APS
2026 APS Annual Convention · 2026
Applying Natural Language Processing to Understand Disengagement In Text-Based Crisis Interventions
- Pablo Rodriguez-Brazzarola
Crisis Text Line - Ignacio Tripodi
Crisis Text Line - Elizabeth Olson
Crisis Text Line - Kelly Zuromski
Crisis Text Line
Abstract
We developed a natural language processing pipeline to analyze and predict disengagement in text-based crisis conversations. Results show that disengagement is common and can be predicted by features such as slow response times, and short messages from the texter. Future work will focus on understanding different reasons for texter disengagement.
