Facebook Hilariously Debunks Princeton Study Saying It Will Lose 80% Of Users
Last week Princeton researchers released a widely covered study saying Facebook would lose 80% of its users by 2015-2017. But now Facebook's data scientists have turned the study's silly “correlation equals causation” …
Context & Ripple Effects
Facebook is answering academic criticism with its own data team rather than a press release. Princeton researchers' widely covered projection of an 80% user collapse by 2015-2017 rested on Google search-volume correlations — the same correlation-equals-causation leap Facebook's scientists now turn around to 'predict' Princeton's own disappearance, in a note by Mike Develin that spread across BGR, Business Insider, the Telegraph and others within days.
The rebuttal lands while the underlying worry hasn't gone away: confirmed reporting just this week has Facebook struggling with declining U.S. teen traffic, even as the stock closed at a record high on January 22 and the company announced tests of an ad network for third-party mobile apps. It echoes a familiar playbook — back in June 2012 Facebook defended itself against skepticism by re-framing the data with its own account of how well its unsold ads performed.
First-order effects
- Facebook gets a public-relations win that blunts the Princeton headline cycle, but it does so by conceding the battleground: the debate over whether teens are leaving shifts from one study's search data to Facebook's own private metrics, which outsiders can't audit.
Second-order effects
- Princeton's researchers face pressure to defend a peer-review-worthy methodology after the satire exposed the search-trend proxy as fragile, raising the bar for any academic group hoping to estimate platform populations without platform cooperation.
Third-order effects
- If the pattern holds, every major platform maintains an in-house data-science operation whose job includes contesting unflattering external research — leaving regulators and academics dependent on either company-disclosed numbers or increasingly contested proxies, with no neutral referee on user counts.
The trend: Platform companies are institutionalizing in-house data teams that publicly rebut external academic studies of their growth and health, making corporate-released metrics the primary evidence in debates over user decline.