The present example illustrates a method for assessing pattern similarity for correlation matrices from populations known to differ in clinically important respects. The data underlying previously published confirmatory factor analyses of neuropsychological traits are presented in two methods for assessing whether the matrices are similar enough to justify their aggregation into a single composite matrix. The significance of this procedure for research in applied neuropsychology is commented upon in terms of the benefits to model diagnostics.
References
1.
GorsuchR. L. (1983) Factor analysis. Hillsdale, NJ: Erlbaum.
2.
HarmanH. H. (1976) Modern factor analysis. (3rd ed.) Chicago, IL: Univer. of Chicago Press.
3.
JöreskogK. G.SörbomD. (1983) LISREL: analysis of linear structural relationships by the method of maximum likelihood—Versions V and VI. (2nd ed.) Chicago, IL: I.E.S.
4.
MacCiocchiS.FowlerP. C.RanseenJ. (1992) Trait analyses of the Luria-Nebraska intellectual processes, motor functions, and memory scales. Archives of Clinical Neuropsychology, 7, 541–551.
5.
SchwagerK. W. (1991) The representational theory of measurement: an assessment. Psychological Bulletin, 110, 618–626.
6.
SteyerR.SchmittM. J. (1990) The effects of aggregation across and within occasions on consistency, specificity and reliability. Methodika, IV, 58–94.
