Diffusion models split RT into drift, caution, and non-decision time

Correct: B — Drift rate (information quality), boundary separation (caution), and non-decision time (encoding + motor)

Building on Ratcliff's diffusion model, Wagenmakers, van der Maas, and Grasman (2007) introduced an EZ parameter-recovery method that maps the mean RT, variance of RT, and proportion of correct responses onto three psychologically interpretable parameters: drift rate (rate of evidence accumulation toward the correct boundary), boundary separation (how much evidence the participant requires before responding — i.e., speed–accuracy caution), and non-decision time (perceptual encoding plus motor output). The framework lets researchers separate 'she's slow because she's cautious' from 'she's slow because the evidence is noisy' — a distinction that simple mean-RT comparisons routinely conflate.