Herbert Marsh argues for a focus on the specific domain of interest rather than on global self-concept (Marsh & Craven, 1997). Marsh (Marsh, 1997; Marsh & Craven, 2006) and Ken Fox (Fox, 1990, 1998; Fox & Corbin 1989) have contributed substantially to our understanding of the physical self. Both used the Shalverson model as a blueprint but found the physical self to be more complex than that which can be captured with two subareas. In Fox’s three-tier hierarchical model, global self-esteem is at the top. Physical self-worth, at the next level, is based on the four subdomains of sport competence, attractive body, physical strength, and physical condition. Physical self-worth mediates the relationship between self-esteem and the four subdomains. Although the hierarchical structure of self-concept is a pervasive perspective today, some research questions it (e.g., Kowalski, Crocker, Kowalski, Chad, & Humbert, 2003).
Fox developed a measurement tool, the Physical Self-Perception Profile (PSPP), for assessing the physical self. It includes the following subscales:
- Sport competence (sport): Perceptions of sport and athletic ability, ability to learn sport skills, and confidence in the sport environment
- Physical condition (condition): Perceptions of physical condition, stamina, and fitness; ability to maintain exercise; and confidence in the exercise and fitness setting
- Body attractiveness (body): Perceived attractiveness of figure or physique, ability to maintain an attractive body, and confidence in appearance
- Physical strength (strength): Perceived strength, muscle development, and confidence in situations requiring strength
- Physical self-worth (PSW): General feelings of happiness, satisfaction, pride, respect, and confidence in the physical self
The inclusion of physical self-worth underscores that the physical subdomains cannot be summed to obtain the physical self-worth score. Fox and Corbin (1989) provided evidence for the sensitivity, reliability, and stability of the subscales; confirmed the subscale factor structure; and reported associations of the subscales with physical activity involvement to provide initial validity support. Sonstroem, Speliotis, and Fava (1992) subsequently found that the PSPP showed strong internal consistency, separated exercisers from nonexercisers, and predicted degree of exercise involvement among adults, and they recommended its continued use.
Marsh (Marsh, 1996; Marsh, Richards, Johnson, Roche, & Tremayne, 1994) used psychometric techniques to develop the Physical Self-Description Questionnaire (PSDQ), a multidimensional physical self-concept measure with 11 scales: strength, body fat, activity, endurance and fitness, sport competence, coordination, health, appearance, flexibility, global physical self-worth, and global esteem. Marsh and colleagues provided good psychometric support for the measure; they confirmed its validity by correlating PSDQ subscales with external criterion measures of body composition, physical activity, endurance, strength, and flexibility.
Consistent with the hierarchical models (e.g., Shavelson et al., 1976), Fox and Marsh contend that the subdomains are further divided into more situation-specific areas. For example, sport competence can be divided into a facet (e.g., basketball ability), a subfacet (e.g., shooting ability), and self-efficacy or situation-specific confidence (e.g., "I can make this free throw"). Further, these more dynamic, lower-level self-perceptions affect their more stable, higher-order counterparts. These lower-level self-perceptions (e.g., confidence and self-efficacy) are discussed later in the chapter.
Sonstroem and Morgan (1989) made the nature of the relationship among the higher- and lower-order self-perceptions more explicit in their model, similar to that of Fox (1990) and Marsh (1990). They proposed the following:
- Physical fitness is more highly related to physical self-efficacy than to physical competence, physical acceptance, and global self-esteem.
- Physical self-efficacy is more highly related to physical competence than to physical acceptance or global self-esteem.
- Physical competence is more highly related to global self-esteem than is physical self-efficacy or physical fitness.
Further developing the understanding of self-esteem, Marsh and colleagues (Marsh, Chanal, & Sarrazin, 2006; Marsh & Craven, 2006) argue that achievement is influenced by both a domain-level self-worth (e.g., physical self-worth) and performance, which are mutually reinforcing. Thus, physical performance influences physical self-worth and physical self-worth influences performance.