Does Meta's science really shed light on audience behaviour?

A host of new social science research into Meta's platforms has landed. Are we to trust the data provided by Meta that informs them?

by Rob Corbidge

Published: 14:07, 03 August 2023

Four new studies were published last week in the social sciences field by researchers who had been given access to user data shared by Meta.

This was part of an initiative launched by Meta back in 2020, the year of a rancorous US election (although rancorous seems to be the new normal), with no less than the name of Nick Clegg, VP of Global Affairs, attached to it.

In broad terms the aim of the project was to assess the impact social media had on democracy, and I assume, given their backing and co-operation with outside researchers, find Meta innocent of all charges. I'm obviously already suggesting bad science here - but there's a better reason than my reactionary anti-big-tech tendencies to believe that is the case.

The four studies broke down areas of research into the following:

Users given a reverse-chronological feed
The usual algorithmic feed was replaced by one on which most recent posts appeared first. Both Facebook and Instagram were used for this experiment.

A user's ability to share another’s post was limited
This was done with the aim of seeing if changes could be enacted in the subject's personal political beliefs, targeting "fake news" viral content.

Limiting the amount of content to which users were exposed by friends or by Facebook pages or groups with which they sympathised
This experiment had the aim of isolating users from existing sources of influence and replacing them with others.

Researchers looked at which news stories arrived on the feeds of Facebook users in the US, then correlated this with how liberal or conservative users were
This study was looking at ideological separation in the US, essentially the echo chamber effect most often identified on social platforms.

After such massed ranks of academia had delivered their verdict on many aspects of his eyeballs-to-screens advertising business, a verdict that many would argue is far from complete, VP Clegg came back triumphant. Meta was Not Guilty.

Yet shade has already been cast on the data Meta made available.

In his post, VP Clegg, said that research process independent observer, Professor Michael Wagner of the University of Wisconsin, in his introduction to the papers, said that the changes made to the subjects' feeds "did not reduce polarization or improve political knowledge during the 2020 US election."

There's nothing Meta can do about it.

However, Professor Wagner also described the entire process as "independence by permission" which is a strange concoction of words.

According to science.org Meta "collaborated with 17 outside scientists who were not paid by the company, were free to decide what analyses to run, and were given final say over the content of the research papers. But to protect the privacy of Facebook and Instagram users, the outside researchers were not allowed to handle the raw data."

And therein lies the rub, so to speak. One would assume that such properly scientific research as Meta would have us believe has been conducted would be conducted to the highest standard. What we don't know is the extent to which Meta prevented exposure of certain aspects of the data. Privacy is their fig-leaf to hide behind, yet we're talking about proper academic research being conducted on what Meta thinks should be available, not what research requires. Can't professional researchers be trusted?

You can't blame the researchers. Some data is better than none at all and they are desperate to study such an unfolding human phenomenon as social media behaviour in all the detail they can.

There are another 12 research papers from the data still to drop.

My own belief is that social media is a reflection of ourselves, and a reflection of our societies, both bad and good. Technology has simply given individuals the ability to reach wider audiences, with whatever content they are reaching those wider audiences with. However, when someone is making money on the back of it, and making money whatever "it" is, then responsibilities, such as those responsibilities every single publishing business has to bear, sit firmly on those platforms too.

However, at the heart of a platform such as Meta's Facebook sits a content paradox. The content they don't want is actually the content they do want, but unlike conventional publishers, they want to fart and run away. It's the thing they can't say.