Application of Multivariate Strategies to Problems of Measuring and Structuring Long-Term Change

ABSTRACT Developmental psychology, as a scientific discipline, has not paid sufficient attention to methodological techniques that permit the examination of change in a multivariate setting. Multivariate analysis techniques and procedures are discussed which are relevant to the investigation of qualitative (structural) and quantitative change–two crucial aspects of the study of long-term change in a life-span developmental context. More specifically, the strategy of comparative factor analysis attempts to disentangle qualitative and quantitative changes as distinct sources of variability evidenced by the empirically determined characteristics of factor loading patterns and factor scores. Suggestions are discussed for assessing the degree of structural invariance in a set of measures, a necessary prerequisite for the examination of quantitative change. Implications of Cattell's “covariation chart,” which defines six covariation techniques ( R, Q, P, O, S, T ), for the study of developmental change are discussed. It is suggested that the typical, cross-sectional, factor analysis procedure ( R technique) may not be the most useful for studying developmental phenomena. Attention is drawn to alternative covariation techniques and some discussion is given of their application to particular developmental issues.