This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: BST 225, STA 225: Final Project Notes, Spring 2006. Projects were graded on the appropriateness and accuracy of the statistical work and on the presentation of the statistical methods and the results. Here are the components I looked for, and how people generally did on each. Study design and planning: Accurate sample size calculation based on information given. • Depending on how much you round off, the sample size for a two-sided test comparing survival curves, assuming log-rank test as primary analysis and exponential survival, gives a sample size around 490. • Most people elected to go with all 500 even if they came up with slightly smaller number. • Two people chose to do one interim analysis, about half through the deaths. Neither interim analysis gave significant findings, so both elected to continue the study. Write-up of statistical methods should look like the examples in clinical trials literature. I expected a separate section that described your methods before seeing your quantitative results. Added features that mark a really good write-up would include a brief description of the power for your design, references, and citation for statistical software used....
View Full Document
- Winter '06
- Regression Analysis, Survival analysis, KIT expressers