Recent studies on the economic effects of works councils in Germany using the European Company Survey estimate a significant negative effect of works councils on establishment productivity and profitability. These results are in stark contrast to studies using the IAB Establishment Panel estimating a significant positive effect of works councils on establishment productivity and profitability. This article scrutinises these empirical approaches. While sample selection and control variables have a substantial impact on the magnitude of marginal effects, the definition of the dependent variable as an objective or subjective measure causes the opposing signs. Beyond that, similar measures in both datasets lead to comparable marginal effects highlighting the relevance of the definition of the dependent variable for inferences and interpretation of studies about the effectiveness of industrial relations institutions and raising questions about the validity of the performance measures.