Extending multinomial processing tree models to measure the relative speed of cognitive processes

Area of Science:

• Cognitive Psychology
• Mathematical Psychology
• Psychometric Modeling

Background:

• Multinomial processing tree (MPT) models explain categorical response data through underlying cognitive processes.
• Integrating response times (RTs) into MPT models is crucial for understanding the speed of these processes.
• Existing methods often require strong assumptions about the shape of latent RT distributions.

Purpose of the Study:

• To propose a general, distribution-free method for incorporating RTs into any MPT model.
• To enable the measurement of the relative speed of hypothesized cognitive processes within MPT frameworks.
• To extend the applicability of MPT models by analyzing both response categories and RTs.

Main Methods:

• A novel approach is presented to include RT data within existing MPT models.
• Observed RT distributions are treated as mixtures of latent RT distributions corresponding to different processing paths.
• RTs are incorporated in a distribution-free manner by binning individual response times into categories (fast to slow).

Main Results:

• The proposed method yields RT-extended MPT models that can be analyzed using existing MPT statistical tools and software.
• This approach allows for the estimation of the relative speed of distinct cognitive processes.
• The method was successfully demonstrated using the two-high-threshold model of recognition memory.

Conclusions:

• The developed method provides a flexible and powerful way to extend MPT models with RT information.
• This integration allows for a more comprehensive understanding of cognitive processes by considering both accuracy and speed.
• The RT-extended MPT models offer enhanced insights into cognitive architecture and performance.