Longitudinal Trajectories of Observed Behavioral Engagement Predict Second Graders’ Reading Achievement

• In Event: Poster Session 14
  In Poster Session: PS 14 Topic - Education, Schooling

Abstract

The importance of learning and achievement in the early elementary years for later academic success is well-documented within the educational, developmental, and clinical literatures (Duncan et al., 2007; Ladd & Dinella, 2009; Watts et al., 2014). Investigations designed to identify processes that facilitate academic success often document the importance of students’ school engagement (Fredricks et al., 2004; Hughes et al., 2008; Wang & Eccles, 2012). Among different types of school engagement, behavioral engagement, which refers to participation in the learning environment (e.g., concentration, attention, persistence) and adherence to classroom rules (Finn, 1993; Skinner & Belmont, 1993), is particularly important for children in the early elementary years (Bryce, et al., 2018; Hughes et al., 2008). However, there are key limitations in the behavioral engagement literature, including the overreliance on concurrent data and use of self- or teacher-reports of engagement. The goals of this study were to overcome these two limitations by identifying developmental trajectories of observationally coded engagement across the early elementary years and by examining if behavioral engagement trajectories were associated with changes in academic achievement. Furthermore, we evaluated whether children’s family socio-economic status (SES) and initial academic performance moderated the associations between behavioral engagement and academic achievement.

Data were collected from 301 children in the United States who were observed each semester from kindergarten (Mage = 65.74 months; 49% boys) to 2nd grade. Children’s behavioral engagement was observed from kindergarten throughout 2nd grade. Reading and math achievement were assessed via standardized tests in kindergarten and 2nd grade. Family SES was measured by parents’ educational attainment and household income. Covariates included child age, sex, and race/ethnicity.

We tested linear growth mixture models with one to four classes and found that the 2-class solution was the most appropriate (see Figure 1). The majority of children (n = 262, 87.0%) were in the high-decreasing class, with behavioral engagement being relatively high in the fall of kindergarten (Mintercept = 2.82, p < .001; Variance = 0.005, p = .012) and decreasing significantly across time (Mslope = -0.034, p < .001; Variance < 0.001, p = .227). Other children (n = 39, 13.0%) were in a moderate-stable class, with behavioral engagement being moderate in the fall of kindergarten (Mintercept = 2.53, p < .001; Variance = 0.058, p = .028) and stable across time (Mslope = -0.012, p = .614; Variance = 0.007, p = .018). After controlling for achievement in kindergarten and demographic variables (i.e., child age, sex, race/ethnicity, and family SES), children in the high-decreasing class displayed higher reading (B = 6.68, p = .016), but not math (B = 2.41, p = .338), than children in the moderate-stable class. Additional analyses revealed that differences in reading achievement between the two classes were present only for children from low SES families or for children low in kindergarten reading achievement (see Figure 2). These findings suggest that children exhibit different developmental trajectories of behavioral engagement, and, compared to their peers, economically or academically at-risk children may benefit more from high behavioral engagement.

Authors

• Longfeng Li, Pennsylvania State University, University Park
  Presenting Author

• Carlos Valiente, Arizona State University
  Non-Presenting Author

• Nancy Eisenberg, Arizona State University
  Non-Presenting Author

• Tracy L Spinrad, Arizona State University
  Non-Presenting Author

• Sarah Katherine Johns
  Non-Presenting Author

• Rebecca H Berger, Urban Institute
  Non-Presenting Author

• Marilyn S. Thompson, Arizona State University
  Non-Presenting Author

• Jody Southworth, Arizona State University
  Non-Presenting Author

• Armando A. Pina, Arizona State University
  Non-Presenting Author

• Maciel M. Hernández, University of California - Davis
  Non-Presenting Author

• Diana Elena Gal-Szabo, Arizona State University
  Non-Presenting Author