Sequential processes and the shapes of reaction time distributions

Area of Science:

• Cognitive Psychology
• Psychometrics
• Experimental Psychology

Background:

• Reaction time (RT) distribution shape is often assumed to be invariant across experimental conditions.
• This assumption simplifies data analysis but may not reflect underlying cognitive processes.

Purpose of the Study:

• To demonstrate that reaction time (RT) distributions are unlikely to have the same shape across conditions.
• To challenge the assumption of shape invariance in RT data analysis.
• To highlight the importance of considering selective influences on sequential processing stages.

Main Methods:

• Theoretical analysis of RT as a sum of independent sequential stages.
• Empirical demonstration using flash-detection experiment RT data.
• Review of studies fitting ex-Gaussian distributions to RT data.

Main Results:

• RT distributions are highly unlikely to share the same shape when RT comprises multiple stochastically independent stages.
• Selective influences on individual stages lead to substantial shape differences in RT distributions.
• Most RT datasets reviewed do not meet even weak requirements for shape invariance.

Conclusions:

• The assumption of RT distribution shape invariance is often violated.
• Selective influences on sequential processing stages are critical for understanding RT variability.
• Methodologies should account for potential shape differences in RT distributions across conditions.