The diffusion model is not a deterministic growth model: comment on Jones and Dzhafarov (2014)

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Decision Making Research

Background:

• Current models of speeded decision making, such as the diffusion model, face challenges regarding their empirical testability.
• Jones and Dzhafarov (2014) proposed that these models are special cases of more general, unfalsifiable models.
• This claim implies that the diffusion model's testability relies on artificial restrictions.

Purpose of the Study:

• To refute the claim that the diffusion model is unfalsifiable.
• To clarify the empirical testability of the diffusion model in cognitive tasks.
• To address the misrepresentation of the diffusion model within broader model classes.

Main Methods:

• Analysis of the Jones and Dzhafarov argument concerning model falsifiability.
• Distinguishing between within-trial and across-trial variability in decision-making models.
• Comparing the standard diffusion model with deterministic or near-deterministic growth models.

Main Results:

• Jones and Dzhafarov's argument relies on an expanded definition of diffusion models, including deterministic ones.
• Unfalsifiable models attribute variability to across-trial factors, unlike the standard diffusion model.
• The standard diffusion model primarily uses within-trial variability, with constrained across-trial variability.

Conclusions:

• The diffusion model remains empirically testable and distinct from unfalsifiable deterministic models.
• Jones and Dzhafarov's critique misrepresents and trivializes the diffusion model.
• The study clarifies the nature of variability in cognitive decision-making research.