Evaluation of Bootstrap and Parametric Percentile Contrasts. Volume 1. Splits Analysis: A Method for Noncentral Tendency Comparisons
Abstract
In some instances group comparisons in terms of upper or lower portions of the score distributions are more informative than comparisons of central tendency. These comparisons can be done by carrying out a split on the data prior to an analysis of variance (ANOVA). The resulting test statistic from ANOVA is not distributed as an F ration however, and requires evaluation for significance relative to an empirical monte-carlo distribution. An example and computer program are presented.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 25, 1990
- Accession Number
- ADA230506
Entities
People
- Jeffrey L. Rasmussen