Details
| Original language | English |
|---|---|
| Pages (from-to) | 461-473 |
| Number of pages | 13 |
| Journal | Communications in Statistics Part B: Simulation and Computation |
| Volume | 32 |
| Issue number | 2 |
| Publication status | E-pub ahead of print - 9 Dec 2011 |
Abstract
For a long time the exact evaluation of parametric multiple comparison procedures was computationally almost infeasible. Resampling based techniques have been proposed instead, aiming (Hochberg, Y., Tamhane, A. C. (1987). Multiple Comparison Procedures. New York: Wiley) to approximate the true underlying distribution function, where standard integration methods failed so far and (Hsu, J. C. (1996). Multiple Comparisons. London: Chapman and Hall) to robustify the parametric test statistics against certain violations of the assumptions. This article compares several resampling based techniques with new and efficient integration methods for multiple comparisons. The goal of the numerical study is to assess, how the procedures compare to each other under a variety of normal and non-normal conditions.
Keywords
- Bootstrap procedures, Multiple comparisons, Multivariate t-distribution, Non-normality, Unequal variances
ASJC Scopus subject areas
- Mathematics(all)
- Statistics and Probability
- Mathematics(all)
- Modelling and Simulation
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Communications in Statistics Part B: Simulation and Computation, Vol. 32, No. 2, 09.12.2011, p. 461-473.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Comparison of exact and resampling based multiple testing procedures
AU - Bretz, Frank
AU - Hothorn, Ludwig A.
PY - 2011/12/9
Y1 - 2011/12/9
N2 - For a long time the exact evaluation of parametric multiple comparison procedures was computationally almost infeasible. Resampling based techniques have been proposed instead, aiming (Hochberg, Y., Tamhane, A. C. (1987). Multiple Comparison Procedures. New York: Wiley) to approximate the true underlying distribution function, where standard integration methods failed so far and (Hsu, J. C. (1996). Multiple Comparisons. London: Chapman and Hall) to robustify the parametric test statistics against certain violations of the assumptions. This article compares several resampling based techniques with new and efficient integration methods for multiple comparisons. The goal of the numerical study is to assess, how the procedures compare to each other under a variety of normal and non-normal conditions.
AB - For a long time the exact evaluation of parametric multiple comparison procedures was computationally almost infeasible. Resampling based techniques have been proposed instead, aiming (Hochberg, Y., Tamhane, A. C. (1987). Multiple Comparison Procedures. New York: Wiley) to approximate the true underlying distribution function, where standard integration methods failed so far and (Hsu, J. C. (1996). Multiple Comparisons. London: Chapman and Hall) to robustify the parametric test statistics against certain violations of the assumptions. This article compares several resampling based techniques with new and efficient integration methods for multiple comparisons. The goal of the numerical study is to assess, how the procedures compare to each other under a variety of normal and non-normal conditions.
KW - Bootstrap procedures
KW - Multiple comparisons
KW - Multivariate t-distribution
KW - Non-normality
KW - Unequal variances
UR - http://www.scopus.com/inward/record.url?scp=0038271900&partnerID=8YFLogxK
U2 - 10.1081/SAC-120017501
DO - 10.1081/SAC-120017501
M3 - Article
AN - SCOPUS:0038271900
VL - 32
SP - 461
EP - 473
JO - Communications in Statistics Part B: Simulation and Computation
JF - Communications in Statistics Part B: Simulation and Computation
SN - 0361-0918
IS - 2
ER -