Academics are still waiting to receive key data originally promised in 2018 from Facebook, which has declined to provide some of the data citing privacy fears
In 2018, Facebook announced a partnership to provide data to academics to “help people better understand the broader impact …
BuzzFeed NewsCraig Silverman
Context & Ripple Effects
The promise dates to 2018, when Facebook announced a partnership to hand academics data on the platform's broader impact; by April 2019 it had scoped a version of that deal — access for 60-plus researchers worldwide — but with a window running January 2017 to February 2019 that conveniently excludes the run-up to the 2016 US election and Brexit. This report is the status check six months later: the data still hasn't arrived, and Facebook is now declining to provide some of it outright, citing privacy fears.
What follows in the corpus shows the pattern hardening rather than resolving. The first real deliverable, a dataset of roughly 38M URLs didn't reach researchers until February 2020, two years after the announcement. And when internal emails surfaced in 2021, they showed the misinformation data Facebook had supplied since 2019 actually covered about half of its US users — not all of them, as the company had claimed at the time.
First-order effects
The named academics in the Social Science One-style partnership remain unable to independently study the platform's impact on elections and misinformation, while Facebook alone decides which slices of data clear its privacy bar.
Second-order effects
Every delay shifts the terms in Facebook's favor: the delivered windows exclude the most contested periods (the 2016 election run-up, Brexit), and the eventual releases turn out narrower than advertised — leaving researchers to negotiate scope with the very company under study.
Third-order effects
If the pattern holds, independent audit of major platforms becomes something platforms grant rather than something outsiders obtain — privacy language functioning as the gatekeeping mechanism, and regulators left weighing mandated data access as the only counterweight.
The trend: Platform transparency is converging on a model where the companies under scrutiny control both the pipeline and the scope of their own external research.
Scoop from me: Facebook's partnership to provide platform data to academics is in trouble. Researchers are still waiting to receive key data, funders are getting nervous, and no one knows when the full data sets will be available: https://www.buzzfeednews.com/ ...
Facebook has wasted the time of dozens if not hundreds of academics who study social media, promising them data to study the platform through the Social Science One initiative. Now they seem to be chickening out. https://twitter.com/...
This is a really important story that highlights the fact that social media is becoming MORE opaque to research. What's missing is the recognition that @BuzzFeedNews and other outlets have effectively been pushing for this outcome by their unbalanced reporting on privacy. https:/…
“Stamos' characterization of the scandal as ‘misuse of an API by an academic’ belies the fact that Facebook intentionally designed a system in which a single user was authorized to consent to the release of data belonging to thousands of other users.” https://gizmodo.com/...
Facebook announced this effort at transparency to provide data to researchers on April 9th, 2018... the day before Zuckerberg testified on Capitol Hill. It was just another cynical PR effort by an utterly venal company. https://twitter.com/...
what's missing from this criticism is proper context, IE: a decade of disregard for user privacy and a profound lack of transparency about how our personal data is used, stretching all the way back to beacon in 2007 and including cambridge analytica, free VPN provider Onavo, etc.…
This project is struggling because they have no credible independent leverage, and now are subject to Facebook's whims and internal politics. This is why credible independence matters and why the data governance design deserved a lot more attention. https://medium.com/... https:/…
FB is emphatic that it's committed to Social Science One. It has 20 people working on it and says the delays are because it's hard and they have set a high privacy and security standard.
Efforts like this truly aren't just “academic”—they're important for bringing tech companies into the realm of accountability to the public. No segment of our societies should be above democratic accountability, and right now the information balance tips entirely to big tech. htt…
“I think the only way to feel reasonably confident about this project is if you ignore what's happened over the past 16 months. If you step back and look where we started and where we are it's a pretty big step-down,” a source said. Ppl want it to work but some are losing faith.
Lots of interesting threads circulating after @CraigSilverman 's piece for Buzzfeed dropped about FB's molasses progress in sharing promised data with outside researchers. Here's one about how hard it is as a researcher to study FB. https://twitter.com/...
It might appear “unbalanced” to former Facebook executives, and to (some) researchers who think they and their students should get access to whatever their institutional ethics boards will nod through. Two billion Facebook users might have other ideas https://twitter.com/...
The issues with Social Science One fit into a pattern of problems with Facebook's big transparency initiatives over the past 2.5 years. Delays, walkbacks, incomplete products, lack of disclosure. Think fact checking, Clear History, ads archive etc. pic.twitter.com/9tyOkLpc3u
Since the project on “2018 Italian election” is also mentioned here, I'll take the chance to deliver a quick update on what @mine_2018 team is doing with data and tools provided by the @ssrc_org @SocSciOne @facebook initiative. https://twitter.com/...