Virtually all candidates involved in competitive races in the 2010 midterm elections in the US used Facebook. Together with my colleague Cristian Vaccari from the University of Bologna, I have collected data on the number of supporters of each Democratic and Republican candidate for the House of Representatives in the 112 most competitive districts (see end for sample).
The numbers are quite interesting, even at a first glance. The figure below shows the distribution of the number of Facebook supporters on November 2 of all the candidates we coded (222 out of 224 had profiles).
The first striking thing is how skewed the distribution of Facebook supporters is in both parties. Michele Bachmann (R-MN) and Alan Grayson (D-FL) are clear outliers, with 139,203 and 30,807 Facebook supporters on Election Day, respectively. Compare this with the average number of supporters for each party, 4223 for Republicans, and 2149 for Democrats. Bachmann, one out of a hundred and ten Republican candidates on Facebook, had a massive 29% of all the Republican Facebook supporters in the sample. Grayson, one out of a hundred and twelve Democrats, had 13% of all the supporters.
The second striking thing is how relatively few supporters most candidates have, in absolute numbers. The median number of Facebook supporters for a Republican running in 2010 in one of these 112 competitive districts was 2156, 1609 for a Democrat. This should be seen in light of the on average about 700,000 people who live in each congressional district.
Politicians like Bachmann and Grayson can leverage their very considerable online following both in terms of votes, attention, and money (Bachmann won re-election, Grayson did not—both raised significant sums online). But the vast majority of candidates—even looking only at the most competitive districts—had a much more limited reach on social networks this election cycle.
The data we have gathered thus runs counter to the frequently discussed notion that Facebook and the like is widely used for “direct communication” between candidates and voters (see for instance this recent report). In a sense, this is not surprising. Social media are based on people opting in, and most Americans do not care very much about electoral politics (and do not care much for most electoral politicians).
This needs to be taken into consideration both by those of us trying to understand the implications new information and communications technologies have for politics, and by those trying to use them.
Note: We included in the sample all congressional districts rated as lean Republican, tossup, or lean Democrat by the New York Times, the Congressional Quarterly, Cook, or Real Clear Politics in August, 2010. This comes to 112 districts.
rasmuskleisnielsen.net 
Great work, Rasmus and Cristian. Your headline finding — that these are surprisingly low overall numbers — is striking.
And yet, it seems to me that the the big unanswered question about Facebook (and Twitter) remains: what about the interpersonal communication networks that allow for the cascading transmission of information down from those who directly “friend” or “follow” candidates to those among their friends/followers who do not follow a candidate? In other words, accidental exposure.
I wanted to explore this question in the 2010 UK poll, but sadly my project fell at the final hurdle and wasn’t funded. 😦
Thanks Andy, I agree that the question of how people communicate amongst themselves about politics and forward informations via SNS and other platforms is absolutely crucial, and would love to see actual research on it.
Another question I’d like to see explored is whether the numbers are more impressive in cases (like the Scandinavian countries) that have high levels of internet use and much higher levels of trust in politicians than what is the case in the U.S.
In Germany we have a similar phenomenon. Although I do not have a comparable sample of German MPs here are a few recent numbers on high profile profiles and their relationship to the whole electorate:
As of today Angela Merkel has 54,685 supporters on Facebook and 61,437 supporters on the VZ-Networks, a comparable social networking platform in Germany. In comparison her challenger in the election of 2008 Frank-Walter Steinmeier has 10,601 supporters on Facebook and 13,430 supporters on the VZ networks. If you put that in relation to the 62.2 million eligible voters in the 2008 general election these numbers appear to be very low.
In Germany MPs and candidates also turned to Facebook and Twitter during recent campaigns. Although I don’t have the exact numbers I think that not nearly all MPs on the national and state levels use Facebook. This has probably two reasons:
1. It seems like candidates mainly turn to web-based communication in the run-up to campaigns. In 2008 Facebook was still in its infancy in the Germany. The political reach on Facebook was not very far. So parties concentrated mainly on other social networks as for example the VZ-networks. But since these networks were up to that point mostly used by students or pupils many candidates, who I talked to during the time, were very cautious of entering these networks because they were not sure how to adapt to the sometimes quite idiosyncratic communication conventions.
2. In Germany there are some reservations among the public and politicians to using Facebook due to data protection issues.
If you examine German MPs on a national level you’ll probably find some outliers at around 4.000 supporters although most MPs should have supporter counts in the hundreds. What might be the reasons?
First, although the numbers of German web users who use the internet to search for political information steadily rises this number is still not comparable to the US. The same goes for the adoption of social networking platforms let alone their political use.
Second, I don’t think that the absolute supporter counts really translate in votes or efficiency of the local campaigns. In German campaigns Facebook proves to be very efficient to hold intensive contact to a certain demographic of supporters. Some candidates choose to use their profiles to regularly engage their supporters, others only use their profiles to collect supporters like stamps. These profiles regularly have only few supporters.
Third, engaging the constituency through Facebook or other social networks costs time. While some outlier candidates might find it very efficient to communicate on Facebook with their supporters (be they eligible to vote or not) or the media. Most candidates will find it probably more rewarding to engage with their constituency offline. This will probably depend on the candidate and her style of campaigning and the internet use in her constituency.
Fourth, in Germany social media or online presences of candidates or parties still do not translate as well to donations as it does for some candidates in the US. So the main direct benefit politicians have from their online activities is the media response they get to them. With that in mind only a few politicians choose to invest time to build active social network presences.
Personally I think these might be reasons that could also translate well to the US. Also this might be another Matthew Effect of online communication. There is a certain number of high traffic profiles, be it because a politician uses a social network profile very efficiently to engage her supporters or be it because she has high visibility in the national media. People flock to these profiles because they expect them to be interesting so these profiles might attract more than just the die-hard supporters. The other profiles remain with much lower supporter numbers since they only hold interest for the local core supporters.
I don’t think that these low numbers speak against the fact that profiles on social networks can be very effective campaigning tools. But I think the results show that their use and success always depends on the candidate and the way the profiles are integrated in the overall campaigning efforts.
Thanks Andreas, very interesting comparison to the German situation, and important notes both on the potential indirect effects and importance of SNS-tools and on practical differences like the absence of large-scale systematic fundraising from supporters in many European countries. I liked this observation in particular:
Well put.
I think you can see a celebrity effect. For instance, Bachmann has more fans than there are likely eligible voters in her district. Many of her fans likely come from people who regularly see her on Fox news. Perhaps having a Facebook page has reached a point of normalization where any professional candidate has one and anyone who doesn’t might be questioned as to whether they are serious. However, who is active online doesn’t seem to have reached a sufficiently representative point as a semantic analysis of Twitter posts showed that Angle was significantly more popular in the Twitterverse than Reid, but of course Reid won the electoral contest (which even the polls didn’t predict). While the pages may not translate into votes, Bachmann (or others), may be able to point to statistics like these in making her case for a leadership position in the House (and she is running). This lack of a link between fans and votes Nick Anstead and I found as well in the UK election as Clegg’s support kept increasing (as did fans of the Lib Dems), yet they vastly underperformed for their electoral expectations.
Thanks Mike, yes, I think there is defintely preferential attachment going on here, both in terms of online and offline attention–and as you say, the top candidates almost certainly attract much of their support from outside their district (still very useful for fundraising, for example, as direct mail operators have long known).
On the role of Fox News, MSNBC and the like in driving online popularity, Sandra Gonzalez-Bailon has an interesting paper documenting that apperances in traditional news media drives online attention, something that probably also applies to attention on, say, Daily Kos and other high-traffic websites.
You know how much I like empirical evidence, and this research is great. Yet, according to Facebook, number of Facebook friends was a strong indicator of who would win the contest:
“An early sample of some of the hottest House and Senate races bodes well for the world’s largest social networking site. The Facebook political team’s initial snapshot of 98 House races shows that 74% of candidates with the most Facebook fans won their contests. In the Senate, our initial snapshot of 19 races shows that 81% of candidates with the most Facebook fans won their contests.” (analysis: http://www.meta-activism.org/2010/11/new-evidence-that-slackitivism-matters/)
So it seems that it matters more the relative difference between candidates in a given race than the absolute distribution of friends across all candidates.
Thanks Mary–I agree with the main point in your post criticizing the cheap pot-shots at “slacktivism”–that various forms of low-cost online engagement can be a first step up the ladder towards other forms of participation, provided both citizen and organizer are genuinely commited to moving forward together (and all too often, one or both sides aren’t).
In this sense, Facebook, Twitter, and email can be gateways to other forms of political activism, just as meetings can.
I don’t agree, however, that Facebook’s short press release offers any evidence that the number of supporters a candidate has matters. It might, but we don’t know. Someone would have to control for a whole range of other variables (incumbency, party, money spend, etc) to establish that the difference in the number of supporters between any two candidates has any indpendent impact. My intuition is that it is with Facebook supporters (and Twitter followers and so on) as it is with other forms of activist support, they can help a candidate win, but they are not strongly correlated with victory.
The point of this post is simply to document why one should not exaggerate how many people–in absolute and relative terms–take even the first step towards potential further engagement on Facebook. In most competitive congressional districts, the number of supporters was quite low.
Pingback: Twitter in the Midterms | rasmuskleisnielsen.net
Pingback: Mundane internet tools and political mobilization | rasmuskleisnielsen.net
Pingback: Direct communication via social media? Not for most politicians, no | rasmuskleisnielsen.net