The Analysis of Profiles

Abstract

In Chapter 3.1 it has been proposed that an operationalization of the effective stimulus for the efferent systems might be obtained by the profile of observed physiological responses (see Figure 1). Profiles of variables contain important information not accessible through the separate variables. The kind of unique information inherent in profiles and how it can be retrieved from the observed response profiles is the topic of this chapter. Another link between previously discussed issues and the present chapter is (1) the conceptualization of activation as a multicomponent process (see Chapters 4.2 and 4.3) and (2) the analysis of physiological response specificities that simultaneously take into account the uniqueness and the consistency aspects of specificity (see Chapter 6.2.2). Profile analysis can provide a means to study both of these issues effectively. As has been noted previously, discriminant analysis (DA) provides a tool for the substantive description of specificity effects of physiological profiles; likewise, DA identifies components within specified sources of variance and covariance of an experimental design that might provide a starting point for the definition of activation components. Therefore, in contrast to approaches that analyze profiles in the context of factor analysis (see, e.g., Horn, 1969; Nunnally & Kotsch, 1983; Ross, 1964) or clustering procedures (see, e.g., Overall & Klett, 1972; Skinner, 1978, 1979), the present exposition is directed towards an introduction into the analysis of profiles in the context of DA.