• Power of Tests for Negative Binomial Regression Coefficients in Count Data

Abstract

In this study we focus on a negative binomial (NB) regression model to take account of regression coefficients in Poisson counts. An algorithm for estimating parameters was obtained and score test versus its alternative tests were presented for testing the significance of regression coefficients in Poisson regression model against NB regression model. The power of the score test is compared with their Likelihood ratio test (LRT) and Wald test via Monte Carlo simulation. The simulation result indicated that the Wald test is superior over the LRT and Score test in terms of its power. An Ethiopian under five children death rate data is used to illustrate the tests.