Conditional Logistic Regression with Sandwich Estimators
(Written by Doug Midthune)
This IBM PC program performs conditional logistic regression on clustered data and calculates a sandwich ("robust") estimator of the covariance of the regression coefficients. Use of the sandwich estimator allows appropriate inferences for either fixed or random effects across clusters. The program uses two modifications to the usual Wald test to adjust for cases where some parameters are estimated from a small number of clusters. First, the middle of the sandwich estimator is calculated using standardized residuals, and second, the test is performed using the t-distribution instead of the normal distribution, where the degrees of freedom are estimated using Satterthwaite's approximation.
For additional information, see Fay, Graubard, Freedman, and Midthune, "Conditional Logistic Regression with Sandwich Estimators: Application to a Meta Analysis", Biometrics Vol. 54 pp. 195-208, 1998.
Also included are the Fortran program used to make the PC version, a data example, and the documentation on the use of the program.
- clogist.zip (zip, 328kb) - Zip file containing PC program, documentation and example.