Full Description
In our study, we conducted groundtruth methods to compare two datasets–the Perverted-Justice (PJ) dataset and the Instagram Data Donation (IGDD) dataset–to explore the similarities and differences of adult volunteers posing as youth versus real youth when engaging in online sexually risky conversations from linguistic and semantic level.
From the study, we found that on the linguistic level, the PJ victim’s and PJ predator’s language are very similar. However, there is a big difference between language of teens disguised by adults and real teens. We will further explore whether nuance cybergrooming stages will emerge from the conversations involving real teens.
