Construct validity of stages of change for exercise behavior

Purpose. Previous research examining the transtheoretical model of behavior change within the exercise do-main has been limited by use of self-report measures exclusively and inconsistent practices with regard to stage of exercise assessment. The present study was designed to partially circumvent these limitations and extend the current literature by determining the degree of association among stage of exercise and body mass index, cardiorespiratory fitness, exercise behavior relapse, barriers, and self-efficacy, after controlling for several potential confounders. Design. A descriptive, cross-sectional study. Subjects. Two hundred thirty-five adults (M age = 34.7 years) volunteered to participate. Measures. Participants were classified by stage of exercise and compared on two behavioral, two biometrical, and three psychological variables while statistically controlling for social desirability and demographic differences. Results. Significant between-stage differences were found for the overall set of dependent variables (p <.0001) and for each dependent variable separately (p <.01). The proportion of variance accounted for by the dependent variables ranged from .06 to .53. Conclusions. This study offers objective support for the stage-of-change model within the exercise domain. By acknowledging and accurately assessing stage of exercise, researchers and clinicians may be able to improve physical activity promotion efforts.