2014; Nordin-Bates et al. 2016) and burnout (Nordin-Bates, Raedeke, and Madigan 2017). The role of PS, however, requires further investigation; for instance, Nordin-Bates et al. (2016) found that not just PC but also PS predicted disordered eating.

The 2 × 2 Model of Perfectionism

Patrick Gaudreau and Amanda Thompson (2010) developed the 2 × 2 model of perfectionism, which allows researchers to examine the interactive effects of PS and PC. In the model, perfectionism is categorized into one of four subtypes: pure personal standards perfectionism (PSP; high PS, low PC), mixed perfectionism (high PS, high PC), pure evaluative concerns (EC; low PS, high PC), and non-perfectionism (low PS, low PC). Alongside the model, four hypotheses were proposed: pure PSP can be either healthier (H1a), less healthy (H1b) or equivalent (H1c) when compared to non-perfectionism; pure EC should be the unhealthiest of all perfectionism subtypes (H2); mixed perfectionism should be healthier than pure EC (H3); and mixed perfectionism should be less healthy than pure PSP (H4).

The 2 × 2 model, though not developed specifically for research into perfectionism in athletes, was eagerly adopted by the sport research community. Studies have supported several of its major tenets; indeed, the extensive review by Hill et al. (Forthcoming) revealed that H2 and H4 are strongly supported by research findings. However, research to support all four hypotheses posed by the 2 × 2 model is lacking. In fact, studies in dance (Cumming and Duda 2012; Nordin-Bates, Raedeke, and Madigan 2017; Quested, Cumming, and Duda 2014) have found no differences between mixed perfectionism and pure EC for a range of outcomes including fear of failure, body dissatisfaction, motivational regulations, and burnout. These findings question the notion that mixed perfectionism would be healthier than pure EC (H3).

Qualitative Perfectionism Research

Much of our current understanding of perfectionism is founded in questionnaire-based research. A disadvantage of questionnaires, however, is that they are limited to yielding brief, numeric answers on the questions asked. By contrast, interview methods have the potential to yield rich insights into real-life lived experience, including causal explanations and vivid examples (e.g. Patton 2002). They also allow for the emergence of unexpected data. Additionally, the small-scale purposive sampling typical of interview methods allows us to focus on the experiences of those with truly high levels of perfectionism (PS and/or PC) to a far greater extent than is possible in the larger-scale recruitments necessary for questionnaire-based work.

Encouragingly, interview methods have had a recent surge in performance perfectionism research (Gotwals and Spencer-Cavaliere 2014; Hill et al. 2015; Mallinson-Howard et al. 2018; Nordin-Bates and Abrahamsen 2016; Sellars, Evans, and Thomas 2016). Across these investigations, perfectionistic interviewees clearly held high standards (a defining characteristic of PS), and experienced frequent dissatisfaction (typical of PC). They typically saw perfectionism as having contributed to their successes, because it made them strive high and not settle for less. Negative effects of PC such as dissatisfaction, anxiety, pressure, and exhaustion were also recognized. They reported interpersonal difficulties such as not wanting to let others down or comparing themselves unfavorably to others, and often struggled with handling mistakes.

It should be noted that the aforementioned qualitative studies into perfectionism in sport and dance have been undertaken with adults (Gotwals and Spencer-Cavaliere 2014; Hill et al. 2015; Nordin-Bates and Abrahamsen 2016; Sellars, Evans, and Thomas 2016) or young athletes at lower levels of competition (Mallinson-Howard et al. 2018). Here, we provide the first qualitative study of perfectionism among high-level adolescent performers. Adolescence is a critical time in the talent development of young dancers: numerous physical, mental and social changes occur as part of maturation and identity formation, and dance training often intensifies. Therefore, understanding experiences of perfectionism among high-level adolescent dancers may make an important contribution to the literature.

Mallinson-Howard et al. (2018) were the first to use Gaudreau and Thompson’s (2010) 2 × 2 model of perfectionism in a qualitative investigation. Focus was on the experiences of young females from a range of sports. It was found that athletes representing different subtypes of perfectionism differed regarding the meaning given to sport participation, and regarding the social-environmental features highlighted as important. For instance, non-perfectionists and pure PSP athletes reported goals relating to social affiliation and personal improvement. The latter was related to achievement goal theory (e.g. Roberts 2012), where such self-referenced goals are considered the hallmark of task-involvement. Importantly, pure PSP and mixed perfectionism groups also strove to outdo others; in achievement goal theory, this is considered typical of ego-involvement.

To date, the only qualitative evidence relating specifically to dancers’ perfectionism is a case study of a recently retired ballerina, who was a self-identified perfectionist (Nordin-Bates and Abrahamsen 2016). This acclaimed international performer attributed much of her success to perfectionism and saw it as a strong driving force. At the same time, she gave examples of how perfectionism had led to crippling self-criticism, anxiety, and injury. Because of its design, however, that study could not clarify the extent to which PS vs. PC led to particular outcomes, and did not link its findings to the 2 × 2 model of perfectionism. Extending qualitative perfectionism work in dance appears important, not least because dancers appear to be highly susceptible to perfectionistic tendencies (Cumming and Duda 2012; Nordin-Bates and Abrahamsen 2016; Nordin-Bates et al. 2011). It is our hope, therefore, that the information gained in the present investigation will be of use to dance teachers and researchers alike, in our shared striving to make dance education promote high levels of both performance and health.

The Present Study

The aim of the present study was to investigate experiences of perfectionism in a high-level dance context, using the 2 × 2 model of perfectionism as a guiding framework (Gaudreau and Thompson 2010). That is, we explored the experiences of dancers identified as representatives of different perfectionism subtypes. Broadly following the procedures of previous studies (Gotwals and Spencer-Cavaliere 2014; Mallinson-Howard et al. 2018), we (a) administered questionnaires to a high-level dance school sample, (b) identified dancers representing all four quadrants of the 2 × 2 model, and (c) used qualitative inquiry to explore their experiences of perfectionism.

Methods

Like previous studies (Gotwals and Spencer-Cavaliere 2014; Mallinson-Howard et al. 2018), we employed a two-phase mixed-methods design where Phase 1 helped identify participants with a particular perfectionism profile for interview. In Phase 2 we explored experiences of perfectionism among dancers identified in Phase 1. Participants were from a top-level, nationally recruiting ballet school in Sweden which provides full-time education (dance and academic schooling 5–6 days/week) for students aged 10–19. Entry is by audition, and dancers must re-audition every three years to keep their place. Students take classes in both classical ballet and modern dance, though the emphasis on ballet is pronounced for those aged 10–16. For the last three years students focus on either ballet or modern dance, but continue to take classes in both.

Participants

Participants in Phase 1 were 77 dancers at a nationally recruiting Swedish ballet school, representing 74% of the total number of students of the target age range (aged 12 years and above). Dancers were 12 to 19 years old (M = 15.52, SD = 2.30), 77% female, and had 9.96 (mean) years of dance experience (SD = 3.29). In Phase 2, eight of these dancers participated; for further information, see Table 1.

Table 1. Interviewee descriptives by quadrant of the 2 × 2 model of perfectionism

Pure Personal Standards Perfectionism	Mixed Perfectionism
n = 1	n = 3
(Girl, 18)	(Three girls: 12, 15 & 18)
● High self-regulation	● High self-regulation
● Task-oriented● Some rigidity	● Some task-orientation, some avoidance goals
● Positive view of mistakes	● Doubts & uncertainty
	● Others' opinions very important
	● Mixed views of mistakes
Non-Perfectionism	Pure Evaluative Concerns
n = 3	n = 1
(Girl, 19; two boys, 12)	(Boy, 12)
● Task-oriented	● Low self-regulation
● Positive view of mistakes	● Low or varied strivings
	● Some rigidity
● Younger: low self-regulation	● Doubts & uncertainty
● Older: high self-regulation	● Others' opinions very
	important
	● Negative view of mistakes

Procedure

The study was granted ethical approval prior to commencement, as well as approval by school leaders. The first author also discussed study design with the schools’ health team and gave information to teachers.

Phase 1: Quantitative Identification of Participants

The 10-item version of the Multidimensional Inventory of Perfectionism in Sport (MIPS; Stoeber et al. 2007) was used to capture striving for perfection and negative reactions to imperfection in training situations, the latter being a proxy for perfectionistic concerns (Stoeber and Madigan 2016). As these authors point out, several subscales on several measures may be used as indicators of perfectionistic strivings and concerns; we chose the 10-item MIPS for its clarity, conciseness, applicability to dance, and evidence of validity and reliability (Stoeber and Madigan 2016). We adapted the MIPS for use in our study by prefacing questions with “During dance training … ” and by replacing original words relating to sport training/competition for dance training/class. Items are scored on a 1–6 Likert scale ranging from 1 (Never) to 6 (Always). Internal reliabilities were high for both subscales (strivings α = .83, concerns α = .84).

The MIPS was translated into Swedish using a translate-backtranslate procedure with two bilingual persons. Following discussions with staff, we intended to recruit students aged 12 and over in order to ascertain understanding. Therefore, we piloted the questionnaire with a sample of 13 high-level gymnasts of a similar age, who were encouraged to ask questions and point out any unclear words or phrases.¹ Their coach also provided written feedback on item clarity. Our intention to recruit participants from age 12 was confirmed as rational because pilot participants understood the MIPS well, and no changes or clarifications were required for the questionnaire.

School staff organized times and classrooms for questionnaire completion, but were not present during data collections. Instead, one or both of the authors held these, depending on class size. Information was given verbally and in writing, and questions (e.g. to clarify an item) were encouraged. In line with Swedish requirements for ethical vetting, informed consent was given by all participants and by parents of students under 15. To enable interview recruitment, students were given anonymous ID-codes. The school counselor provided contact details and room access, but was not informed on what grounds a student was identified for interview.

Phase 2: Interviews

Interviews were founded in a relativist, social constructionist approach; as such, we acknowledge that the results obtained and presented here are not unbiased or ultimately “true” in any objective sense (Smith and McGannon 2018). Instead, multiple realities and stories may co-exist, yet are all valuable for what they are: that is, examples of possible experiences. To help obtain in-depth accounts, an interview guide² was constructed based on existing literature, including notable gaps. Interviews were semi-structured, comprising open-ended questions regarding background (views on dance, characteristics of successful dancers; e.g. “Do you have a particular goal or dream for your dancing?”), goals and perfectionism (strivings, reactions to mistakes, experiences of perfectionism; e.g. “How would you describe a really good class?”; “Sometimes we make mistakes, do things wrong and don’t reach our goals. Can you describe such a situation?”), and two sections on creativity (the results from which are included in a separate paper; Nordin-Bates 2019). Clarification- and elaboration probes were used to elicit deeper descriptions and examples (Patton 2002).

Interviewees were first asked a series of open-ended questions to encourage descriptions of perfection, perfectionism, and their experiences around these phenomena in their own words. Thereafter, they were shown short lists of what might comprise perfectionistic strivings (have very high goals; want to be perfect; working very hard) and perfectionistic concerns (worried, concerned over mistakes; feeling of not being good enough; unsure of one’s choices), and asked whether they agreed. Importantly, interviewees were informed that these were some people’s views and that they were free to disagree with some, all, or none of it. Introducing these concepts during interviews was done to help ensure rich material was obtained on these topics.

Pilot interviews were held with two dancers, two teachers, and one dance researcher. These helped test question wording and flow, and improved familiarity with the interview guide. As a result, minor alterations were made. The researchers also listened to and discussed the recordings of each other’s pilot interviews to further hone interview skills.

Potential interview recruits were contacted via e-mail and/or mobile phone. The second author conducted six of these interviews and the first author two. Seven interviews were conducted in person and one was via Skype. Interviews were conducted in a quiet space, digitally recorded, and lasted 85–137 minutes (M = 108.38).

Interviewees were encouraged to be honest and to ask if anything was unclear, informed that they had a right to stop at any time or not answer a particular question, and that recordings would be treated confidentially. It was emphasized that there were no right or wrong answers but that we were simply interested in individual experiences. At the end of each interview the researcher summarized main points, asked whether this summary adequately covered the respondents’ true experiences and opinions, and whether they wanted to add anything.

Analysis

Phase 1: Quantitative Identification of Interview Recruits and Baseline Data

Means and standard deviations for the full sample and for interviewees are provided in Table 2. In line with the 2 × 2 model, participants were identified as potential recruits for Phase 2 if they recorded especially high or low perfectionistic strivings (PS) and perfectionistic concerns (PC). To identify a realistic number of interviewees, a number of criteria were considered, including the 33^rd and 66^th percentiles (as per Gotwals and Spencer-Cavaliere 2014). Because this generated a large number of potential recruits (N = 34) and because we wanted to make interviewees more distinct (i.e. more extreme scores, rather than close to the population average), the 20^th and 80^th percentiles were inspected. This identified 13 dancers, which was deemed realistic. Of these 13, five were mixed perfectionists (i.e. high PS, high PC) while six were non-perfectionists (low PS, low PC). Only one dancer was identified as pure PSP (high PS, low PC), and one as pure EC (low PS, high PC). We were able to recruit three mixed perfectionists, three non-perfectionists, and both the pure PSP and the pure EC dancers (see Table 1).

Table 2. Descriptive statistics for the full sample and by quadrant of the 2 × 2 model of perfectionism

	Full Sample	Pure Personal Standards Perfectionism	Mixed Perfectionism	Non-Perfectionism	Pure Evaluative Concerns
	M (SD)	M	M	M	M
Perfectionistic Strivings	4.04 (1.09)	5.60	5.53	2.27	2.80
Perfectionistic Concerns	3.00 (1.05)	2.00	5.27	1.80	4.00

Descriptive statistics are displayed with means (M) and standard deviations (SD) for the overall sample, but only with means (M) for the four quadrants (interviewees) because these represent only 1–3 persons each. Perfectionism questions were scored on a scale from 1–6.

Phase 2: Interviews

Interviews were transcribed verbatim and analyzed using the qualitative software NVivo. Meaning units relevant to the study topics were labeled with words close to the participants’ own and then assigned to the relevant quadrants of the 2 × 2 model of perfectionism. Within each quadrant, meaning units were then grouped deductively into two main categories: high or low PS and PC, as appropriate. Within each of these, inductive reasoning was used to categorize themes (Côté et al. 1993; Patton 2002). For instance, no interview questions focused on self-regulation, but when numerous quotes relating to structured goal setting, hard work and self-reliance emerged, it made sense to group them under this theoretical label. Theoretical saturation was deemed as reached when meaning units fitted adequately into the overall results structure (Côté et al. 1993). The results of this analysis are illustrated in Table 1. Interviews, transcription and analyses were undertaken in Swedish, and quotes were translated into English by the first author, who is bilingual, when selected to appear in this paper.

In line with recent recommendations (Smith and McGannon 2018) and the relativist approach, trustworthiness criteria were used in a list-like rather than a rigid, universal manner. That is, they were chosen specifically for their suitability and relevance to our study. Our first trustworthiness-initiative was the triangulation obtained by collecting both quantitative and qualitative data. For instance, asking students in-depth questions regarding their experiences led to a comprehensive view of the extent to which the two forms of data concurred. Second, extensive piloting helped inform our interview guide and technique, optimizing clarity and flow. Third, participants were assured that responses would be kept confidential and data only presented in anonymized forms, encouraging them to speak freely. Fourth, the first author has a background in dance, psychology research and of consulting with dancers, while the second has some dance experience in addition to an extensive background in elite esthetic sports as performer, coach, and journalist. Hence, we believe that we were able to empathize, establish good rapport and use appropriate prompts. At the same time, we are aware of the potential for subjective bias; indeed, the results are necessarily affected by the views and experiences of researchers as well as participants. Fifth, the second author separately coded two transcripts to ensure agreement on coding density. She also critically examined several versions of the emerging results structure (i.e. critical friend-technique; Smith and McGannon 2018). Sixth and finally, exact quotes are presented, allowing readers to form their own interpretations (Sparkes 1998).

Quantitative Results

The overall sample reported moderate to somewhat high perfectionism scores, with PS being higher than PC (see Table 2). As expected, PS and PC correlated strongly and positively (r = .49, p < .001). Also shown in Table 2 are mean scores for interviewees. Their distinct differences in PS and PC means compared to one another as well as to the overall sample gave confidence that interview recruits were indeed different in a manner congruent with the 2 × 2 model (Gaudreau and Thompson 2010).

Qualitative Results

Participants all agreed that the descriptions of PS and PC shown during interview were accurate and made sense, and that the 2 × 2 model was applicable to dance and to their experiences. This gave us further confidence in presenting results using this model.

Pure Personal Standards Perfectionism (Pure PSP)

The high PS of the dancer representing pure PSP was characterized by high self-regulation (see Table 1), as illustrated via accounts of high goals, a task-oriented and self-reliant attitude, and the value put on demanding teachers and disciplined hard work. For instance, she expressed that “I must always stay [after class] and practice because that is very important if I want to improve. But you have to plan your time and when you should do it.” Her goals were both short- and long-term, and seen as crucial: “ … very, very important … Everyone has to have a goal to become what they want to become. There is no meaning in life either, I think, if you have no goal.”

She was clear that striving for perfection was key to get anywhere, and consequently did so – though while seemingly retaining a task-oriented focus:

… every day. Really. Because I always have in my head this “you must improve; you must be even better than what you are; you must improve this; you must walk nicer or run nicer”. … and “if you want to be as good as possible you must work, work, work, work, and every day in class you must do your best.”

Her self-reliant, task-oriented attitude was further evident in statements about how it was fun and informative to watch others, and of how she typically relied on her own “feel” to assess goal progress.

Despite her positive, self-regulating characteristics, rigidity was evident in the way she expressed a strong respect for authority, a wish that things should be “just right,” frequent use of words such as “ought” and “must,” and an apparent need to work constantly, almost obsessively: “ … always. Actually always, always; every second, minute and hour is important to me … ” She also felt that very high levels of PS: “ … might wear you down, that you … break down from maybe working well and hard and might be trying and trying and trying. And then you might get injuries, or you might get really tired.”

Her low PC were illustrated via a constructive view of mistakes: “if you maybe make a mistake it definitely doesn’t affect the whole rehearsal. Yes, you made a mistake, you just carry on, so it doesn’t matter [waves her hand].”

Mixed Perfectionism

These dancers also exhibited high levels of self-regulation, for instance by setting high goals and working hard in a task-oriented manner. They also indicated that PS could lead to success: “I mean you get further, I think, if you are perfectionistic … I have always been this way, striving for perfection. I think 99% of dancers are very much like that. You can’t get anywhere if you’re not.” At the same time, they reported that very high levels of PS could lead to being ill:

I mean you can get depressed or something … I think it goes with not letting yourself live, you know; that you live under some sort of condition that everything you do must be according to some ideal. … I think you can be a perfectionist but that you should also be able to let it go a bit, so that you feel well, I think.

The mixed perfectionists also gave examples of avoidance goals (“I always try to think that I shouldn’t do worse than yesterday”) and several descriptions of how goals must not be too high. It was not possible to discern whether the latter reflected healthy flexibility or fear of disappointment.

The heightened levels of these dancers’ PC was particularly evident in the emphasis on the opinions of others. Teacher approval was important, resulting in a longing for praise or tendency to see teachers as the main source of goal evaluation. Moreover, peer approval mattered: “Sometimes you don’t want to do your best because they think it turns out odd if you do your best. Because if you do your best it might look ‘different’.” There were also many examples of how they perceived other people, or dance itself, to have heightened their perfectionism.

The effects of their elevated PC included problems with motivation (“If you don’t feel good enough … then it is sort of like you give up a bit”) and avoidance (“It is usually when I am thinking negatively, because then I avoid trying because I am worried about making a mistake.”). Generally, they held mixed views of mistakes: several examples of mistakes being undesirable emerged, but so did examples of how mistakes were no big deal. One dancer explained that this may vary depending on the degree of support experienced:

We’re very supportive, in this class at least, so nobody would give you strange looks or anything … you don’t have to be ashamed to do things wrong, so that is really good. … if you are surrounded by people who would judge you if you did it wrong then you don’t dare do it …

A number of responses related to strategies used to re-focus after making mistakes, such as rationalizing or making an effort to think positively. However, the perceived success of these strategies varied. As a final point, there was considerable diversity in the degree of PC expressed: many accounts of high-PC characteristics were from a single person, who was also the youngest. An older dancer described how she had experienced more PC before, but learnt to temper them:

I listened and did everything the teacher wanted, was very afraid of doing anything wrong and was incredibly focused – but I feel I still am that, but I … I have begun to understand how you can handle mistakes. Because everyone makes mistakes, even professionals, and you have to learn how you handle [them] and how you can get yourself out of it, and not just stand there and get completely frustrated. But, I as a perfectionist now, hmm; it is another kind of perfectionism I think. Before, it was a bit sick really, I think, if I think back: I was so very “I must keep within the boundaries” and “I must absolutely not miss anything”. … but now it is like, yes I must be perfect – or not must, but I do want to be perfect. But there is like another way to get there.

Pure Evaluative Concerns (Pure EC)

While the strivings of the dancer representing pure EC were not perfectionistic, they were more varied than simply low. Goals were generally vague (“ … to get as good as I can”), dependent on others (“To get a ‘good’”; i.e. positive teacher feedback) or absent (“ … I have probably had some goals but I can’t think of anything right now”). This was despite recognition that “That is when you work the hardest, I think, when you really have a goal”. Thus, self-regulation appeared to be low. Strivings varied with motivation but it appeared that higher strivings, even if temporary, could easily become rigid:

When I actually do it, it feels like I have to; then it has to be perfect … and then it can sometimes feel good because then I feel “wow, I really want this,” and that is good. But sometimes it can feel hard … that it has to be perfect, and that can be really tough.

This rigidity, or inability to set high yet achievable goals, appeared closely linked to his elevated PC. For instance, he gave the following example of doubts and uncertainty: “ … it’s better to think ‘what should I do next time so that what I did before will get better?’, but often I think ‘if I’d done that, it would have been better.’” His PC also included a somewhat negative view of mistakes and a strong focus on the opinions of others. For instance, he addressed feedback to obtain a positive comment from a teacher rather than for his own sake. If his work was not recognized, this led to disappointment and a feeling that the effort was not worthwhile. His elevated PC also led to a fear of “sticking out”: “It happens a lot, that when I do certain things I do them maybe not exactly the way I would like because … you don’t know what the class would say. Or what the teacher would say … ” He felt influenced by both teachers and peers to become more perfectionistic, despite teacher efforts to curb his PC:

… they’ve said to me a few times that they think I am perhaps hard on myself, so then they are encouraging me to be less perfectionistic … but it is more encouragement that I should become more perfectionistic, even though they don’t mean it … because in the lessons we have every day, they can bring up all the time “ah, but those in Russia” or “those in China” … But they never bring up “don’t be hard on yourselves.”

Non-Perfectionism

These dancers differed considerably in self-regulation, though shared a generally task-oriented and positive attitude. They could, for instance, give examples of being pleased about having learned new things. Their low PC were evident in the lack of quotes on this theme, and via their positive view of mistakes. Speaking of falling over in class, one dancer explained:

… if it happens that you do a jump and fall you just feel like “ok, I fell; I’ll stand up and carry on, nothing has happened.” It is more that everyone just “oh, how did it go?” – that’s a bit annoying because everyone is just “oh my god, are you ok?” And I’m just like “it’s fine, take it easy.”

Beyond such attitudes, dancers within the non-perfectionism quadrant were starkly heterogeneous; in particular, it appeared that self-regulation differed between the youngest and oldest respondents. The low self-regulation of the youngest dancers was evident in their lack of reflection (“I don’t think that much about classes once I’ve done them”) and their often vague or non-existent goals. They agreed that goals would likely lead to greater progress and gave examples of striving more at particular times (e.g. before shows), but found this hard: “ … it is very hard to stand there for half an hour and try to be flawless”.

In stark contrast to her younger peers, one young woman emerged as the most reflective of all interviewees, and she clearly denounced perfection as a worthwhile goal. Asked about the degree to which she strove for perfection, she replied:

Not much! (laugh). I don’t find it very interesting. … I mean, I think perfection is different for everyone. … it is more interesting to search for other qualities … perfection, that feels like you are aiming for something that is the best, like. I think it is a bit flatter than that; [other] things are equivalent, or just as good.

Interestingly, she outlined how her own strivings had changed over time:

When I was little … I wanted so dearly, like, always to perform at my peak, whatever I did. And I was a bit of a competitor: I always wanted to be the best at everything I did. But then I realized that there were more interesting things … I became a bit more secure in myself, too: I didn’t have to do it … I am good as I am.

Having reflected extensively on these matters, she also gave several examples of how friends with high PS appeared “restricted” or “inhibited,” and of how a perfectionistic friend at school had developed an eating disorder.

Her high self-regulation was evident in extensive goal-setting, which included high long-term goals (e.g. contributing to the modern dance scene) and specific, short-term goals (e.g. shoulder blade placement). She spoke of how these goals helped her stay motivated, gave her something to work toward, and led to satisfaction. She was also highly self-reliant, for example in how she worked in a class where the teachers’ style contrasted with her own:

She gave a bit of an atmosphere in the class that it was a lot about performing, to produce cool steps and things she thought looked nice. And I’m not so interested in that (laughs). So, I don’t know; I guess it made me not listen to everything she said, but rather I worked more for myself.

Discussion

Using the 2 × 2 model of perfectionism as a guiding framework (Gaudreau and Thompson 2010), the aim of this study was to investigate experiences of perfectionism in high-level dance. More specifically, we explored the experiences of dancers identified as representatives of the four different perfectionism subtypes outlined in the 2 × 2 model.

High-level Adolescent Dancers’ Experiences of Perfectionism

The 2 × 2 model indicates that the four subtypes of perfectionism differ in their motivational underpinnings (Gaudreau and Verner-Filion 2012). Just like Mallinson-Howard et al. (2018), we found support for the notion that this extends beyond motivational regulations (see e.g. Nordin-Bates, Raedeke, and Madigan 2017; Quested, Cumming, and Duda 2014) into broader conceptualizations of goals and purposes of dance or sport participation. In this particular study, the following four themes describing such characteristics emerged: self-regulation, achievement goals, views on mistakes and the role of others.

Self-regulation is an effortful, deliberate process involving initiative and planning how to reach goals, focusing attention and effort, and being persistent and determined (Duda, Cumming, and Balaguer 2005). Although it emerged inductively, self-regulation appeared to be an ideal label for many of our findings. In particular, pure PSP and mixed perfectionist dancers could be characterized by their high self-regulation: they set high, specific goals and worked hard to reach them, and reported engaging in planning and preparation. The pure EC dancer and his similarly young, non-perfectionistic peers instead appeared to self-regulate to a low extent, with low, vague or absent goals.

Notably, there was great variation among non-perfectionists, with this quadrant also comprising an older, highly self-regulating dancer. This highly reflective individual was striving for excellence, with many of the hallmarks of PS described in the literature (e.g., high striving, lofty goals, hard work). At the same time, she plainly stated that perfection was not a worthwhile goal. This contrasts previous accounts of non-perfectionists, who characterized them in terms of narcissism, low self-discipline, and high distractibility (e.g. Parker 1997). It is, however, in line with recent findings from Mallinson-Howard et al. (2018), who described non-perfectionistic young athletes as participating for the sake of learning, self-development, and social affiliation. What we add to this emerging picture of non-perfectionists in the literature is that such participation motives can sit alongside self-regulation in the pursuit of excellence. From an applied perspective, it also highlights that striving for perfection is by no means the only way to strive high. Therefore, dance teachers should guide dancers toward goals that are high yet realistic and, not least, individually meaningful (see Table 3). When goals are reached, but also when simply working well toward their goals, dancers should be encouraged to feel proud and satisfied.

Table 3. Characteristics and Potential Support Strategies for Dancers with Different Combinations of Perfectionistic Strivings and Concerns

	Characteristics		Potential Support Strategies
	Perfectionistic Strivings	Perfectionistic Concerns	Encourage Healthy Striving	Alleviate Concerns
Pure Personal Standards Perfectionism	• Sets very high goals • Wants to be perfect • Holds self to veryhigh standards • Works very hard		Help set realistic, task-oriented and meaningful goals Nurture intrinsic motivation Watch out for goals and practice habits becoming rigid Point out importance of rest and recovery Encourage work-life balance
Mixed Perfectionism		Concerned about making mistakes Reacts negatively when feeling imperfect Uncertain about “the right thing to do” Highly dependent on others for approval Perceives pressures to be perfect		Role model a positive view of mistakes Focus class structure and feedback on self-improvement Encourage dancers to give themselves feedback and not always rely on teachers Help dancers focus on the present moment (e.g., artistry) Include creative exploration Meet avoidance with personalized tasks
Pure Evaluative Concerns

Characteristics are derived from perfectionism literature (e.g., Gaudreau and Thompson 2010; Stoeber and Madigan 2016; Stoeber and Otto 2006). Potential support strategies are formulated on the basis of findings from the present study alongside literature specifically focused on perfectionism in dance (e.g., Karin and Nordin-Bates 2019; Nordin-Bates 2019; Nordin-Bates and Abrahamsen 2016).

Duda, Cumming, and Balaguer (2005) outlined how there are logical links between self-regulation and achievement goals. In our findings, the achievement goals of both pure PSP and non-perfectionist dancers appeared to be task-oriented: for instance, they spoke of the pleasure in learning and progressing. While mixed perfectionist dancers described their strivings in partly task-oriented terms, they also gave examples of avoidance goals. John Gotwals and Nancy Spencer-Cavaliere (2014), too, found that “healthy perfectionists” were more approach-oriented and “unhealthy perfectionists” more avoidance-oriented.

Interestingly, Gotwals and Spencer-Cavaliere (2014) reported primarily on performance avoidance among athletes while our work (Nordin-Bates and Abrahamsen 2016 and the present results) has focused on mastery avoidance among dancers. Mallinson-Howard et al. (2018) further extended this picture, outlining both pure PSP and mixed perfectionist athletes as striving for performance approach goals (e.g. outperforming others). Future research may be able to discern whether these differences are study-specific or perhaps represent differences between dance and sport. In the studio, dance teachers may wish to keep a watchful eye on emerging avoidance, as it may signal underlying PC. When avoidance is observed, teachers can perhaps offer multiple, personalized entry points into an exercise so that students feel seen and challenged at a suitable level (see Table 3).

Pure PSP and non-perfectionistic dancers spoke little about mistakes and, when they did, described them in largely positive terms. It is possible that this is because they, like the “healthy perfectionists” interviewed by Gotwals and Spencer-Cavaliere (2014), got over their mistakes quickly. In contrast, pure EC and mixed perfectionist dancers held mixed or negative views of mistakes, and spent more time speaking about mistakes and about other people. They cared greatly about what others thought of them, and placed an importance on teacher approval, which at times appeared unhealthily needy. By contrast, the pure PSP dancer relied more on sensory feedback. This difference in reliance on oneself vs. others between dancers representing different perfectionism subtypes was distinct in our data, and has, to our knowledge, not been reported in sport. Mallinson-Howard et al. (2018) have, instead, described how pure EC athletes may desire minimal coach involvement. It would be interesting to examine whether over-reliance on teacher opinions is a result of the teacher-led instruction style often associated with classical ballet. Regardless, we would encourage teachers with such “needy” students to try and reduce teacher reliance (see Table 3). For instance, they could tell students to listen to their bodies, what feels right, and to trust their own judgment rather than always looking to the teacher first.

The Present Findings and the 2 × 2 Model

It became apparent that PS can energize and drive a performer forward, but are not entirely positive. For instance, rigidity and perhaps even obsessiveness appeared to accompany the work ethic of the pure PSP dancer, supporting findings by Andrew Hill et al. (2015). She, and several others, also spoke of how PS might lead to problems such as exhaustion, injury through overwork, and in various ways “restricting” a person, which resembles other findings in the performing arts (Hill et al. 2015; Nordin-Bates and Abrahamsen 2016). Such findings are also in line with the mixed evidence presented for hypotheses H1 and H3 in the quantitative perfectionism literature (Hill et al. Forthcoming). In Table 3, we suggest a number of strategies which may help teachers support a healthy striving for excellence, rather than perfection.

It was interesting that the highly self-regulating non-perfectionist dancer strongly questioned perfection as a suitable goal for dancers, and linked PS to low self-esteem. In sum, we have reservations about Gaudreau and Thompson’s (2010) hypothesis that pure PSP is healthier than non-perfectionism (H1a) and do not agree that PS should be promoted (Sellars, Evans, and Thomas 2016). Instead, we propose that PS may be positive to a point (e.g. by contributing to hard work), beyond which it becomes counterproductive (e.g. by leading to rigidity, overwork or injury) – that is, a curvilinear relationship. Quantitative researchers may wish to test such a relationship between PS and its potential outcomes, but it will need to be done via novel methods, because there is no way to identify such a relationship using the standard statistical techniques upon which the majority of perfectionism studies rely.

Interviewees with elevated levels of PC reported problems with motivation, avoidance, and over-dependence on other people, all of which are in line with previous qualitative (Gotwals and Spencer-Cavaliere 2014; Hill et al. 2015; Nordin-Bates and Abrahamsen 2016; Sellars, Evans, and Thomas 2016) and quantitative literature (Hill et al. Forthcoming). Yet while unhealthy perfectionists in sport have also reported worries over the perceptions of important others (Gotwals and Spencer-Cavaliere 2014; Sellars, Evans, and Thomas 2016), the problem might be more pronounced in dance, where how one moves is not only important but also subjectively assessed and linked to largely unchangeable aspects (e.g. appearance). This differs from activities where an outcome (i.e. ball in goal) is more important than how one gets to that outcome, or looks while doing so. As such, dancers with PC may be especially likely to conform rather than stay true to their own movement style – let alone innovate. Further research is required to test this speculation. For now, our suggested strategies for alleviating perfectionistic concerns (see Table 3) include helping dancers focus on present-moment aspects such as artistry, and to include creative exploration in classes as doing so may help reduce perfectionistic thinking (Karin and Nordin-Bates 2019; Nordin-Bates 2019).

Methodological Considerations

It is striking that all qualitative studies to date have struggled to recruit persons with pure PSP (or equivalent). In the present study, just one out of 77 dancers met criteria for representing this quadrant, as compared to five for mixed perfectionism and six for non-perfectionism. Paul Sellars, Evans, and Thomas (2016) was not able to identify any so-called “healthy perfectionists” in a sample of 67 elite athletes, and Gotwals and Spencer-Cavaliere (2014) had to widen their sampling frame considerably to find enough “healthy perfectionists”. Even when considering their sampling frame (33^th and 66^th percentiles), we only identified three dancers with pure PSP/“healthy perfectionism”; that is, a prevalence of between 1% (using our sampling frame) and 4% (using the frame of Gotwals and Spencer-Cavaliere 2014) in a population often considered prone to perfectionistic tendencies. Yet this state of affairs is no surprise because PS and PC are almost always strongly and positively correlated. Further research, such as a clinical case study of a pure PSP individual, would be of great interest. In particular, how do they manage not to experience concerns, despite their high strivings? Are they able to keep this healthy striving also over time, and when encountering difficulties? And, not least, how can a teacher help dancers with high levels of PS from succumbing to PC over time?

Accounts from several interviewees agreed with previous investigations that perfectionistic tendencies are commonplace among dancers (Cumming and Duda 2012; Nordin-Bates and Abrahamsen 2016; Nordin-Bates et al. 2011). Moreover, interviewees recruited for high PS did strive for perfection, unlike participants in Gotwals and Spencer-Cavaliere (2014). These findings, coupled with the comparative lack of research in dance, strengthens the rationale for further study of perfectionism in this domain. The key limitations of our study include the single interviewees representing pure PSP and pure EC; authors of future studies may wish to recruit larger numbers of such individuals, though such recruitment must be balanced against the desired width of the sampling frame. Another limitation is the relative (not absolute) scores upon which interviewees were selected; a necessary consequence of the lack of cutoff scores for the MIPS (Stoeber et al. 2007). Still, the fact that interviewees identified via MIPS scores differed in a manner congruent with the 2 × 2 model suggests that it was suitable for our purposes.

While not a test of the 2 × 2 model in a traditional, quantitative, hypothetico-deductive sense, this study aligns with and extends an emerging trend in the literature (e.g. Gotwals and Spencer-Cavaliere 2014; Mallinson-Howard et al. 2018; Sellars, Evans, and Thomas 2016) which opens up possibilities for models to be examined and questioned by the real-life persons they are designed to represent. And while recruiting non-perfectionists for a study of perfectionism may not sound ground-breaking, the heterogeneity of this quadrant helps us question how the terms perfection and perfectionism are used and measured. That is, might persons like our highly self-regulating non-perfectionist be (mis-)identified as perfectionists in questionnaires which do not explicitly use the word “perfection,” but rather ask about striving for excellence or high standards, as several popular instruments do? (e.g. Frost et al. 1990; Gotwals and Dunn 2009).

Issues such as these, the difficulties in finding persons with pure PSP and pure EC, and the mixed support for the 2 × 2 model hypotheses, must all be kept in mind when designing future research in this area. Interestingly, however, all interviewees agreed that Gaudreau and Thompson’s (2010) 2 × 2 model of perfectionism made sense to them and for dance. Therefore, dance faculty may wish to introduce the concepts of PS and PC to students and facilitate a discussion of what these might feel like and lead to, short-term as well as long-term. Teachers may also wish to use Table 3 as a tool in their own practice.

In summary, results from the present investigation suggest that the experiences of dancers representing different quadrants of the 2 × 2 model of perfectionism (Gaudreau and Thompson 2010) differ in ways which may have bearing on performance success as well as well-being. Findings relating to PC were largely in line with previous research, in that they illustrate how persons with PC frequently suffer from a range of problems. Findings relating to PS were more nuanced, and as such they add to the complex, emergent picture of what it is like for a performer to strive for perfection.

Additional information

Funding

The financial support of The Swedish Research Council for Sport Science is gratefully acknowledged [P2015-0047].

Notes

1. Our intended participants are in many ways unique, not least because they attend the only school in the country providing pre-professional dance training integrated with academic schooling. Therefore, there are no pilot participants who were very similar to our intended participants. The key criteria on which we wanted to match pilot participants were instead simply age and their being involved in an esthetic activity at a level high enough to ensure that the questionnaire made sense for them. This gave us important information about readability, time taken, and so on.

2. A copy of the interview guide is available from the first author on request.