USE OF NESTED ORTHOGONAL CONTRASTS IN ANALYZING RANK DATA

• 成果类型: Article
• 署名作者: MARDEN, JI
• 刊物名称: JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
• ISSN/ISSBN: 0162-1459
• DOI: 10.2307/2290261
• 发表日期: 1992
• 页码: 307-318
• 关键词: models

摘要

A data set consisting of 143 rankings of 10 occupations from a survey of Goldberg has been analyzed in a number of recent papers. The purpose of this paper is to use so-called nested orthogonal contrasts of the occupations to gain further insight into the data. A contrast is a comparison of subsets of the occupations based on their relative ranks; contrasts are orthogonal if the comparisons they represent are not confounded. Various models based on sets of orthogonal contrasts-including contingency table models, models analogous to those of Fligner and Verducci, and latent class models-are applied to the data. It is found that there are three main groups of occupations based on overall prestige, and within each group are distinctions between managerial and technical occupations.