Matrix Spreadsheet

Matrix spreadsheets are simply spreadsheets with special contents. Therefore, they do everything that spreadsheets do. The only time their "special contents" make a difference are as follows: When you save a matrix spreadsheet, the default extension is changed to .smx instead of .sta, so that you can distinguish between them when opening them. If they are in a workbook, you cannot tell the difference. When you are in the Multidimensional Scaling module, only matrix spreadsheets can be used.

The "special contents" of matrix spreadsheets are defined as follows: If you have a correlation matrix of size N, the file must have N variables and N+4 cases. The matrix itself is in the first N rows/N variables, and then the final four cases contain the following: Means (one for each variable), Std.Dev. (one for each variable), No.Cases (value stored in first variable), and Matrix (defines matrix type). See Matrix file format for further details.

While a matrix spreadsheet looks identical to a regular spreadsheet, analyses that require matrix input (such as multidimensional scaling) will only consider input spreadsheets that are also matrix spreadsheets. You can create a matrix spreadsheet from any analysis that allows you to create a matrix of results [e.g., Basic Statistics - Correlations, Canonical Analysis, Cluster Analysis, Distributions & Simulation, Factor Analysis, Fixed Nonlinear Regression, Multidimensional Scaling, Multiple Regression, Reliability Analysis, and Structural Equation Modeling (SEPATH)].

Note that when you create a matrix file from one of the above analyses, the number of cases will be set to the minimum number of cases across all selected variables. This only becomes an issue when pairwise deletion of missing data is used.

When you save a matrix spreadsheet (by selecting Save or Save As from the File menu), STATISTICA will detect that the spreadsheet is in this special matrix form, and set the default file extension to .smx.