This study examined the validity and reliability of a Volleyball Skills Assessment Test (VSAT) as a measure of volleyball skill and as a predictor of team success in Special Olympics International (SOI) volleyball competition. Test-retest reliability data from 130 SOI volleyball players with mental retardation (101 males and 29 females) in the sixth week of an SOI volleyball training program yielded intraclass reliability coefficients (R) above .80 for all VSAT subtests (forearm pass, spike, set, serve) across gender with the exception of the set test for females (R = .75). Multivariate test battery test–retest reliability, examined using canonical correlation analysis, yielded moderate total redundancy estimates ranging between 62.5 and 66.1%. A high degree of concurrent validity was evidenced when correlating VSAT scores with judges’ ratings of performance on the four skills: r = .93 (r 2 = .86) serve, r = .94 (r 2 = .88) pass, r = .98 (r2 = .96) spike, and r = .86 (r2 = .74) set. Contingency table analysis, multiple regression, and discriminant function analysis revealed that the predictive validity of the VSAT as the primary determinant for allocating teams to pools of equal ability is questionable.
Steve B. Downs and Terry M. Wood
Emily Cole, Terry M. Wood and John M. Dunn
Tests constructed using item response theory (IRT) produce invariant item and test parameters, making it possible to construct tests and test items useful over many populations. This paper heuristically and empirically compares the utility of classical test theory (CTT) and IRT using psychomotor skill data. Data from the Test of Gross Motor Development (TGMD) (Ulrich, 1985) were used to assess the feasibility of fitting existing IRT models to dichotomously scored psychomotor skill data. As expected, CTT and IRT analyses yielded parallel interpretations of item and subtest difficulty and discrimination. However, IRT provided significant additional analysis of the error associated with estimating examinee ability. The IRT two-parameter logistic model provided a superior model fit to the one-parameter logistic model. Although both TGMD subtests estimated ability for examinees of low to average ability, the object control subtest estimated examinee ability more precisely at higher difficulty levels than the locomotor subtest. The results suggest that IRT is particularly well suited to construct tests that can meet the challenging measurement demands of adapted physical education.
Margaret J. Safrit, Terry M. Wood and Rod K. Dishman
Sonstroem's psychological model for physical activity offers a testable theory for understanding certain aspects of involvement and outcomes among adolescent boys. The usefulness of the model for other populations cannot be clarified, however, until the psychometric properties of its technology, the Physical Estimation and Attraction Scales (PEAS), are known for the groups studied. As a step in this direction, the factorial validity of PEAS responses among college males (N = 488) and females (N = 347) was examined. An independent group of college females (N =413) was also sampled to examine the general ability of the initial findings. These results revealed a robust factor of items that apparently tap perceptions of general physical competence and a perceived strength factor. These emerged across samples and analyses and were not gender-specific. Investigators using the PEAS with adult populations should consider its unique factor structure in the process of testing Sonstroem's physical activity model. Psychometric research regarding revision of the PEAS for adult populations is recommended with the aim of reducing instrument length while maintaining construct validity and measurement precision.
Robert W. Schutz, Frank L. Smoll and Terry M. Wood
Simon and Smoll's (1974) inventory for assessing children's attitudes toward physical activity (CATPA) has been used in numerous studies of children's at-titudinal dispositions and their relationships to a variety of situational and dispositional variables. Recent research revealing low attitude-behavior relationships and instability across time has raised questions about the psychometric properties of the CATPA inventory. The purpose of this research was to psychometrically analyze the six attitude subdomains of this semantic differential inventory and derive recommendations for its modification. The first of three studies reported herein included a four-phase analysis of the CATPA scores of 1,752 children, the results of which indicated that (a) three of the original eight bipolar adjectives were not good discriminators, (b) internal consistencies were high and were not improved by reciprocal average reweighting, and (c) a seven-factor structure emerged, differing from the underlying six-factor theoretical model. In Study 2 a revised CATPA inventory was administered to 1,895 boys and girls. The findings supported the inventory revisions and suggested the necessity for dichotomizing one of the six original attitude sub-domains. Study 3 incorporated the derived rescoring procedures in the reanalysis of earlier attitudinal investigations. Results revealed that modifying the scales neither changed the nature or strength of attitude-behavior relationships nor did it affect the intraindividual stability of CATPA over a period of time. The revised CATPA inventory was deemed to be an improvement over the original instrument because of its superior psychometric characteristics and reduced length, thereby making it more efficient for administrative purposes.