A Bayesian approach to students’ perceptions of teachers’ autonomy support

Background: How to best operationalize teachers’ autonomy support, an instructional style aiming to satisfy students’ psychological need for autonomy, is unclear because teachers can support the whole class and/or individual students. Students might perceive inequalities concerning the autonomy support they receive relative to classmates, which might undermine their motivation and engagement. Aims: The current study aims to contribute to the conceptualization of autonomy support. We investigated students’ perceptions of teachers’ autonomy support (individual, class-directed, and perceived differences), concerning choice provision, fostering relevance, stimulating interest, and acknowledging frustration, and associations with students’ motivation and engagement. Sample: 446 Dutch primary school students (age_range = 9–14) from 22 mathematics classes. Methods: With Bayesian Multilevel-CFA and -SEM, we examined the factorial structure of students’ perceptions of teachers’ autonomy support and associations with motivation and engagement. We evaluated whether individual and class-directed autonomy support were distinct constructs, both concerning individual ratings at the student level, and regarding the whole-class-aggregated assessments at the class level. Results: Individual and class-directed autonomy support was differentiated at the student level. At the class level, one factor (overall autonomy-supportive atmosphere) was found. Regarding perceived differences, we revealed three student-level factors (e.g., relative lack of autonomy support). At the student level, individual and class-directed autonomy support positively predicted intrinsic motivation, effort, and persistence; perceived relative lack of autonomy support positively predicted extrinsic regulation. Conclusions: Both individual and class-level support should be high to yield optimal results for students’ motivation and engagement. Focusing solely on class-directed autonomy support may omit essential information.