**StATS: **SPSS
dialog boxes for logistic regression (July 22, 2002)

This page is moving to a new website.

This handout shows some of the dialog boxes that you are likely to encounter if you use logistic regression models in SPSS.

Before we can use the bf data set for binary logistic regression, we need to create a variable which is coded as binary. In this case, we want to compare exclusive breast feeding as one category with partial, supplement, and none combined as a second category. To do this, select

Transform | Recode from the SPSS menu.

Be sure to define your output variable name and label and click on the CHANGE button. Then click on the

Old and New Values button.

We convert the values

- Exclusive (1) to 1;
- Partial (2), Supplement (3), and None (4) to 0;
- All other values to System-missing.
To fit a logistic regression model, select

Analyze | Regression | Binary Logistic from the SPSS menu. Shown below is the dialog box you will see.

In this dialog box, add your binary outcome variable to the

Dependent field. Add all of your independent variables to theCovariates field. If one or more of your independent variables is categorical, click on theCategorical button. Shown below is the dialog box you will get.

Add any of your categorical covariates to the

Categorical Covariates field. By default, SPSS will use one or more indicator variables to represent your categorical variables and will reserve the last category as a reference level. This is a reasonable choice for most situations. In this data set, feed_typ has two levels, 1=bottle, and 2=ng tube, and SPSS will set up a single indicator variable which equals one for the bottle group and 0 otherwise.Click on the

Continue buttonto close this dialog box and return to the earlier dialog box.SPSS offers additional analysis options for the logistic regression model. Click on the

Options buttonto see your choices.

This dialog box shown above illustrates all of the additional options for the logistic regression model. The one option I would always recommend is the

CI for exp(B) option box. This gives you a confidence interval for all of the odds ratios produced in this logistic regression analysis.The

Hosmer-Lemeshow goodness of fit option boxallows you to evaluate the assumptions of your logistic regression model.The

Classification plots option boxand theClassification cutoff fieldare useful when you are using a logistic model to evaluate a diagnostic test.Click on the

Continue buttonto close this dialog box and return to the earlier dialog box.SPSS also lets you choose whether to save some information from this logistic regression model in your data set. Click on the

Save buttonto see your choices.

The dialog box shown above lists all of the information that SPSS can save in your data set. The most useful information is in the

Probabilities option box.The

Group membership option boxis useful when you are evaluating a diagnostic test.The various measures under the

Influence options groupand theResiduals options groupare useful for examining the assumptions of your logistic model and for identifying individual data values which the model is highly sensitive to.Click on the

Continue buttonto proceed and then click on theOK buttonin the earlier dialog box.

This page was written by Steve Simon while working at Children's Mercy Hospital. Although I do not hold the copyright for this material, I am reproducing it here as a service, as it is no longer available on the Children's Mercy Hospital website. Need more information? I have a page with general help resources. You can also browse for pages similar to this one at Category: Logistic regression.