Repeat problem 9.8, except this time create your own dummy variables (i.e., do not use a CLASS statement). Make the reference level for SES ‘2-Medium’ as before.
Problem 9.8
Run the following program to create a SAS data set called SMOKING. Outcome values of 1 represent cases (those with lung disease) and outcome values of 0 represent controls (those without lung disease). For this problem, generate a two by two table for outcome (OUTCOME) by smoking status (SMOKING). What is the chi-square and the odds ratio? Now, run a logistic regression with only SMOKING as the predictor variable. HINT: PROC LOGISTIC uses the lower value formatted value of the outcome variable (if there is a format), and since ‘Case’ comes before ‘Control’, you don’t need to include the DE~ SCENDING option).