CoChat
Sign Up
Cited by 892Ethics and Social Impacts of AINames, Identity, and Discrimination ResearchMedia, Gender, and Advertising

Algorithmic Bias? An Empirical Study of Apparent Gender-Based Discrimination in the Display of STEM Career Ads

Open access
Lambrecht, Anja, Tucker, Catherine·2019·Management Science
DOI 10.1287/mnsc.2018.3093
Open this paper in CoChat — free accountOpen PDF

Abstract

We explore data from a field test of how an algorithm delivered ads promoting job opportunities in the science, technology, engineering and math fields. This ad was explicitly intended to be gender neutral in its delivery. Empirically, however, fewer women saw the ad than men. This happened because younger women are a prized demographic and are more expensive to show ads to. An algorithm that simply optimizes cost-effectiveness in ad delivery will deliver ads that were intended to be gender neutral in an apparently discriminatory way, because of crowding out. We show that this empirical regularity extends to other major digital platforms. This paper was accepted by Joshua Gans, business strategy.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1287/mnsc.2018.3093

Cites

26 of 26 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Gender bias can occur without any conscious or unconscious attempt to produce a biased outcome, using only an unbiased algorithm.

    “Researchers have also shown that gender bias can occur without any conscious (or unconscious) attempt to produce a biased outcome -using only an unbiased algorithm (Lambrecht and Tucker 2019).”

    From A strategic framework for artificial intelligence in marketing (2021) · cites this paper as Lambrecht and Tucker 2019

    The Implications section says a cost-optimizing, apparently neutral ad algorithm can produce a gender-imbalanced outcome without deliberate intent. But the authors say they did not observe the actual algorithm, cannot determine whether it was itself biased, and regard the proposed economic explanation as suggestive rather than conclusive. The claim therefore goes beyond the evidence by saying the outcome occurred using an unbiased algorithm. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

  • The collaborative filtering-based recommender system approach can be problematic in terms of fairness and privacy.

    “However, this approach can be problematic in terms of fairness and privacy [13-15, 18, 19].”

    From A Deep Dive into Fairness, Bias, Threats, and Privacy in Recommender Systems: Insights and Future Research (2024) · cites this paper as [19]

    The Introduction and §10 (Implications; page not provided) say that automated ad distribution can produce gender imbalances and apparent discrimination, supporting the fairness concern. The paper examines ad distribution, not a collaborative-filtering recommender system, and the passages do not establish a privacy concern with that approach. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • Even after eliminating all bias from the data, an algorithm's results may still vary across different data groups due to algorithmic bias.

    “Even after eliminating all bias from the data, an algorithm's results may still vary across different data groups due to algorithmic bias [19, 82].”

    From A Deep Dive into Fairness, Bias, Threats, and Privacy in Recommender Systems: Insights and Future Research (2024) · cites this paper as [19]

    The Introduction reports that a gender-neutral STEM ad was shown to over 20% more men than women, and the Implications section says cost-effective ad delivery can produce an apparently discriminatory outcome. The paper does not say that all bias was eliminated from the data; it offers suggestive evidence for an economic cost mechanism and says it cannot determine whether the algorithm itself was biased. Page numbers were unavailable in our copy. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • Lambrecht and Tucker discovered fewer women than men were shown ads about STEM careers even though the ads were gender neutral by design.

    “Lambrecht and Tucker (2019), for example, discovered fewer women than men were shown ads about STEM careers even though the ads were gender neutral by design.”

    From Algorithmic management in a work context (2021) · cites this paper as Lambrecht and Tucker, 2019

  • Bias has been reported in algorithmic advertisement, with opportunities for higher-paying jobs and jobs within science and technology advertised to men more often than to women.

    “Bias has also been reported in algorithmic advertisement, with opportunities for higherpaying jobs and jobs within the field of science and technology advertised to men more often than to women (Datta et al. 2015;Lambrecht and Tucker 2019).”

    From The ethics of algorithms: key problems and solutions (2022) · cites this paper as Lambrecht and Tucker 2019

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 6

  1. Athey, S. and D. Nekipelov (2010). A structural model of sponsored search advertising auctions. In Sixth ad auctions workshop, Volume 15.
  2. Datta, A., M. C. Tschantz, and A. Datta (2015). Automated experiments on ad privacy settings. Proceedings on Privacy Enhancing Technologies 2015 (1), 92-112.
  3. Ghosh, A., R. P. McAfee, K. Papineni, and S. Vassilvitskii (2009). Bidding for representative allocations for display advertising. In WINE, Volume 5929, pp. 208-219. Springer.
  4. Gunderson, M. (1989). Male-female wage differentials and policy responses. Journal of Economic Literature 27 (1), 46-72.
  5. Silverstein, M. J. and K. Sayre (2009). The female economy. Harvard Business Review 87 (9), 46-53.
  6. Sweeney, L. (2013). Discrimination in online ad delivery. ACMQueue 11 (3), 10.

Cited by

Sorted by citations

Related

Authors
Lambrecht, A., & Tucker, C.
Affiliations
Boston College (US)
ORCID
Venue
Management Science
Publisher
Institute for Operations Research and the Management Sciences (INFORMS)
Published
2019-07
Volume
65
Issue
7
Pages
2966-2981
ISSN
  • 0025-1909 (ISSN-L)
  • 1526-5501
DOI
10.1287/mnsc.2018.3093
OpenAlex
W2550925836
Cited by
854
References
26
Citations by year
2026: 108 · 2025: 128 · 2024: 139 · 2023: 163 · 2022: 109 · 2021: 120 · 2020: 73 · 2019: 12 · 2018: 2
FWCI
137.66
Type
journal article
Language
en
Open access
Locations
  • Management Science (publishedVersion)
  • London Business School Research Online (London Business School) (submittedVersion)
  • DSpace@MIT (Massachusetts Institute of Technology) (submittedVersion, cc-by-nc-sa)
Topics
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Social Sciences › Sociology and Political Science › Names, Identity, and Discrimination Research
  • Social Sciences › Gender Studies › Media, Gender, and Advertising
Keywords
algorithmic bias, gender discrimination, online advertising, STEM careers, crowding out, demographic targeting, digital platforms, women in STEM

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review