synr

synr analyzes synesthesia consistency test data to calculate consistency scores from repeated color-choice responses and to classify participant responses for objective verification of synesthetic experiences.

Key Features:

• Implementation: Implemented as an R package for analysis of synesthesia consistency test data.
• Consistency score calculation: Calculates consistency scores from repeated color-choice responses to quantify within-participant response consistency.
• Participant classification: Classifies participant data as valid or invalid based on response patterns.
• Identification of non-interpretable scores: Identifies participants with limited response variability (e.g., consistently choosing only black or red) using density-based spatial clustering of applications with noise (DBSCAN) combined with a measure of spread in 3D space.
• Validation against human ratings: Data validation procedure demonstrates high correspondence with human ratings and can outperform human raters in scenarios prone to error.

Scientific Applications:

• Objective verification of synesthesia: Provides a standardized, objective method to verify synesthetic experiences from consistency test results.
• Enhanced data validation: Detects and flags non-interpretable or invalid responses to improve the quality of synesthesia datasets.
• Application in psychological research: Applicable to psychological studies that require assessment of consistency and variability in participant responses.
• Comparison with human ratings: Enables systematic comparison between algorithmic validation and human ratings to assess reliability and identify cases where human judgment may be misleading.

Methodology:

Calculates consistency scores from repeated color-choice responses; classifies participant data as valid or invalid based on response patterns; identifies non-interpretable scores using density-based spatial clustering of applications with noise (DBSCAN) together with a measure of spread in 3D space; and compares data validation outcomes with human ratings.