LOG-LINEAR MODELS

by David Knoke

• ISBN13: 9780803914926
• ISBN10: 080391492X
• Format: Paperback
• Copyright: 1980-08-01
• Publisher: Sage Publications, Inc

Summary

Introduces methods for quantitative assessment of relationships among categoric variables in multivariable crosstabulations. Procedures to estimate and interpret effect parameters for hierarchical models are described for both the general loglinear model and its logit version.

Table of Contents

Editor's Introduction	p. 5
Relationships in Crosstabulations	p. 8
The Log-Linear Model	p. 11
Specifying Models	p. 11
Saturated Models	p. 12
Nonsaturated Models	p. 17
Fitting Marginals	p. 19
Generating Expected Frequencies	p. 22
Analyzing Odds	p. 24
Testing for Fit	p. 30
How To Evaluate Models Fitted to Data	p. 30
Comparisons of Different Models of the Same Data	p. 31
Independence Hypothesis	p. 32
Equal Marginal Distributions Hypothesis	p. 33
More Complex Models: Polytomous Variables	p. 33
More Complex Hypotheses	p. 37
An Analog to Multiple R[superscript 2] for Large Samples	p. 40
Applications to Substantive Problems	p. 42
Causal Models for Log-Linear Models	p. 42
Analyzing Change Over Time	p. 47
Comparative Cross-Sections	p. 47
Two-Wave Panels	p. 48
Markov Chain Models	p. 54
Age, Period, and Cohort Models	p. 57
Special Techniques with Log-Linear Models	p. 63
What To Do About Zero Cells	p. 63
Fixing Start Values	p. 65
Analyzing Ordered Data	p. 67
Collapsing Polytomous Variables	p. 70
Nonhierarchical Models	p. 72
Conclusions	p. 76
References	p. 77
Table of Contents provided by Syndetics. All Rights Reserved.