MAPPING OF GENDER INEQUALITY IN INDONESIA BASED ON INFLUENCING FACTORS USING GEOGRAPHICALLY WEIGHTED ORDINAL LOGISTIC REGRESSION

Abstract

Gender inequality is a condition of discrimination between men and women that results from unequal social systems and structures. Gender inequality is measured based on the gender inequality index (IKG). This research aims to map gender inequality in Indonesia based on influencing factors and compare classification accuracy results between the GWOLR and ordinal logistic regression model. Data was obtained from the Indonesian Central Statistics Agency (BPS RI) and KemenPPPA in the year of 2022. The Gender Inequality Index data as the response variable is categorized using an ordinal data scale, namely IKG (1) Low, IKG (2) Middle, and IKG (3) High with ten predictor variables from the dimensions of health, education, human empowerment, socio-culture, and employment, with the amount of data is 34 observation data. The research method uses geographically weighted ordinal logistic regression (GWOLR) based on exponential kernel weighting. In the data analysis stage, ordinal logistic regression is performed before applying GWOLR, and after the model is formed, the classification accuracy will be calculated. The results of this study indicate that mapping gender inequality in Indonesia based on influencing factors using the GWOLR model forms three groups. The first mapping location labeled as low inequality is influenced by women whose birth was attended by a health worker (X1), women who have a pre-employment card (X7), women who are employed (X8), and the percentage of women who married before the age of 17 (X10). The second mapping location labeled with middle inequality is influenced by women whose delivery is attended by a health worker (X1), women's net enrolment in higher education (X2), and women married before the age of 17 (X10). The three locations categorized as high inequality are influenced by female birth attendance by health personnel (X1), Women's Human Development Index (X3), female rape offenses (X4), female domestic violence offenses (X6), and female marriage under the age of 17 (X10). Modeling the Gender Inequality Index using the GWOLR model resulted in higher classification accuracy than the ordinal logistic regression model, which was 94.11%.