ABSTRACT
THE data now to be examined were collected by H. P. Hildebrand during a course of studies for a Ph.D. at this Institute. His findings, based mainly on a centroid factor analysis, were published first in his thesis (1953) and later in an article in the British Journal of Psychology (1958). Eysenck has quoted them extensively (e.g. 1957) and asked me to make some further analyses of parts of the data, using the methods described in Part II. Most of this was done during 1956–7; the electronic computation by the National Physical Laboratory under the supervision of Mr. J. G. Hayes. L. H. Storms, who was pursuing a course in statistics here at that time, used other parts of the data not too bulky to be treated by electric calculator for a parallel analysis, and published his findings in the J. Ment. Sci. in 1958. Anyone who goes to the trouble of reading all three texts is likely to find differences in the authors’ interpretations of the results and may even come to the opinion that the different methods of analysis conflict. In this account I shall attempt to show where the differences originate and suggest a synthesis.
