Purpose: Recent teacher evaluation reforms across the globe can improve student outcomes, though doing so depends on implementation by school leaders. This study aims to describe how teacher reports about classroom observation processes (TROPs) vary across within-teacher observations, teachers and schools and depend on contextual, teacher and school characteristics. Design/methodology/approach: A researcher-practitioner partnership collected TROPs using surveys administered multiple times during one semester and linked results to Tennessee administrative data. Random effects quantify TROP variation within examined clusters and estimate relationships between TROPs and the examined characteristics. Findings: Although observation processes are well implemented on average, TROPs vary considerably across teachers and observations within teachers over a short period. TROPs also depend on several examined characteristics, but not necessarily in ways predisposed toward teacher development. Originality/value: The study extends the understanding of the degree and extent to which teacher evaluation implementation by school leaders varies. Leadership preparation programs and in-service support providers might use the evidence herein to improve the extent to which evaluation-related observation processes are implemented consistently and in ways supporting teacher development.