"I used improper factor analysis, and now there isnt a g factor. Spearman was wrong!"
www.emilkirkegaard.com
"I should point out that a number of factor analytic studies of Piagetian tests along with other measures commit an egregious psychological error by orthogonally rotating the factors (or principal components) by some method such as varimax, which prohibits the emergence of the large general factor in all such tests. About the only wholly correct factor analysis of Piagetian tests I have found in the literature is the one by Philip Vernon (1965), a well-known expert in factor analysis and psychometrics. Many developmental psychologists, with no special training in factor analysis or psychometrics, simply select the most popular computer program, Kaiser’s varimax, for doing their factor analyses. As applied to factor extraction in the abilities domain, this is flatly wrong, not mathematically, but psychologically and scientifically. In the abilities domain, either oblique rotation should be done to permit the hierarchical extraction of g, or the g factor should be extracted (as the first principal factor) prior to rotation of the remaining factors. (In the latter procedure, one additional factor should be extracted prior to rotation.) It will be a great day for psychology when we no longer have to read studies in which the author automatically applies the varimax computer program (which is expressly intended to “ rotate away” a general factor) and then points out that “ factor analysis” fails to reveal a general factor in his test data!"
"I used improper factor analysis, and now there isnt a g factor. Spearman was wrong!"
"I used improper factor analysis, and now…
"I used improper factor analysis, and now there isnt a g factor. Spearman was wrong!"
"I should point out that a number of factor analytic studies of Piagetian tests along with other measures commit an egregious psychological error by orthogonally rotating the factors (or principal components) by some method such as varimax, which prohibits the emergence of the large general factor in all such tests. About the only wholly correct factor analysis of Piagetian tests I have found in the literature is the one by Philip Vernon (1965), a well-known expert in factor analysis and psychometrics. Many developmental psychologists, with no special training in factor analysis or psychometrics, simply select the most popular computer program, Kaiser’s varimax, for doing their factor analyses. As applied to factor extraction in the abilities domain, this is flatly wrong, not mathematically, but psychologically and scientifically. In the abilities domain, either oblique rotation should be done to permit the hierarchical extraction of g, or the g factor should be extracted (as the first principal factor) prior to rotation of the remaining factors. (In the latter procedure, one additional factor should be extracted prior to rotation.) It will be a great day for psychology when we no longer have to read studies in which the author automatically applies the varimax computer program (which is expressly intended to “ rotate away” a general factor) and then points out that “ factor analysis” fails to reveal a general factor in his test data!"