Skip to main content
Conspirafy
Technology & SurveillanceConfirmed100% confidenceQuick answer: is it true?

Facebook Emotional Manipulation Experiment

Written by:Conspirafy editorial team· Apr 27, 2026

TL;DR

Facebook secretly altered 689,003 users' news feeds for a week in 2012 to test emotional contagion. Published in PNAS 2014; ethics outcry; no retraction, no compensation. Primary-source confirmed.

Analysis

Claim Map

Core claim

Facebook secretly manipulated the news feeds of 689,003 users for one week in January 2012 to test whether emotional contagion could be achieved through social media. Published in PNAS in 2014 without informed consent from subjects.

Documented fact

The PNAS paper itself

Unsupported inference

Effect size was tiny

Evidence that would change this

This would only weaken if leaked program documents, court filings, and Snowden-era disclosures were authoritatively shown to be fabrications — which has not happened in over a decade of independent journalism.

Current verdict

confirmed, 100% confidence

Evidence Strength Matrix

A compact map of what is documented, where the claim leaps, and what evidence affects the verdict.

Adjacent documented fact
Documented: The PNAS paper itself
Unsupported: The adjacent fact does not by itself prove coordination, motive, scale, or concealment.
Counter-evidence: Effect size was tiny
Verdict impact: Sets the baseline for what is real before broader claims are tested.
Claim mechanism
Documented: Any proposed mechanism must be tied to records, physical evidence, technical limits, or named procedures.
Unsupported: A mechanism remains weak when it depends on inference from coincidence, visual artifacts, or anonymous claims.
Counter-evidence: ToS arguably did cover it
Verdict impact: Determines whether the claim is testable or mainly narrative pattern-matching.
Verdict movement
Documented: This would only weaken if leaked program documents, court filings, and Snowden-era disclosures were authoritatively shown to be fabrications — which has not happened in over a decade of independent journalism.
Unsupported: A claim does not move the verdict by repeating suspicion without new primary evidence.
Counter-evidence: Confirmed by the primary source — the published PNAS paper written by Facebook's own data scientists. Facebook did not dispute the facts, only defended the ethics.
Verdict impact: confirmed, 100% confidence

Claim Family Context

View all families
Synthetic Media and Platform Claims

Internet-era claims built from bots, AI images, screenshots, engagement manipulation, surveillance fears, or algorithmic amplification.

Evidence question: Is there technical provenance, platform data, or forensic analysis, or is the claim based on a screenshot and suspicion?

Rumor Lifecycle

How this claim moves from origin to amplification, record check, verdict, and recurrence.

1
First appearance

2012

2
Amplification

Peaked June 2014 when the PNAS paper was published and immediately retracted amid public backlash, sparking global debate about platform consent and A/B testing ethics.

3
Record check

The PNAS paper itself

4
Verdict boundary

Confirmed by the primary source — the published PNAS paper written by Facebook's own data scientists. Facebook did not dispute the facts, only defended the ethics.

5
Recurrence risk

Often recurs through the synthetic media and platform claims claim family.

Source Health

Source guideURLs complete
Sources
12/12
Credibility Mix
9 high / 3 medium / 0 low
Missing URLs
0
Latest Source Year
2025

Why this page is still being upgraded

This page is below one or more content-quality gates: body depth (273/1200 words). Editors are expanding the narrative, source base, and related reading before marking the page complete.

What would change our verdict

This would only weaken if leaked program documents, court filings, and Snowden-era disclosures were authoritatively shown to be fabrications — which has not happened in over a decade of independent journalism.

1 min readDifficulty: 4/5First emerged: 2012Fact-checked: Apr 2026
Body 273/1200 wordsSources 12/12Freshness Apr 2026, review Oct 2026Evidence 6 supporting / 4 counter
Confirmed
100% confidence / methodology →

Confirmed by the primary source — the published PNAS paper written by Facebook's own data scientists.

8 min read273 wordsUpdated 4 Nov 2026
6 supporting4 debunking12 sources

Also Read

Browse all

What do you think?

Sign in to cast your vote.

What happened

Between January 11-18, 2012, Facebook's data science team — in collaboration with Cornell University researchers — altered the News Feed algorithm for 689,003 users. Half saw fewer posts with positive words; half saw fewer posts with negative words. The researchers measured how users' own subsequent posts shifted in emotional tone.

The findings

Users exposed to more negative content posted slightly more negative content themselves (and vice versa). This was described as evidence of "massive-scale emotional contagion" via the platform.

The exposure

The study was published in PNAS on June 2, 2014. Within hours, researchers and journalists noted that the study violated basic research ethics norms: subjects had not provided informed consent, and the researchers had used the Facebook terms-of-service as a substitute for IRB approval.

The response

PNAS issued an "expression of concern." Facebook's Chief Technology Officer acknowledged the research was "poorly communicated." Facebook did not retract the study nor issue compensation to affected users. The case became foundational in debates about tech-company research ethics and informed consent in the social media era.

Why this is a confirmed conspiracy

Unlike many tech conspiracy theories, this one has a primary source: the peer-reviewed PNAS paper itself, written by Facebook's own employees, explicitly describing what they did. Facebook didn't deny the experiment — they published it.

What is documented, and where the claim outruns it

The published PNAS paper, university statements, privacy investigations, and platform-policy debate establish the experiment and the consent controversy.

The unsupported leap is claiming the study proves every user emotion, vote, or social movement is directly controlled by Facebook or any single platform.

Evidence Filters10

The PNAS paper itself

Supporting

Primary source: Kramer, Guillory, Hancock (2014), published and never retracted.

Effect size was tiny

Debunking

The actual emotional contagion measured was very small — statistically significant but practically modest.

Facebook admissions

Supporting

Facebook CTO acknowledged the research, defending it as ToS-permitted.

ToS arguably did cover it

Debunking

Facebook's terms of service did include a research clause, though broadly and without specific consent.

PNAS expression of concern

Supporting

Editor-in-chief issued concern statement about informed consent.

Not ongoing

Debunking

This specific experiment was one week in 2012 — there's no evidence of ongoing emotional manipulation experiments.

Cornell IRB controversy

Supporting

Cornell IRB had not reviewed the study because data was "pre-collected" by Facebook.

Many feed changes are legitimate

Debunking

Not every A/B test on news feeds is unethical — the specific violation was emotional manipulation without consent.

Adam Kramer apology blog post

Supporting

First author Adam Kramer posted a Facebook note acknowledging the harm — effectively a confession.

UK ICO response

Supporting

UK Information Commissioner's Office investigated and criticized Facebook's approach.

Evidence Cited by Believers6

The PNAS paper itself

Supporting

Primary source: Kramer, Guillory, Hancock (2014), published and never retracted.

Facebook admissions

Supporting

Facebook CTO acknowledged the research, defending it as ToS-permitted.

PNAS expression of concern

Supporting

Editor-in-chief issued concern statement about informed consent.

Cornell IRB controversy

Supporting

Cornell IRB had not reviewed the study because data was "pre-collected" by Facebook.

Adam Kramer apology blog post

Supporting

First author Adam Kramer posted a Facebook note acknowledging the harm — effectively a confession.

UK ICO response

Supporting

UK Information Commissioner's Office investigated and criticized Facebook's approach.

Counter-Evidence4

Effect size was tiny

Debunking

The actual emotional contagion measured was very small — statistically significant but practically modest.

ToS arguably did cover it

Debunking

Facebook's terms of service did include a research clause, though broadly and without specific consent.

Not ongoing

Debunking

This specific experiment was one week in 2012 — there's no evidence of ongoing emotional manipulation experiments.

Many feed changes are legitimate

Debunking

Not every A/B test on news feeds is unethical — the specific violation was emotional manipulation without consent.

Quick Talking Points

  • PNAS issued an "Editorial Expression of Concern" over the study's consent framework.
  • Facebook and Cornell published a 2014 PNAS paper documenting an experiment on 689,003 users' news feeds.
  • Users did not provide specific informed consent for the experiment — only broad ToS coverage.
  • Set precedent for ongoing debate about platform A/B testing ethics and user autonomy.

Timeline

  1. Facebook begins emotional manipulation experiment

  2. Experiment concludes (7 days)

  3. PNAS paper published

  4. Public outcry begins

  5. Facebook CTO acknowledgment

  6. PNAS expression of concern published

Notable Quotes

“The experiment was conducted for one week in 2012 and we never meant to upset anyone. Having looked at this more carefully, we recognize that there are concerns about how we communicated about this study. Those concerns are valid.”

— Sheryl Sandberg, Facebook COO, press statement following PNAS publication backlash, 2014

Verdict

Confirmed100% confidence

Confirmed by the primary source — the published PNAS paper written by Facebook's own data scientists. Facebook did not dispute the facts, only defended the ethics.

What would change our verdicti

This would only weaken if leaked program documents, court filings, and Snowden-era disclosures were authoritatively shown to be fabrications — which has not happened in over a decade of independent journalism.

Frequently Asked Questions

Did Facebook really manipulate users' emotions?

Yes. Facebook's own data scientists published a peer-reviewed paper documenting it.

How many people were affected?

689,003 Facebook users in January 2012.

Were users compensated?

No. Facebook issued no compensation and did not retract the study.

Is Facebook still doing this?

There is no evidence of ongoing emotional-manipulation experiments specifically, but Facebook continues to conduct A/B tests on feed algorithms. The 2012 experiment prompted internal policy changes.

Why does this matter?

It demonstrates that social media platforms can — and have — manipulated user emotions at massive scale without consent. The ethical framework for platform research has been materially reshaped by this case.

Sources

Kramer et al. PNAS paper
PNAS·Apr 2025·Archive
High Credibility
PNAS Expression of Concern
PNAS·Apr 2025·Archive
High Credibility
Adam Kramer's public response
Facebook·Apr 2025·Archive
High Credibility
The Atlantic coverage
The Atlantic·Apr 2025·Archive
High Credibility
NYT coverage
New York Times·Apr 2025·Archive
High Credibility
Show 7 more sources
Guardian analysis of the Facebook emotional contagion ethics controversy
The Guardian·Jul 2014
Medium Credibility
Grimmelmann analysis
James Grimmelmann (University of Maryland Law)·Apr 2025·Archive
High Credibility
UK ICO statement
The Guardian·Apr 2025
High Credibility
Frances Haugen testimony (context)
US Senate·Apr 2025
High Credibility
Wired coverage
Wired·Apr 2025
Medium Credibility
Cornell Chronicle coverage of the Facebook emotional contagion study
Cornell University·Jun 2014·Archive
High Credibility
Facebook reveals news feed experiment to control emotions
The Guardian·Jun 2014
Medium Credibility

Further Reading

In Pop Culture

book2010

The Facebook Effect

David Kirkpatrick

Authorised biography of Facebook's growth, providing context for the corporate culture that enabled the 2012 emotional-contagion experiment to proceed without ethical review.

Conspiracy Glossary·Confirmed Conspiracies·Browse All Theories

Related Theories

Technology & SurveillanceUnsubstantiated
Bitcoin Was Created by the NSA
The claim that Bitcoin was secretly built by the NSA (or another intelligence agency) as a surveillance tool, honeypot, or stepping stone to a controllable cashless currency. Proponents point to two real facts: the NSA designed the SHA-256 hash function Bitcoin uses, and NSA cryptographers co-authored a 1996 paper on anonymous electronic cash. But no evidence connects the agency to Bitcoin's creation, the technology it uses is public and standardized, and the code and history point to an independent cypherpunk lineage.
15 sources87% confidencefully sourced
Technology & SurveillanceUnsubstantiated
AI existential risk / 'AI doomer' debate (2014-present)
The claim that artificial intelligence poses an existential or catastrophic risk to humanity has been advanced by Nick Bostrom ('Superintelligence,' 2014), Stuart Russell ('Human Compatible,' 2019), Eliezer Yudkowsky and the LessWrong rationalist community, and — in more measured form — by Geoffrey Hinton and Yoshua Bengio. The Future of Life Institute open letter (22 March 2023) calling for a six-month pause on large AI experiments attracted over 30,000 signatories. The Center for AI Safety statement (30 May 2023) was signed by Hinton, Bengio, Sam Altman, and Demis Hassabis. The AI Safety Summit at Bletchley Park (1-2 November 2023) convened 28 countries plus the EU. Counter-arguments from Andrew Ng, Yann LeCun, and Melanie Mitchell hold that existential framing overstates near-term risk and distracts from concrete algorithmic harms including bias and labour displacement.
17 sources66% confidencefully sourced
Technology & SurveillancePartially True
Meta/Facebook knowledge of Cambridge Analytica before Mar 2018 (2015-18)
Facebook discovered in December 2015 that Aleksandr Kogan's 'thisisyourdigitallife' app had harvested data from approximately 87 million users and transferred it to SCL Group/Cambridge Analytica. Facebook contacted Cambridge Analytica, obtained certifications of data deletion, and did not notify affected users. Internal documents produced in the 2018 FTC investigation showed Facebook had identified the policy enforcement gap but treated it as resolved by the deletion certifications. Christopher Wylie's whistleblower disclosures to the New York Times and The Guardian (17 March 2018) revealed that the data had not been deleted. Facebook paid a $5 billion FTC settlement (24 July 2019) and a $100 million SEC fine. The claim that Facebook knowingly concealed this data harvest from regulators and users is partially supported by the documentary record.
16 sources78% confidencefully sourced