Objectives: One important issue in sport and exercise psychology is to determine to what 1 extent sports and exercise can help to increase self-esteem, and what the underlying 2 mechanism might be. Based on the exercise and self-esteem model (EXSEM) and on findings 3 from the sociometer theory, the mediating effect of physical self-concept and perceived social 4 acceptance on the longitudinal relationship between motor ability and self-esteem was 5 investigated. Design: Longitudinal study with three waves of data collection at intervals of ten 6 weeks each. Method: 428 adolescents (46.3 % girls, Mage = 11.9, SD = .55) participated in the 7 study, in which they performed three motor ability tests and completed paper-and-pencil 8 questionnaires for physical self-concept and perceived social acceptance, as well as for self-9 esteem, at all three measuring points. Results: Using structural equation modelling 10 procedures, the multiple mediation model revealed both physical self-concept and perceived 11 social acceptance to be mediators between motor ability and self-esteem in the case of boys. 12In girls, on the other hand, the mediation between motor ability and self-esteem only takes 13 place via physical self-concept. Conclusions: Gender differences in the relationship between 14 motor ability and self-esteem suggest gender-specific interventions aimed at promoting self-15
concept. 16Key-words: sport competence, self-perceptions, adolescents, gender, structural 17 Global self-esteem is traditionally seen as a central indicator for mental health and an 22 explanatory variable for human behaviour (Rosenberg, 1965). People with high self-esteem 23 are more emotionally stable, less prone to experiencing depression and display higher 24 academic achievements (Marsh & O'Mara, 2008). From a multidimensional perspective on 25 the self, which is now widely accepted in many psychological disciplines, global self-esteem 26 marks the apex of the hierarchically organized and multidimensionally structured self-concept 27 (Shavelson, Hubner, & Stanton, 1976). Moving from the top to the bottom, the term 'self-28 concept' is reserved for evaluations in discrete domains such as academic, social, emotional 29 and physical domains, and thus refers to domain-specific self-perceptions or self-conceptions 30 (Harter, 2012). In recent years, various scientific disciplines (e.g. educational, sport or 31 developmental psychology) have adopted a multidimensional perspective by the fact that 32 domain-specific self-perceptions are more suitable for predicting specific behaviour (Marsh & 33 O'Mara, 2008), that it is easier to influence specific facets through interventions (Schmidt, 34Valkanover, Roebers, & Conzelmann, 2013), and that they are more strongly related to 35 corresponding external criteria (Möller, Pohlmann, Köller, & Marsh, 2009). In addition, a 36 multidimensional perspective of the self allows the relationship between domain-specific self-37 perceptions and global self-esteem to be examined, as well as their inter...