Examining Longitudinal Associations between Internalizing Problems, Body Mass Index, and Language during Childhood

• In Event: Poster Session 13
  In Poster Session: PS 13 Topic - Developmental Psychopathology

Abstract

Body mass index (BMI), internalizing problems, and language are all important and interrelated constructs that impact development (Ames et al., 2015; Bornstein et al., 2013; Braet & Crombez, 2003; Clegg et al., 2015; Deighton et al., 2018), although relations between BMI and these constructs are not as robust in the early childhood period (Krombholz, 2012; Patalay & Hardman, 2019). Although some studies have explored the co-development of these constructs over time, they have typically utilized cross-lagged panel modeling (CLPM), which does not tease apart within- and between-person differences, and as a result, may produce upwardly biased estimates (Hamaker et al., 2015). Further, previous studies have not included BMI, internalizing problems, and language in the same model. The present study builds on previous literature by reexamining associations between these constructs across five time points (36 and 54 months, first, third, and fifth grades) by utilizing a newer approach to testing co-development among constructs (CLPM with random intercepts; RI-CLPM) that accounts for between-person stability and time-specific variations. It was hypothesized that random intercepts would be correlated between BMI and internalizing problems (positively), language and internalizing (negatively), and BMI and language (negatively). Further, we expected language would predict internalizing problems at all time points (Bornstein et al., 2013), that bidirectional associations between BMI and internalizing would emerge at first to third and third to fifth grades (Bradley et al., 2008), and that bidirectional associations between language and BMI from would be present from third to fifth grade (Holcke et al., 2008).
Data for this study come from the NICHD Study of Early Child Care and Youth Development. Participants included 1,364 children and one of their parents. A majority of the sample was White (80%) and 48% were female. All measures were collected when children were 36 and 54 months, and when they were in first, third, and fifth grade. Raw BMI scores were used. Internalizing problems were assessed using the anxious/depressed and withdrawn scales from the Child Behavior Checklist (Achenbach, 1992; Achenbach, 1991). Receptive language was assessed using the Reynell Development Language Scale at 36 months (RDLS; Reynell, 1969, 1990), and the Picture Vocabulary subtest of the Woodcock Johnson was used subsequently (McGrew et al., 1991).
A RI-CLPM was run that included several covariates: income-to-needs ratio, gender, maternal education, maternal vocabulary, maternal depression, father in the home, and race. Results (see Figure 1) showed that the internalizing problems and language random intercepts were significantly correlated. BMI was concurrently related to language at 36 months and internalizing problems at 54 months. Language at 36 months predicted internalizing problems at 54 months, language at 54 months predicted internalizing problems at first grade, and bidirectional relations emerged among language and internalizing problems between third and fifth grade. Finally, internalizing at third grade predicted BMI at fifth grade.
These findings shed light on previous research documenting transactional associations between BMI, internalizing problems, and language suggesting that some of these previous associations may have been due to between child differences that were not accounted for.

Authors

• Lindsey Bryant, Purdue University, West Lafayette
  Presenting Author

• Robert Duncan, Purdue University, West Lafayette
  Non-Presenting Author

• Kristine Marceau, Purdue University, West Lafayette
  Non-Presenting Author

• Sara Anne Schmitt, University of Oregon
  Non-Presenting Author