Heiss, R., & Freiling, I. (2026). Addressing social media platforms’ influence on academic research. Humanit. Soc. Sci. Commun., 13, 192. https://doi.org/10.31235/osf.io/ny2tx_v2
Summary
This commentary argues that academic research on social media is uniquely vulnerable to industry influence because platforms hold exclusive control over the data required to study them. Unlike food, pharmaceutical, or tobacco research—where independent data generation is at least possible—platform research cannot proceed without industry-controlled access to information on algorithms, content flows, and engagement. Drawing an explicit analogy to the “commercial determinants of health” literature, Heiss and Freiling identify four channels through which industry bias can enter the research cycle and warn that these threaten the independence of the evidence used in regulation and policymaking. They advocate a dual response: regulatory reform (exemplified by the EU’s Digital Services Act) paired with stronger ethical norms and “research on research” within the social sciences.
Key Contributions
- Transposes the “commercial determinants of science” framework from health-related fields onto social media platform research.
- Provides a structured four-challenge typology of risks in platform–academic collaborations.
- Offers a concrete policy and community-level agenda: regulated data access, independent funding intermediaries, ethical guidelines, and meta-research on industry-funded studies.
- Frames the DSA as simultaneously a model for and a cautionary tale about researcher data-access regimes.
Methods
A conceptual, analytical commentary rather than an empirical study—no datasets are generated or analyzed. The argument proceeds through comparative analogical reasoning drawing on prior literature on industry influence in pharmaceutical, tobacco, and food research, plus case-based illustration (the Meta–academic 2020 US election partnership, Social Science One, the Chan Zuckerberg Initiative, and Jigsaw). The synthesis is organized around a four-challenge framework mapping each challenge to lessons and recommended actions.
Findings
- Four key challenges in platform–academic collaborations: (1) restrictive data access, (2) selective funding of researchers and topics, (3) difficulty detecting subtle influence, and (4) institutionalization of influence via long-term partnerships.
- Meta’s 2020 election studies illustrate how exclusive data access, undisclosed algorithm changes during the study, and platform framing of results can shape policy-relevant narratives.
- Evidence from other industries shows industry-funded studies systematically report more industry-favorable findings.
- Early implementation of DSA Article 40 shows platforms interpret eligibility narrowly, delay or reject applications, and provide inadequate documentation, limiting its effectiveness.
- Reciprocity dynamics mean even small gifts or access provisions can subconsciously bias researchers, undermining safeguards like preregistration.
Connections
This piece speaks directly to the empirical and legal literature on the DSA and platform data-access regimes, connecting to work examining Article 40 implementation and researcher access frictions Rieder2026-pp, Rieder2025-ju, and to critiques of the reliability and independence of platform-provided data infrastructures Gonzalez-Bailon2024-rq, Efstratiou2025-gs. Its concern with meta-research and the trustworthiness of evidence used in regulation also resonates with methodological audits of platform data quality Bruns2026-yv and with debates over the collapse of independent access after CrowdTangle’s deprecation Bak-Coleman2025-pm.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify · Apple Podcasts