3. Suppose, at any Variable Selection problem, we are employing the Forward Selection approach. Suppose we have 7 covariates in our dataset of 50 observations, but we like to have at most 3 covariates in the regression equation. At one step of the algorithms, we have 2 covariates x1 and x5 in the regression equation. We are trying to see if another covariate can be added to the model. We tried each of the remaining 5 covariates separately, and the computed t-test statistics as follows: t-stat for x2 = t-stat for x3 t-stat for x4 t-stat for x6 t-stat for x7 = 2.5 0.4 0.6 2.57 6.28 We need to compare the significance of these test statistics by looking at their distance from an appropriate cutoff interval. Suppose a = 10%. a) From the below table, choose the appropriate one and then construct the cutoff interval at level a. (Note that the interval includes two cut off points). t (0.95, df = 42) t (0.95, df = 46) t(0.95, df = 48) t(0.05, df = 46) = 1.6819 = 1.6786 = 1.6772 = -1.6786 t (0.90, df = 42) t (0.90, df = 46) t(0.05, df = 48) t(0.10, df = 42) = 1.3020 = 1.3002 = -1.6772 = -1.3002 b) Which covariate will be added in the next step