|
|
|
John S. Witte Associate Director Head, Division of Cancer Epidemiology Co-Leader, Program in Cancer Genetics, Helen Diller Family Comprehensive Cancer Center
UC San Francisco
Diller Bldg, Room 388 (UCSF Internal Box: 3110) 415-502-6882 |
SoftwareSoftware for Hierarchical Modeling There are a number of ways one can fit hierarchical models. For example, some software packages have been developed specifically for hierarchical modeling. However, these packages are generally designed for social science applications (e.g., in education research). Another approach is to use software supporting matrix algebra to fit hierarchical models. In Witte et al. (Epidemiology 1998;9:563-566), we provide a detailed discussion of software for hierarchical modeling. The following links should connect you to software appropriate for your own analyses. A SAS IML procedure from the manuscript "Software for Hierarchical Modeling of Epidemiologic Data" (Witte et al., Epidemiology 1998;9:563-566).Example SAS code and Macro from the manuscript "Multilevel Modeling in Epidemiology with GLIMMIX" (Witte et al., Epidemiology 2000;11:684-688). R code from the manuscript "Hierarchical modeling of linkage disequilibrium: genetic structure and spatial relations" (Conti and Witte, Am J Hum Genet 2003;72:351-363). R code from the manuscript "Enriching Genome-wide Association Studies with Hierarchical Modeling" (Chen and Witte, Am J Hum Genet 2007;81:397-404). GAUSS procedures written by Sander Greenland. This is an extensive set of procedures for undertaking many different analyses. It requires a recent version of GAUSS and understanding of GAUSS programming. One can use logregp.g for hierarchical modeling. |