Modeling response time with power-law distributions

TL;DR

This paper reviews models assuming power-law distributions to describe response time data, highlighting their properties, implications, and advantages for group comparisons in cognitive science.

Contribution

It provides a tutorial overview of power-law models for response times, demonstrating their application in comparing groups such as children with and without dyslexia.

Findings

Power-law models effectively capture skewed heavy tails in response time data.

Larger sample sizes improve the accuracy of distribution analysis.

Power-law assumptions facilitate group comparisons in cognitive studies.

Abstract

Understanding the properties of response time distributions is a long-standing problem in cognitive science. We provide a tutorial overview of several contemporary models that assume power law scaling is a plausible description of the skewed heavy tails that are typically expressed in response time distributions. We discuss several properties and markers of these distribution functions that have implications for cognitive and neurophysiological organization supporting a given cognitive activity. We illustrate how a power law assumption suggests that collecting larger samples, and combining individual subjects' data into a single set for a distribution-function analysis allows for a better comparison of a group of interest to a control group. We demonstrate our techniques in contrasts of response time measurements of children with and without dyslexia.