Using Information on Ordering for Loglinear Model Analysis of Multidimensional Contingency Tables.

Abstract

Many different authors have claimed that the loglinear model approach to the analysis of contingency table data is appropriate only for nominal variables and does not make use of information on the ordinal nature of some categorical variables (i.e. the ordering of the categories). This paper, reviews a variety of loglinear model methods which do take into account, either explicitly or implicitly, such information on ordering. The focus is on methods involving maximum likelihood estimation, but other methods of estimation can be used with these models. Some additional models for ordered categorical data are considered.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 1982
Accession Number
ADA121923

Entities

People

  • Stephen E. Fienberg

Organizations

  • Carnegie Mellon University

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Biometrics
  • Classification
  • Computations
  • Contracts
  • Data Science
  • Data Sets
  • Information Science
  • Maximum Likelihood Estimation
  • Military Research
  • Multivariate Analysis
  • New York
  • Statistical Inference
  • Statistics
  • Surveys
  • Two Dimensional
  • Universities

Fields of Study

  • Mathematics

Readers

  • Regression Analysis.