EloChoice

EloChoice quantifies perceived characteristics of visual stimuli using Elo-rating-based pairwise comparisons to support quantitative perception research.

Key Features:

• R implementation: Implemented as an R package enabling Elo- and Bradley-Terry-based analyses from pairwise comparison data.
• Pairwise comparison framework: Employs the Elo rating system to update stimulus ratings from pairwise comparisons.
• Statistical outputs: Computes Elo scores, mean Elo (mElo) scores, and Bradley-Terry scores.
• mElo characteristics: Produces mean Elo (mElo) scores that exhibit reduced variability relative to traditional Elo scores and are faster to compute than Bradley-Terry scores.

Scientific Applications:

• Perception studies: Applied to studies of perceived physical attributes, including perceived strength from images of male bodies.
• Correlation with objective measures: EloChoice scores have shown strong correlations with grip strength and body size.
• Sample-size efficiency: mElo's reduced variability supports comparable results with fewer participants than Likert rating tasks.
• Response-time findings: Pairwise comparison tasks using EloChoice produced quicker participant response times compared to traditional rating tasks.

Methodology:

Implements Elo rating updates from pairwise comparisons and computes Elo, mean Elo (mElo), and Bradley-Terry scores within R.