Observed and Predicted Frequencies



The goodness-of-fit statistics are based on the cells of the observed and predicted frequencies table, the top half of which is shown here. Cells in the table represent the "covariate patterns" defined in the Statistics dialog box. In this case, they are the cross-classifications of the factors and dependent variable specified on the main dialog box. Thus, the first row of the table pertains to inactive males under the age of 31 who preferred the breakfast bar.
- The observed column reports the number of cases observed in the data file that are in the covariate pattern.
- The predicted column reports the number of cases you would expect to see in the covariate pattern if the model is correct.
- The Pearson residual is a measure of the difference between the observed and predicted values. Large Pearson residuals can indicate covariate patterns that are not well fit by the model.
When subpopulations are based on covariates, goodness-of-fit tests generally are not particularly useful because there will be many cells with 0 frequencies.