Prediction of Athletic Participation in Middle School Males

Abstract

During the first ten days of the academic year the Physical Estimation and Attraction Scales (PEAS) and Bialer's Locus of Control Scale were administered to seventh and eighth grade middle school males. Subsequently, calls were issued for the fall sport season (flag football, cross-country). Stepwise discriminant analysis produced a function containing Attraction first and Estimation second, which accurately predicted athletic participation, canonical r = .35 p < .001. Use of smaller, previously factored collections of PEAS items improved the prediction, canonical r = .40, p < .001. Prediction of a single, more specific activity, cross-country, by means of the more specific factor scores produced a canonical coefficient of .47, p < .001. PEAS Factor 5, referring primarily to activities of running and hiking, developed the largest relationship with cross-country participation of any study variable. Control did not enter either of these two equations. No significant relationship was found between any study variable and adherence (remaining on the cross-country team for the complete season). It was concluded that PEAS variables are capable of predicting initial exercise involvement. In terms of developing optimal psychometric predictions of exercise recruitment and adherence, it is recommended that attitude statements be developed which are more specific and congruent to corresponding desired behaviors.