Generalized Cluster Analysis Results - Advanced tab

Select the Advanced tab of the Generalized Cluster Analysis results dialog box to access options to review details of the results for the continuous and categorical variables selected for the analyses.

Categorical Variables

These options pertain to results for the categorical variables included in the cluster analyses. See also the Introductory Overview for details on how categorical variables are handled in the different clustering methods.

All Variables

You can specify to compute results for all variables, or just choose the variable that you want to see results for from this drop-down list.

Frequency table

Click the Frequency table button to show for the categorical variable(s) the pattern of frequencies of the observations across the clusters to which they were assigned, and the classes specified in the respective categorical variables.

Graph of frequencies

Click the Graph of frequencies button to display a multiple line graph showing the patterns of frequencies of the observations in each cluster across the classes specified in the respective categorical variables.

Chi-square test

Click the Chi-square test button to display Chi-square and G-square (maximum likelihood Chi-square) tests for the two-way crosstabulation tables of frequencies for each categorical (class-) variable by cluster. These tests reflect on the magnitudes of the differences in the distributions of observations over the respective groups (classes) across the clusters (i.e., how different the clusters are with respect to the class variables).

Continuous variables

These options pertain to results for the continuous variables included in the cluster analyses. See also the Introductory Overview for details on how continuous variables are handled in the different clustering methods.

All Variables

You can specify to compute results for all variables, or just choose the variable that you want to see results for from this drop-down list.

Descriptive statistics

Click the Descriptive statistics button to display a spreadsheet containing the cluster means, standard deviations, and minima and maxima for each continuous variable, and the number of cases in each cluster.

Graph of distributions

Click the Graph of distributions button to display a summary line graph for each continuous variable in the analysis, showing the expected distributions for the respective variable in the different clusters. This graph is particularly informative in the context of EM clustering, which will estimate mixtures of distributions to assign classification probabilities (cluster assignment probabilities) for each case, and for each cluster. Use this graph to review the fitted distributions for each variable and for each cluster. See also the Introductory Overview for details.

Analysis of variance

Click the Analysis of variance button to display ANOVA results pertaining to the differences in the means for the continuous variables across the clusters. Note that the significance (p-) values reported in this table cannot be interpreted in the usual manner, i.e., as the statistical significance of a priori hypotheses. Instead, the clusters were specifically constructed to yield maximum differentiation between clusters, so highly "significant" effects should be expected in this table.