We note here that there are a multitude of valid and reasonable approaches that investigators may take when conducting a genetic study of a multiethnic, family-based sample with longitudinal measures of phenotype, behaviors and the environment. For example, Add Health researchers may be interested in imputing genotypes for purposes of combining association signals across multiple data sources genotyped on different platforms. Add Health researchers may also be interested in incorporating Add Health sampling weights and/or taking into account phenotypic clustering particularly when studying genetic risk factors within an environmental context. Furthermore, the family-based structure of the data would allow for more specific analyses using informative sibling pair family units through a variety of family-based association approaches (e.g. FBAT, Laird et al., 2000).
