Factor Analysis - Options
main topic     example     see also 

Stat > Multivariate > Factor Analysis > Options

Allows you to specify the matrix type and source, and the loadings to use for the initial extraction.

Dialog box items

Matrix to Factor

Correlation: Choose to calculate the factors using the correlation matrix. Use the correlation matrix if it makes sense to standardize variables (the usual choice when variables are measured by different scales).

Covariance: Choose to calculate the factors using the covariance matrix. Use the covariance matrix if you do not wish to standardize variables. The covariance matrix cannot be used with a maximum likelihood estimation.

Source of Matrix:

Compute from variables: Choose to use the correlation or covariance matrix of the measurement data.

Use matrix: Choose to use a stored matrix for calculating the loadings and coefficients. (Note: Scores can not be calculated if this option is chosen.) See To perform factor analysis with a correlation or covariance matrix.

Loadings for Initial Solution

Compute from variables: Choose to compute loadings from the raw data.

Use loadings: Choose to use loadings which were previously calculated, then specify the columns containing the loadings. You must specify one column for each factor calculated. See To perform factor analysis with stored loadings.

Maximum Likelihood Extraction

Use initial communality estimates in: Choose the column containing data to be used as the initial values for the communalities. The column should contain one value for each variable.

Max iterations: Enter the maximum number of iterations allowed for a solution (default is 25).

Convergence: Enter the criterion for convergence (occurs when the uniqueness values do not change very much). This number is the size of the smallest change (default is 0.005).