What's the Controversy Over the Facebook Social Contagion Study?

Teresa Torres defends Facebook's 2014 emotional contagion study, published in PNAS, arguing that experimenting on users is standard practice for internet products and that the backlash is overblown. She contends the study was meant to answer a fair question about whether emotional newsfeed content affects users' own posts, and that Facebook's data use policy and the study's likely IRB approval address the consent issue in a technical sense. She acknowledges real ethical gray areas and says companies should be cautious and held to high standards. The piece matters for product teams because it frames how experimentation, user trust, and ethics intersect as data-driven products grow.

01Key takeaways

• Nearly all internet products run continuous experiments on users, so experimentation itself is not inherently controversial.
• Emotional experience affects whether users return, making feed composition a core product and retention concern.
• Terms of service compliance and informed consent are different standards; teams should not treat one as proof of the other.
• Publishing experiment results publicly is rare and valuable for learning, but it invites scrutiny that teams should anticipate.
• Ethical review of experiments should scale with the potential for harm, and companies should hold themselves to high standards as data grows.

02Key sections

Experimentation is how internet products improve

Torres argues every internet company runs experiments on users, from simple before-and-after comparisons to large-scale regression testing. The sophistication varies, but the underlying practice is the same.

What the study found and why it matters to Facebook

The study filtered newsfeed posts by emotional word content and measured changes in users' own status updates. Torres argues that emotional experience drives whether people return, so the findings bear directly on the product.

Informed consent and policy compliance

Torres notes Facebook's data use policy mentioned research, so the company was within its terms, though that falls short of formal informed consent. She points out informed consent is an academic ethics standard, not a legal requirement, and that tech companies rarely seek it for A/B tests.

Whether PNAS should have published it

She examines whether the journal's editorial standards were met, noting the authors' affiliations and a reported IRB approval. She suggests that if approval was missing, publication would be questionable.

Balancing ethics with product learning

Torres argues the reaction exaggerates manipulation, since algorithmic feed selection already happens constantly and the effect was small. She still concedes genuine gray areas and calls for caution and high ethical standards as internet data grows.

03From the post

“A lot of people are up in arms about a research study that was published in the latest issue of the Proceedings of the National Academy of Sciences (PNAS). I’m talking about the Facebook social contagion study. But Facebook didn’t do anything wrong. And neither did PNAS by”

04Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.