This paper investigates the role of works councils in job satisfaction. Using the recently developed Linked Personnel Panel, we consider both the direct and indirect impact via further training. Basic estimates on an individual level do not reveal clearly direct effects, but on an establishment level, the existence of a works council increases the average job satisfaction in a company. In more extended approaches, we also find a positive, weakly significant link on an individual level accompanied by positive training with regard to job satisfaction if we control for personal characteristics, working conditions, firm size, collegiality variables and industry dummies. Firms with industry-wide bargaining agreements drive this result. The effects are stronger if the firm carries the training costs and if the share of trained workers within the firm measures training. The direct impact of works councils remains positive but becomes insignificant if Lewbel's instrumental variables estimator is applied.