|Friday, February 22|
|CS08 Theme 2: Data Modeling and Analysis #3||
Fri, Feb 22, 1:30 PM - 3:00 PM
Resampling: No Assumptions Needed!View Presentation *Dennis L Eggett, Brigham Young University - Department of Statistics
Keywords: resampling, regression, ANOVA, mixed models
When you are not sure of the distribution of your data, when you have outliers, when you can’t rely on the usual statistical methods, what can you do? Resampling in the modern age of computers is not only possible, but it is also practical and relatively fast to accomplish. The key is to make sure you understand the underlying structure of your data and to do the resampling in a logical way that reflects that structure. You cannot just push a button and get results. With resampling you can better approximate the permutation distribution that common statistical methods are trying to approximate. A two sample t-test is easy. But, you can go way beyond that. How about regression or even mixed models. If you want to, you can use resampling for any situation, as long as you do the resampling to replicate your data structure.