Levi Kaplan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Levi Kaplan has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
Alan Mislove
Northeastern University, Boston, Massachusetts 02115
1 shared publication- Alan Mislove
Northeastern University, Boston, USA
1 shared publication - Aleksandra Korolova
Princeton University, Princeton, New Jersey, USA
1 shared publication - Catherine Tucker
Massachusetts Institute of Technology (MIT) - Management Science (MS)
1 shared publication - Nico Neumann
Melbourne Business School, University of Melbourne, Carlton, Victoria 3053, Australia
1 shared publication Piotr Sapiezynski
Northeastern University, Boston, Massachusetts 02115
1 shared publication
Latest Publications
- Open Access
- 7 November 2024
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 8 (CSCW2) , 1-31
On the Use of Proxies in Political Ad Targeting
Detailed targeting of advertisements has long been one of the core offerings of online platforms. Unfortunately, malicious advertisers have frequently abused such targeting features, with results that range from violating civil rights laws to driving division, polarization, and even social unrest. Platforms have often attempted to mitigate this behavior by removing targeting attributes deemed problematic, such as inferred political leaning, religion, or ethnicity. In this work, we examine the effectiveness of these mitigations by collecting data from political ads placed on Facebook in the lead up to the 2022 U.S. midterm elections. We show that major political advertisers circumvented these mitigations by targeting proxy attributes: seemingly innocuous targeting criteria that closely correspond to political and racial divides in American society. We introduce novel methods for directly measuring the skew of various targeting criteria to quantify their effectiveness as proxies, and then examine the scale at which those attributes are used. Our findings have crucial implications for the ongoing discussion on the regulation of political advertising and emphasize the urgency for increased transparency.
- 7 November 2024
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 8 (CSCW2) , 1-31
Open AccessOn the Use of Proxies in Political Ad Targeting
Detailed targeting of advertisements has long been one of the core offerings of online platforms. Unfortunately, malicious advertisers have frequently abused such targeting features, with results that range from violating civil rights laws to driving division, polarization, and even social unrest. Platforms have often attempted to mitigate this behavior by removing targeting attributes deemed problematic, such as inferred political leaning, religion, or ethnicity. In this work, we examine the effectiveness of these mitigations by collecting data from political ads placed on Facebook in the lead up to the 2022 U.S. midterm elections. We show that major political advertisers circumvented these mitigations by targeting proxy attributes: seemingly innocuous targeting criteria that closely correspond to political and racial divides in American society. We introduce novel methods for directly measuring the skew of various targeting criteria to quantify their effectiveness as proxies, and then examine the scale at which those attributes are used. Our findings have crucial implications for the ongoing discussion on the regulation of political advertising and emphasize the urgency for increased transparency.
- 1 November 2024
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 70 (11) , 8003-8029
Data Deserts and Black Boxes: The Impact of Socio-Economic Status on Consumer Profiling
Data brokers use black-box methods to profile and segment individuals for ad targeting, often with mixed success. We present evidence from 5 complementary field tests and 15 data brokers that differences in profiling accuracy and coverage for these attributes mainly depend on who is being profiled. Consumers who are better off—for example, those with higher incomes or living in affluent areas—are both more likely to be profiled and more likely to be profiled accurately. Occupational status (white-collar versus blue-collar jobs), race and ethnicity, gender, and household arrangements often affect the accuracy and likelihood of having profile information available, although this varies by country and whether we consider online or offline coverage of profile attributes. Our analyses suggest that successful consumer-background profiling can be linked to the scope of an individual’s digital footprint from how much time they spend online and the number of digital devices they own. Those who come from lower-income backgrounds have a narrower digital footprint, leading to a “data desert” for such individuals. Vendor characteristics, including differences in profiling methods, explain virtually none of the variation in profiling accuracy for our data, but explain variation in the likelihood of who is profiled. Vendor differences due to unique networks and partnerships also affect profiling outcomes indirectly due to differential access to individuals with different backgrounds. We discuss the implications of our findings for policy and marketing practice. This paper was accepted by David Simchi-Levi, marketing. Funding: Financial support from the National Science Foundation [CAREER Award 6923256] and an anonymous panel company is gratefully acknowledged. Supplemental Material: The web appendix and data files are available at https://doi.org/10.1287/mnsc.2023.4979 .
- 1 November 2024
- Published by Institute for Operations Research and the Management Sciences (INFORMS) in Management Science
- Vol. 70 (11) , 8003-8029
Data Deserts and Black Boxes: The Impact of Socio-Economic Status on Consumer Profiling
Data brokers use black-box methods to profile and segment individuals for ad targeting, often with mixed success. We present evidence from 5 complementary field tests and 15 data brokers that differences in profiling accuracy and coverage for these attributes mainly depend on who is being profiled. Consumers who are better off—for example, those with higher incomes or living in affluent areas—are both more likely to be profiled and more likely to be profiled accurately. Occupational status (white-collar versus blue-collar jobs), race and ethnicity, gender, and household arrangements often affect the accuracy and likelihood of having profile information available, although this varies by country and whether we consider online or offline coverage of profile attributes. Our analyses suggest that successful consumer-background profiling can be linked to the scope of an individual’s digital footprint from how much time they spend online and the number of digital devices they own. Those who come from lower-income backgrounds have a narrower digital footprint, leading to a “data desert” for such individuals. Vendor characteristics, including differences in profiling methods, explain virtually none of the variation in profiling accuracy for our data, but explain variation in the likelihood of who is profiled. Vendor differences due to unique networks and partnerships also affect profiling outcomes indirectly due to differential access to individuals with different backgrounds. We discuss the implications of our findings for policy and marketing practice. This paper was accepted by David Simchi-Levi, marketing. Funding: Financial support from the National Science Foundation [CAREER Award 6923256] and an anonymous panel company is gratefully acknowledged. Supplemental Material: The web appendix and data files are available at https://doi.org/10.1287/mnsc.2023.4979 .