Methodology

Between Nov. 13 to Nov. 22, 2023, NewsGuard analyzed 30 tweets advancing misinformation about the war posted by 10 accounts previously identified by NewsGuard as being a super-spreader of Israel-Hamas war-related misinformation. NewsGuard defined “misinformation super-spreader” as those who have spread three or more false or egregiously misleading claims related to the war, and have more than 100,000 followers on X. Many of the accounts have spread far more than three false or egregiously misleading war-related claims.

To simulate the programmatic ad experience of different users, NewsGuard used Virtual Private Networks (VPN) — a tool that enables users to browse the internet as if they were in another country — and repeatedly refreshed the page of each of the 30 tweets. For each tweet, NewsGuard browsed in U.S., U.K., Germany, France, and Italy and refreshed the page a maximum of 25 times. NewsGuard scrolled to “reload” the feed a maximum of five times, because not all replies to a given tweet are shown on the screen of a given tweet.

In reviewing the 10 accounts, analysts looked for advertisements featured in the “replies” section of posts containing the war-related misinformation. Replies — and the ads featured among them — are shown directly below tweets. Analysts took screenshots of each advertisement.

X has recently sued Media Matters for America for an article that the progressive advocacy group published about prominent brands’ ads appearing alongside accounts carrying extremist content. In that case X has alleged that Media Matters created the pairing of these brands with that content by having the X accounts it was using follow only a selected group of prominent brands and only a selected group of accounts with extremist content, creating what X’s suit called an “inorganic” scenario in which the ads for those brands would appear next to that content, thereby producing “maliciously manufactured side-by-side images.”

Without speculating on the veracity of those allegations, we should note that NewsGuard’s methodology was different than the methodology that X claimed Media Matters for America used:

• NewsGuard did not manipulate the sample by selectively following certain brands or accounts. Instead, NewsGuard analysts reviewed viral content that was naturally occurring on the platform shown to them on their own personal X accounts.
• The tweets Media Matters for America analyzed received relatively few impressions. As noted above, the 30 tweets NewsGuard analyzed have cumulatively reached an audience of over 92 million viewers, according to X data. On average, each tweet was seen by 3 million people.
• NewsGuard capped the number of times that it refreshed each tweet in each country it analyzed (25), and scrolled to unfurl the feed (five). (X only shows a limited amount of replies to a tweet in a given feed.)

A list of the 30 tweets and the 10 accounts used in NewsGuard’s analysis is available here.