Abstract: Typically, the meta-frontier network data envelopment analysis model can be used to evaluate the technological gaps in individual stages simultaneously. However, the technological gap in each stage could also have a biased effect, i.e. one which improves the productivity of a subset of input or output factors. Since decision-making units may face different operational technologies, and have two-stage network structures, this study develops a new approach to investigate the favored direction of technology shift. The novelty of this approach is to identify that the technological gaps could affect the overall production function, both with the input technological bias in the first stage and with the output technological bias in the second stage. By comparing the difference between meta-technology and group technology, the favored direction of technology shift for individual decision-making units can be judged. The proposed approach is illustrated using data from 109 Taiwanese tourist hotels in 2015.