This research sought to improve precision in sex trafficking victim identification and investigations.
The study first examined whether there are indicators that differentiate online escort ads related to sex trafficking from ads for non-trafficked sex work, and then it determined which indicators are most likely to predict whether the ad is a case of sex trafficking. Recommendations are made about how and when online escort ads are most useful in identifying trafficking cases. The key contribution of this work is that the indicators found to be predictive were tested against a counterfactual (ads known not to be related to trafficking). indicators were tested based on previous literature and three sets of focus groups: sex-trafficking survivors, non-trafficked sex workers, and criminal justice/victim advocate professionals. Language indicators associated with a higher risk of human sex trafficking cases were trustworthy provider language, obscured phone number (one that uses techniques to avoid collection by technology or web scraping) mention of provider ethnicity, and language suggesting the youth of the provider (survivor). Traffickers may focus more on racialized descriptions in their marketing of victims than non-trafficked sex workers.