We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
Stepdown analysis of variance: A refinement.
- Authors
Koslowsky, Meni; Caspy, Tamir
- Abstract
This article presents a refined procedure of the multivariate analysis of variance (MANOVA). Multivariate techniques for analyzing data that contain more than one dependent variable are numerous and well-documented. The most commonly used procedures include discriminant analysis, canonical analysis, and the MANOVA. The last procedure is an extension of univariate analysis of variance to the case of several dependent measures. As with all multivariate techniques, the advantages of MANOVA include the ability to simultaneously analyze a series of dependent measures while keeping the experiment-wise error rate constant. In addition, multiple significant findings may actually represent one common under-lying factor. By identifying a small set of explanatory factors to replace the original variable list, the researcher obtains a clearer picture of the overall results. However, the advantage of MANOVA as an overall indicator of significance is somewhat limited. The researcher who has obtained a significant MANOVA is usually confronted with the rather difficult task of identifying the precise location of the significant finding.
- Subjects
MULTIVARIATE analysis; ANALYSIS of variance; STATISTICAL hypothesis testing; EXPERIMENTAL design; RANDOM variables; STATISTICAL correlation; ERROR analysis in mathematics
- Publication
Journal of Organizational Behavior, 1991, Vol 12, Issue 6, p555
- ISSN
0894-3796
- Publication type
Article
- DOI
10.1002/job.4030120608