| Abstract: | Considers that J. P. Guilford's (see PA, Vol 53:Issue 1) reply to the authors' criticisms of his factor analysis of his structure-of-intellect model has failed to come to grips with evidence indicating that if the number of targeted variables for each factor is 3 or fewer, Procrustes factoring provides no better evidence for structure-of-intellect theory than for any of an infinity of other arbitrarily determined theories. Under these conditions factors can be rotated in such a way as to suggest support for almost any desired solution, and one such solution can be that labeled a replication. Hence the factorial invariance claimed under such conditions can indicate only consistency in the researcher and his methods, not in the phenomena studied. (16 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
