PENERAPAN REGRESI POISSON UNTUK MENGETAHUI FAKTOR-FAKTOR YANG MEMENGARUHI JUMLAH SISWA SMA/SMK YANG TIDAK LULUS UN DI BALI

Abstract

This study aims to determine factors that influence the number of high/vocational school students that do not pass the examination in Bali in 2011. Factors that used in this study are the external factors include the proportion of  public high/vocational school, proportion of the classrooms high/vocational school have damaged, the proportion of high/vocational school teachers have certified, and the number of participants high/vocational school national examination. This study uses Poisson regression analysis. Poisson regression is included in the Generalized Linear Models and this is one of the regression analysis for count data models with Poisson distributed response variables. Poisson regression analysis methods assume the mean and variance of the response variables are equal. Response variable is said to be Poisson distributed and overdispersion not happen if the value of the ratio between the variance and the mean close to one and less than 2.5. Test results show the response variable in this study Poisson distributed and there is no overdispersion. The analysis showed that all four factors are independent variables in this study affect the response variable is the number of high school/vocational school students which do not pass the examination.