Did Twitter followers predict the Boston At-Large election?

Did Twitter followings translate into votes?

In August, when we explored Boston political twitter, we wrote that:

If the percentage of Boston electeds following a candidate translated to votes, next year’s At-Large Council seats would include Michael Flaherty, Julia Mejia, Ruthzee Louijeune, and David Halbert.

 Of course it’s not that simple, and this is not a prediction of who will win. But it will be interesting to check after the election to see how Twitter matches up with what the voters decide. It may be that there’s no connection and that these follows don’t reflect the larger electorate. 

It could also be that elected officials following a candidate on Twitter is an early indicator of established political connections, support, and name recognition. This could be used to parse out a race before polling and endorsements are available.

Now that the preliminary election has happened, we have votes to check the theory.

The numbers

In the preliminary election, citizens can choose up to four candidates (for the four At-Large seats), and the top eight move on to the general election in November. Below are the results for the candidates in terms of what percentage of Boston’s elected officials (state and local) follow their Twitter accounts, and their share of the total preliminary votes.

Of course, this is only one election, with a small sample size, and some of the candidates did not have a Twitter handle.

Nonetheless, it is remarkable that the two data sets were so closely correlated, with an R-squared of 0.9, meaning that about 90% of the variation in elected Twitter followers can be explained by preliminary vote. 

Of course, there is not a direct relationship between the two (we are not suggesting that people vote for a candidate because elected officials follow them on Twitter). But it does indicate that elected officials following a candidate on Twitter may be a proxy for their chances in an election, whether that is because of incumbency advantages, name recognition, government connections, or online presence. 

While we are not suggesting that this replace polling, interviews, or other analysis, it does seem like a candidate’s Twitter followers are a useful piece of information to be taken into account when looking at a race.