Statistical computing underlies much of the work we do as researchers. However, students tend to be left to their own devices to develop their programming abilities.
While teaching oneself is an indispensable part of the process, new researchers may be better equipped to do so provided a base from which to develop their computing skills. This website is intended to serve as such an introductory resource, targeted at incoming graduate students at the Brown University School of Public Health.
While the content in some sections focuses on R so as not to repeat material covered in detail by the SAS modules you’re required to complete upon entry into your program, we provide SAS code for statistical analyses, programming techniques such as loops or functions, and data visualization.
Main objectives
Install R (and RStudio)
- Include R packages that will be helpful to download as the tutorials progress
Understand some basic data types in R
Understand how each language handles:
- Variable names
- Value formats
- Missing data
- Objects
- Environments (as applicable)
Be able to conduct the following tasks in R and SAS:
- Import/export data to and from common filetypes
- Use common machine-readable filetypes
- Manipulate variables
- Assign variable names and value formats
- Change variable names and value formats
- Create new or derived variables
- Statistical analysis
- Generate contingency tables and conduct appropriate statistical tests on them
- Implement basic regression analyses
- Linear
- Logistic
- Relative risk (log-binomial, Poisson)
- Simulate data
- Write simple loops
- Write simple functions/macros
Produce publication-quality plots
- Review image formats (JPG, PNG, EPS, WMF, EMF, PDF)
- Know where to find information on principles of data visualization
Consider ways to make your analyses reproducible
- RMarkdown
- Pseudorandom number seeds
Resource repository
- Excellent, mostly free resources on programming and data visualization
Acknowledgements
I’d like to thank Prof. Alison Field for the opportunity to develop this resource and Brown University School of Public Health for providing the resources with which to do so. In addition, I’d like to thank the following individuals, who provided valuable feedback during development: Adam Sullivan, Rachel Yorlets, Christopher Barry, and Hongseok Kim.
