Examining the determinants of student academic performance in Somaliland: estimating unobserved effects at student and school levels using multi-level logistic regression

Abstract

This study investigates the factors influencing student academic performance in Somaliland, including school type, gender, region, and residence type. It utilizes a dataset of 14,342 students and employs a multi-level logistic regression model to analyze how these variables affect students’ performance in national secondary exams. The preliminary analysis reveals interesting trends, such as girls generally outperforming boys in most regions and rural students scoring higher than their urban counterparts. Private school students also achieve better grades than those in public schools. The multi-level logistic regression analysis uncovers unobserved heterogeneity at the school level, suggesting the presence of hidden factors influencing student outcomes. The regression analysis suggests that school type, residence type, and region significantly impact academic performance at the school level, with urban students in public schools having a higher likelihood of excelling. Surprisingly, gender does not appear to be a significant factor in student performance. These findings highlight the complex interplay of factors influencing education outcomes in Somaliland and emphasize the need for targeted interventions at the school level to enhance educational quality. However, it’s important to acknowledge the study’s limitations and the necessity for further research to uncover additional hidden variables affecting student outcomes.

Keywords

• Academic performance
• multi-level logistic regression
• ANOVA
• t-test
• Somaliland
• Adult Education and Lifelong Learning