The main purpose of this article is: given a score matrix called S, find out the joint proportional contribution of factors due to persons (conditions, situations, and so forth) and factors due to variables, for any sij observed score, where i identifies persons, and j, variables. This approach makes it possible a) to show that the same score in a given variable may have a different quantitative interpretation in terms of persons or conditions, and b) to find out how subjects differ in the way in which they relate variables.

We present the results of a principal-components analysis of the Personality Adjective Check List (PACL) at the item level, rather than at the scale level. The PACL was constructed by Strack (1987) to evaluate the personality scales proposed by Millon (1981). Data were obtained from a sample of normal adolescents. A scree test was applied to determine the number of factors to be extracted, and the retained factors were then rotated using a varimax procedure. Using the data from this second analysis, congruence coefficients were calculated by correlating the factor weights from the two analyses of men and women. Results reveal the existence in the whole sample of five unipolar personality factors, which we have called: Aggressive, Conscientious, Inhibited, Impulsive and Gregarious. These are not similar to those obtained from a sample of normal adults.