top of page

Performance Anxiety: Can Established Methods in Sports Science be adapted to Dance? A Review of Rele


There are large quantities of published research in the area of general anxiety in populations, with many of the methods used being adopted in investigations specific to various sports. The aim of this paper is to demonstrate whether some of these methods are applicable to dance performance situations by describing, comparing and evaluating the appropriate information and research available on the subject. The implications of selected studies will be discussed with judgements on what they show and how the knowledge can be used for further study in a dance environment. Using various techniques and data bases (manual library search, Dance on Disk, Psychological Bulletin, Psych Info, Social Sciences Citation Index, Sports discus, Zetoc) several prominent journals where identified which provided thirteen articles of interest. From this selection, six original investigations and one conference paper were chosen. This review will cover broad information on the subject of anxiety, narrowing the field to specific articles related to athletic performance anxiety and current research in dance and performing arts.


As stated by Woods, B. (1998. P.98) anxiety can be defined as a state of uneasiness about what might happen. It is a negative emotional state related to immediate concerns of a situation perceived as threatening or just to a vague fear or apprehension. The symptoms may include physiological, cognitive and emotional responses. Performance anxiety involves concerns, thoughts and feelings associated with an approaching performance; the danger is that these responses have the potential to undermine preparation, talent, ability and self-esteem. This type of anxiety is normally referred to as state anxiety however there is also enduring anxiety, which is a personality trait and part of an individual’s pattern of behaviours (Spielberger, C.D. 1966). For athletes and dancers to be able to reach their optimal performance levels research and information into the subject needs to be obtained in order to manage anxiety effectively.

Dance science is still very much in the shadows of sports science, it is important to take on all the information available from current sports research and evaluate how applicable and adaptable it can be to the dance world. Anxiety in athletes has become one of the most common topics of sports psychology research. It is this research that shows that the ability to cope with pressure and anxiety is an integral part of sports, particularly among elite athletes. (Hardy, L., Jones, G., and Gould, D. 1996. P.140). However anxiety is not just a simple measurable response and many models over the years have been developed to find better ways of evaluating and understanding the phenomenon.

Anxiety models.

There is a relationship between arousal, anxiety and stress, psychologists have tried to differentiate between them in order to gain a better understanding but it is impossible to avoid an overlap. (Woods, B. 1998. p.90). One of the earliest models that attempted to explain this relationship was the inverted-U hypothesis (Yerkes, R.M., Dodson, J.D. 1908)

The inverted-U hypothesis stated that as arousal increased performance would increase as well; but if arousal became too great performance would deteriorate. Basically, as stress began to build an individual still felt confident in their ability to control it and performance would improve. However, once a stressor became so great that the individual started to doubt their ability to cope with it, performance would decline.

An individualistic approach was added to this hypothesis when researchers developed the concept of individualized zones of optimal functioning (ZOF) (Hanin, Y.L., 1980, 1986) According to this theory, each individual has an optimal level of performance anxiety. If the athlete is in this “zone”, peak performances will result. However if anxiety is too high or low then performance will be impaired. ZOF can be determined by repeatedly measuring anxiety and performance through the athlete’s recall of anxiety levels prior to performance with the use various anxiety state questionnaires. Criticisms of this model are quite simply that it may be too simplistic and does not consider anxiety as multidimensional, it suggests that performers can “tune in” and control their ZOF.

In 1998 a new model emerged, the catastrophe theory. (Fazey, J., and Hardy, L. 1998) They argued that somatic or physiological and cognitive anxiety, do not just have different effects on performance but also interact, having an effect on each other to change performance. The interactions of many factors take into account the multidimensions of anxiety. It is based on the inverted-U but this model shows that, if a performers physiological arousal is high and then they start to experience an increase in cognitive anxiety, a large and sudden deterioration in performance could occur. (Biddle, J.H. 1995.p.144-146)

Methods of Measurement.

Various measures have been used throughout research on performance anxiety each with their own benefits and drawbacks. Physiological measures monitor somatic responses such as heart rate and try to develop a relationship between the body’s response and perceived anxiety; however this can be impractical and expensive.

The Weinberg, R.S., and Hunt, U.V. (1976) study observed high and low trait anxious students throwing tennis balls at a target. They observed individual behaviour such as facial expression and negative self-talk along with aspects of the performance including accuracy of the throw. The drawbacks to this can be that just the presence of the observers can increase arousal and it can be difficult to recreate an actual competition environment, If a “real” performance is being monitored human error can result in portions of the event being missed which may affect results of the study. None of the studies in this review use these techniques but used modified versions of behaviour observations and self-reported questionnaires.

It is popular for researchers to use questionnaires that ask participants how they feel in certain situations; examples of these are the Sports Competition Anxiety Test (SCAT) and the Competitive State Anxiety Inventory (CSAI-2). (Martens, R., Burton, D., Vealey, R.S., Bump, L.A and Smith, D. 1990 p. 117-190). Information on these and other scales can be obtained from Biddle, S.J.H. (1995 p.131-139). However there are still drawbacks to this method Woods, B. (1998 p.108-109) suggest honestly of the participants must be taken into account even if they are only trying to give socially desirable answers. In addition the wording may have confusing indications, especially if used in the context of different sports, a particular point that may be relevant when some of these scales are used for dancers.

Dance Related Research.

Six papers relating to dance and dancers were retrieved during the initial search for research on performance anxiety. One particular point that immediately stands out is the titles. Every study looks at anxiety within the context of another variable, it is interesting what implications this implies to any basic level of research on performance anxiety in dance; however this are will be discussed later.

Excluding one study, Singer, K. (2004) all the papers were original cross-sectional studies that used relatively large numbers of participants, from age pools and dance standards appropriate for the their aims. The layouts are clear and concise, with sub-headings and the use of prose and tables to display results. For the purposes of this review Singer, K. was excluded on the grounds that the investigation only looked at three dancers (one of which dropped out during the study). Their aim was to see if they could influence anxiety through neurofeedback, which did not completely tie in with idea of how previous sports research using various models and measurements compared to dance studies.

Of the remaining studies various methods of measurements where employed to investigate anxiety, adopted from previous sports studies. Although they all covered various other aspects, the main focus here is the techniques used to research, record and evaluate the performance anxiety aspect.

Krasnow, D., Mainwaring, L., Kerr, G. (1999) study focussed on stress and perfectionism in relation to injuries in dancers and gymnasts rather than anxiety. Although not entirely the same, as demonstrated earlier arousal, anxiety and stress are very closely linked. They defined stress as the relationship between a person and their environment that is appraised as taxing personal resources and threatening well-being. (p.51). A modified survey modelled after the Life Experience Survey (LES) (p.53) was used to access stressors unique to dancers, although examples of how the survey was structured were given they were not particularly comprehensive and published under the slightly miss-leading heading of “ Dance Experience Survey” (p.53). Without examples of the types of questions used it is difficult to see if the questionnaire would be useful in inquiring about actual anxiety. Analysis was performed to determine the validity of the multidimensional perfectional scale (MPS) used to evaluate perfectionism, which they report as positively correlating to other published perfectionism scales. (p.53). However, there is no mention of validity research regarding their modified LES. There is also no information in the study as to when or how the questionnaires were administered which would lead to difficulties in replicating the research. An interesting point raised from their results was that they found dancers had higher total stress than gymnasts an intriguing area, as gymnasts would presumably be involved in more of a competition environment rather than performance, although the gymnasts had lower levels of positive stress. The authors put this down to gymnasts’ “lives and identities revolving around gymnastics: (p.56), without the research to back this up it would seem that dancers would surely be in a similar situation; perhaps an area for further enquiry.

Another paper to look at similar components is Carr, S., and Wyron, M. (2003) who looked at perfectionism but also achievement goals and trait anxiety within the context of motivational climate for dance students. The authors were interested in trait anxiety and whether teachers and environment affected participants’ behaviors, not just in the studio but also in a wider context of shaping their personalities. As noted earlier this is different to state anxiety that is usually seen in terms of performance anxiety and “stage fright”. However Spielberger, C.D. (1966) stated that higher levels of trait anxiety may predispose an individual to perceive situations as threatening, responding with a disproportionate reaction of state anxiety. Because of this I believe the paper could hold useful insights into method and measurements et cetera. Trait anxiety was measured using a slightly modified version of the Sport Anxiety Scale (SAS). Their research lead them to believe context specific assessments would provide more indicative responses then generalized measures (p.108). A notion not discussed in Krasnow, D., Mainwaring, L., Kerr, G. (1999) study. How questions were adapted and scored are included along with references to the properties and validity of the original measurement scale but not whether they tested their own version. A MPS similar to Krasnow, D., et al was administered to research perfectionism, conversely more information is provided on adaptations to questions and scoring methods in comparison. Procedures on how and when the questionnaires were carried out were fully reported, along with reliability test scores and the specific statistical analysis used. The results showed that higher levels of perfectionism and cognitive trait anxiety where significantly linked to higher perceptions of a performance related climate. The findings also suggest that high levels of trait anxiety support the notion that a predisposition to perceive circumstances as threatening may stem from perceptions of performance orientated climates. Participants tended to worry about performances more when their perceptions of punishment for mistakes were high. These findings are similar to research conducted with an elite group of swimmers. Jones, G., Hanton, S., Swain, A.B.J. 1994) found that anxiety intensity levels were higher in subjects who interpreted their anxiety as harmful, than those who reported it as being an aid. This was also found to be true of gymnasts. (Jones, G., Swain, A.B.J., Hardy, L. 1993). Carr, S. et al report that it may be important to identify variables that predispose individuals to develop such dispositions especially as psychological responses in sport have shown a link between goal orientations, trait anxiety and perfectionism. (p.110). The discussion section of the paper clearly states that assumptions about psychometric properties to the dance specific modifications made to the inventories employed should be interpreted with caution; (p.110) an extremely vital observation, which none of the other papers remark on.

The SAS was also employed in Barrell, G.M., and Terry, P.C. (2003) study researching trait anxiety and coping strategies among ballet dancers. The paper clearly states how the inventory was administered and the questions involved. Keeping the method consistent to previous research they discus, the only change involved the wording of “competition” to “performance” to make the items more appropriate to dancers. (p.61). Agreeing with Carr, S., et al. (2003) the SAS was chosen over the SCAT due to its multidimensional structure. It is also reported that previous research showed satisfactory psychometric characteristics for the SAS among dancers (p.61). Researchers were not present during the administrating of the questionnaire, relying on the teachers and psychologists associated with the schools and companies. A comment not raised in any of the previous studies but could possibly lead to validity worries. It was found that scores from the SAS were significant predictors of dancers coping strategies, however there was no significance between gender or status. One of the limitations of the study was that gender and status comparisons were based on uneven cell sizes, (p.63) possibly why Carr, S., et al did not differentiate between genders in their study although this is not stated.

High trait-anxious dancers reported to use more coping strategies more frequently but they were emotional based, seeking social support, self-blame et cetera rather than problem solving. The paper states that previous research shows problem-focused coping to be more effective for athletes experiencing performance difficulties (p.62) an area for possible future research in dance. Another import issue raised in this paper is the differences between companies and schools with regard to their research and possible further studies, was that the times of term, performance schedules touring and so forth must be taken into account as these would undoubtedly incur different responses to the questionnaires (p.63). Something not discussed in other studies, but would be useful information. The authors suggest that a relationship exists between trait anxiety and the way dancers cope in performance situations which aids the argument that psychological education should be considered, especially in reducing trait anxiety and improving problem solving coping strategies. They believe interventions should be employed during the early stages of professional development, before habits are formed to improve mental skills and coping strategies, as no differences were found between professional and student dancers (p.63). This is contradictory to Gould, D., Petchlikof, L., and Weinberg, R.S. (1984) who used the CSAI-2, they found on a study of tennis players that the strongest predictor of anxiety was years of experience, the more experience an athlete had the lower their level of cognitive anxiety, This seems to make sense, because as an athlete gains experience they learn the “tricks of the trade”, proper stress management techniques and the added experience reduces the likelihood of encountering new and stressful events.

Reinforcing Gould, D., Petchlikof, L., and Weinberg, R.S. (1984), the results of Monsma, E.V., and Overby, L.Y. (2004). Looking into the relationship between the use of imagery and its effects on competitive anxiety in an audition setting, suggest that there was a correlation between dancers who were previously successful having lower cognitive anxiety levels and their years in training which would suggest that anxiety was also related to experience. It was realised that auditions produce similar responses to that of athlete competition (p.11). The authors felt that from their research the CSAI-2 could effectively be adapted for use during an audition context and stated it that it showed “excellent internal consistency” (p.16) a table (p.14) represents their validity results along side previous sports research involving the CSAI-2. Original instructions were followed for the questionnaire with the content remaining the same except for wording being changed to reflect the nature of dance and audition, no examples are given. Ignoring the aspect of imagery in the paper, the results show that there is a significant relationship between anxiety self-confidence and performance, with successful dancers reporting lower cognitive anxiety. This seems to follow the catastrophe theory model with performance being predicted by cognitive anxiety levels. The paper discusses that their findings are limited by the fact that only females ballet dancers were used and further research should consider genders and styles. It was found that the positive use of imagery seemed to reduce anxiety and boost self-confidence (p.17). This could mean information from sports research such as Holm, J.E., Beckwith, B.E., Ehde, D.M and Tinius, T. P. (1996), which looked not only at reducing stress and anxiety levels, but also improving self-confidence through imagery, relaxation and cognitive intervention could hold interesting possibilities for dance research.

Lucinda, S and McLean, N. (2003) provided an interesting presentation looking at performance anxiety also using the CSAI-2; however without a full published paper, just presentation notes from the 13th Annual Meeting of the International Association of Dance Medicine and Science, it is unclear as to how the questionnaire was structured or adapted. Their findings suggested that personality traits and thought processes could have an effect on performance anxiety (p.113). They were interested in seeing if the perceived threat of an assessment would increase anxiety levels. Results showed that physiological arousal significantly predicted self-perceived threat, participants who associated this increase as unfavourable felt physical sensations to be threatening in subsequent performances (p.115). These findings completely agree with Spielberger, C.D. (1966) and Carr, S., et al. (2003) studies discussed earlier, which could mean that they have significant implications for the design of methods and models when used in a dance specific context. Lucinda, S. et al stated that their research could help in the development of psychological interventions aimed at controlling performance anxiety, in particular the central role of cognitive appraisal (p.115), which would tie in with Carr, S., et al’s conclusions. Also agreeing with Barrell, G.M., et al. (2003) they note that early education especially about physical response will help dancers.


Clearly, anxiety levels can have a variety of effects on performance, which are based on the type of activity and the individual. The aims of this review were to determine if models and methods used in researching sport performance anxiety could be successfully adopted to use in a dance environment. Although each study found valid and interesting results, one important theme emerged, is dance science just following sports science and taking their previous results as an absolute? All the papers reviewed here combined anxiety with another variable, to my knowledge no published studies have focussed directly on the subject of performance anxiety in dancers, nor was there enough evidence here to show that adaptations to sports questionnaires entirely work for dancers. The fact that Barrell, G.M., et al. (2003) study found contradictions to the findings of Monsma, E.V., et al. (2004) could quite simply be due to the fact that they used different methods of measure, which were both individually adapted for their research. Perhaps there is scope to take a slight step back and investigate the development of a measure solely looking at dance performance anxiety, rather then trying to fit dancers into groups which have already been devised in sports. My own opinion is that these scales are useful, but it must be insured that they are applicable and that a standardized inventory would make research studies more comparable.

An area for future reviews is the current research investigating performance anxiety in musicians and other performers. In recent years dancers have been seen more as elite athletes, but it must not be forgotten that dance is principally an art form and that a way of combining all fields of study needs to be adopted in order to get a balance. Of cause all scientific methods should be considered, as stated by Carr, S., et al (2003 p. 112), who suggest that a longditudal study would be valuable, obviously this has its own drawbacks such as expense but would provide a wealth of knowledge. With both dance and anxiety seen as multi-dimensional to a degree, the adoption of both quantitative and qualitative methodologies could provide a more detailed and clearer perspective on the experience and effect of anxiety. Also another interesting line of inquiry would be the development of a means of measuring anxiety during actual performance. I personally believe dance science still has a long way to go to catch up with the sports world, but maybe we should not just be treading in their footsteps.


Featured Posts
Recent Posts