The purpose of this paper is twofold. First, using data on Belgian railroad operations, we provide the first application of hedonic output aggregation to the railroad industry. Second, we comapre the traditional homogeneous output approach with the use of these hedonic aggregates and carefully evaluate differences in estimates of input substitution possibilities, returns to scale, and productivity growth. It is found that ignoring the role of operating characteristics in cost analyses implies substantial bias in estimates of railroad technology.