On the Relationship Between Dummy Variable Regression and Multiple Classification Analysis

The functional relationship between the methods of Dummy Variable Regression (DVR) & Multiple Classification Analysis (MCA) is investigated. Both methods assume a single metric predicted (criterion) variable, & both estimate additive, recursive models. MCA focuses upon testing the overall st...

Full description

Saved in:
Bibliographic Details
Published in:Mid-American Review of Sociology 1978-04, Vol.3 (1), p.83-93
Main Author: Jacobsen, Christian Wells
Format: Article
Language:English
Subjects:
Analysis/Analyses/Analyzing
Analytical estimating
Classification/Classified/ Classificatory
Coefficients
Degrees of freedom
Estimation methods
Modeling
Multi-/Multiple
Multiple regression
NOTES AND COMMENTS
Regression
Regression coefficients
Statistical variance
Terminology
Variable coefficients
Variable/Variables/ Variableness
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The functional relationship between the methods of Dummy Variable Regression (DVR) & Multiple Classification Analysis (MCA) is investigated. Both methods assume a single metric predicted (criterion) variable, & both estimate additive, recursive models. MCA focuses upon testing the overall strength of a factor in predicting the criterion variable. DVR tests the pattern of factor category effects. An example is given by application of both methods to the same data. The data come from the National Opinion Reseach Center's 1976 General Social Survey. Tested are the relations between "tolerance" (criterion variable) & the factors, region & type of place, with the covariate, years of education. MCA addresses the effects of region & type of place on tolerance; DVR tests the significance of difference in predicted value between categories of factors. By conversion of MCA into DVR coefficients, the relationship between the two methods is clarified. 3 Tables. W. Zimmerman.
ISSN:0732-913X
1094-5830
2469-8466