Doctor of Philosophy (Ph.D.)

Part 1 of this three-part dissertation investigates the patterns of relation between high-ability and average-ability students‟ beliefs in intelligence (entity and incremental beliefs), achievement goal orientations (mastery-approach, performance-approach, and performance-avoidance goals), and four student outcomes (achievement, persistence, effort regulation, and subjective well-being) in the context of mathematics learning in Singapore. Part 2 of the dissertation examines the relations of goal orientations, affect (interest and boredom) and student outcomes in the same context. Part 3 combined Part 1 and 2 to investigate the associations between belief in intelligence, goal orientations, affect variables and student outcomes. The sample consists of 1355 Secondary One students recruited from eleven secondary schools in Singapore. The instruments administered included the Raven‟s Advanced Progressive Matrices, the Attitude in Learning Math Survey and a standardized math achievement test. The sample was divided into high-ability and average-ability groups based on the Raven‟s score.

Each part of this empirical study consists of three phases of analyses. Phase 1 of Part 1 and 2 built the basic path models by establishing (a) the direct relations between belief in intelligence and student outcomes; and (b) the direct relations between belief in intelligence and goal orientation. Since the basic path models have been established in Part 1 and 2, these procedures were not necessary in Part 3. Phase 2 generated the meditational models in separate path models of the relevant variables in each part of the study. Phase 3 of Part 1 through 3 investigated the structural fit indices using multiple-group analyses and test of significance using Bootstrap Method for the path models.

Several key findings of this study have contributed to the literature in understanding belief about intelligence, goal orientations and the associated affect and student outcomes between average and high-ability students. The relation between entity belief and mastery-approach goal was positive for high-ability students but it was negatively related for average-ability students. Although the results for average-ability students were consistent with Dweck‟s (1986) theory, the results for high-ability students were not.

The relations between entity belief and mastery-approach goals; and performance-approach goal were positive for high-ability students but the relations were negative for average-ability students. The findings for average-ability students were consistent with studies in the literature, the relations between incremental belief and approach goals in high-ability students were not.

The relation between incremental belief and math achievement was mediated by mastery-approach goals for both ability groups. For high-ability students, decreased boredom further enhanced math performance. The negative relations between entity belief and math achievement were mediated by mastery-approach and performance-avoidance goals for average-ability students.

In line with Harackiewicz, Barron, Pintrich‟s (2002) findings on the adaptive potential of performance-approach goals, the positive relations between incremental belief and persistence; and effort regulation, were mediated by both the mastery and performance-approach goals for both ability groups. There were negative relations between entity belief and persistence; and effort regulation which were mediated by performance-avoidance goal for both ability groups. Specifically, for average-ability students, entity belief decreased approach goals but increased avoidance goals, in turn decreased the level of interest and increased boredom, which decreased the level of persistence. For high-ability students, however, the positive relations between entity belief and persistence were mediated by both mastery-approach and performance-avoidance goals, in which further mediated by interest, in turn increased level of persistence.

While there was no relation between belief about intelligence and subjective well-being for average-ability students, the relation was mediated by mastery-approach and interest for high-ability students. This finding echoes Elliot and Thrash‟s (2002) findings on positive associations with mastery-approach goal. These findings provide insights to differential learning experiences of learners with different abilities. Consistent with Harackiewicz et al.‟s (1997) findings, for both ability groups (a) interest did not mediate the relation between goal orientations and math achievement; (b) the positive relations between interest and approach goals; and (c) the negative relation between interest and avoidance goals. It was found that there was positive relation between entity belief and interest which mediated by approach goals for high-ability students. However, it is the reversed pattern for average-ability students.

Similarly, boredom demonstrated between-group variations in the meditational relations between goal orientation and student outcomes. For instance, the negative relations between approach goals and math achievement were mediated by boredom for high-ability students. This relation does not appear in the path models for average-ability students. Boredom plays an important role in mediating the relations between entity belief and persistence for average-ability students, in which increased boredom decreased the level of persistence. Boredom, however, was not part of the path model for high-ability students. On the contrary, decreased boredom increased the level of persistence when average-ability students endorsed incremental belief. As for the relation between incremental belief and effort regulation, decreased boredom in turn increased the level of effort regulation for both ability groups.

These findings underscore the variations of the role of affect between the ability groups in this study.

Parenting and children’s academic outcomes are among the most important concerns for parents not only in Singapore but all over the world. In this study, key components of parenting were explored and their effects on learning outcomes of the student were tested and verified, using a nationally representative sample of more than 4000 grade 9 students from 39 secondary schools in Singapore. The learning outcomes comprised of students’ perceived metacognitive self-regulation, attitude towards homework, individual and group engagement, classroom misbehaviour, and academic achievement. These variables were selected because they had been found to be closely related to academic performance.

Scientific research on parenting and its consequences started in early 20th century. Baumrind’s three-fold typology of parenting styles might be the most influential research work on parenting. Authoritative parenting was endorsed by many researchers including Steinberg and his colleagues. Grolnick, Deci and Ryan also proposed parenting dimensions that could satisfy the three basic human needs in accordance with the Self-Determination Theory. Noting the shortfall of Western theories to cater to Asian culture, three components of parenting practices suited for students in Singapore were identified and tested in this study, using both exploratory and confirmatory factor analyses. Guided by Eccles’ social-cognitive expectancy-value model of achievement motivation, two measures of student’s motivational beliefs – self-efficacy and task value – were used as mediators in trying to explain the relationships between parenting and learning outcomes. Multi-level Structural Equation Modelling (SEM) technique was used to explore the direct and indirect effects of parenting components on students’ learning outcome variables.

From the analyses of the structural models using two sets of samples in learning of Mathematics (N=2094) and English (N=1941), parental autonomy support, parental monitoring of children’s activities, and parental involvement in learning emerged as three clear factors. When studying the effects of each component of parenting on the learning outcomes, children’s motivational beliefs were found to mediate the effects of parental autonomy support and parental involvement in learning. However, the effects of parental monitoring of children’s activities were found only to be directly affecting students’ learning outcomes.

Results from the study suggests that the influence of parenting has a moderate effect on metacognitive self-regulation, attitude towards homework, and both individual and group engagement in learning. The effects on classroom misbehaviour and academic achievement were very small, except for a moderate direct effect of parental monitoring of children’s activities on classroom misbehaviour. The effects of parenting on learning outcomes seem to be consistent for the two set of samples. The findings from this study have helped to address the gap in parenting research in the Asian context especially for children and parents in Singapore.

Student engagement has gained substantial attention in education research, policy, and practice. This is largely due to its purported associations with desired scholastic and non-scholastic outcomes like academic achievement and subjective well-being (SWB), both of which are considered as key indicators of effective schooling. While ample evidence suggests that student engagement is indeed empirically related to both academic achievement and SWB, the magnitude of the reported effect sizes is inconsistent between individual studies. Furthermore, there is also a dearth of systematic review and meta-analysis research that seeks to ascertain the strength of association between student engagement and the outcomes of interest.

One likely explanation for this observation is that student engagement is not a singular or unidimensional construct. Instead, it is a meta-construct consisting of multiple subcomponents that operate in varying contexts and timescale. In addition to its multifaceted and complex nature, student engagement is also plagued with the issue of conceptual haziness. As a result, it is difficult for researchers to interpret, synthesise, generalise, extend, and apply the findings obtained from student engagement research. To understand the current state of the construct and ascertain the widely proclaimed benefits of student engagement, the present thesis aims to conduct a systematic review and meta-analysis to address three overarching research questions: (1) What is student engagement? (2) To what extent is student engagement associated with academic achievement? (3) To what extent is student engagement associated with SWB?

A total of 124 quantitative studies, which consisted approximately 144,521 participants, were retrieved through a systematic review. Upon analysing the student engagement measures that were employed in these studies, it was revealed that each type of student engagement dimension (i.e., affective, behavioural, cognitive) has been operationalised disparately in the literature. In fact, they could be further differentiated into (at least) seven conceptually distinct engagement subtypes. A series of meta-analyses were subsequently conducted on studies involving 3rd to 12th grade students. Using meta-regression with robust variance estimation, it was revealed that student engagement has a large average correlation with academic achievement (r = .34) and SWB (r = .36). Upon closer inspection, academic achievement was found to have the strongest association with behavioural engagement (r = .40), followed by cognitive (r = .33) and affective engagement (r = .27). SWB was most closely related to affective engagement (r = .40), followed by cognitive (r = .35) and behavioural engagement (r = .32). Moderation analyses indicated that the magnitude of these effect sizes was influenced by the way engagement was operationalised in the primary studies, as well as other factors like the informant source of engagement and type of achievement measure.

While the present study showed that student engagement is associated with positive student outcomes, it also illustrated how student engagement is, at the current point in time, overgeneralised and in need of conceptual refinement. To advance the field of student engagement, therefore, the thesis concluded with the proposition of the Dual Component Framework of Student Engagement. Informed by extant student engagement perspectives, the framework posits that student engagement is composed of two separate but related constructs—learning activity engagement and school community engagement—that differ in their conceptual definitions, boundaries, object, and dimensionality. The theoretical, research, and applied implications of the newly proposed framework were discussed.