The criminal career evolution of child exploitation websites: identification, survival, and community

The distribution of child sexual exploitation (CE) material has been transformed by the emergence of the Internet. Efforts to combat distribution have been hindered by the prevalence and graphic nature of the material. One way to aid combating is to use automated data collection techniques to scan websites for CE-related criteria. Another is to contribute to proactive combat strategies by developing a theoretical framework to explain the evolution of CE distribution. Within this dissertation I develop a custom-designed webcrawler to collect data on hyperlinked networks containing CE websites and compare them to non-CE website networks. I then begin to develop a theoretical framework based on the criminal career paradigm and social network analysis to explain the evolution of website entities. Through the first study, I assess the effectiveness of a police CE-images database and 82 CE-related keywords at distinguishing websites within 10 CE-based networks from 10 sexuality and 10 sports networks. In the second study, I use a repeated measures design to compare baseline survival rates across the 30 collected networks. I then conduct Cox regression models, using criminal career dimensions adapted to website characteristics, to predict failure in CE-based networks. In the third study, I use the faction analysis to explore the formation of communities within CE-seeded networks and the characteristics that bind those communities. Results show that a) automated data collection tools can be effective, provided that the appropriate inclusion criteria is selected; b) a modified criminal career framework can be applied to CE websites, and their surrounding networks, to explain their evolution; c) individual-based criminal career dimensions can be transitioned to entity-based offenders (websites); d) websites within CE-seeded networks differ from non-CE-seeded networks in composition, survival, and network structure. The findings in this dissertation have implications for law enforcement strategies, private data-hosting services, CE researchers, and criminologists. Future research will refine inclusion criteria, expand to the Deep Web, and continue to develop an online criminal career framework.