All models in data envelopment analysis (DEA) have been built on the foundation of performance factors. Performance factors in DEA are divided conventionally into the input and output measures. In some positions, we confront with dual-role factors which can play simultaneously input and output roles. Traditionally, all performance factors are considered as precise values, while in some real-world problems they characterized as imprecise values. In this paper, we evaluate the performance of 18 third-party reverse logistics (3PL) providers in the presence of dual-role factors and under uncertainty. We illustrate the superiority of the employed DEA approach over a suggested approach in the literature.