Rethinking negotiation research through large(r) data sets

Negotiation research is increasingly moving beyond small, tightly controlled laboratory simulations toward larger and more diverse data sources. Here, we review research on negotiations that uses large datasets and show how utilizing such datasets can advance contemporary negotiation research. We define large(r) data broadly by volume (i.e., number of observations), variety (i.e., different data modalities), and velocity (i.e., temporal density of observations). Our review is organized around four shifts for negotiation research through such data. First, platform data can make impasses more visible as a central negotiation outcome and thereby contribute to understanding negotiation dynamics beyond classic findings, which may have been biased towards agreements. Second, high-density process data, including response times, conversational turns, pauses, gaze, and physiological signals, allow researchers to analyze negotiation as an unfolding interpersonal process. Third, large field, administrative, market, and crowdsourced data sets can reveal heterogeneity and boundary conditions, clarifying which negotiation strategies work for whom, under which social and institutional conditions, and at what cost. Fourth, AI-generated negotiation data can create synthetic spaces for exploring tactics, agent characteristics, and AI-AI or human-AI negotiation dynamics. Finally, and based on the discussion of these developments, we argue that larger data are not inherently better data. Their value depends on whether they capture the outcome, process, context, or mechanism of theoretical interest. The future of negotiation research will therefore benefit most from combining large-scale evidence with theory-driven designs that preserve psychological validity.