The Binary Logistic Regression is used to perform logistic regression for binary response variable with one or more predictors.
1. Choose ProcessMA > Statistics > Regression > Binary Logistic Regression.
2. In Response, select the column containing the response values.
3. In Event, select the reference event of the response.
4. In Frequency (Optional), select the column containing frequency data.
5. In Model, select the columns containing variables to be included in the model.
6. In Factors (Optional), select the columns containing variables in the model that are factors.
7. In Link functions, select the link functions to fit the response model.
8. Check Diagnostic plots, to plot Delta chi-square, Delta deviance, Delta beta (Standardised) and Delta beta.
9. In Diagnostic plots VS, choose the x-axis variable for the diagnostic plots.
10. Click OK.
Note To select a column of data into a textbox, double-click on any of the column names shown in the list on the left of the dialog box while in the textbox.
Response: Numeric.
Frequency: Integer; Must be greater than zero; Must have the same number of data points as the variable.
Model: Text or Numeric; Must have the same number of data points as the variable.
Factors: Text or Numeric; Must have the same number of data points as the variable.
You want to understand the effect of number of hours of sleep and smoking on blood pressure. You classified the blood pressure into Good and No Good.
1. Open worksheet ProcessMA > Tools > Data Files > Regression.xls.
2. Choose ProcessMA > Statistics > Regression > Binary Logistic Regression.
3. In Response, select J – Pressure.
4. In Event, choose Good.
5. In Model, select K – Sleep Hours, L – Smoker.
6. In Factors (Optional), choose L – Smoker.
7. Click OK.
Interpretation
From the logistic regression table, you can obtain the coefficients for the regression model. In additions, the p-values for Sleep Hours and Smoker are less than 0.05 which indicates that there is sufficient evidence that their coefficients are not zero.