ABSTRACT
Purpose: We tested the effects of an autonomy-supportive intervention in physical education (PE) on high-school students’ autonomous motivation in PE, and their autonomous motivation, intentions, and physical activity (PA) behavior in a leisure time guided by the trans-contextual model. Method: PE classes in two schools were assigned to receive either an autonomy-supportive intervention and/or a control intervention via random allocation by the school. The PE teacher of the school assigned to the autonomy-supportive intervention was trained to provide autonomy support while the PE teacher of the school assigned to the control intervention received no training. Students (N = 256) in all classes completed measures of perceived teacher autonomy support, autonomous motivation in PE and leisure time, and beliefs, intentions, and PA in leisure time before and immediately after the intervention. Results: Results revealed direct effects of the autonomy-supportive intervention on changes in perceived autonomy support. However, there were no direct intervention effects on change in intentions and PA behavior. The intervention also had indirect effects on changes in autonomous motivation in PE and leisure time. Additionally, change in perceived autonomy support had direct effects on change in autonomous motivation in PE and indirect effects on change in leisure-time autonomous motivation. Changes in autonomous motivation in leisure time had direct effects on changes in beliefs and indirect effects on changes in intentions and PA behavior through changes in beliefs. Conclusion: The study provides valuable information on the effect of autonomous supportive climate on students’ beliefs toward PA in PE lessons and in their leisure time outside of school.
Notes
1 Data files, analysis scripts, and output from the data analysis including reliability, statistical power, and path analyses are available online: https://osf.io/b4t9c.
2 Path analyses were conducted without FIML imputation, patterns of effects were unchanged, full results are available online: https://osf.io/b4t9c.
3 Where possible Omega reliability coefficients (Revelle, Citation2019) were computed. For two-item scales, the Spearman–Brown inter-item correlation was computed. For the relative autonomy index in leisure time at baseline, the Omega reliability calculation did not converge due to a non-positive definite matrix, so the standard Cronbach alpha is reported.