The gap between teacher grades and standardised test scores by student-ascribed characteristics is often interpreted as evidence of teacher bias. This gap raises concerns about teachers’ roles as fair evaluators of academic merit and their contribution to educational inequality, as posited long ago by Cultural Capital Theories (CCT). Conversely, Rational Action Theories (RAT) emphasise family decision-making, often overlooking teachers as relevant actors. Identifying whether teacher assessments are fair or biased is critical for equal opportunity but methodologically challenging. Discrimination in education is underexplored compared to the labour market, with research hindered by omitted variable bias (e.g., effort and classroom behaviour), measurement error (e.g., test scores), and underpowered, low-validity experiments. These issues complicate determining whether teacher bias is a substantial inequality mechanism or a statistical artefact. Grounded in interdisciplinary theories—including implicit bias, status characteristics beliefs, statistical discrimination, CCT, and RAT—this seminar integrates NEPS panel data and a large-scale factorial experiment to analyse whether teacher assessments are biased by student ascribed characteristics, particularly socioeconomic status. Observational findings reveal that adjusting for measurement error reduces teacher bias in grading and tracking recommendations by 40%, challenging prior (over)estimations. Experimental findings unveil bias favouring highbrow cultural capital in essay grading and statistical discrimination against working-class students in long-term expectations. The seminar discusses the relevance of teacher bias as a mechanism of educational (in)equality alongside families.