ABSTRACT
This chapter captures students’ sense of school belonging (SBL) and teaching engagement (EGM) as well as the affective learning outcomes in mathematics – ATM. Three TIMSS 2015 scales are adapted to measure the above latent variables. Confirmatory factor analyses are run to check the fit of the latent variable models to the data. Descriptive statistics are conducted to identify the overall patterns and variation of the measures. Then, two hypothesised teaching mechanisms are tested using multilevel modelling and multilevel structural equation modelling. Overall, we find that master maths teachers’ students on average give a high rating on SBL, EGM and ATM. The multilevel analyses indicate that 12% of the variance in ATM is situated at the class level and that EGM explains 96.7% of the class-level variance. The multilevel structural equation modelling uncovers a direct effect of SBL on ATM at the student level and an indirect effect of SBL on ATM via its direct effect on EGM at both the student and class levels.
