This paper presents an econometric mediation analysis. It considers identification of production functions and the sources of output effects (treatment effects) from experimental interventions when some inputs are mismeasured and others are entirely omitted.