=+Try conducting the analysis using the uncentered data (and cross product based on uncentered data), as was done in Kranzler et al. (1999). Compare the coefficients and correlations from the two analyses. Would your interpretation be the same? You should find that the intercepts are not statistically significantly different without centering, but that they are different using the centered continuous variable and the interaction term created from the centered continuous variable. Compare the two printouts; see if you can develop a sense of why this difference occurs. (Hint: I recommend you focus on the correlation matrices and then the standard errors of the regression coefficients. Reread the section discussing the advantages of centering.)