Academic journal article Research Quarterly for Exercise and Sport

Physical Self Description Questionnaire: Stability and Discriminant Validity

Academic journal article Research Quarterly for Exercise and Sport

Physical Self Description Questionnaire: Stability and Discriminant Validity

Article excerpt

Construct validation is an ongoing process that involves evaluation of internal structure, stability over time, and relations to external validity criteria (Marsh, in press; Marsh, Richards, Johnson, Roche, & Tremayne, 1994; Nelson, 1989; Ostrow, 1990; Vealey, 1986). Thus, for example, Gill, Dzewaltowski, and Deeter (1988) argued for the construction of multidimensional sport and exercise instruments based on theory, followed by item and reliability analysis, exploratory and confirmatory factor analyses, tests of convergent and divergent validity, and application in research and practice. The usefulness of confirmatory factor analysis (CFA) and structural equation modeling for these purposes is widely endorsed but these techniques are just beginning to be widely used in the sports sciences. Schutz and Gessaroli (1993), for example, claimed that this is the statistical tool for the 1990s in many disciplines but lamented its nonuse in sports psychology. They noted that one important reason for this relative scarcity was a dearth of good introductory demonstrations of CFA applications in the sport sciences. The purpose of this article is to extend evaluations of the construct validity of physical self-concept responses based on the Physical Self-Description Questionnaire (PSDQ) by evaluating the stability of responses collected on four occasions over a 14-month period. In pursuing this purpose I also demonstrate the application of confirmatory factor analytic (CFA) models of multitrait-multimethod (MTMM) data to test for the convergent and discriminant validity of PSDQ responses.

Physical Self-Concept

In his review of physical-self-concept measurements, Marsh (in press) noted that physical self-concept has typically been inferred from responses to relatively ill-defined global scales or responses to one or two physical scales on multidimensional self-concept scales designed to measure a wide variety of self-concept facets (e.g., physical, social, emotional, academic). However, Marsh, Richards et al. (1994; Marsh, in press) argued that these global physical scales may confound distinguishable physical components reflecting, for example, health, physical attractiveness, body composition, and physical activity (also see similar comments by Fox & Corbin, 1989). Concerns such as these led to the development of multidimensional physical-self-concept scales such as the PSDQ that measures nine specific components of physical self-concept (Strength, Body Fat, Activity, Endurance/Fitness, Sports Competence, Coordination, Health, Appearance, Flexibility), Global Physical Self-concept, and Global Esteem. The theoretical rationale for the PSDQ is based on the Marsh/Shavelson self-concept model, previous SDQ research (see Marsh, 1990), and CFA tests of the structure of physical fitness (Marsh, 1993b). Research thus far (see review by Marsh, in press) has provided good support for the PSDQ factor structure and its generalizability over different samples and over gender (Marsh, Richards et al., 1994). Support for construct validity also comes from studies relating PSDQ responses to two other physical-self-concept instruments (Marsh, Richards et al., 1994) and to a set of external validity criteria including body composition, physical activity levels, and physical fitness tests of cardiovascular endurance, strength, and flexibility (Marsh, in press). The present investigation extends this construct validity research by examining the stability of PSDQ responses over time and using this stability data to evaluate further support for the discriminant validity of PSDQ responses.

Marsh, Richards et al. (1994) conducted an MTMM analysis of responses to the PSDQ and two other physical-self-concept instruments. The MTMM design is the most widely used construct-validity paradigm specifically developed to assess convergent, discriminant, and construct validity, and is widely recommended in self-concept research (Shavelson et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.