Identification of children at risk of dropping out of primary school using a probit model

The purpose of this study was to provide an indicator to estimate the probability of primary school children being exposed to school dropout based on the probit regression model and using the information of 8678 school dropouts and the same number of children in school. These variables can be divided into three categories: economic variables, non-economic variables and environmental variables. The estimation results showed that the signs of the coefficients of the variables are in accordance with theoretical and experimental expectations. For example, increasing the age of the child caregiver, the family size, or the fact that the child caregiver is a woman increases the chances of the child being left out of school. In the section on economic variables, the results show that the low monthly purchase rate, the low household income decile and the lack of a fixed income by the head of the household will also increase this probability. The results of the model estimate also showed that households covered by support institutions will have a lower chance of the child dropping out of school, which indicates the role of support institutions in improving social indicators. In poorer and less developed provinces, the child has a better chance of surviving school. From the point of view of good fitting statistics, the results of the model with a power of 83% is able to distinguish a child who is dropping out of school from a child who is not about to drop out of school.