Antigone Davis, Facebook's global head of safety, is scheduled to testify in front of the Senate Commerce Subcommittee on Thursday. The company is clearly worried about how it's going to go. Worried enough, in fact, that it published two slide decks showing internal research on how teens experience Instagram, some of which informed the Wall Street Journal's Facebook Files investigation.
- The first deck, entitled Hard Life Moments—Mental Health Deep Dive, is available here.
- The second, entitled Teen Mental Health Deep Dive, is here.
Along with the internal presentations, Facebook also published annotations of each slide in the two decks. There are two clear overarching takeaways, from Facebook's point of view: The WSJ overstated the extent to which teen girls' mental health is affected by Instagram, and Facebook's research isn't very useful. Facebook's annotations call out mistakes in graphs, reframe arguments and take pains to note that "the methods employed in the study are perceptual in nature and not suitable for inferring how Instagram impacts these changes." Facebook argues repeatedly that these were internal studies, presented with context and additional information, and that to read them as proven fact is to misread them.
The studies Facebook's researchers did, the company said in its annotations, were designed to be directionally useful to product teams rather than scientifically credible for public consumption. "Moreover," the company wrote, "all results are based entirely on the perceptions of participants and are not designed to evaluate causal claims between Instagram and health/well-being." And there are the sheer lawyer-isms: "'Mental health' should not be mistaken for a clinical, formal or academic definition," one slide's annotation reads.
Of course, Facebook didn't provide data to the contrary, opting to just throw cold water on the public information rather than offer anything better. The company said that the survey data included only those users who had reported having hard moments or mental health issues, and thus, over-reports most results, compared to a study of the overall Instagram population. But Facebook didn't dispute the experience that those users are having on the platform. "We invest in this research to proactively identify where we can improve — which is why the worst possible results are highlighted in the internal slides," head of research Pratiti Raychoudhury wrote. "That's why the most important thing about this research is what we've done with it."
Facebook has maintained that the Journal misrepresented its data, and shared the decks with Congress ahead of Davis' hearing to prove it. Facebook also apparently shared the slide decks after receiving notice from the Journal that they planned to publish even more of the documents that underpinned the Facebook Files reporting. Those documents include some not found in Facebook's disclosure.