Search Results
Title: A Dynamic Mixture Biometric Model of Cognitive Development in the NLSY Children
Resulting in 1 citation.
| 1. |
Hunter, Michael D. Bard, David E. Beasley, William H. Meredith, Kelly M. Rodgers, Joseph Lee |
|
A Dynamic Mixture Biometric Model of Cognitive Development in the NLSY Children Presented: Charlottesville VA, Behavior Genetics Association Annual Meeting, June 2014 Cohort(s): Children of the NLSY79 Publisher: Behavior Genetics Association Keyword(s): Digit Span (also see Memory for Digit Span - WISC); Genetics; Kinship; Modeling, Multilevel; Peabody Individual Achievement Test (PIAT- Math); Peabody Individual Achievement Test (PIAT- Reading); Peabody Picture Vocabulary Test (PPVT); Siblings Permission to reprint the abstract has not been received from the publisher. A novel method of combining within-person and between-person variability in a biometrically informed model was used to examine nonlinear cognitive development in the National Longitudinal Survey of Youth-Child/Young Adult (NLSYC/YA) dataset. Entirely within person biometric models (e.g. Molenaar 2010) can be fit, but generally assume that all persons are heterogeneous. By contrast, conventional between-person biometric models (e.g. Martin & Eaves 1977) make the opposite assumption: that the sample is uniformly homogeneous. State space mixture modeling (SSMMing) is a middle ground. SSMMs make a within-person longitudinal biometric model for each pair of genetically related participants to account for the idiographic nature of genetic and developmental variability (Nesselroade, Gerstorf, Hardy, and Ram 2007; Molenaar, Boomsma, and Dolan 1993). Simultaneously, SSMMs allow for a finite number of groups that are within-group homogeneous and between-group heterogeneous to allow for uniformity in development among some people. The longitudinal model in SSMMs has both autoregressive and linear slope components with individually estimated growth trajectories. Hence, nonlinear patterns of change are allowed in the context of linear modeling. Five longitudinally measured cognitive variables (PIAT Reading Recognition, Reading Comprehension, and Math; PPVT; and Digit Span) from the NLSYC are used both to illustrate SSMMs as a method and to provide insight into this important process. The finding that cognitive ability is highly heritable between individuals was replicated in cross-sectional subsets of the NLSYC. However, the within-person longitudinal model showed minimal contribution from additive genetic variance across the five cognitive variables. A SSMM with two groups found a small subgroup in which cognitive ability was heritable within persons, but for the majority of individuals studied the intraindividual variance was dominated by common and specific environmental factors. The structure of intraindividual heritability of cognitive ability thus appears quite different from that found in conventional between person biometric modeling. |
|
|
Bibliography Citation
Hunter, Michael D., David E. Bard, William H. Beasley, Kelly M. Meredith and Joseph Lee Rodgers. "A Dynamic Mixture Biometric Model of Cognitive Development in the NLSY Children." Presented: Charlottesville VA, Behavior Genetics Association Annual Meeting, June 2014. |