Analyzing facial expression decision times: Reaction time distribution matters

Area of Science:

• Cognitive Psychology
• Perception Science
• Human Behavior

Background:

• Facial expression recognition is rapid, often under 1 second.
• Research indicates facial attributes like sex influence expression perception.
• Inconsistencies exist regarding participant sex differences in these perceptions.

Purpose of the Study:

• To clarify participant sex differences in facial expression recognition.
• To investigate the influence of analytical methods on reaction time (RT) data.
• To identify robust methods for analyzing RT data in facial perception studies.

Main Methods:

• Multiverse and distributional analyses were employed.
• Outlier removal methods and specific distribution models were compared.
• The Diffusion Model, ex-Gaussian, ex-Wald, and shifted Wald distributions were considered.

Main Results:

• Participant sex differences in RT were contingent on outlier removal and distribution modeling choices.
• The Face Sex × Expression interaction pattern varied for female participants across stimulus sets.
• Distributional models like ex-Gaussian offer improved RT analysis over mean RTs.

Conclusions:

• Analytical choices significantly impact findings on sex differences in facial perception.
• Distributional models are recommended for more accurate reaction time analyses.
• Understanding these analytical nuances is crucial for reliable conclusions in facial expression research.