11/11/2020

Return to "The use of R as a Toolbox for Mathematical Statistics Explortation" by Nicholas Horton, Elizabeth R. Brown, and Linjuan Qian, The American Statistician, 2004

Small group project #3
- Option 1 - Monte Carlo Power Comparison of at least two competing test procedures (see samples in !Monte Carlo Power Study Samples)
- Option 2 - Application of Markov chains (see articles and samples)
- Option 3 - Introduce new material from Modern Data Science with R
Complete disussion of power (see horton-tas.R)
- Power and sample size calculations
- Empirical power calculations (useful for comparing competing test procedures)

There is a major error in the code for comparing brain volume for schizophrenics and their relatives.
Work with your partner(s) to identify and rectify the error.
How close is your estimated power to the value of 0.89 provided in the article?
If you were to expand this power study (or a another power study), what factors would be important for you to consider. That is, what would you vary as you come up with ways to compare competing tests? The samples in !Monte Carlo Power Study Samples may be very helpful.