Q analysis, the factor analysis of person-by-person correlation matrices, becomes unmanageable with large samples. There is, however, a one-step approach for computing principal components of the matrix of correlations among an unlimited number of individuals, so long as the number of response items is limited.
Get full access to this article
View all access options for this article.
References
1.
EckartC. and YoungG., “The Approximation of One Matrix by Another of Lower Rank,” Psychometrika, 1 (September 1936), 211–8.
2.
HsuP. L., “On a Factorization of Pseudo-orthogonal Matrices,” Quarterly Journal of Mathematics, 17 (1946), 162–5.
3.
JohnsonR. M., “On a Theorem Stated by Eckart and Young,” Psychometrika, 28 (September 1963), 259–63.
4.
KellerJ. B., “Factorization of Matrices by Least Squares,” Biometrika, 49 (1962), 239–42.
