Above we can see that Bernie Sanders leads in the polls nationally and in his Facebook and Twitter followings.  The Twitter audit column shows an estimate of the percentage to Twitter Followings are real (ie not bots). The last column shows the estimate or the real number of Twitter followers.  Bernie Sanders does have a low percentage of real followers but still leads in the overall number of real followers.  Tom Steyer is included even though he dropped out of the race yesterday.  The graph below shows how Twitter followings are associated with national support.  

The R squared statistics shows the percentage of the variability in the poll averages that is accounted by the variability in the twitter following.  This value is 0.8546 or 85%.  In January, this relationship accounted for 51% of the variability in the poll averages.  This was mostly due to Biden having high poll averages but low a Twitter following and to there being more candidates in the race.

When I limit the data to real twitter followers, the relationship with the RCP poll average becomes even stronger accounting for 91% of the variability in the RCP average.  If 100% of the variability were accounted for, all of the candidates would fall right on the line in the graph.

Looking at the relationship between the Facebook followings and the RCP averages, a weaker relationship that that with Twitter emerges.  Sixty-three percent of the variability is accounted for which is higher than the 42% in January.  These higher percentages in the Twitter and Facebook relationships suggest greater engagement by consumers of social media with the candidates and its potential influence on voters now that primaries are heating up.


Trump now has a following of 73,242,775 on Twitter (79% real) and 26,091,741 Facebook.  It remains to be seen what impact it will have on the general election.