European researchers want to study the harms of social media but can’t get the data

Iamnitchi is one of several disinformation researchers who told WIRED that while they have a legal right to platform data under EU law, they face hurdles, confusion, and, in some cases, court battles to obtain it. Last year, the law was expanded to increase researchers’ access, but it’s the implementation, not the scope of the law itself, that researchers say is the problem.

In recent years, social media companies have shut down public access tools e.g dead CrowdTangle and replace it With content librariesor, like X, put a paywall on it Data APIsIt forces academics to pay “hundreds of dollars a month,” said Duncan Allen, a research officer at Democracy Reporting International (DRI) in Germany. Without access to an API, what Emenichi describes as “black holes” in what the community knows about how these platforms recommend content or handle reports of sensitive content will only grow, researchers say.

Others, like TikTok, limit the number of posts any given research account can pull each day, which Allen said can make it “impossible to study anything on a large scale.”

Daily subsistence allowance This problem was supposed to be solved by allowing vetted researchers at credible institutions access to API data if they could demonstrate that it would help study systemic risks, from illegal content to threats to fundamental rights. But two years after the law took effect, researchers say they are struggling to meet its security requirements and face platforms’ narrow interpretations of what counts as systemic risk.

Application forms vary by platform, but most require data to be stored on untouchable infrastructure — such as a device physically disconnected from the Internet — a resource that most universities lack, Yamanichi said.

Even for researchers who get consent, “there is no guarantee that the data is good,” said LK Seling, the study’s coordinator. DSA40 Cooperativea German initiative tracking 46 web-based digital accessibility applications. It is often difficult for a colleague to reproduce API data, so a researcher’s work cannot be checked for errors, which Emenichi described as a “fundamental requirement of science.”

DSA40 data It shows that of the 46 requests tracked, 20 were approved and 14 were rejected. But approval rates vary widely: TikTok approved 11 of 13 requests, while X rejected 11 of 23. The true rejection rate is likely higher because the tracker relies on voluntary reporting, Selling said.

“There is no structured advantage for researchers to use this route,” Seeling said. “Data access as it is set up now tries to discourage researchers.”

A TikTok spokesperson told WIRED that the company has given more than 1,500 research teams access to its tools, approved 130 apps in the EU in the second half of last year, and that its daily quota of 1,000 API requests allows researchers to pull up to 100,000 video recordings and comments per day, or up to 2 million follower records. TikTok said it considers its search tools to be DSA compliant but “welcomes more public guidance.”

A Meta spokesperson said CrowdTangle covers only a portion of the company’s public data, while the meta-content library and API it replaces are “the most comprehensive research tools to date.” It covers Facebook, Instagram, WhatsApp, and threads with “strong privacy protections,” and qualified nonprofit researchers, including journalists, can apply.