Elizabeth Letourneau | Searcharxiv

arXiv2026

Deterring Searches for Child Sexual Abuse Material on Google Search and Promoting Help-Seeking

Google Search deploys a "Onebox" feature at the top of the results page when users conduct searches for Child Sexual Abuse Material. This study evaluates the impact of a strategic shift in this feature, comparing a revised intervention, focused on repercussions and therapeutic resources, to a previous iteration that focused on reporting. Using a difference-in-differences analysis of internal Google Search logs data, we found the new messaging resulted in a 3.8 percentage point reduction as compared to the status quo in subsequent CSAM-related queries within the same Search session. We found an average click through rate of 0.73% on any of the hyperlinked buttons to help-providing resources. Together, this research presents convergent evidence that a subset of individuals can be deterred from ongoing CSAM-seeking and redirected to therapeutic services.

cs.HC↗

arXiv2010

Changing approaches of prosecutors towards juvenile repeated sex-offenders: A Bayesian evaluation

Existing state-wide data bases on prosecutors' decisions about juvenile offenders are important, yet often un-explored resources for understanding changes in patterns of judicial decisions over time. We investigate the extent and nature of change in judicial behavior toward juveniles following the enactment of a new set of mandatory registration policies between 1992 and 1996 via analyzing the data on prosecutors' decisions of moving forward for youths repeatedly charged with sexual violence in South Carolina. To analyze this longitudinal binary data, we use a random effects logistic regression model via incorporating an unknown change-point year. For convenient physical interpretation, our models allow the proportional odds interpretation of effects of the explanatory variables and the change-point year with and without conditioning on the youth-specific random effects. As a consequence, the effects of the unknown change-point year and other factors can be interpreted as changes in both within youth and population averaged odds of moving forward. Using a Bayesian paradigm, we consider various prior opinions about the unknown year of the change in the pattern of prosecutors' decision. Based on the available data, we make posteriori conclusions about whether a change-point has occurred between 1992 and 1996 (inclusive), evaluate the degree of confidence about the year of change-point, estimate the magnitude of the effects of the change-point and other factors, and investigate other provocative questions about patterns of prosecutors' decisions over time.

stat.AP↗