Indiana University
University Information Technology Services
  
What are archived documents?
Login>>
Login

Login is for authorized groups (e.g., UITS, OVPIT, and TCC) that need access to specialized Knowledge Base documents. Otherwise, simply use the Knowledge Base without logging in.

Close

In Stata, how does .logit handle a non-binary dependent variable?

The Stata .logit command fits a binary logit model for a dichotomous dependent variable. This command, however, fits a model even when the dependent variable is not binary. All values are a success, except for 0, which .logit treats as failure.

Suppose your dependent variable dv has values of -1, 0, 1, and 2. You may recode the variable into a dichotomous variable (0 versus all other values) and then fit a binary logit model as follows:

. recode dv -1=1 0=0 1/2=1 *=., gen(dv_new) . logit dv_new x1 x2 x3

It is possible to execute the .logit command without recoding:

. logit dv x1 x2 x3

However, to avoid unwanted results, correctly recode all variables before running .logit.

For more about statistical and mathematical software, email the UITS Stat/Math Center, visit the center's web page, or phone 812-855-4724 (IUB) or 317-278-4740 (IUPUI). The center is located in Bloomington at 410 N. Park Avenue, and is open for consultation by appointment Monday-Friday 9am-5pm.

This is document awxj in domain all.
Last modified on March 17, 2011.

Comments/Questions/Corrections

Use this form to offer suggestions, corrections, and additions to the Knowledge Base. We welcome your input!

If you are affiliated with Indiana University and would like assistance with a specific computing problem, please use the Ask a Consultant form, or contact your campus Support Center.

Contact Information

Note: We will reply to your comment at this address. If your message concerns a problem receiving email, please enter an alternate email address.