Logistic Regression Exercises
Research Question: What factors predict that someone is in good health? Using Polit2SetB data set, run a logistic regression using Good Health (Health) as the outcome variable (Y) and Respondent’s age (age), Smoking status (smoker), Current employment (worknow), Body Mass Index (bmi), and Frequency of High stress in the past month (stressed) as the independent variables (X). When conducting any regression analysis, the dependent (outcome) variable is always (Y) and is placed on the y-axis, and the independent (predictor) variable(s) is always (X) and is placed on the x-axis. However, with logistic regression, the outcome variable MUST be a dichotomous variable (e.g., yes/no). Your independent variables may be continuous or dichotomous.
Follow these steps when using SPSS:
1. Click on Analyze, then Regression, then Binary Logistic. 2. Move Health into the Dependent box. Remember, in logistic regression the dependent variable is dichotomous. 3. Click on your predictor variables (age, smoker, worknow, bmi, and stressed) and move each of them into the box labeled “Covariates.” 4. For method, be sure “Enter” is displayed. 5. Click on Options and check the following: Hosmer-Lemeshow goodness of fit and CI for Exp(B) (should be 95%). Click on Continue. 6. Click on OK.
Assignment: Analyze the SPSS output and answer the following:
1. In one or two paragraphs, summarize the findings from the data. 2. Did you note any improvement in the percentage of classified cases from the null model (block 0) to the new model (block 1)? Indicate what the improvement was, if any. 3. Interpret (in words) the meaning of the odds ratio for age. 4. Interpret (in words) the meaning of the odds ratio for worknow. 5. Interpret (in words) the meaning of the odds ratio for bmi.