Class 9: homework
1. Estimate a regression model having q32 (reported earnings) as the dependent variable and degree, gender (the variable q8), age (the variable q9age) and age squared as the independent variables.
Caution: values less than zero or equal to 99998 of the variable q32 should be recoded as missing data.
2. Are earnings statistically significantly related to education? If so, what is the relationship between earnings and the degree of education?
3. Do men report greater earnings than women? If so, what is the average difference in reported earnings between male and female respondents
4. Use coplot() [see class 08] to create a chart showing the predicted relationship between prudence and earnings in categories of education and gender.
5. The variable size in the dataset denotes the category of a subject's place of residents. Recode the variable so that value 1 corresponds to "rural area"; values 2, 3 and 4 correspond to a "small city"; and values 5 thru 8 correspond to a "city". Create a factor from the recoded variable. Set small town as the reference category. Add the variable to your model. Does that significantly improve the fit?