Reference no: EM132403122
Assignment Task :
Use the Stata voter data set voter.dra. This is a data set dealing with choice of presidential candidate in the 1992 presidential election.
First carry out an ordinal logistic regression to predict the highest degree the respondent achieved. The degree categories arc 0 n less than high school, 1 = high school, 2 = junior college, 3 - bachelor's degree, 4 - graduate degree. The independent variables should be sex (M = 1, F = 2) and age. I recommend that you recode binary variables so that they take on the values 0 and 1.
Ask for an asymptotic correlation matrix for the estimated coefficients, and a test of parallel lines. What do you learn about the determinants of educational attainment? tN.B. you can obtain the asymptotic correlation matrix of estimated coefficients with the post-estimation command estat vet, eon]
If your output says that it is appropriate to do so:repeat the analysis treating the dependent variable as nominal rather than ordinal.
Manipulate the results so that you obtain coefficients telling the probability of having a particular educational degree relative to the next lowest degree.
Create a classification table that cross-classifies the predicted degree category and the actual degree category.
On the basis of this table, how well does the model predict people's degree categories? What do you learn from this analysis that you did not learn from the ordinal logit?