ABSTRACT
Regional CFA models differ from the global models discussed in Chapter 5 in one fundamental aspect. Whereas global models assign all variables the same status, regional models distinguish between groups of variables. For instance, regional CFA models allow one to explore the relationships between the groups of the motivational and the cognitive variables, or they allow one to discriminate between patterns of leisure behaviors in men and in women. The type of analysis is still exploratory. However, there is a stronger explanatory component in regional CFA than in global CFA. The researchers specify the variable groups before CFA. The method of CFA can be used to analyze groups of variables (see Sections 6.1.1 and 6.1.2). If no prior knowledge exists about the composition of variable groups, cluster analysis (Hartigan, 1975) or correspondence analysis (Greenacre, 1984) can also be used to create variable groups. The following sections introduce readers into regional CFA of existing groups of variables. We begin with Interaction Structure Analysis (ISA).
