ABSTRACT
The idea of an interaction being carried by an IV in MRC analysis was intro duced in Chapter 5, in the discussion of contrast coding for a 2 x 2 design (Section 5.5.4). We saw there that the coding of the interaction contrast required the multiplication of the coding coefficients for the two main effects (Xx and X2) and that the resultingX 3 could be interpreted as any other IV in an MRC analysis. (A review of Section 5.5.4 at this point probably would be helpful.) It was also stated there that “ throughout MRC, whether we are dealing with nominal or quantitative IVs or combinations of these, interactions are carried by products of variables (or variable sets),” and a footnote claimed that the implications of this statement for data analysis are far-reaching. In this chapter we seek to draw these implications out fully, and deliver on that promise.
