PubH 7402 Biostatistics Models and Methods Spring 2009 -
Syllabus
- Intro to course
Notes for Intro to Course.
- General Linear Models (4 weeks) Suggested Readings: Chapters 8-11 in HH, Chapter 4 in Vittinghoff et al. (2005)
- Generalized linear model (4 weeks)
- Automated model selection
- MATERIAL COVERED ON REVIEW DAY - March 10, 2009
- Methods for correlated data (4 weeks)
- Clustered data
(e.g. patients within clinics, clinics within health plans), longitudinal data repeated measures, mixed effects modeling and GEE
- Notes for repeated measures (updated 4-23-09)
- Pancreatic enzyme example using SAS
- Pancreatic enzyme example using R
- Wide to long format in R demo written by David Solberg
(PH7402 student 2009)
- A now classic dataset used as an example for introducing hierarchical linear models is from the 1982 "High School and
Beyond" survey on Math Achievement of 7185 studetns from 160 schools. The data was used in Bryk and Raudenbush's first edition
1992 text Hierarchal Linear Models, a
step-by-step analysis of the data using SAS was
done by Judith Singer in Journal of Educational and Behavioral Statistics and a
step-by-step analysis of the data using R was done by John Fox as an appendix to his text
An R and S-plus Companion to Applied
Regression
- Pre-post Homeless data example .
- Read and discuss paper: Tu YK, Blance A,
Clerehugh V, and Gilthorpe MS (2005) "Statistical power for analyses of changes in randomized controlled trials",
Journal of Dental Research, 84(3): 283-287, 2005.
- Multiple time points
- Two ways of modeling intraclass correlation in error structure
- GEE for alcohol abstain by clinic
- Methods for censored data (3 weeks)
- Survival data models, Kaplan Meier, Cox proportional hazards
- Notes for survival Adapted from David Glidden's notes at
http://www.epibiostat.ucsf.edu/dave/biostat209/lect12.pdf
- Notes for Kaplan Meier, comparing survival curves (log-rank and Wilcoxon),
and also Notes for Cox proportional hazards were
handed out in class adapted from Dr. Will Thomas's 6451 notes.
- Using SAS and R for K-M and logrank
- Extra topics - last day Heteroskedasticity,
Equivalence testing
- REVIEW FOR FINAL EXAM!!!!!! Topics reviewed for final