A test of the diffusion model explanation for the worst performance rule using preregistration and blinding

Area of Science:

• Cognitive Psychology
• Psychometrics
• Computational Neuroscience

Background:

• Higher IQ scores correlate with better performance on elementary cognitive tasks.
• The worst performance rule (WPR) posits this IQ-task association is strongest with slowest responses.
• The Ratcliff diffusion model previously explained WPR via a shared 'drift rate' parameter.

Purpose of the Study:

• To test four predictions derived from the WPR's diffusion model explanation.
• To evaluate the robustness and ubiquity of the worst performance rule.

Main Methods:

• A diffusion model was fitted to data from 916 participants on a perceptual task.
• A working memory capacity test score was initially shuffled to prevent bias during model fitting.
• Confirmatory Bayesian hypothesis tests evaluated predictions after unshuffling the data.

Main Results:

• The study found evidence contradicting the worst performance rule.
• The diffusion model's explanation for the WPR was not supported by the data.
• The WPR may be less robust and ubiquitous than previously believed.

Conclusions:

• The worst performance rule's validity is questioned.
• The findings necessitate a re-evaluation of the WPR's role in cognitive research.
• Alternative explanations for the IQ-task performance relationship may be required.