Log-Linear Analysis

Graham Tall   G.E.Tall@bham.ac.uk

Log-Linear Analysis is a statistical process equivalent to MANOVA, but designed to analyse  categorical/ordinal rather than interval data.  In research terms, the analysis is important because it can assess the affect of categorical variables on answers to attitude and other multiple choice questions.  Unfortunately the analysis is much more complex than those required for for one and two-way Chi-square and the expected frequencies for Log-Linear Analysis need to be obtained through a process of estimation (Gilbert (1981)).

Gilbert (1981) describes the statistical logic using an example reasonably clearly, deliberately avoiding details of the mathematical process.  If the process is to be used, it is essential that the researcher reads the literature so that they feel able to justify the statistical process  involving estimation and interpret the results of the analysis.

SPSS Help File Description:

The General Loglinear Analysis procedure analyzes the frequency counts of observations falling into each cross-classification category in a crosstabulation or a contingency table. Each cross-classification in the table constitutes a cell, and each categorical variable is called a factor. The dependent variable is the number of cases (frequency) in a cell of the crosstabulation, and the explanatory variables are factors and covariates.....

You can select up to 10 factors to define the cells of a table. A cell structure variable allows you to define structural zeros for incomplete tables, include an offset term in the model, fit a log-rate model, or implement the method of adjustment of marginal tables. Contrast variables allow computation of generalized log-odds ratios (GLOR).

SPSS automatically displays model information and goodness-of-fit statistics. You can also display a variety of statistics and plots or save residuals and predicted values in the working data file.

Gilbert, G.N. (1981)  Modelling Society:  An Introduction to Loglinear Analysis for Social Researchers.  London:Allen&Unwin

Home Page   Research Introduction   Quantitative Advice    Index   Statistical Tests     Research and Statistics Courses