Related Experiment Videos
Analysis of group differences in processing speed: where are the models of processing?
Roger Ratcliff1, Daniel Spieler, Gail McKoon
1Department of Psychology, Ohio State University, Columbus, OH 43210, USA.
Psychonomic Bulletin & Review
|December 8, 2004
Summary
Brinley functions are quantile-quantile plots, not just reaction time distributions. Their slope interpretation for processing speed is model-dependent, and speed-accuracy effects partially explain slopes greater than 1.
Area of Science:
- Cognitive Psychology
- Psychometrics
- Human Aging Research
Background:
- This study addresses a debate regarding the interpretation of Brinley functions, specifically concerning their relationship with reaction time (RT) distributions and processing speed.
- Myerson et al. (2003) contested previous interpretations, focusing on methodological and terminological aspects of Brinley function analysis.
Discussion:
- The authors clarify that Brinley functions are indeed quantile-quantile (QQ) plots of mean RT distributions.
- They argue that the slope of a Brinley function is best estimated by the ratio of standard deviations due to variability in RT distributions.
- The interpretation of Brinley function slopes as a direct measure of processing speed is presented as model-dependent.
Key Insights:
- Brinley functions are quantile-quantile plots, not merely distributions of mean reaction times.
- The ratio of standard deviations is the appropriate method for estimating Brinley function slopes.
- Speed-accuracy trade-offs can explain Brinley slopes exceeding 1 in certain experimental contexts, but are not the sole factor.
Outlook:
- Future research should prioritize model-based explanations of cognitive processing that incorporate both correct and error reaction time distributions, alongside accuracy.
- A comprehensive understanding of cognitive aging requires analyses that account for the full spectrum of response data, not just mean RTs.