Whose School Feels Safe? Student Voice, Relational Trust, and Everyday Inclusion in Diverse Learning Communities

Abstract

School safety is often monitored through incidents, rules, and physical security, but these indicators may not capture whether different students feel protected in everyday interactions. This convergent mixed-methods study integrated a survey of 450 students from four Indonesian secondary schools with participatory safety mapping and focus-group evidence. Five scales demonstrated acceptable to strong internal consistency (Cronbach's alpha = .82–.91). Physical safety was rated high (M = 4.12, SD = 0.68), whereas emotional safety (M = 3.24, SD = 0.91), relational trust (M = 3.45, SD = 0.82), belonging (M = 3.58, SD = 0.79), and student voice (M = 2.98, SD = 0.95) were weaker. Emotional-safety and teacher-trust scores were significantly lower for girls than boys, students with disabilities than their peers, and low-SES than high-SES students. A regression model explained 54.2% of variance in perceived safety, F(3, 446) = 175.8, p < .001. Relational trust was the strongest predictor (beta = .412), followed by belonging (beta = .325) and student voice (beta = .148). Participatory maps located insecurity in poorly supervised back areas and toilet corridors, while focus groups showed that dismissive adult responses made formal reporting socially risky. Integration produced a central meta-inference: safety is relational and uneven before it becomes an incident statistic. The article proposes a relational–spatial model of everyday inclusive safety and recommends disaggregated monitoring, credible confidential response, participatory hotspot audits, and documented feedback to students.