The topic for my PhD project is assessment of music performance by social moderation in upper secondary school. My aim with the study is to develop knowledge of how teachers perceive and meet the challenges of performance assessment, with specific focus on the intersubjective teacher team. Research in the field of music performance assessment has traditionally focused on interrater reliability, comparing scores and assessment statements from individuals judging classical music performance. A number of recent studies have explored how social moderation is practiced in authentic situations, but these have not investigated how teachers assess performance across different instruments and genres in an educational context. My study seeks to fill this gap by observing the ways groups of teachers discuss and arrive at evaluative judgments of students’ performances at a Norwegian upper secondary school. This case study describes some specific methodological challenges that I encountered in the course of the data collection, related to researching authentic situations using video recording. The ways in which these challenges illustrate issues of identifiability, unpredictability, and reflexivity in qualitative case studies are discussed.