Misinterpreting Moderation in the Age of AI Summarization


Project status in progress
Project time since 13.01.2026
Keywords moderation; interaction effects; conditional effects; research communication; AI summarization; large language models; meta-science

Many of marketing's most valuable findings hold only under conditions, and those conditions tend to get lost on the way from the analysis to the decision. Our project examines where this loss occurs: in how empirical marketing research itself handles conditional effects, and in how findings are relayed to readers at a time when machines increasingly do the reading. What is at stake is the reliability of the advice the field gives, because a conditional effect read as a general one points managers in the wrong direction. The project documents the problem empirically on both sides and derives implications for authors, reviewers, and readers.