Bridging the Knowledge-Behavior Gap: A PLS-SEM Analysis of Cybersecurity Awareness Among Indonesian Teachers

Abstract Despite growing digital integration in education, teachers' cybersecurity practices remain inadequately understood, particularly in developing contexts. This study examines the relationships between cybersecurity knowledge, ethics, and behavior among Indonesian K-12 teachers using partial least squares structural equation modeling (PLS-SEM). Survey data from 233 teachers reveal a paradox: while knowledge significantly predicts protective behaviors (β = 0.388, p < 0.001), ethical awareness shows no significant relationship with behavior (β = 0.040, p = 0.623). The model explains 16.8% of variance in cybersecurity behavior. Measurement model analysis demonstrates excellent reliability (Cronbach's α > 0.93) and discriminant validity (HTMT < 0.42), though convergent validity issues (AVE < 0.50) necessitated item purification. These findings challenge prevailing assumptions that ethical frameworks automatically translate into secure practices. We theorize this disconnect reflects a moral licensing effect, wherein ethical awareness paradoxically reduces behavioral vigilance. The study contributes to cybersecurity education theory by (1) empirically demonstrating the insufficiency of ethics-based interventions, (2) establishing knowledge-as-capability as the primary behavioral determinant, and (3) introducing context-specific measurement instruments validated in non-Western educational settings. Practical implications suggest redirecting teacher professional development from values-based to competency-based cybersecurity training.