Thread by @conspirator0

It's been asserted that the #Resistance and associated hashtags are a Kremlin social media operation. @ZellaQuixote and myself decided to try testing this statistically.

We performed a few statistical comparisons between the hashtags alleged to be Kremlin-related (#BlueWave2018, #FBR, #NotMyPresident, #Resist, #Resistance, #TheResistance) with the hashtags that were pushed by the Internet Research Agency per Mueller's February indictment.

First, are these hashtags being pushed by botnets? We tested for some traits associated with bots (24/7 operation, use of automation services, default profile pic). The hashtags from the indictment score higher across the board than the #Resistance ones.

Testing the accounts for links to the Kremlin is obviously more difficult than testing for automation, but there are a few things we can check.

We took a random sample of 500 accounts that used each hashtag and tested for the presence of multiple links to Russian media sites. The accounts using the hashtags from the indictments have far more than the #Resistance accounts.

Next, we counted the number of tweets in the dataset of content from banned Internet Research Agency accounts published by NBC. The pattern continues to hold up: the hashtags from the indictments show up far more frequently than the #Resistance ones.

(Previous thread containing a link to the dataset of banned troll tweets)

One more test related to the banned IRA trolls: we checked a sample of 500 random accounts that used each hashtag for replies to or mentions of the banned troll accounts. Once again, the hashtags from the indictments show far stronger signal than the #Resistance ones.

While we can't conclude for certain that the Kremlin isn't behind the #Resistance and associated hashtags, the statistical evidence in this thread doesn't support that conclusion; rather, it strongly suggests the opposite.

@ZellaQuixote and I are firmly committed to the resistance. Dub us RIS if you dare. @FBIWFO @NSA