Principal Components and Classification Analysis Startup Panel - Advanced Tab

Principal Components and Factor Analysis

Select the Advanced tab of the Principal Components and Classification Analysis Startup Panel to access options to opt for not only the computations of the principal components and the related statistics through the correlation or the covariance matrix, but also for sample- or the population-based computations.

Variables

Click the Variables button to display the standard variable selection dialog box in which you can select Variables for analysis (active variables), Supplementary variables (optional), a Variable with active cases (optional), and a Grouping variable (optional). If you select a Variable with active cases, you must specify a Code for active cases (see below). Note that overlapping of the selected variables in four lists is not permitted. Also, note that the variable with active cases as well as the grouping variable need to be qualitative or grouping variables. See the Principal Components and Classification Analysis Startup Panel - Quick tab for additional information about the types of variables and their role in the computation of principal components.

Code for active cases

If you select a Variable with active cases, you must choose the code for active cases. To do this, double-click the Code for active cases edit box (or press the F2 key on your keyboard) to display the Variable code dialog box, or type in the respective code value for the active cases.

Analysis based on

Use the Analysis based on group box to choose to compute principal components from the correlation matrix or the covariance matrix of the active variables.

Correlations

Select the Correlations option button if you need to standardize data (center the data about the means and scale it by the standard deviations). To standardize data, you need to compute principal components through the correlation matrix.

Covariances

Select the Covariances option button to center the data about the means, and for being able to do so, you need to compute principal components through the covariance matrix.

Compute variances

Use the Compute variances group box to choose among the two ways in which you can compute the variances: as the sums of squares of deviations divided by N (the valid number of cases) or N -1. If divided by N -1, the resulting variances are estimates of the population (Sigma); if divided by N, the resulting variances are descriptions of the sample only. The same argument holds for the computation of covariances. In this group box, choose to compute variances and covariances for the PCCA from the sample (select the SS/(N-1) option button) or population (select the SS/N option button).