A randomized controlled study was performed to assess the efficacy of financial incentives on quitting smoking. A total 878 employees (who smoked) of a multinational company based in the United States were randomized to either: - receive information about smoking-cessation programs (442 employees) : the control group - receive information about programs plus financial incentives (436 employees): the intervention group The primary outcome for this study was quitting smoking within 12 months after randomization. (we will refer to this simply as "quitting smoking") The results from a simple logistic regression of the study results are as follows: In(odds of quitting smoking) = -2.95 + 1.20x1, where x, =1 for the intervention group, and 0 for the control group. The standard error of the intercept is 0.20, and the standard error of the slope for x, is 0.25 The researchers also collected additional information on the enrollees include demographic and lifestyle predictors. Suppose the researchers want to use multiple logistic regression the build a predictive model to allow them to identify smokers who are most likely to quit. - How should the predictive power of such a model be assessed? a. Make sure that the p-values for all individual slopes in the model are less than 0.05. b. Measure how well the model predicts "quitting smoking" for the dataset used to estimate the multiple logistic model. c. Include only predictors that were statistically significantly associated with quitting smoking in simple logistic regression models. d. Measure how well the model predicts "quitting smoking" for a data sample from the same population, but that was not used to estimate the multiple logistic model