Abstract
ABSTRACT some variables may have nominal or ordinal scale, this causes a failure of analyzing the Pearson's correlation matrix by factor analysis. However, it was suggested to replace the elements of the correlation matrix by correlation coefficients which should be calculated according to the scales of the two correlated variables. The suggested correlation matrix might become an indefinite matrix which can not be analyzed by factor analysis, so the suggested methods were established to eduction a positive definite matrix or positive semi-definite matrix.
