Multivariate Regression Look at the results of a regression

Multivariate Regression. Look at the results of a regression analysis of opposition to Obamacare (Table 2).

Logistic Regression. Assume that the dependent variable in Table 2 is a binary variable (0=support Obamacare, 1= oppose Obamacare) and that the coefficient estimates are from a logistic regression (Table 2).

What is the predicted probability of opposing Obamacare for a Hispanic female from the highest family income quartile with the lowest political expertise/sophistication?

What are the estimated odds of opposing Obamacare for the same female?

Table 2: Opposition to Obamacare

__________________________________

N = 1,000

R^2 = .473

Regression (Total) Sum of Squares = 360.00

__________________________________

Race: 0 if Hispanic, 1 if otherwise
Gender: 0 if male, 1 female
Income: 1-4 quartile, 1 = lowest
Political Expertise: 0-3 scale, where higher score indicates more sophistication and expertise

Solution

Multivariate Regression. Look at the results of a regression analysis of opposition to Obamacare (Table 2).

Logistic Regression. Assume that the dependent variable in Table 2 is a binary variable (0=support Obamacare, 1= oppose Obamacare) and that the coefficient estimates are from a logistic regression (Table 2).

What is the predicted probability of opposing Obamacare for a Hispanic female from the highest family income quartile with the lowest political expertise/sophistication?

What are the estimated odds of opposing Obamacare for the same female?

Table 2: Opposition to Obamacare

Coefficient

Standard Error

Constant

8.56

0.74

Gender

-7.59

2.04

Race

1.37

0.21

Political Expertise

0.41

0.22

Family Income Quartile

0.05

0.18

The regression equation is

Log(P/1-P) = 8.56-7.59*gender +1.37*race+0.41* Political Expertise +0.05* Family Income Quartile

What is the predicted probability of opposing Obamacare for a Hispanic female from the highest family income quartile with the lowest political expertise/sophistication

The value for the variables are

Gender female =1

Race = Hispanic =0

Political Expertise = lowest political expertise =0

highest family income quartile =4

Log(P/1-P) = 8.56-7.59*1 +1.37*0+0.41* 0 +0.05* 4 = 1.17

P/1-p =exp(1.17) =3.222

P= 3.222/( 1+3.222) =0.763

Predicted probability of opposing Obamacare for a Hispanic female from the highest family income quartile with the lowest political expertise/sophistication = 0.763

Estimated odds of opposing Obamacare for the same female = P/(1-P) =3.222