A researcher used stepwise regression to create regression models to predict CarTheft (thefts per 1,000) using four predictors: Income(per capita income), Unem (unemployment percent), Pupil/Tea(pupil-to-teacher ratio), and Divorce (divorces per 1,000 population) for the 50 U. S. states. Regression Analysis—Stepwise Selection (best model of each size) 50 observations CarTheft is the dependent variable p-values for the coefficients Nvar Income Unem Pupil/Tea Divorce Std. Err Adj R2 R2 1 0.0004 167.482 0.218 0.234 2 0.0018 0.0000 152.362 0.353 0.379 3 0.0013 0.0157 0.0001 144.451 0.418 0.454 4 0.0007 0.0323 0.0003 0.1987 143.362 0.427 0.474 (a) Which model (Nvar 1, 2, 3, or 4) best balances fit and parsimony? The (Click to select) variable model best balances fit and parsimony.
(b) Does the addition of Divorce improve the model with respect to the R2adj?
Yes
No
(c) Which two variables appear to be the most significant? Income and Unem Income and Pupil/Tea Unem and Divorce Unem and Pupil/Tea