PENGGUNAAN MODEL REGRESI QUASI LIKELIHOOD UNTUK MENGATASI MASALAH OVERDISPERSI PADA REGRESI POISSON

Abstract

One of the main issue in the poisson regression model is over dispersion. The existence of over dispersion causes the estimated regression coefficients tend to produce estimates that deviate from the actual parameter values. The main purpose of this study is to show how to overcome the dispersion problem in this regressiom model using Quasi Likelihood regression method.  This study used secondary data with sample size 316. The results of this study indicated that the value of Pearson Chi-square and deviance in the testing of the goodness of fit for quasi likelihood  regression model were more closed to the value of chi-square table with degree of freedom 313. Therefore, the application of quasi likelihood model regression is more adequate than the Poisson regression model to handle over dispersion.