What Lies Behind the Answers: Applying Machine Learning to Understand Tenants’ Reaction to Eviction

Background: Eviction is one of the most pressing challenges facing millions of households today, particularly in urban areas, as filings have rebounded to or even surpassed pre-pandemic levels. While nonpayment of rent is frequently cited as the primary reason for eviction, the specific subcategories of nonpayment have not been thoroughly examined. Methods: This study analyzes eviction court files from 2014 to 2017, obtained from Pierce County, Washington, and applies advanced data science techniques, including layout analysis and natural language processing, to investigate renters’ responses to summonses and their stated reasons related to nonpayment. Results: The findings reveal that the most frequently mentioned responses include “payment plan or rent repayment request,” and “demand for hearing or court review.” Furthermore, among cases that provide explanations for eviction, the most commonly cited subcategories related to nonpayment were “repair problems or unlivable condition,” “job loss or reduced work income,” and “medical hardship or disability.” Conclusion: Based on insights from the answer filings, we propose interventions such as mediation and tenant rights education. Finally, we discuss the study’s limitations and suggest directions for future research.