Drift Diffusion Modelling of Decisions

Description

Fits choice and full response-time distributions to separate the quality of evidence, caution, bias and non-decision time behind users' decisions.

The drift diffusion model treats a two-choice decision as noisy evidence accumulating toward one of two boundaries. Its parameters have psychological meaning: drift rate (how clear the evidence is), boundary separation (how cautious the user is), starting point (prior bias toward one option) and non-decision time (perception and motor). Fitting the model to accuracy and the full RT distributions, rather than mean RT, disentangles explanations that look identical in averages. For UX, it answers whether a design makes information clearer (higher drift), makes users more careful (wider boundary), or biases them (starting point), which is directly relevant to defaults, warnings and dark-pattern assessment.

Attributes

Loss type
Unmigrated
Effort
Weeks

Steps

  1. Design a two-choice task representing the real decision (e.g. accept/reject a suggestion, safe/phishing).
  2. Collect many trials per participant (ideally 100+ per condition) with precise RT logging.
  3. Clean data by removing implausibly fast responses (e.g. < 200 ms) and extreme slow outliers.
  4. Fit the DDM with established tools (e.g. HDDM, fast-dm, or EZ-diffusion for quick estimates).
  5. Compare parameters across design conditions using hierarchical Bayesian estimation.
  6. Interpret: drift changes imply clearer information, boundary changes imply caution, starting-point shifts imply bias or nudging.

Sources

  • Ratcliff, A theory of memory retrieval (1978), Psychological Review
  • Ratcliff & McKoon, The diffusion decision model: theory and data for two-choice decision tasks (2008), Neural Computation
  • Wiecki, Sofer & Frank, HDDM (2013), Frontiers in Neuroinformatics
  • van der Maas, H., Complex Systems Research in Psychology (2024)

Also known as

  • DDM
  • Diffusion model
  • Sequential sampling models

Categories

  • User Research
  • Cognitive Psychology & Perception