Question 3 For this question, use the data crimel from the wooldridge package. datatcrimel , package= ' wooldridgc ') 'We want to use a Poisson regression to explain the determinants of the number of times an i1'1dividual was arrested in 1986 (narr86}. 2. 5. Explain the benefits of estimating the model by Poisson regression in this case \"re want to test if there are some racial discrimination, so we want to estimate the model using black and hispan as regressors. Estimate the model with glm using the log link. Interpret the coefcients. Estimate the same model using the square root link and interpret the coefficients. Explain why they are different from the ones obtained in the previous question. For the square root link. you need to set family to poissonlfliansqrt}. Try to compute the marginal effect of black and hispan manually for the two models estimated in the previous two questions. Explain the difference. Compare your result with the ones obtained with the margins package. lGiven the results from the previous question. can you conclude that there is racial discrin'iination? Explain. In order to better test the presence of racial discrimination. control for the proportion of prior convictions (pcnv), time in prison since the age of 18 (tottime) and the number of quarters employed in l986 (qempBB). For this question, only use the log link. Interpret the marginal effects of the coefficients. Using the previous model. do you detect discrimination? Do you think the model is better to detect discrimination than the ones from questions 2 and 3? Do you think is it a biased measure of discrimination? Explain