4. Use data in le ass5w18.dta at the course Canvas website. In answering the following it can be helpful (though not necessary) to uses estimates store estimates table to present model results side by side. Example: regress y x estimates store MODEL1 regress y z estimates store MODEL2 regress y w estimates store MODEL3 estimates table MODEL1 MODEL2 MODEL3, b se t p stat(N r2) (a) Fit linear, linear-log, log-linear and log-log models for regression of sales on tv . (b) Which model, if any, do you prefer? Explain. (c) Compute the average marginal elect (AME) and the marginal elect at the mean (MEM) for the linear-log model