The great open access journal International Journal of Communication just published a paper I’ve written with Cristian Vaccari called “Do People “Like” Politicians on Facebook? Not Really. Large-Scale Direct Candidate-to-Voter Online Communication as an Outlier Phenomenon”.
In the paper, we analyze the presence of 224 major-party candidates for the House of Representatives across the 112 most competitive districts in the 2010 U.S. Congressional Elections across Facebook, Twitter, YouTube, and their campaign websites and find that attention is highly concentrated with just a few candidates attracting vastly more supporters, followers, video views, and website visitors than everyone else.
Though a few politicians stand out, the median candidates count their followers etc in the hundreds or at best the low thousands. It is a hit economy out there, and most politicians are not a hit with the wider population, at least in the U.S. (I’ve blogged about this data before, and TechPresident wrote a bit about it back in 2010—the article had a tortured review process, so it has taken a bit of time to get this simple but I think pretty important argument out there in its final form.)
Are our findings still relevant three years after the data was gathered, three years in which the spread of smartphones and tablets and the growing popularity of new social media have changed the web?
I very much thing so—the four basic patterns we identify seems stable. They are:
(1) limited reach in terms of the number of people who follow most campaigns on various platforms; (2) high levels of concentration of attention across all platforms, with a few politicians drawing many people, and most drawing few; (3) considerable correlations between visibility on each platform, where candidates who do well on one also tend to do well on the others; and (4) noticeable growth in the total number of people following candidates in the course of the campaign period without any change in the overall pattern of highly skewed distributions.
Data from our follow-up work on 2012 confirm all four basic patterns (here for a first cut of that data), and several of them are familiar from earlier work by Matt Hindman and others on “web 1.0” online politics.
As we write in the paper,
The few candidates with significant online audiences are not as much ahead of the curve as they are on top of the curve. If the limited number of politicians in our sample who attracted a lot of attention were distinguished by being early adopters of particular platforms, perhaps others could do likewise and achieve similar results. But with adoption rates of the four platforms considered here ranging between 91% and 100% among candidates [in 2010], and large parts of the adult population already using them regularly [again, in 2010], the highly uneven distributions are clearly not the outcome of uneven levels of use.
Ultimately, the online environment, partially discounting online marketing, is a pull environment in which people opt in and self-select. And at least in the U.S., interest in politics is limited and unevenly distributed, trust and confidence in politicians is limited, and while many people talk about politics online, most people do not connect with most politicians online.
It is an open question how relevant these findings are in countries where more people are more interested in politics (many countries have far higher turnout than the U.S.) and have more faith in politicians and the political process (as in some Northern European countries, for example). (On the other hand, they may be even more true in countries with lower levels of engagement and trust, or in supra-national political systems like the E.U., where very few European-level politicians have build significant popular followings online.)
But at least in the U.S., what we wrote in conclusion about our 2010 data still seems accurate to me–
As long as competition for attention is so fierce and levels of interest so low and uneven, only a few politicians will attract large online audiences that allow them to communicate directly to the electorate to any significant degree via various social networking sites. The rest will have to find other ways, including both traditional means, such as direct mail, field operations, and television advertising, as well as new forms of push marketing online. The topology we have mapped here is one dominated by a few exceptional outliers who attract tens of thousands of supporters and viewers, but where the great majority of candidates—even in well-funded, competitive, high-stakes contests—labor in relative obscurity online.
rasmuskleisnielsen.net 
A really interesting post. It seems that the majority of voters still see politics linked to the ‘real world’ and don’t want it to encroach on their digital life (a desire for autonomous self-governance as opposed to apathy?) but as we spend more time online the line between the two will have to be blurred. Somehow.
Did you notice a difference in content or approach between the politicians with a significant online presence, and those who who were less successful?
In a separate paper, we’ve found some traits that correlate with online popularity, including being a challenger, a candidate in a relatively competitive district, and being frequently discussed on political blogs. Surprisingly, news coverage does not correlate with online popularity, nor does campaign spending or popularity as measured in polls. Like a lot of other people, we are quite keen to identify aspects of what candidates actually do, in terms of political substance and online behavior, that might help explain the huge differences that leave a few politicians with large online followings, and most with very small ones.