3. Compute a multiple regression model using number of times drunk as the dependent variable. Include the measures for age, sex, race (recoded as a dummy variable), and at least two other variables that you think are related to the number of times drunk. This will be the baseline model in the following questions. (a) Compute the tolerance statistic for each of the independent variables in the baseline model. Does it appear that there is a problem with collinearity in the regression model? Explain why

(b) Compute an interaction term for sex with age. Add this term to the baseline model, and rerun the regression command. Is the effect of age on number of times drunk significantly different for males and females (i.e., is the interaction effect statistically significant)? If so, interpret the effect of age on the number of times drunk for males and females. • Compute the tolerance statistic for each of the independent variables in this model. Does it appear that there is a problem with collinearity in this model? Explain why. (c) Compute an interaction term for race (which should be coded as a dummy variable) and age. Add this term to the baseline model, and rerun the regression command. (The interaction effect from part (a) should no longer be included in the analysis.) Is the effect of age on number of times drunk significantly different for these two race groups? If so, interpret the effect of age on the number of times drunk for each race group. • Compute the tolerance statistic for each of the independent variables in this model. Does it appear that there is a problem with collinearity in this model? Explain why. (d) If one of the additional variables that you have added to your regression model is measured at the ratio level of measurement, compute an interaction term between this variable and either the sex or the race variable. Add this term to the baseline model (there should be no other interaction terms included in this analysis), and rerun the regression command. Is the effect of this variable on number of times drunk significantly different for the two groups? If so, interpret the effect of this variable for each group. (e) Compute the tolerance statistic for each of the independent variables in this model. Does it appear that there is a problem with collinearity in this model? Explain why.