Quantifying the benefits of using decision models with response time and accuracy data.

Area of Science:

• Cognitive Psychology
• Behavioral Science
• Computational Neuroscience

Background:

• Response time and accuracy are key behavioral measures, but speed-accuracy trade-offs (SATOs) can obscure true participant abilities.
• Traditional analyses often fail to adequately address SATOs, potentially leading to inaccurate conclusions.

Purpose of the Study:

• To demonstrate and quantify the benefits of using decision-modeling approaches, specifically the drift diffusion model (DDM), for analyzing experimental data.
• To compare the sensitivity and specificity of DDM-derived parameters against traditional measures (accuracy, reaction time) for detecting group differences.

Main Methods:

• Simulated experimental data using plausible parameters for the DDM, incorporating both real and null group differences and systematic and null SATOs.
• Applied the DDM to fit the simulated data to recover SATO-unconfounded decision parameters.
• Compared the performance of DDM drift rate against accuracy and reaction time in detecting simulated group differences.

Main Results:

• The DDM provides a principled way to recover decision parameters, effectively addressing SATOs.
• DDM-based analysis allows for significantly more efficient data collection by reducing required sample sizes for a given statistical power.
• Accuracy, when informed by DDM parameters, can be a more sensitive indicator of group differences than reaction time alone.

Conclusions:

• Decision modeling offers a powerful framework for experimentalists to gain deeper insights into cognitive processes.
• Implementing DDM analysis can lead to more efficient experimental designs and more robust conclusions in behavioral science research.
• The study highlights the utility of established decision models for enhancing the rigor and efficiency of empirical research.