4] Using our height eweight data ,estimate a linear probability model that predicts female on the basis of weight a) Report the results and interpret them b ) Plot the predicted values on the y , axis (and weight on the x , axis ) . How many are > 1 or 0.50 and zero otherwise . Form the 2x2 table comparing this prediction to the true value of female .What percentage of observations are correctly predicted ? 5] Explain how the RESET test works . Can it detect all types of omitted variables .7 Suppose a regression equation fails the RESET test : what changes can one make to the functional form to try to solve this problem .7 6 ] Using the dataset INFMRT .DTA , run the regression illustrated in equation 9.43 [Example 9.10 ] . (Note The example in the hook uses the data from 1990 only ; you should use the data for both years , so you will get different results . ) a ] Conrm that you get a positive and signicant coefcient for [physio . What does this mean , and why does it occur ? (Read the discussion in Example 9.10 for an explanation .] b ] Generate a scatter plot of in fm o r t v e r s u s lphysic . Do you see the outliers .7 G] Add the dummy variable D C to the regression and interpret the new results ; does this solve the problem 7 d ] Instead of adding the dummy variable D C to the regression , run the regression leaving out those two observations { " if DC : : " ] . Does this solve the problem ? e ] Instead of the approach in c) or d ] J try using the " qreg \" com mand to run the Least Absolute Deviations version of this regression [ include the two DC cases , and leave out the DC dummy ] . Does this solve the problem as well as c] or cl)