Quantitative Analysis of Narrative Reports of Psychedelic Drugs - Matthew Baggott

Psychedelic experience reports from Erowid are analyzed using a bag-of-words approach and random forest classification to identify specific substances. This methodology demonstrates that as few as 100 characteristic words can accurately predict which substance a user ingested based on their written narrative. The classification results reveal distinct linguistic differences between substances, suggesting that user narratives capture genuine, substance-specific phenomenological patterns rather than solely subjective or noisy accounts. These findings highlight how large-scale text analysis can serve as a valuable, low-cost research tool alongside traditional clinical studies for exploring human psychedelic experiences.