(e)(3 points )Based on Pseudo R2measures ,which model is preferred between the Probit and Logit models ? Discuss your logic . (f) (3 points) Without performing any additional calculations, compared to your results in part (c) please qualitatively anticipate the change in probability of an increase in ?iira?i? from 0.4 to 0.5 . Explain your rationale C ONTINUED BEL 0W Free -Response #3: A researcher examines wage and education data for a sample of men in 1976 . Data is collected based on the following model (where educ : number of years of education and exper : number of years of experience ): log(wage) : B0 +B1educ + zexper + [33exper2 + u a.) (3 points )Assuming exper is exogenous ,what potential problems arise when estimating this model using ordinary least squares? Explain your rationale. b.) (3 points) One of the variables in the dataset is called nearfcollegei4 , which equals 1 if the person grew up near a 4year college , and 0 if not. Why is near_college_4 not plugged into the above model instead of educ ? What is the intuition behind collecting data on near_college_4? Explain your rationale. c.) (3 points )Assuming nearicollegejl is uncorrelated with the error term , state the necessary conditions for the validity of using near_college_4 as an instrument. Explicitly state: i.) Any relevant regression (s) that have to be run to check the validity of the instrument. ii.) The null hypothesis and the relevant statistical test to check the validity of the instrument. d.) (3 points) The researcher's estimations are presented in the table below .As usual , each column in the table presents the results of a different regression. Aside from classical hypothesis testing, how can we use the below results to evaluate if the coefficient estimate on educ in column (2) is individually significantly different from zero at a 5% level