Clarification of Cliff and Caruso (1998).

In response to N. Cliff and J. C. Caruso (1998), the author clarifies that it is the sum of the reliabilities of the components that remains invariant under rotation in reliable component analysis.

A Legitimate Case of Component Analysis of Ipsative Measures, and Partialling the Mean as an Alternative to Ipsatization.

Ipsatizing variable prior to component analysis, by subtracting the man score of each individual from all scores of that individual, has met with serious criticisms, both strategic and technical. On the other hand, the procedure is still popular in the study of personality, as a means of removing acquiescence variance. An attempt is made at reconciling these facts. Technical objections to component analysis of ipsatized variables, recently restated by Dunlap and Cornwall, are shown to be based on erroneous premises. A strategic objection leveled by Clemans does apply in a majority of cases, but has no bearing on the specific case of ipsatizing responses to personality questionnaires made up of opposite item pears. As an alternative to ipsatizing by subtraction of the mean, partialling the mean component is suggested as a more elegant procedure, allowing a breakdown of explained variance into variance due to the mean and to subequent components uncorrelated with the mean.

Some Relationships Between Descriptive Comparisons of Components from Different Studies.

This paper considers descriptive methods of comparing components from different studies, based on correlations between columns of component scores matrices, on congruence coefficients between columns of pattern, structure or component scores coefficient matrices, or on coefficients of invariance. Contrary to common belief, it is shown that coefficients of invariance are unrelated to correlations between component scores. On the other hand, having a perfect coefficient of invariance is shown to be equivalent to having a perfect congruence between corresponding columns of the component scores coefficient matrices. A similar but weaker relationship between the latter congruence and congruence between columns of pattern matrices is demonstrated.

Rotation to Perfect Congruence and the Cross Validation of Component Weights Across Populations.

This paper deals with strategies of congruence studies, aimed at evaluating recoverableness of a given set of components from a first population in a second population where the same variables have been used. Five decisions inherent to congruence studies are analysed in detail. Confirmatory evidence with respect to recoverableness can be obtained from an independent component analysis for the second population, parallel to that of the first population. Disconfirmatory evidence requires oblique rotation to perfect congruence, which can always be attained. Rotation to perfect congruence is advocated as a new strategy, in which amounts of variance explained are of major concern. The perfect congruence strategy can be applied to variable-component correlations and to weights. The latter approach is to be preferred for two reasons. First, rotating weights to perfect congruence can be easily understood as a cross-validation method, closely related to the well-known multiple group method. Second, this approach appears to give more satisfactory results in practical applications than are obtained from rotating variable-component correlations to perfect congruence.

The Kaiser, Hunka and Bianchini Factor Similarity Coefficients: A Cautionary Note.

Kaiser, Hunka and Bianchini (1971) have given a method to compare two matrices of factor loadings based on the same variables, but different groups of individuals. They suggest expressing the factor similarity by the elements of the rotation matrix that optimizes the sum of inner products between corresponding columns or rows of the matrices. The optimal rotation involved in this method is examined from a mathematical point of view and, on the basis of this, the method is shown to be invalid.

Scale Construction on the Basis of Components Analysis: A Comparison of Three Strategies.

Cattell suggested constructing scales on the basis of components analysis by assigning weights 1 to variables with high positive loadings on the components and -1 to variables with high negative loadings on the components. This strategy is compared with other strategies of scale construction, which assign weights 1 or -1 to variables with high weights for the components. The latter strategies appear to yield scales which have higher correlations with the corresponding component scores and lower correlations among them, but also are less homogeneous then scales constructed by the former strategy. It is concluded that the former strategy, which is very popular among practitioners, can continue to be used if homogeneity of scales is preferred to orthogonality of scales.

Convergence Properties of Certain Minres Algorithms.

Levin (1988) has challenged the convergence properties of the Harman and Jones (1966) method of Minres factor analysis. Levin claimed that convergence of the Harman and Jones method is not guaranteed and that a modified version of this method, with proven convergence, is to be preferred. In the present article it is shown that both claims are invalid. Monotone convergence of the Harman and Jones method is guaranteed whereas the modified version, proposed by Levin, may converge to an incorrect solution. Levin has also claimed that the rank-one version of the Harman and Jones method, as implemented in a method by Zegers and ten Berge (1983) lacks a valid convergence proof, and that a method suggested by Comrey and Ahumada (1964, 1965) should be used instead. It is shown that these claims, too, should be reversed.

A Fast and Simple computational Method of Minimum Residual Factor Analysis.

Some new procedures for minimum residual factor analysis are presented. First a successive method developed by Comrey is modified in order to guarantee convergence and to provide a way to handle Heywood cases. Next, this modified Comrey procedure is extended to a simultaneous procedure which is computationally simpler and faster than the Minres method developed by Harman and Jones. This latter method, however, satisfies a stronger necessary condition for the minimum of the sum of squared off-diagonal residuals. Some empirical results are presented. These are in accordance with the theoretical considerations; that is, the Harman and Jones procedure tends to be slower, but attains in general a lower value for the sum of squared off-diagonal residuals.

OKT3-induced nephrotoxicity is associated with release of group II secretory phospholipase A2.

Administration of the murine IgG2a CD3 monoclonal antibody OKT3 exerts a transient nephrotoxic effect. Increased levels of group II secretory phospholipase A2 (sPLA2-II) might account for this nephrotoxicity as sPLA2-II induces the biosynthesis of prostaglandins, vasoactive lipid mediators that influence glomerular haemodynamics and renal function. Furthermore, extracellular phospholipases seem to be involved in proximal tubular cell injury. We studied plasma sPLA2-II levels in relation to circulating creatinine, tumour necrosis factor alpha, interleukin 6 and C-reactive protein levels in 15 renal allograft recipients receiving rejection treatment with OKT3. As a control group, we studied 15 renal allograft recipients receiving rejection treatment with methylprednisolone. A maximal fourfold increase in sPLA2-II levels was observed 48 h after the first OKT3 administration, preceded by increased tumour necrosis factor alpha and interleukin 6 levels and accompanied by increased C-reactive protein levels. Creatinine levels reached a maximal increase 72 h after initiation of treatment. During methylprednisolone treatment no increase in any of the studied parameters was observed. Thus, administration of OKT3 induces increased sPLA2-II levels, presumably via generation of cytokines. We hypothesize that sPLA2-II may contribute to the nephrotoxic effect of OKT3 by inducing vasoconstrictive prostaglandins and renal tubular cell injury.