Hierarchical Self

Researchers  and  practitioners  have  long  believed that how people feel about and describe themselves can strongly influence motivated behavior in sport and  exercise.  Two  key  constructs  studied  in  the sport  and  exercise  psychology  literature  are  self-esteem  and  self-concept.  Since  many  people  have argued that judgments about the self will influence the  selection  and  maintenance  of  physical  activity  behavior,  it  is  critical  to  distinguish  between self-concept  and  self-esteem.  Self-esteem  or  self-worth refers to how a person feels about oneself. It is a self-evaluation that provides a sense of how an  individual  assesses  worth.  Self-concept  refers to  self-perceptions  of  personal  attributes  such  as skills, abilities, and physical characteristics.

Historically,   researchers   studied   both   self-esteem and self-concept from a very broad global perspective.  Modern  work,  however,  recognizes that self-esteem, and especially self-concept, is best viewed  from  a  multidimensional  and  hierarchical perspective.  Starting  with  the  work  of  Richard J.  Shavelson  and  colleagues  in  the  1970s,  self-concept has been conceptualized as a multifaceted hierarchy  consisting  of  the  global  self  at  the  top, with  specific  domains  (e.g.  academic,  social,  and physical) and subdomains nested underneath. For example,  one  important  domain  of  global  self  is the physical self, which can be further divided into several subdomains, such as flexibility, endurance or conditioning, sport skills, strength, and appearance.  Each  subdomain  can  be  further  differentiated into more discrete components. For example, physical  flexibility  can  be  further  separated  into facets  such  as  leg  flexibility,  back  flexibility,  hip flexibility,  and  shoulder  flexibility.  Typically,  subdomains  of  physical  self-concept  have  moderate to strong interrelationships. Aspects of the physical  self,  especially  body  appearance,  demonstrate moderate  correlations  with  perceptions  of  global self across the life span.

Although there is strong support for the multidimensional  view  of  self-concept,  support  for  the hierarchical  perspective  is  less  compelling.  Work by Herb Marsh and others has found that relationships among some specific domains or subdomains are low or non-existent. Furthermore, global self-concept also has low correlations with specific self-concept  domains  or  subdomains.  A  hierarchical model also implies either a top-down or bottom-up direction of causal flow. The bottom-up approach proposes  that  situation-specific  and  task-oriented experiences  influence  perceptions  in  specific  subdomains, which in turn will influence domain specific  self-concept  and  overall  global  self-concept. Researchers  conducting  interventions  using  this model would target a specific facet of physical self (e.g.  leg  strength)  and  would  expect  changes  in the  facet  to  effect  perceptions  of  a  specific  physical subdomain (e.g. physical strength). Changes in the specific physical subdomain would then result in expected changes to global physical self, which would effect change in global self-concept.

The top-down model hypothesizes that changes in  global  self-esteem  will  influence  lower  order domains  (e.g.  physical  self),  which  will  impact specific  subdomains  and  in  turn  would  influence specific physical activity behaviors. Therefore, top down  interventions  target  global  self-esteem  as  it is assumed that individuals with higher self-esteem would  be  more  confident  to  engage  in  physical activity  behaviors  regardless  of  experience  and perceived competence in that domain. A top-down model  implies  that  changes  in  global  self-concept (or  self-esteem)  should  cause  changes  in  lower order  domains  and  influence  subsequent  situation  specific  experiences.  Unfortunately,  longitudinal  and  experimental  research  has  found  little evidence  of  either  top-down  or  bottom-up  causal flow between the domain level and global level.

Measurement of Hierarchical Self

Investigating the causal relationship between self-esteem,  self-concept,  and  physical  activity  behaviors has been greatly enhanced by multidimensional models and subsequent measurement instruments. Many  instruments  have  been  developed,  evaluated,  and  modified  for  various  populations  and age  groups.  Within  sport  and  exercise  research, there are a few instruments that have consistently demonstrated   strong   measurement   properties. Susan  Harter  and  colleagues  developed  the  Self Perceptions Profiles for various age groups. These instruments  were  based  on  hierarchical  models with global self-esteem and various domains (e.g., social, athletic, appearance) depending on the age group.  Separate  work  by  Kenneth  Fox  and  Herb Marsh sought to capture the physical self in more detail.  Kenneth  R.  Fox  and  Charles  B.  Corbin developed  the  Physical  Self-Perception  Profile, which consists of four subdomains of physical self-concept  and  one  global  domain  of  physical  self-worth. The subdomains include sport competence, attractive  body,  physical  strength,  and  physical conditioning. Marsh and colleagues developed the Physical  Self-Description  Questionnaire,  which specifies physical self-concept as nine unique physical  self-dimensions  (strength,  body  fat,  activity, endurance  or  fitness,  sports  competence,  coordination, health, appearance, and flexibility) nested under  general  physical  self-concept  and  global self-esteem.  Marsh  and  colleagues  also  developed the Elite Athlete Self-Description Questionnaire to assess  physical  self-concept  in  elite  athletes  using six different dimensions. Elite athlete self-concept dimensions  include  skill,  body,  aerobic  fitness, anaerobic fitness, mental competence, and overall performance.

Self-Esteem, Self-Concept, and Physical Activity Behavior

Researchers  in  sport  and  exercise  psychology recognize  the  importance  of  self-esteem  and  self-concept  as  not  only  significant  mental  health outcomes,  but  also  as  potential  determinants  of physical activity behavior. There is solid evidence, in  both  youth  and  adults,  that  physical  activity behavior is correlated with aspects of physical self-concept  and  sometimes  with  global  self-concept and  self-esteem.  For  example,  many  researchers have found that conditioning and endurance self-perceptions  predict  objective  physical  fitness  and physical  activity  levels.  Similarly,  self-perceptions of  body  fat  or  physical  appearance  self-concept predict objective measures of body mass index or body composition. Physical self-concept measures are far superior compared to global self-concept or self-esteem  measures  in  predicting  physical  activity  behaviors.  However,  self-perceptions  of  body appearance  and  body  fat  are  often  unrelated  to physical activity behaviors.

Many  physical  activity  intervention  studies in  children,  youth,  and  adult  populations  have primarily  focused  on  global  self-esteem  or  self-concept and physical activity. In studies with children and youth, the majority of these studies have found  a  modest  short-term  significant  intervention  effect.  However,  many  of  these  studies  have been  criticized  for  being  poor  quality.  In  adults, John Spence, Kerry McGannon, and Pauline Poon conducted  a  meta-analysis  of  over  100  exercise intervention  studies  and  reported  that  exercise intervention seemed to have a small but significant effect on either global self-concept or self-esteem. Overall,  researchers  suggest  that  physical  activity interventions can produce small changes in global self.  However,  there  is  little  consistent  evidence about  the  characteristics  of  the  physical  activity intervention (intensity, duration, types) required to produce changes.

Physical activity intervention research targeting changes  in  the  various  levels  of  the  hierarchical self is more limited. In one review of intervention studies in various populations, Kenneth Fox concluded  that  physical  activity  conferred  significant improvements in some components of the physical self. Typically, the intervention effect was stronger for  physical  self-perceptions  compared  to  global measures.  More  recent  work  in  adults  suggested that interventions have a stronger impact on physical self-concept compared to global self.

Based on a mix of findings, researchers seem to suggest that physical activity can cause changes in both global and physical self. Intervention studies often  indicate  a  bottom  up  model  in  that  physical activity will cause a change in self-concept or self-esteem. However, recent work by Herbert W. Marsh  and  Rhonda  G.  Craven  and  Marsh  and other colleagues strongly suggests there is a reciprocal relationship between achievement behaviors and domain level self-concept. Using longitudinal designs  and  complex  modeling,  they  found  that self-concept and physical activity are both a cause and  effect  for  each  other.  The  reciprocal  model suggests  that  previous  physical  activity  behavior  would  influence  physical  self-concept,  which would  in  turn  impact  physical  activity  behavior even after controlling for previous physical activity behavior. This reciprocal effect was found in several studies using various physical activity contexts, such  as  elite  swimming,  physical  education,  and extracurricular activities. The work of Marsh and colleagues in various populations and achievement domains indicates that the reciprocal effect occurs primarily  at  the  domain  level.  The  use  of  the reciprocal  model  has  implications  for  research and applied practice. For instance, it implies that interventions  should  be  targeted  at  both  the  self-concept level and at specific physical skills or perceptions related to the targeted domain. Also, the level at which self-concept is measured needs to be considered, as reciprocal effects are stronger when the  domain  in  self-concept  is  congruent  with  the achievement domain.

Gender and Cultural Differences

Researchers  have  consistently  found  that  males have   more   positive   global   and   physical   self-perceptions than females in both youth and adults, although  there  are  large  variations  within  each gender.  This  parallels  the  findings  in  the  physical activity literature. However, there is no consistent evidence  of  the  relationship  between  self-concept and  physical  activity  behaviors  being  different across genders. Some researchers have argued that gender stereotypes, cultural beliefs, and the dominance of traditional boys’ sports in school systems and  community  organizations  results  in  sports being more highly valued for males and thus more prominent in the development of their self-concept and self-esteem. There is, however, no reliable evidence that physical activity interventions are more effective  in  enhancing  self-perceptions  for  either females or males.

Understanding cultural differences in the global and  physical  self-concept  is  challenging.  Most studies  comparing  samples  from  various  countries show similar results in terms of physical self-concepts  being  meaningfully  related  to  physical activity behaviors and there is strong evidence that males  have  more  positive  self-perceptions  across groups.  However,  most  studies  compared  nationality  (e.g.,  Turkey  vs.  United  Kingdom)  rather than assessing cultural values, such as filial piety, settings. Males tend to have more positive perceptions  of  physical  self  as  well  as  higher  levels  of physical activity behaviors. Physical activity interventions have a stronger impact on enhancing the physical self-concept compared to the global self. There is growing evidence, however, that there is a reciprocal effect between physical self-concept and physical activity behavior over time. Cultural differences are not well-understood.


Self-esteem and self-concept are best understood from a multidimensional perspective. Physical self-concept is an important dimension in understanding motivated behavior in sport and exercise settings. Males tend to have more positive perceptions of physical self as well as higher levels of physical activity behaviors. Physical activity interventions have a stronger impact on enhancing the physical self-concept compared to the global self. There is growing evidence, however, that there is a reciprocal effect between physical self-concept and physical activity behavior over time. Cultural differences are not well-understood.


  1. Crocker, P. R. E., Kowalski, K., & Hadd, V. (2008). The role of self and identity in physical (in)activity. In A. Smith & S. J. Biddle (Eds.), Youth physical activity and inactivity: Challenges and solutions (pp. 215–237). Champaign, IL: Human Kinetics.
  2. Fox, K. R. (Ed.). (1997). The physical self: From motivation to well-being. Champaign, IL: Human Kinetics.
  3. Fox, K. R. (2000). The effects of exercise on self perceptions and self-esteem. In S. J. H. Biddle, K. R. Fox, & S. H. Boutcher (Eds.), Physical activity and psychological well-being (pp. 88–117). London: Routledge.
  4. Marsh, H. W., & Craven, R. G. (2006). Reciprocal effects of self-concept and performance from a multidimensional perspective: Beyond seductive pleasure and unidimensional perspectives. Perspectives on Psychological Science, 1, 133–162.
  5. Spence, J. C., McGannon, K. R., & Poon, P. (2005). The effects of exercise on global self-esteem: A quantitative review. Journal of Sport & Exercise Psychology, 27, 311–334.

See also: