Analisis Regresi Probit Biner Bivariat

Abstract

Bivariate binary probit regression is a regression analysis that uses two dependent variables and each has two categories. This regression analysis is used on education index data and expenditure index of district/city on Kalimantan island in 2017. The best model obtained in this regression analysis is a model that uses 4 independent variables namely APS 16-18 years, percentage of poor population, open unemployment rate, and GRDP ACMP (Gross Regional Domestic Product at Current Market Prices). The parameters that significantly influence the two dependent variables are the APS 16-18 years in models 1 and 2 and the percentage of poor people in model 2. In Samarinda, every change of the APS 16-18 years, the percentage of poor people, and the open unemployment rate of 1 the unit will increase the probability of Samarinda entering the education index and high expenditure index categories by 0,33 percent, 0,42 percent and 0,07 percent, respectively. Every change of GRDP ACMP by 1 unit will reduce the probability of Samarinda entering the education index and the high expenditure index by 1,63 percent.