Using data-driven analytics and ecological systems theory to identify risk and protective factors for school absenteeism among secondary students

Abstract

Chronic absenteeism is an administrative term defining extreme failure for students to be present at school, which can have devastating long-term impacts on students. Although numerous prior studies have investigated associated variables and interventions, there are few studies that utilize both theory-driven and data-informed approaches to investigate absenteeism. The current study applied data-driven machine learning techniques, grounded in “The Kids and Teens at School” (KiTeS) theoretical framework, to student-level data (N = 121,005) to identify risk and protective variables that are highly associated with school absences. A total of 18 risk and protective variables were identified; all 18 variables were characteristics of the microsystem or mesosystem, emphasizing school absences' proximity to variables within inner ecological systems rather than the exosystem or macrosystem. Implications for future studies and health infrastructure are discussed.