Abstract
Teacher collaboration is widely promoted as a resource for professional learning, teacher efficacy and retention, but research and policy often treat collaboration as a single practice and use frequency as a proxy for professional value. This risks overlooking differences between routine exchange, coordination and deeper forms of collaborative inquiry. Using TALIS 2024 teacher data from eight education systems with both lower and upper secondary participation (N = 33,137), this study examines how two content-defined observed practice indices—practice-focused and exchange/coordination—are associated with perceived professional learning impact, teaching self-efficacy and job satisfaction. Teacher-weighted controlled path models with country fixed effects and school-cluster robust standard errors were estimated. A formal within-model contrast showed that the practice-focused index had a larger standardised association with perceived professional learning impact than the exchange/coordination index. Professional learning impact was associated with teaching self-efficacy, which was associated with both teaching-profession and work-environment satisfaction. Formal interaction tests found no statistically detectable ISCED 2/3 differences in the five core path coefficients or in the practice-focused versus exchange/coordination coefficient gap. Country moderation was detected for the practice-focused index association but not for the exchange/coordination index association. Because professional learning impact was a single self-report item whose observability depended on questionnaire-form allocation, participant-specific non-administration, routing and response, the results are interpreted as theory-guided indirect associations rather than causal mediation. The findings suggest different associations for the two observed indices; weak measurement diagnostics preclude interpreting them as clearly separated collaboration types.