Power and Type I Error Rates for Rank-Score MANOVA Techniques

Robert Pavur, Ravinder Nath

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

A Monte Carlo simulation study was performed to compare the power and the Type I errors of the Wilks's lambda statistic and the Puri and Sen (1971) statistic on transformed data in a oneway MANOVA. The behavior of the Wilks's lambda statistic and the Puri and Sen statistic are studied using normal scores and ranks of the data. Both chi-square approximations and F approximations are used to estimate the distribution of these statistics. Various combinations of sample sizes and numbers of variables are used in the simulation study. For the situations considered in both normal and nonnormal homoscedastic conditions, the Wilks's lambda statistic calculated using normal scores and used with the F approximation appeared to be somewhat more robust with respect to maintaining its stated significance level than its competitors. Also, the F approximation for the Puri and Sen statistic is also shown to be more robust with respect to maintaining its stated significance level than the standard chi-square approximation.

Original languageEnglish
Pages (from-to)477-501
Number of pages25
JournalMultivariate Behavioral Research
Volume24
Issue number4
DOIs
StatePublished - 1989
Externally publishedYes

Fingerprint

Multivariate Analysis of Variance
Type I Error Rate
Statistic
Approximation
Significance level
Chi-square
Simulation Study
Type I error
Statistics
Sample Size
Monte Carlo Simulation
Estimate

All Science Journal Classification (ASJC) codes

  • Arts and Humanities (miscellaneous)
  • Statistics and Probability
  • Experimental and Cognitive Psychology

Cite this

Power and Type I Error Rates for Rank-Score MANOVA Techniques. / Pavur, Robert; Nath, Ravinder.

In: Multivariate Behavioral Research, Vol. 24, No. 4, 1989, p. 477-501.

Research output: Contribution to journalArticle

@article{78ff6db60941468cb00bfbf0cf9c9449,
title = "Power and Type I Error Rates for Rank-Score MANOVA Techniques",
abstract = "A Monte Carlo simulation study was performed to compare the power and the Type I errors of the Wilks's lambda statistic and the Puri and Sen (1971) statistic on transformed data in a oneway MANOVA. The behavior of the Wilks's lambda statistic and the Puri and Sen statistic are studied using normal scores and ranks of the data. Both chi-square approximations and F approximations are used to estimate the distribution of these statistics. Various combinations of sample sizes and numbers of variables are used in the simulation study. For the situations considered in both normal and nonnormal homoscedastic conditions, the Wilks's lambda statistic calculated using normal scores and used with the F approximation appeared to be somewhat more robust with respect to maintaining its stated significance level than its competitors. Also, the F approximation for the Puri and Sen statistic is also shown to be more robust with respect to maintaining its stated significance level than the standard chi-square approximation.",
author = "Robert Pavur and Ravinder Nath",
year = "1989",
doi = "10.1207/s15327906mbr2404_6",
language = "English",
volume = "24",
pages = "477--501",
journal = "Multivariate Behavioral Research",
issn = "0027-3171",
publisher = "Psychology Press Ltd",
number = "4",

}

TY - JOUR

T1 - Power and Type I Error Rates for Rank-Score MANOVA Techniques

AU - Pavur, Robert

AU - Nath, Ravinder

PY - 1989

Y1 - 1989

N2 - A Monte Carlo simulation study was performed to compare the power and the Type I errors of the Wilks's lambda statistic and the Puri and Sen (1971) statistic on transformed data in a oneway MANOVA. The behavior of the Wilks's lambda statistic and the Puri and Sen statistic are studied using normal scores and ranks of the data. Both chi-square approximations and F approximations are used to estimate the distribution of these statistics. Various combinations of sample sizes and numbers of variables are used in the simulation study. For the situations considered in both normal and nonnormal homoscedastic conditions, the Wilks's lambda statistic calculated using normal scores and used with the F approximation appeared to be somewhat more robust with respect to maintaining its stated significance level than its competitors. Also, the F approximation for the Puri and Sen statistic is also shown to be more robust with respect to maintaining its stated significance level than the standard chi-square approximation.

AB - A Monte Carlo simulation study was performed to compare the power and the Type I errors of the Wilks's lambda statistic and the Puri and Sen (1971) statistic on transformed data in a oneway MANOVA. The behavior of the Wilks's lambda statistic and the Puri and Sen statistic are studied using normal scores and ranks of the data. Both chi-square approximations and F approximations are used to estimate the distribution of these statistics. Various combinations of sample sizes and numbers of variables are used in the simulation study. For the situations considered in both normal and nonnormal homoscedastic conditions, the Wilks's lambda statistic calculated using normal scores and used with the F approximation appeared to be somewhat more robust with respect to maintaining its stated significance level than its competitors. Also, the F approximation for the Puri and Sen statistic is also shown to be more robust with respect to maintaining its stated significance level than the standard chi-square approximation.

UR - http://www.scopus.com/inward/record.url?scp=24944440426&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=24944440426&partnerID=8YFLogxK

U2 - 10.1207/s15327906mbr2404_6

DO - 10.1207/s15327906mbr2404_6

M3 - Article

VL - 24

SP - 477

EP - 501

JO - Multivariate Behavioral Research

JF - Multivariate Behavioral Research

SN - 0027-3171

IS - 4

ER -