Robust Standard Error Estimate for Cluster Sampling Data: A SAS/IML Macro Procedure for Logistic Regression with Huberization

Data sets with cluster structure are very common in practical business and research, one has to take into account the intra-cluster correlation in data analysis. This paper systematically discusses the Huber/White standard error estimate for cluster sampling data in logistic regression, and presents a user-friendly SAS/IML macro procedure which can automatically fit logistic model, calculate robust standard errors and produce confidence intervals for odds ratio. The robust standard error calculated in this procedure also has a finite sample-adjustment feature which is now available in most updated version of some statistical software. The syntax for the procedure is simple and easy. One data example is shown to illustrate how to actually use the procedures. The SAS system products included in this work are SAS, SAS/STAT and SAS/IML . This procedure can be run on any IBM compatible personal computer, MVS/UNIX system and any other computer platforms with a working SAS system.