Validating a Two-Factor School Discipline Measure for Adolescents: An Unrestricted Factor Analysis in Indonesian Samples

This study examined the dimensional structure of the School Discipline Scale by comparing two exploratory models: the bifactor model and the traditional two-factor model. Data were obtained from 1,300 secondary students who responded to 23 items on a 5-point scale assessing perceptions of punitive and non-punitive disciplinary practices. Exploratory analyses were conducted using FACTOR (Lorenzo-Seva & Ferrando, 2023) with polychoric correlations, MORGANA extraction, and robust Promin rotation. Results from the bifactor model indicated strong saturation of a general discipline factor (ECV = .69; ωₕ = .82) but insufficient unidimensionality (UniCo = .76; MIREAL = .33), suggesting an overarching discipline climate with two residual domains. In contrast, the two-factor model yielded a simpler, interpretable structure with adequate fit (CFI = .95; RMSEA = .052) and moderate correlation between punitive and non-punitive factors (r = .51). Reliability was high across models (ω = .85-.91), and external validation showed non-punitive discipline correlated more strongly with lower academic procrastination (r= -.25, p < .01). Findings highlight that school discipline represents both a unified moral perception and two distinct operational approaches. The bifactor model offers theoretical depth, while the two-factor model supports applied use. A hybrid scoring approach is recommended to preserve both general and specific variance for comprehensive evaluation of school discipline practices.