Séminaire LICAÉ : Christophe Blaison

Drawing from Kurt Lewin’s field theory, prior work on spatialized affective judgments (e.g., pleasantness ratings) showed they follow a complex "affective field," where influence from hotspots dissipates in a psychological gradient (Blaison, 2022; Blaison & Hess, 2016). A key gap remained: do people use this complex computation for active decision-making, or do they revert to simpler pure distance-based (geometric) heuristics?
We used a computational modeling approach to arbitrate between four competing models (focal-geometric, global-geometric, focal-psychological, and global-psychological). We tested these models against human behavior in two online experiments. Participants chose the "most pleasant" or "most unpleasant" location on 60 2D maps featuring only negative hotspots (“dangerous neighborhoods,” Study 1, N=51) or only positive hotspots ("parks," Study 2, N=50).
Results decisively showed that psychological models, which compute a gradient of influence, significantly outperformed all geometric models. Furthermore, we found people flexibly adapt their strategy based on goals. For both positive and negative hotspots, the task of selecting the "most unpleasant" location was best fit by models with a steep influence gradient, while selecting the "most pleasant" location was fit by a smoother gradient. Thus, the affective field is a robust, flexible mechanism for spatial decision-making, not a mere epiphenomenon of the judgment tasks used in prior research.