An Application of Linear Mixed Models to the China Health and Nutrition Survey Data

Linear mixed models offer the basis for a methodology that provides flexibility in fitting models with various fixed and random elements in the presence of correlation among random effects and nonconstant variance. This thesis attempts to specify and fit linear models for growth data from two-year-old children in China using a unified maximum-likelihood approach. This thesis estimates variance components and selects appropriate covariance models. Further, it makes inferences about the mean growth curve and compares growth curves between genders, and between urban and rural areas.
THES
Li, Li
1999
48 leaves
Iowa State University
Ames, Iowa
188