Web Privacy Measurement: Early Results, Engineering Challenges and Selected Applications

Report ID: TR-979-14
Author: Eubank, Christian
Date: 2014-06-01
Pages: 50
Download Formats: |PDF|
Abstract:

Web Privacy Measurement experiments have heavily in uenced privacy debates by shedding light into practices such as third party online tracking and price discrimination. However, these research experiments have typically been one-o projects, with di erent groups encountering similar methodological and engineering challenges without forming an institutional knowledge base. As an illustrative example of this trend of repeating e ort through self-contained experiments, we present the results of one study comparing mobile and desktop tracking and discuss how challenges from this study and other works in the literature shaped the formation of general design principles intended to improve the eciency of such experiments. We present a robust and modular web measurement platform that enables scalable and repeatable experiments while avoiding many common pitfalls observed by the research community. We describe case studies performed on this framework, including the detection of unique cookie identi ers, the detection of third-party synchronization of these IDs and an examination into the personalization of the content on news sites based on user history.