A solution to the problem of separation in logistic regression

Citations Over TimeTop 10% of 2002 papers

Abstract

The phenomenon of separation or monotone likelihood is observed in the fitting process of a logistic model if the likelihood converges while at least one parameter estimate diverges to +/- infinity. Separation primarily occurs in small samples with several unbalanced and highly predictive risk factors. A procedure by Firth originally developed to reduce the bias of maximum likelihood estimates is shown to provide an ideal solution to separation. It produces finite parameter estimates by means of penalized maximum likelihood estimation. Corresponding Wald tests and confidence intervals are available but it is shown that penalized likelihood ratio tests and profile penalized likelihood confidence intervals are often preferable. The clear advantage of the procedure over previous options of analysis is impressively demonstrated by the statistical analysis of two cancer studies.

Related Papers

• → Likelihood Ratio Testing of Variance Components in the Linear Mixed-Effects Model Using Restricted Maximum Likelihood(1998)167 cited
• → Modifications of REML algorithm for HGLMs(2011)16 cited
• → An alternative derivation of Harville's restricted log likelihood function for variance component estimation(2018)1 cited
• → Maximum Likelihood (ML) and Restricted Maximum Likelihood (REML)(1992)6 cited
• → ML Inference in the Presence of Incidental Parameters(2011)