ABSTRACT
THE publication of the M.P.I. and various factorial studies connected with it aroused a certain amount of interest, and a number of critical discussions appeared, which raised fundamental problems obviously requiring an answer. Two of the main problems to appear were raised expressly by Carrigan (1960) in her excellent review of ‘Extraversion/ Introversion as a Dimension of Personality’. The first problem to be raised by Carrigan was the relation between extraversion and adjustment; her final comment was that ‘A clear-cut answer cannot be given’ (p. 351). It will be remembered that studies with the M.P.I, have usually shown a correlation of between 0·15 and 0·20 between neuroticism and introversion, with rather higher values for maladjusted samples; this might be interpreted as showing a lack of orthogonality between the two factors. An alternative explanation of course would be that this correlation is an artefact produced by a faulty selection of items. It is clearly impossible to find items having loadings only on one or the other of the two type factors with which we are concerned; in the ordinary way each item will have at least some slight loading on the factor which it is not supposed to be measuring. If now in the selection of items there is a lack of balance in the positive and negative loadings on the factor which is not supposed to be measured by a given item, then an artificial positive or negative correlation between extraversion and neuroticism can easily appear. The possibility exists that in the M.P.I. there are too many items from the dysthymic quadrant, i.e. having loadings both on neuroticism and introversion, and that it is due to the presence of these items that the two factors appear as correlated. This possibility, and the questions raised by the existence of the positive correlation between introversion and neuroticism in the M.P.I., were one of the reasons for carrying out the research underlying the production of the Eysenck Personality Inventory (E.P.I.) which is essentially an improved version of the M.P.L (Eysenck and Eysenck, 1964).
