Changes in coordinated and non-coordinated account activity and interactions after Musk’s purchase.

We use 3 different coordination metrics: hashtag sequence [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0313293#pone.0313293.ref032" target="_blank">32</a>], co-repost [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0313293#pone.0313293.ref032" target="_blank">32</a>, <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0313293#pone.0313293.ref033" target="_blank">33</a>], and activity [<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0313293#pone.0313293.ref032" target="_blank">32</a>, <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0313293#pone.0313293.ref033" target="_blank">33</a>]. Accounts are coordinated if they share a sequence of at least 7 hashtags in a post or have a co-retweet or activity cosine similarity in the top 0.01 percentile. (a) Percent change in posts per day, (b) change in mean reposts per post, and (c) change in mean likes per post before versus after Musk purchased X. Changes are for significant for all data (Mann-Whitney U test p-values <0.05) except for co-repost coordinated accounts (Mann-Whitney U test p-values = 0.38). No changes are statistically significant for (b) or (c) (Mann-Whitney U test p-values > 0.1) except for the change in reposts per post for hashtag based coordinated accounts (Mann-Whitney U test p-value = 0.002). Error bars are standard errors. Because of the differences between how standard error and Mann-Whitney U test are calculated, there can be standard errors that overlap with zero and the results could still be statistically significant. (PDF)