Alert: Failed to converge(step-halving-limit) error. What am I missing?

[#1]

Hi,

When I do the below steps with a categorical Y variable:

Select Analyze > Fit Model.

Select the response column and click Y, Response.

Select the predictor colulmns and click Add or use one of the Macros if you want to include higher order effects.

Click the button next to Personality and select Stepwise.

Click Run.

Click Minimum BIC and select P-Value Threshold.

Click Forward and select Mixed.

Click Go.

I get Alert: Failed to converge(step-halving-limit) error

I want to have a mixed stepwise regression.

What am I missing?

[#2 · Accepted solution]

The error you are encountering in the Stepwise platform indicates that there is complete or quasi-complete separation. This is a phenomenon that can occur for binary responses, when one predictor (or a combination of predictors) can perfectly predict the outcome. In this case the maximum likelihood estimates don't exist. There is a SAS note that addresses this issue with logistic models on our website at http://support.sas.com/kb/22/599.html. Although the note addresses SAS, the concepts discussed there are applicable to any software.