Types of Matrices

Correlation (value of the last case [Matrix] is 1). Square correlation matrices are symmetrical and contain the correlation coefficients for all pairs of specified variables.  You can create and then save correlation matrices in the appropriate dialogs (for example, see the option Matrix on the Product-Moment and Partial Correlations dialog box - Advanced/Plot tab). You can also manually create a correlation matrix by entering the correlations into a regular spreadsheet (see Creating a New Spreadsheet) and including in the file the last four cases which describe the matrix (see Matrix File Format).

Similarities (value of the last case [Matrix] is 2). The similarities between objects (e.g., variables) are expressed in this matrix. You can import or manually create this type of matrix file by entering the correlations into a regular spreadsheet (see Creating a New Spreadsheet) and including in the file the last four cases which describe the matrix (see Matrix File Format). Similarities matrices can be used in Multidimensional Scaling analyses.

Dissimilarities (value of the last case [Matrix] is 3). The dissimilarities (distances) between objects (e.g., variables) are expressed in this matrix. You can create this matrix manually or it can be created for you by using the Matrix option on the Cluster Analysis - Joining Results - Advanced tab. Dissimilarities matrices can be used in Multidimensional Scaling analyses.

Covariance (value of the last case [Matrix] is 4). Square covariance matrices contain the covariances for all pairs of specified variables on the off-diagonal and the variances for each variable on the diagonal of the matrix. Covariance matrices can be saved in the Structural Equation Modeling module, or you can manually create a covariance matrix by entering the covariances into a regular spreadsheet (see Creating a New Spreadsheet) and including in the file the last four cases which describe the matrix (see Matrix File Format).