Carl T. Bergstrom: So if I understand correctly, a Facebook-funded study brought together Facebook employees and with academics funded by Facebook to use data provided by Facebook to show the Facebook didn't affect the election, and then via PNAS's contributed track chose their own reviewer—also funded by Facebook.

So if I understand correctly, a Facebook-funded study brought together Facebook employees and with academics funded by Facebook to use data provided by Facebook to show the Facebook didn't affect the election, and then via PNAS's contributed track chose their own reviewer—also funded by Facebook.

The effects of Facebook and Instagram on the 2020 election: A deactivation experiment | Proceedings of the National Academy of Scienceswww.pnas.org We study the effect of Facebook and Instagram access on political beliefs, attitudes, and behavior by randomizing a subset of 19,857 Facebook users...

To be clear, I like and admire many of the authors and am not alleging any malfeasance. I'm only pointing out the degree of industry capture in this area of research and the blind eye journals turn to conflicts. Below, the Competing Interests statement from the paper and a Pubpeer comment on it.

I have also collaborated with industry on journal publications. This is not necessarily a bad thing—it's a powerful model for applied and translational research. The important question is whether *independent work* is also possible in the research space. In this area, it's becoming nigh impossible.

Tech Policy Press just published a good story on all this. www.techpolicy.press/the-politics...

The Politics of Social Media Research: We Shouldn’t Let Meta Spin the Studies It Sponsors | TechPolicy.Presswww.techpolicy.press Meta would like to disarm its critics, but Justin Hendrix and Paul Barrett argue that would be unwise.

Their research agreements are the worst. (I wrote and re-wrote this reply a dozen times trying to figure out how I won't get fired and this is the best I can come up withn that's on a middle ground. rip freedom of speech)

Here’s a good read/example:

TTP - Google Academics Inc.www.techtransparencyproject.org Google has paid scholars millions to produce hundreds of papers supporting its policy interests, following in the footsteps of the oil and tobacco industries

It doesn't need to be intentional malfeasance, just take advantage of the belief of a wide range of academics that they're too smart to be played or fooled by Facebook.

(Or other similar entities, I should add.)

Goes one further too - "the research group was provided ethical counsel by the independent company Ethical Resolve to inform the study designs" Ethical Resolve was paid by, you guessed it...

We might think of it this way: All widely distributed media sent messages of varying quality out to voters. Did FB have more of an effect than "truffe social" (pardon my French) or MSNBC? Yes, I assume, because more people see FB than TV. FB is able to manipulate messages better than broadcast TV.

Yeah, just comparing this with "The authors have no competing interests to declare" really makes me go 😐

📌

None of the academics involved in the project were paid by Facebook and they had processes in place to keep independence. @prowag.bsky.social outlines all of that in this piece. www.science.org/doi/full/10....

Google Scholarscholar.google.com

They were not paid for their work on this particular study. Nine of the academics on the paper have financial relationships with Facebook. I go on, in my second and third posts in the thread, to explain that (1) I’m not alleging malfeasance but (2) doing this work independently is impossible.

For a company that depends on advertising for revenue I'm not convinced that "our platform doesn't influence anything" is a sensible argument.

Also the study’s conclusions, given its parameters, seem insufficiently substantiated. 6wks off Facebook, after however many years of prior use, by opt-in participants, at the end of the election cycle—of course they didn’t find a large impact. Poor dosing and selection bias.

we'd like to strongly suggest, that the data that we skimmed from our application was not used by the people that we sold it to, for them to target users with specialized disinformation to influence the election that we were not being paid to influence. Based on our own rigorous self-examination.

more like PNASTY amirite

Part of their selling point in 2016 was that they could help influence elections. Feels like they have either found morals over money or moved into big tobacco style case studies... Hmmm

"First, both Facebook and Instagram deactivation reduced an index of political participation (driven mainly by reduced participation online)." So their first major result is that taking away Facebook & Instagram reduced online activity?! D'oh!

Well, Facebook is going to have a hard time competing with AI this year...though I imagine they will serve as a superspreader. 😖

Post