Individual differences on speeded cognitive tasks: comment on Chen, Hale, and Myerson (2007)

Area of Science:

• Cognitive psychology
• Psychometrics
• Computational neuroscience

Background:

• The difference engine model posits a relationship between standard deviation (SD) and mean reaction time (RT).
• Chen et al. (2007) tested this model, evaluating if SD is proportional to processing amount (mean RT) in speeded tasks.
• This evaluation aimed to validate the model's core assumption regarding processing limitations.

Discussion:

• The evaluation by Chen et al. (2007) does not constitute a valid test of the difference engine model.
• The observed correlation between SD and mean RT is an artifact of data structure, not specific model properties.
• Any model with increasing RT as a function of difficulty and individual slopes will yield this correlation.

Key Insights:

• The relationship between standard deviation (SD) and mean reaction time (RT) is not exclusive to the difference engine model.
• A correlation between SD and mean RT is a necessary outcome for models exhibiting increasing RT with difficulty and varying individual slopes.
• The diffusion model (Ratcliff, 1978) also produces this correlation, demonstrating its general nature.

Outlook:

• Re-evaluation of existing models is necessary, considering the generalizability of observed correlations.
• Future research should focus on developing more specific tests to differentiate between cognitive models.
• Understanding the underlying data structures is crucial for accurate model testing in cognitive tasks.