Dawoud-Kibria Estimator for the Logistic Regression Model: Method, Simulation and Application

The logistic regression model (LRM) is used when the response variable is in a binary form. In the presence of multicollinearity, the variances and standard errors of the maximum likelihood estimator (MLE) become high. As a remedy caused by multicollinearity, we introduce a new estimator called Dawoud-Kibria (DK) estimator for the LRM and compared its effectiveness with some traditional biased estimators, i.e., ridge and Liu estimators. For investigating the performance of the new estimator, we conduct a Monte Carlo simulation study with different evaluated conditions. Findings proved that the performance of our proposed DK estimator is quite better as compared to the traditional MLE, ridge and Liu estimators. At the end, an empirical application is being considered. Both simulation and application results clearly demonstrate that the proposed DK estimator is a robust and reliable option as compared to the other considered estimators.