Poster #78 - Identifying and Predicting Highly Able Children’s Achievement Trajectories: Effects of Teachers, Peers, and Child Characteristics

• In Event: 2-110 - Poster Session 07
  In Poster Session: PS 07 Section - Education, Schooling

Integrative Statement

Research has shown that highly able children can underachieve (e.g., Baker et al., 1998). In this research, underachievement is usually defined as a discrepancy between children’s ability and achievement (Reis & McCoach, 2000). However, longitudinal research in broad community samples of elementary school children that examines the trajectories of highly able children’s achievement is missing. Our first aim is to examine achievement and underachievement trajectory classes of these children. The second aim is to examine predictors of these classes, more specifically, the role of children’s socio-emotional and learning-related characteristics and their relationships with teachers and peers. We expected that all of these antecedents would have an effect, based on previous research and theory on underachievement in general and highly able samples (e.g., McCoach & Siegle, 2003; Obergriesser & Stoeger, 2015). However, we had no specific expectations regarding their relative importance, given the limited amount of longitudinal research in highly able samples and the lack of studies including all of these antecedents.
A sample of 344 highly able students was used to answer our research questions. We selected the children with the 10% highest scores on cognitive abilities in a representative sample of 3,444 third-graders, aligned with Gagné (2004). Mathematics achievement was assessed using validated standardized tests in Grade 4, 5, and 6. In Grade 4, we assessed aggressive, hyperactive, and anxious/fearful behavior, school liking, self-confidence, independence, mathematical self-concept, closeness and conflict with the teacher, and peer relationships.
We conducted Latent Class Growth Analysis in Mplus to determine the trajectory classes in math achievement, while controlling for cognitive abilities. The three-class solution showed the largest BIC-drop, highest entropy, best average posterior probability estimates and a significant BLRT. We defined a high, an average, and a low achieving class, see Figure 1. The groups did not differ significantly on cognitive abilities. Afterwards, equality of means analyses were conducted to determine predictors of these classes, see Table 1. We tested both for overall effects, showing differences among the classes in general, and for differences between classes, two-by-two. There were overall significant differences in independence, self-confidence and mathematical self-concept. All effects were in the expected direction, i.e., higher achievement was associated with more independence, more self-confidence and better mathematical self-concept. Also, average achieving children showed more hyperactive and anxious/fearful behavior than high achieving children.
Our results confirm the notion in literature that there are well-achieving and underachieving highly able children. In addition, our analyses show that the group of low achieving students seems relatively small. However, we find that there is also a relatively large group of average performing highly able students, who score slightly above class average. It seems that they are also underperforming, based on their cognitive capacities. Further, our study shows that independence, self-confidence, mathematical self-concept, and hyperactive and anxious/fearful behavior are key predictors to differentiate between achievement groups. Unexpectedly, we find no relation with teacher-student relationship quality or peer status. Further research could explore if other aspects of teacher-student and peer interactions are predictive for achievement of highly able students.

Authors

• Tessa Weyns, KU Leuven
  Presenting Author

• Bieke De Fraine, KU Leuven
  Non-Presenting Author

• Karine Verschueren, KU Leuven
  Non-Presenting Author