Decision Field Theory (DFT), developed by Jerome Busemeyer and James Townsend (1993), is a dynamic, stochastic model of multi-attribute, multi-alternative choice. It describes the deliberation process as a sequence of momentary evaluations driven by attention switching among attributes, with preferences evolving over time as a multivariate diffusion process.
The DFT Process
P(t) = preference state vector (one element per alternative)
S = feedback matrix (self-feedback and lateral inhibition)
v(t) = valence vector (momentary evaluation, depends on attended attribute)
At each moment, the decision maker attends to one attribute (stochastically selected with probability proportional to attention weight). The attended attribute produces a valence vector — positive for alternatives that are good on that attribute, negative for poor alternatives. The feedback matrix S introduces distance-dependent inhibition: similar alternatives inhibit each other more than dissimilar alternatives.
Context Effects and Predictions
DFT naturally produces all three major context effects (similarity, attraction, compromise) through the interplay of attention switching, lateral inhibition, and the time course of preference accumulation. It also predicts that context effects should be modulated by deliberation time — short deadlines favor different effects than long deliberation — a prediction that has received empirical support.