Education & Training

Handling Missing Data by Maximum Likelihood Using SAS

This paper makes the case that maximum likelihood is usually better than multiple imputation for handling missing data and demonstrates how maximum likelihood for missing data can readily be implemented with the following SAS procedures: MI, MIXED, GLIMMIX, CALIS and QLIM.

User rating: 
SSRI rating: 
4
Duke SSRI

Durham, NC 27708 | 919.681.6019

SSRI

Gross Hall, 2nd Floor
140 Science Drive
Durham, NC 27708

SSRI

Erwin Mill

Erwin Mill Building
2024 W. Main St.
Durham, NC 27705

Social Media