Applied biostatistics, Spring 2009
All material presented in the course is available to download on this page. No additional resources are necessary, unless you want to go your way beyond what is being covered during the classes.
All data sets have been compiled into an R package, biostats.tar.gz, with additional functions. They are also available as a single archive: biostats_src.tar.gz. To install the package in shell mode, use something like
$ R CMD install biostats.tar.gzWe will make use of the following packages: Hmisc, Design, ggplot, multtest, car, vcd. As some of them are not part of R base packages, please ensure that they are installed on your machine and that they are working properly.
Remark. As a general advice, always try to work with the latest R version (currently, 2.8.1, 2008-12-22), read the R news and save all your commands in dedicated text file(s); also don't forget to save your R workspace when you're done. If you're going to work on large projects, use a versioning system (e.g., RCS or SVN); otherwise, keep down all modifications made on your source files.
Lectures slides
Here are the slides used during the courses. "Printing" version is a printable handout (2x2 per page, BW).
- Exploratory data analysis: screen | printing
- Basic linear modeling: screen | printing
- Advanced linear modeling: screen | printing
- Generalized linear model: screen | printing
- Additional topics: screen | printing
Homeworks
Homeworks are practicals about theoretical and/or applied ideas tackled during the course. They are not mandatory but (I hope) should provide a deeper understanding into what you've learned so far.
- HW1: Multivariate descriptive statistics
- HW2: Linear regression
- HW3: Design of experiment
- HW4: Mixed-effects models
Tutorials
- The R programming language
- R extended Reference card, including plotting commands
- Emacs, ESS and Sweave
- Matrix algebra for LMs
- Statistical considerations for Mixed-effects models