Modeling the interaction of numerosity and perceptual variables with the diffusion model

Inhan Kang, Roger Ratcliff

Research output: Contribution to journalArticlepeer-review


Ratcliff and McKoon (2018) proposed integrated diffusion models for numerosity judgments in which a numerosity representation provides evidence used to drive the decision process. We extend this modeling framework to examine the interaction of non-numeric perceptual variables with numerosity by assuming that drift rate and non-decision time are functions of those variables. Four experiments were conducted with two different types of stimuli: a single array of intermingled blue and yellow dots in which both numerosity and dot area vary over trials and two side-by-side arrays of dots in which numerosity, dot area, and convex hull vary over trials. The tasks were to decide whether there were more blue or yellow dots (two experiments), more dots on which side, or which dots have a larger total area. Development of models started from the principled models in Ratcliff and McKoon (2018) and became somewhat ad hoc as we attempted to capture unexpected patterns induced by the conflict between numerosity and perceptual variables. In the three tasks involving numerosity judgments, the effects of the non-numeric variables were moderated by the number of dots. Under a high conflict, judgments were dominated by perceptual variables and produced an unexpected shift in the leading edge of the reaction time (RT) distributions. Although the resulting models were able to predict most of the accuracy and RT patterns, the models were not able to completely capture this shift in the RT distributions. However, when subjects judged area, numerosity affected perceptual judgments but there was no leading edge effect. Based on the results, it appears that the integrated diffusion models provide an effective framework to study the role of numerical and perceptual variables in numerosity tasks and their context-dependency.

Original languageEnglish
Article number101288
JournalCognitive Psychology
Publication statusPublished - 2020 Aug

Bibliographical note

Funding Information:
This work was supported by the National Institutes of Health grants R01-AG041176 and R56-AG057841-01 .

Publisher Copyright:
© 2020 Elsevier Inc.

All Science Journal Classification (ASJC) codes

  • Neuropsychology and Physiological Psychology
  • Experimental and Cognitive Psychology
  • Developmental and Educational Psychology
  • Linguistics and Language
  • Artificial Intelligence


Dive into the research topics of 'Modeling the interaction of numerosity and perceptual variables with the diffusion model'. Together they form a unique fingerprint.

Cite this