Key points are not available for this paper at this time.
A method is proposed for the evaluation of the degree to which trait measures show stability across diverse methods of measurement. Each mono method block is separately factored and orthogonally rotated so that only one trait measure is highly loaded on a given dimension. The component score intercorrelation matrix is computed implicitly, and this matrix is re-factored and rotated by varimax. The technique is illustrated using multitrait-multimethod matrices from personality assessment, which yield trait-specific factors. By reversing the procedure and separately factoring monotrait-heteromethod blocks followed by the re-factoring of implicit component scores, it is demonstrated that method factors may be isolated.
Douglas N. Jackson (Tue,) studied this question.