Across European countries, sustaining students’ motivation for mathematics remains a persistent challenge. International studies document that students in many European countries report low motivation for the subject (OECD, 2023), with a particular decline across secondary education (Hannula, 2019). These trends are particularly concerning in light of European policy ambitions to broaden participation in STEM programs (e.g., as outlined in the European Commission's STEM Education Strategic Plan). Understanding how mathematics teachers can promote students’ motivation for mathematics is therefore of key importance.

The present study investigates this issue from the viewpoint of Self-Determination Theory (SDT; Ryan & Deci, 2017). In short, SDT posits that students are more likely to develop high-quality motivation when their basic psychological needs for autonomy, competence, and relatedness are met. Teachers can nurture these needs through three need-supportive teaching practices (Niemiec & Ryan, 2009): autonomy support (e.g., acknowledging students’ perspectives and offering meaningful choices), involvement (e.g., warmth, care, and emotional availability), and structure (e.g., clear explanations, guidance, and informative feedback). To date, a substantial body of research has documented positive associations between these practices and students’ motivational functioning (Stroet et al., 2013). However, much of this research implicitly assumed that teaching practices influenced student motivation in a unidirectional manner, attributing (typically cross-sectional) associations to an effect of teaching on motivation (Howard et al., 2025). However, there are good theoretical reasons to assume also a reverse path, in which student motivation would affect subsequent need support. For example, need‑supportive teaching requires teachers to be flexible, emotionally available, and attentive to students’ needs. Investing such effort is more likely when teachers already experience some success in motivating students, as this would satisfy teachers’ own need for competence (Taylor et al., 2008). In addition, motivated students may elicit more need-supportive teaching through their engaged behavior. For example, when students ask a lot of questions and spontaneously signal when they get stuck with a problem, this may give teachers a better idea of students’ needs to which then can adjust their instruction (Jang et al., 2024). Teachers may also consciously decide to provide more need support to motivated students because they believe such practices are only efficacious to such students. For example, teachers tend to believe that granting autonomy only works for motivated students, but that coercion is needed to enforce cooperation from disaffected students (Hornstra et al., 2015).

This study thus investigates bidirectional relations between mathematics teachers’ provision of need-support (autonomy support, involvement, and structure) on one hand, and two aspects of students’ motivational functioning in mathematics (i.e., student autonomous motivation and behavioral engagement) on the other. We drew on a large longitudinal sample from Flanders in which students were surveyed four times across Grades 7 and 8. To differentiate between-student and within-student effects (Orth et al., 2021), we estimated both conventional cross-lagged panel models (CLPMs) and random intercept cross-lagged panel models (RI-CLPMs). In general, we hypothesized all three dimensions of need-supportive teaching to positively affect subsequent students’ autonomous motivation and behavioral engagement; the other way round, we expected student motivation to positively affect subsequent need-supportive teaching. We expected to find these associations both at the between-student level (i.e., in the CLPMs) and at the within-person-level (i.e., in the RI-CLPMs). Because of its more overt nature, we expected that behavioral engagement would exert a more pronounced influence on subsequent need-supportive teaching than autonomous motivation. Finally, because of changes in math teachers between Grades 7 and 8, we expected effects to be somewhat more pronounced within grades (i.e., between W1 and W2 and between W3 and W4) than across grades (i.e., between W2 and W3).

The study used data from a large longitudinal project conducted in Flanders. The sample consisted of 3,409 Grade 7 students (50.2% male, Mage = 12.4 years) nested in 166 classes across 27 secondary schools, followed across four measurement waves spanning Grades 7 and 8 with 6-month intervals. Participation rates per wave exceeded 90%. At each wave, students completed validated self-report measures assessing their autonomous motivation for mathematics (interest, enjoyment, and personal importance; Eccles & Wigfield, 1995), their behavioral engagement in mathematics classes (effort, attention, and persistence; Skinner et al., 2008), and perceptions of mathematics teacher autonomy support, involvement, and structure (Belmont et al., 1988). All items were rated on five-point Likert scales. For all measures, measurement invariance across time was established. To examine reciprocal relationships over time, both cross‑lagged panel models (CLPMs) and random‑intercept cross‑lagged panel models (RI‑CLPMs) were estimated. CLPMs capture how high levels of need-supportive teaching at one time point predict high subsequent student motivation (and vice versa), while accounting for prior levels and concurrent relationships, thus associating rank-order changes in variables across time. RI-CLPMs separate stable between-student differences from time-specific within-student fluctuations by including random intercepts for each construct. Hence, these models investigate whether a temporal fluctuation in need-supportive teaching at one time predicts a similar fluctuation in motivation at the next time point (and vice versa). Both methodological approaches offer complementary insights (Orth et al., 2021): in combination, they provide a robust perspective on reciprocal associations at both the between-student and within-student level (Jang et al., 2024). After comparing the fit of several unconstrained and (partially and fully) constrained models, a partially constrained model was selected, in which cross‑lagged and autoregressive paths were set to be invariant over time, with the exception of paths across the transition from Grade 7 to Grade 8 as students typically had a different mathematics teacher before and after the transition. Analyses accounted for the nested data structure, with students clustered within classes. Gender was included as a covariate at each wave, and full‑information maximum likelihood estimation was used to handle missing data.

Model results showed that cross-lagged associations between autonomy support and involvement and student motivation tended to be bidirectional, whereas structure appeared more as a consequence than as an antecedent of student motivation. In line with SDT, these findings thus point at a potentially virtuous cycle, with autonomy support and involvement boosting student motivation, and motivated students eliciting further need support. First, teachers’ autonomy support predicted subsequent autonomous motivation and behavioral engagement, and both motivational variables predicted greater subsequent autonomy support. These paths were consistently observed across CLPM and RI‑CLPM specifications, attesting to their robustness. Second, teacher involvement demonstrated reciprocal relations primarily with behavioral engagement, again across both specifications. However, involvement did not consistently predict autonomous motivation over time, although autonomous motivation did predict later involvement. Finally, students’ motivation and engagement predicted later structure (although only in the CLPMs), but structure did not predict later student motivation. Arguably, the provision of structure seems to be less vital to improve students’ motivation. Across all analyses, behavioral engagement predicted subsequent teaching practices in a stronger fashion than did autonomous motivation, confirming that teaching practices respond more readily to students’ observable effort than to students’ less visible internal motives. Finally, reciprocal dynamics were observed to be stronger within the academic year than across grade transitions, indicating that teacher–student influences mainly develop when teachers and students work together over time.

Belmont, M., Skinner, E., Wellborn, J., & Connell, J. (1988). Teacher as social context: A measure of student perceptions of teacher provision of involvement, structure, and autonomy support. Eccles, J. S., & Wigfield, A. (1995). In the mind of the actor: The structure of adolescents' achievement task values and expectancy-related beliefs. Personality and Social Psychology Bulletin, 21(3), 215-225. Hannula, M. S. (2019). Young learners’ mathematics-related affect: A commentary on concepts, methods, and developmental trends. Educational Studies in Mathematics, 100(3), 309-316. Hornstra, L., Mansfield, C., Van der Veen, I., Peetsma, T., & Volman, M. (2015). Motivational teacher strategies: the role of beliefs and contextual factors. Learning environments research, 18(3), 363-392. Howard, J. L., Slemp, G. R., & Wang, X. (2025). Need support and need thwarting: A meta-analysis of autonomy, competence, and relatedness supportive and thwarting behaviors in student populations. Personality and Social Psychology Bulletin, 51(9), 1552-1573. Jang, H.-R., Basarkod, G., Reeve, J., Marsh, H. W., Cheon, S. H., & Guo, J. (2024). Longitudinal reciprocal effects of agentic engagement and autonomy support: Between-and within-person perspectives. Journal of Educational Psychology, 116(1), 20. Niemiec, C. P., & Ryan, R. M. (2009). Autonomy, competence, and relatedness in the classroom: Applying self-determination theory to educational practice. Theory and Research in Education, 7(2), 133-144. OECD. (2023). PISA 2022 Results (Volume I): The State of Learning and Equity in Education. OECD Publishing, Paris. Orth, U., Clark, D. A., Donnellan, M. B., & Robins, R. W. (2021). Testing prospective effects in longitudinal research: Comparing seven competing cross-lagged models. Journal of Personality and Social Psychology, 120(4), 1013. Ryan, R. M., & Deci, E. L. (2017). Self-determination theory: Basic psychological needs in motivation, development, and wellness. Guilford Publications. Skinner, E., Furrer, C., Marchand, G., & Kindermann, T. (2008). Engagement and disaffection in the classroom: Part of a larger motivational dynamic? Journal of Educational Psychology, 100(4), 765. Stroet, K., Opdenakker, M.-C., & Minnaert, A. (2013). Effects of need supportive teaching on early adolescents’ motivation and engagement: A review of the literature. Educational Research Review, 9, 65-87. Taylor, I. M., Ntoumanis, N., & Standage, M. (2008). A self-determination theory approach to understanding the antecedents of teachers’ motivational strategies in physical education. Journal of Sport and Exercise Psychology, 30(1), 75-94.