Abstract
In this note, we outline a simple to use yet powerful bootstrap algorithm for handling correlated outcome variables in terms of either hypothesis testing or confidence intervals using only the marginal models. This new method can handle combinations of continuous and discrete data and can be used in conjunction with other covariates in a model. The procedure is based upon estimating the family-wise error (FWE) rate and then making a Bonferroni-type correction. A simulation study illustrates the accuracy of the algorithm over a variety of correlation structures.
| Original language | English |
|---|---|
| Pages (from-to) | 129-134 |
| Number of pages | 6 |
| Journal | Computer Methods and Programs in Biomedicine |
| Volume | 73 |
| Issue number | 2 |
| DOIs | |
| State | Published - Feb 2004 |
Keywords
- Bonferroni-type correction
- Bootstrap algorithm
- Marginal models
Fingerprint
Dive into the research topics of 'A semiparametric bootstrap approach to correlated data analysis problems'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver