Abstract
Adopting deep approaches to learning can have a profound impact on learning outcomes. The extent of change in the learning approach could be attributed to the effect of contextual factors (e.g., instruction). After a substantive review, it was found that research interested in investigating the longitudinal effect of deep approaches to learning on learning outcomes were rarely directly concerned with the longitudinal state and trend of the approach itself. Moreover, the limitations of past analytical methods (e.g., using a single indicator to represent deep approaches to learning at each point in time), has not been appropriately acknowledged. This study examined deep approaches to learning in students from kindergarten to third grade using a multilevel structural equation modeling indicator-specific growth model (MSEM-ISGM). This was used to highlight the methodological issues (e.g., consider four indicators separately at each wave) to investigate the effect of context and the states, stability, and changes in deep approaches to learning over this developmental period. The current study comprised of a large-scale sample of 21,490 kindergarteners in kindergarten. Results showed: (1) there was a contextual effect on the longitudinal changes of deep approaches to learning; (2) deep approaches to learning was high but showed considerable individual differences; (3) most indicators relating to deep approaches to learning declined (however, one increased), whereas the trend were relatively stable over time. Corresponding suggestions were proposed at the end of this article.