Does a reply on Twitter from Elon Musk help amplify misleading information?
What does the data say?
In this post (originally posted on Twitter), I explore to what extent Musk's endorsement of a misleading and racist tweet from @EndWokeness helped amplify it
As many of you know, Musk replied to the below tweet showing a misleading and partial interpretation of violent incident statistics in the US.
The original tweet by @EndWokeness was sent on May 5th at 19:14 Coordinated Univeral Time UTC). Musk replied on 6 May at 06.57 UTC
(I have written UTC in red as our browsers convert UTC into local time).
In order to determine the influence of Musk's reply, we should explore the engagement with the End Wokeness tweet both before and after Musk's intervention.
I downloaded all retweets and replies of and to the End Wokeness tweet (over 30k). The below time series graph shows the overall engagement the tweet received since 5 May and 11 May. As you can see, after an initial surge, the tweet declined in popularity
However, despite this decline a sudden increase in popularity of the tweet occured. I highlight this below in green.
If we zoom in on this transition we can see that after the initial decline, a renewed surge occurred at 06:59 on 6 May, two minutes after Elon Musk sent his reply at 06:57. The almost instantaneous boost indicates Musk's reply was the proximate cause in this renewed activity
If we zoom out again we can see the tweet activity after Musk's intervention (blue) and the activity before his intervention (pink).
We can even calculate the total amount of retweets, replies, mentions before and after his intervention. Before, it's 8k while after it is 65k. (Of course had he not said anything it is difficult to know what the figure would be, or what other influencers would have
promoted the tweet). However, Musk is the most followed person on Twitter, so as 'Chief Twit' his influence is outsized. The below double graph shows a time series graph (red), but below that interventions based on number of followers (green). As we can see,
Musk's followers dwarf the followers of others, and when he tweets it seems to really have a substantial impact on the reach of whoever he annoints, in this case, a person spreading misleading and arguably racist misinformation
So to answer the original question. It seems Musk's interventions really do amplify misinformation, and that his huge influence amplify them significantly. This does not mean everytime he speaks to someone it will have the same degree of influence, as there are other factors at play (timing, newsworthiness, current events etc). However, it is useful to be able to try and quantify his impact on bad information, and adds weight to the argument that Musk is the Rupert Murdoch of Twitter - editor in chief and chief algorithm curator