A new measure of nominal-ordinal association

被引:22
作者
Piccarreta, R [1 ]
机构
[1] Univ L Bocconi, Ist Metodi Quantitat, I-20136 Milan, Italy
关键词
D O I
10.1080/02664760120011635
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 [统计学]; 070103 [概率论与数理统计]; 0714 [统计学];
摘要
A new measure for evaluating the strength of the association between a nominal variable and an ordered categorical response variable is introduced. The introduction of a new measure is justified by analysing the characteristics of a measure of the nominal-ordinal association proposed by Agresti (1981), especially with respect to the problem of the 'choice' of a predictive variable. The sample-based version of the index is studied, and its asymptotic standard error and asymptotic distribution are derived. Simulations are considered to evaluate the adequacy of the asymptotic approximation determined, following Goodman & Kruskal (1963).
引用
收藏
页码:107 / 120
页数:14
相关论文
共 12 条
[2]
MEASURES OF NOMINAL-ORDINAL ASSOCIATION [J].
AGRESTI, A .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1981, 76 (375) :524-529
[3]
Agresti A., 1990, Analysis of categorical data
[4]
MEASURES OF ASSOCIATION FOR CROSS CLASSIFICATIONS .4. SIMPLIFICATION OF ASYMPTOTIC VARIANCES [J].
GOODMAN, LA ;
KRUSKAL, WH .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1972, 67 (33) :415-&
[5]
MEASURES OF ASSOCIATION FOR CROSS CLASSIFICATIONS .2. FURTHER DISCUSSION AND REFERENCES [J].
GOODMAN, LA ;
KRUSKAL, WH .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1959, 54 (285) :123-163
[6]
MEASURES OF ASSOCIATION FOR CROSS CLASSIFICATIONS .3. APPROXIMATE SAMPLING THEORY [J].
GOODMAN, LA ;
KRUSKAL, WH .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1963, 58 (302) :310-&
[7]
GOODMAN LA, 1954, J AM STAT ASSOC, V49, P723
[8]
HEDLUND RD, 1978, PUBLIC OPIN QUART, V41, P498
[9]
Lebart L., 1984, Multivariate Descriptive Statistical Analysis
[10]
A MEASURE OF ASSOCIATION BASED ON GINI MEAN DIFFERENCE [J].
SCHECHTMAN, E ;
YITZHAKI, S .
COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, 1987, 16 (01) :207-231