<p>Meta-frontier analysis (MFA) is widely used across various fields to assess performance and technology gaps. This paper explores MFA within a two-stage framework using a multiplier-formed data envelopment analysis (DEA) model. Our findings reveal that in the two-stage DEA model, the stage efficiency relative to the group frontier may be lower than that relative to the meta-frontier, thereby violating the basic MFA property and resulting in the meta-technology ratios (MTRs) of stages greater than 1. Then, we discuss theoretical conditions under which MTRs of stages could exceed unity and when they remain no greater than unity. To ensure the basic MFA property and also maintain fairness for evaluating stage efficiency relative to group frontiers, we introduce the concept of fairness degree. Building upon this, we propose a modified two-stage DEA model relative to group-frontiers, accompanied by a bisection algorithm. To delve deeper into the sources of inefficiency, we define the revised MTR and inefficiency indexes for both system and stages, respectively, and decompose them within the two-stage network structure. A numerical example and an empirical example strongly support the practicality and effectiveness of the proposed MFA approach.</p>