Methods: This study included data from 444 children and 103 teachers. Teachers’ belief was measured by Modernity Scale (Schaefer & Edgerton, 1985) and children’s self-regulation was measured by Head-Toe-Knees-Shoulders (Ponitz et al., 2009). Children’s academic achievement was measured by PPVT-III (Dunn & Dunn, 1997), PALS-preK (Invernizzi et al., 2000), and WJ-III (Woodcock et al., 2001; Applied Problems and Quantiative Concepts). Multilevel structural equation models were established for litearcy and math.
Results: The model fit of the literacy model was good, x2=5.34, df= 5, p = .38; CFI= 1.00, RMSEA= 0.01, SRMR= .02. At the class-level, teachers’ belief was only significantly associated with self-regulation (β = .17). At the individual-level, children’s self-regulation was significantly associated with literacy (PPVT: β= .27; PALS: β= .34). The model fit of the math model was good x2=21.84, df= 3, p = .00; CFI= .96, RMSEA= .13, SRMR = .05. At the class-level, teachers’ belief was significantly associated with self-regulation (β = .30), which in turn was significantly associated with math abilities (AP: β = .90; QC: β = .77). Mediation effect was significant for AP (β = 1.34) and marginal for QC (β = 1.12). At the individual-level, children’s self-regulation was significantly associated with math abilities (AP: β = .42; QC: β = .46).
Conclusions: The results show that teachers’ child-centered beliefs may contribute to children’s self-regulation and cognitive outcomes. Math abilities were associated with both class- and individual-level self-regulation, while literacy was only associated with individual-level self-regulation. This suggests the importance of considering teachers’ beliefs as a quality indicator for promoting children’s school-readiness through self-regulatory development.