Module 4, Week 2, Paper and Pencil Assignment 8

Module 4, Week 2, Paper and Pencil Assignment 3 1. You are interested in what factors influence whether someone living in the United States smokes. You specify the following linear regression model: y = .30 + lincome + zsouth + Bgage + e y = is a dummy variable that =1 ifsomeone smokes, and =0 if someone does not smoke. income = income measured in thousands south = a dummy variable that =1 if someone lives in the south, and =0 if they do not. age = age measured in years You collect a random sample of 120 individuals. The model above is estimated using least squares, producing the following results: 5? = 0.67 - 0.009income + 0.1lsouth 0.0003age a) What type of regression model is this? b} Interpret the estimated coefficients. c) Calculate the predicted probability that someone with an income of $55,000, who lives in the south, and is 40 years old smokes. d} Calculate the predicted probability that someone with an income of $100,000, who does not live in the south, and is 50 years old smokes. e) What are some ofthe main drawbacks of this type of regression model? 2. Next, you estimate a logistic regression model using the same variables as above and obtain the following results: 8 (2.120.12income+0.42.south 0.002a.ge) y = 1 + e[2.120.12income+0.4230uth-0.002age] a) Are the estimated coefficients in this model interpreted in the same way as the ones in question 1? b) Calculate the change in the odds that someone is a smoker for a $1,000 increase in income. c) Recalculate the predicted probabilities from question 1c and 1d