Rape narratives analysis through natural language processing: Survivor self-label, narrative time span, faith, and rape terminology

Objective: Rape-survivor identity is a sign of recovery and positive therapeutic progress among rape victims. This study is one of the few to focus on factors predicting self-labeling as a survivor among self-acknowledged rape victims by evaluating the contribution of the time span of the rape narrative, rape terminology, and expression of faith. Method: The data were elicited from 3,794 rape narratives published on Brave Miss World, a website established, inter alia, to allow rape victims to share their stories. To identify the key concepts within rape narratives, we performed text mining based on natural language processing (NLP) and developed a model. Structural equation modeling (SEM) was used to test the model's goodness-of-fit analysis. Results: The results indicated a good fit between the theoretical model and the data. Time predicted expressions of faith and the use of rape terminology, and there was a positive correlation between faith and the use of rape terminology. Both expressions of faith and the use of rape terminology mediated between time span and self-labeling as a survivor, with a significant positive effect on self-labeling as a survivor. Conclusions: The results suggest that addressing rapevictim faith and the use of rape terminology in the depiction of their sexual trauma may be beneficial for the therapeutic process and postrape recovery.