In this study, performances of Iranian thermal power plants have been assessed with respect to human factors such as job stress, satisfaction and security through data envelopment analysis (DEA). In the first step performance of power plants (DMUs) have been evaluated considering traditional production factors including installed capacity, internal consumption and fuel consumption as inputs and gross production as output (model 1). In the next step, model 1 is expanded to include three above mentioned human factors (model 2). Power plants have been ranked based on their relative efficiency scores in both models. Pearson and Spearman correlation coefficients between relative efficiency scores and rankings of power plants in two models have been calculated. The results show that efficiency scores and rankings of power plants in two models are quite different indicating that mentioned human factors have direct impact on overall performance of generation companies and that if we were to use just traditional factors it would not reflect the optimum efficiency of the DMUs. Finally, job security has been found as the most effective human factor in Iranian power plants. This is the first study that presents an integrated approach for assessment of electric power plants considering job stress, security and satisfaction. And it is shown that these human factors play important role in overall performance of generation companies. (C) 2014 Elsevier Ltd. All rights reserved.