Monitoring Adolescent Peer Victimization Across PISA 2022 and 2025: Measurement Comparability and Implications for Trend Interpretation

Background/Objectives: Comparable measurement is essential for monitoring adolescent peer victimization. We examined how linking assumptions affected the estimated change between PISA 2022 and PISA 2025. Methods: Three- and four-class models were fitted to four common peer items for 1,173,352 students in 75 matched education systems. An exploratory extension included online experiences in 33 systems. Survey-weighted intervals, alternative anchors, fixed country contributions, and classification-adjusted external correlates were examined. Results: Across the three principal three-class specifications, the occasional profile increased by 1.95–3.81 percentage points. Frequent multiform change was compatible with zero under two specifications and positive under full invariance. In four-class models, severe-profile change was −0.57 points (95% CI [−1.10, −0.04]) with cycle-specific definitions and +0.49 [+0.29, +0.68] with invariant definitions; their paired difference was +1.06 [+0.58, +1.53]. These estimates describe differently defined groups. Fixed country contributions preserved the occasional-profile increase. Conclusions: Estimated trends depend on how profiles are defined across cycles. Monitoring reports should present sampling uncertainty and sensitivity to measurement, population composition, and questionnaire coverage.