Localizing structure in individual differences: A visual illusion case study

Area of Science:

• Cognitive Psychology
• Visual Perception
• Psychophysics

Background:

• Previous research suggests weak correlations in susceptibility to different visual illusions.
• The extent to which these low correlations reflect true relationships versus measurement error is unclear.

Purpose of the Study:

• To investigate the relationship between susceptibility to different visual illusions.
• To assess the impact of measurement error on observed correlations.
• To identify potential common factors underlying illusion susceptibility.

Main Methods:

• Collected data from 149 participants on two variants of five visual illusions using an adjustment paradigm.
• Developed novel data visualizations and hierarchical models to assess measurement error.
• Employed Bayesian hierarchical analysis and latent variable analysis.

Main Results:

• Demonstrated low trial-noise and high reliability (≈0.93) with strong between-subject variability.
• Observed low correlations (≈0.22 ± 0.07) across different illusion susceptibilities.
• Latent variable analysis identified a common factor explaining 23.3% of the variance in illusion susceptibility, with minimal attenuation from measurement error.

Conclusions:

• Susceptibility to one visual illusion does not strongly predict susceptibility to others.
• A shared latent factor contributes to individual differences in susceptibility across various illusions.
• Measurement error minimally attenuates the observed low correlations between illusion susceptibilities.