Log Linear Model Applications.
Abstract
Methods for multivariate data analysis have been developed and received wide usage in the past 30 years. The advent of the computer in the early 1950's made possible an acceleration in the number of multivariate studies. The first to profit was factor analysis which became a commonplace in exploratory studies. Subsequently, linear discriminant analysis, classification methods and then cluster analyses became part of the emerging excitement in attempting and doing studies either not considered possible before or not thought about at all before because of the lack of computer resources. Another multivariate method has also profited from the computer era. It is multidimensional contingency table analysis through use of the log-linear model. This document looks at some applications of the log-linear model, or as it sometimes is called, logistic response analysis. It may be instructive to view one situation explored about 25 years ago through classification techniques before computer capability to handle the log-linear model and software was available, and then analyze the same data base by contingency table analysis. After this comparison, other and more recent direct applications of the log-linear model are discussed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 30, 1984
- Accession Number
- ADA146620
Entities
People
- H. Solomon
Organizations
- Stanford University