Path Analysis of a Self-Esteem Model across a Competitive Swim Season

Article excerpt

An adaptation of the previously developed Exercise and Self-Esteem Model (sonstroem & Morgan, 1989) was tested longitudinally with 93 male interscholastic swimmers from nine high schools who were evaluated at pre-, mid-, and postseason. Swimmers completed three self-perception scales that ranged from evaluations of specific swim skills (SKILL) through broader perceived physical competence (PC) to global self-esteem (SE). Performance scores (PERF) were calculated across events by standardizing swim times to a mean of 0 and a standard deviation of 1. Structural modeling analysis across the three time waves indicated an excellent data fit ([[chi].sup.2] = 25.46 p [greater than] .15). The model was able to explain 84, 83, and 80% of PC, SKILL, and PERF, respectively, at the third testing period. Swim improvement tended to be small (median change = 2.9%). Relationships among variables at each testing period and among the smae variables at different testing periods were large and as hypothesized. Several significant relationships occurred between different variables across different testing periods. These were not extinguished when social desirability was added to the model. Social desirability failed to significantly influence measures of any model variable.

Key words: perceived physical competence, self-esteem model, physical self-perceptions

Current theory on the self-concept tends to regard it as a multidimensional and hierarchically organized system, exemplified by research designed for school settings (Shavelson, Hubner, & Stanton, 1976). General self-concept at the apex of the structure is divisible into academic and nonacademic components. In turn, the former can be separated into the more specific self-concepts of English, history, math, and science. Each of these is divisible further into conceptions of specific behaviors related to the respective self-perception facet. Component stability is hypothesized to be positively related to hierarchical level, with lower level, more situationally specific components believed to be more susceptible to environmental influence. Although the Shavelson model has been modified by subsequent investigation (Marsh, Byrne, & Shavelson, 1988), itt has provided a major contemporary stimulus to research and theory in self-esteem. The value of considering self-concept as a self-system rather than as a single unidimensional construct can be seen in reserch results of Marsh, Richards, and Barnes (1986). Their replications found that Outward Bound experiences influenced self-regard facets relevant to course activities significantly more than they influenced components less relevant to training activities.

The recently developed Exercise and Self-Esteem Model proposes that sport and exercise experiences can influence self-esteem along a dimension of perceived competence (sonstroem & Morgan, 1989). Components of self-regard are arranged on a vertical continuum of specificity-generality, with situation-specific assessment at the lower end of the continuum and broad, content-free, global self-esteem measurement at its upper end. As people develop increased self-efficiacies (expectancies) at performing tasks specific to a sport or exercise setting, these perceptions of ability transfer to more general evaluations of overall physical competence. Perceived physical competence (PC) has been shown to be related to self-esteem (SE) and is regarded as one of the components of global self-esteem (Fox & Corbin, 1989); Sonstroem, 1974 1978). Earlier research by Sonstroem (1978) found that physical fitness in adolescent boys was associated with PC rather than with SE. SE is considered to be the evaluative component of self-concept (Rosenberg, 1979), broadly defined as individuals' cognitive perceptions of themselves (Shavelson et al., 1976). Global SE at the top or more general end of the self-perception continuum is often identified as the variable bestt indicative of favorable life adjustment (Rosenberg, 1979; Wylie, 1979). …