Causal Models of Physical Exercise Participation: Testing the Predictive Power of the Construct "Pressure to Change"

This study examines 2 different causal models to predict physical exercise motivation and behavior under a longitudinal perspective. The first model includes 5 latent variables that were hypothesized to have an impact on exercise intention and behavior: behavior-specific social support, exercise self-efficacy, perceived health benefits, perceived barriers, and subjective vulnerability to cardiovascular disease. The second model was based on all variables of the first model, but additionally included the new variable "pressure to change." Pressure to change was defined as the extent to which a person feels the necessity that specific personal life circumstances (e.g., health status, social relations) may not remain as they are and ought to be changed. It was hypothesized that the inclusion of health-related pressure to change would result in a better prediction of exercise intention. The proposed causal models were tested separately at the stages of exercise adoption and maintenance. Covariance structure analyses (LISREL) confirmed that pressure to change may be an important factor in the motivational process that leads to the adoption of regular physical exercise. Adding this latent variable to the basic model improved the amount of explained variance in exercise intention by 6%. Furthermore, the results did not support the assumption that cognitive control is critical especially during the acquisition of exercise behaviors, but may be less influential once the behavioral routines have been established. Our data rather indicate that regular physical exercise, even if performed on a regular basis for years, always remains a behavior that requires a high level of cognitive guidance.