With the US midterms fast approaching purveyors of online disinformation are very busy indeed spreading their hyper-partisan junk on Facebook.
Their goal: Skewing democratic outcomes by putting out misleading, deceptive or incorrect information that’s packaged as real news about politics, economics or culture — yet presented in a way that panders to prejudices and is more likely to get virally spread on mainstream social media platforms where it has the chance to influence people’s views.
In the meanwhile, what’s to be done? Arming yourselves and your friends with smart digital and news literacy tools to help shine a light on the kind of ridiculously over-inflated political nonsense that’s being passed around on all sides (albeit, not necessarily equally) seems like a good place to start.
Step forward, Oxford University’s Oxford Internet Institute (OII), which has just launched an aggregator tool which tracks what it terms “junk” political views being shared on Facebook — doing so in near real-time and offering various ways to visualize and explore the junk heap.
What’s “junk news” in this context? The OII says this type of political content can include “ideologically extreme, hyper-partisan, or conspiratorial news and information, as well as various forms of propaganda”.
This sort of stuff might elsewhere get badged ‘fake news’, although that label is problematical — and has itself been hijacked by known muck spreaders. (So ‘online disinformation’ tends to be the label of choice in academic and policy circles, these days.)
The OII is here using its own political propaganda content categorization — i.e. this term “junk news” — which is based on what it describes as “a grounded typology” derived through analyzing a large amount of political communications shared by US social media users.
Specifically it’s based on an analysis of more than 2.5 million tweets sent in the period September 21-30, 2018 — applying what the Institute dubs “rigorous coding and content analysis techniques to define the new phenomenon”.
This involved labelling the source websites of shared links based on “a grounded typology that has been tested over several elections around the world in 2016-2018”, with a content source getting coded as a purveyor of junk news if it failed on 3 out of 5 of criteria of the typology.
(Examples of sources that are being judged junk via this method include the likes of Breitbart, Dailycaller and Dailywire to name just a few.)
Now to the tool itself:
The Visual Junk News Aggregator does what it says on the tin, aggregating popular junk news posts into a bipartisan thumbnail wall of over-inflated (or just out and out) BS.
Complete with a trigger warning for the risk of graphic images and language. Mousing over the thumbnails brings up any title and description that’s been scraped for the post in question, plus a date stamp and full Facebook reaction data.
Another tool — the Top 10 Junk News Aggregator — shows the most engaged with English language junk news stories posted to Facebook in the last 24 hours, in the context of the 2018 US midterm elections. (With engagement being based on total Facebook reactions per second of the post’s life.)
While the full aggregator tool supports keyword searches of the junk heap (by content and/or publisher), and also by time — allowing for sifting of junk posts published to public Facebook pages as recently as the last hour or up to a full month old.
Returned search results can be further sorted by time and reaction — across all eight types of possible Facebook reactions.
“The Junk News Aggregator is an interactive tool for exploring junk news stories posted on Facebook, particularly useful right now in the lead-up to the US midterms,” the Institute writes. “It is a unique tool for systematically studying misinformation on Facebook in real time. It make visible the depth of the junk news problem, displaying the quantity and the content of junk news, as well as the levels of engagement with it.
“Junk news content can be sorted by time and by engagement numbers, as well as via keyword search (such as for a candidate, district, or specific issue). It also offers a visual overview and a top-10 snapshot of the day’s most engaged-with junk news.
“Our goal is to help shed light on the problem of junk news on social media, to make this issue more transparent, and to help improve the public’s media literacy. It also aims to help journalists, researchers, policy-makers, and social media platforms understand the impact of junk news on public life.”
It sent us a case study example to help demonstrate the “functionality and usefulness” of the tool (based on a search it conducted at 11:00am GMT, October 31, 2018).
For this example it used the search keyword “caravan”, selecting posts from the last day and filtering for the most shared posts — which served up several posts.
The most shared post was this one, below, from junk news source Chicks on the Right:
The Institute doesn’t make any comment on why it chose to track junk news on Facebook, specifically, vs other social media platforms (e.g. Twitter) — though there’s little doubt that Facebook’s platform remains the kingpin where skewing political views is concerned, given its massive user-base.
Meanwhile the company’s ongoing attempts to dampen the virality of democracy-denting junk shared on its platform continue — and continue to yield underwhelming results, given the size and gravity of the problem.
Also unconvincing: Facebook’s extremely recent attempts to install systems that verify the actual identity of political advertisers on its platform. Yet these self-imposed checks look to be off to a terrible start — as Facebook has just been shown hosting (and spreading) yet more fake information… ouch…
Putting your faith in Facebook to sort its shit out on the political front — and fast — looks about as sensible as trusting your pet turtle to a shark to babysit.
Much better to tool up and seek to stay on top of the junk heap yourself — at least until the world’s political representatives sort their shit out and get a proper handle on regulating social media.
In the meanwhile, don’t forget to vote.
This post was updated with a correction after it originally stated that the typology is based on analysis of 21.8 million tweets sent during the 2016 Presidential campaign period up til the 2018 State of the Union Address in the United States — in fact that was referencing a previous study carried out by an OII research group; for the Aggregator project the researchers used a smaller number of tweets (2.5M+) that were posted ahead of the midterms and included a midterms-related hashtag. These tweets were collected from September 15 to September 19, 2018