Evaluation of innovative projects is essential for scientific and technological advancement. While these projects require assessment of multiple dimensions—including impact, novelty, and feasibility—the cognitive demands of joint evaluation may compromise thorough assessment of implementation challenges. We investigate how the evaluation format affects evaluators’ identification of feasibility issues through a field experiment at a leading research university. The experiment randomly assigned evaluators of grant proposals to either a multi-criteria assessment (impact, novelty, and feasibility) or a single-criterion, feasibility-only review. Feasibility-only evaluators assigned systematically lower feasibility scores and identified more potential implementation challenges than their multi-criteria counterparts. Analysis of evaluators’ qualitative comments using a human-large language model (LLM) approach reveals that focusing attention on a single criterion leads to more comprehensive assessments of implementation challenges. These findings suggest that focused feasibility reviews could be particularly valuable for high-risk, novel projects where early identification of feasibility issues is crucial.
