DSpace Repository

Pairwise multiple comparisons: New yardstick, new results

Pairwise multiple comparisons: New yardstick, new results

Show full item record

Title: Pairwise multiple comparisons: New yardstick, new results
Author: Cribbie, Robert A.
Abstract: Behavioral science researchers often wish to compare the means of several treatment conditions on a specific dependent measure. The author used a Monte Carlo study to compare familywise error controlling multiple comparison procedures (MCPs; Tukey, Bonferroni) with MCPs that were not developed to control the familywise error rate on the probability of correctly identifying the true underlying population mean configuration (true model rate). Recently proposed MCPs that are not intended to control the familywise error rate had consistently larger true model rates than did familywise error controlling MCPs. Furthermore, of the familywise error controlling MCPs investigated, the popular Tukey and Bonferroni MCPs had consistently lower true model rates than did other familywise error controlling MCPs.
Sponsor: Social Sciences and Humanities Research Council
Subject: false discovery rate
model testing
multiple comparison procedures
true model rate
Type: Article
Rights: http://psycnet.apa.org/record/2003-05800-002
URI: http://dx.doi.org/10.1080/00220970309602065
http://hdl.handle.net/10315/34617
Published: American Psychological Association
Citation: Cribbie, R. A. (2003). Pairwise multiple comparisons: New yardstick, new results. Journal of Experimental Education, 71, 251-265. doi: 10.1080/00220970309602065
Date: 2003

Files in this item



Article Level Metrics




This item appears in the following Collection(s)