The U.S. midterm election season is upon us. Midterm races typically focus more on state and local offices and issues, and attract lower discussion and voter turnout. But new Pew research hints that they may offer a surprising opportunity for state and local civic engagement via social media.

Social media is often dismissed as a cesspool of political polarization, but in fact the majority of U.S. adult internet users indicate little or no political interest in how they use social media. In the presentation Media Zones Where People Live, Lee Rainie, director of the Pew Research Internet Project noted that fully 50% of U.S. adults display no connection to the world of politics at all in their social networking usage.

Furthermore, 44% of social networking users (one in four U.S. adults) only have occasional light exchanges about politics on social networks. Plus, nearly 30% of social networking users report that they've trimmed contacts from their social networks due to political content. A mere 4% of social networking users (2% of U.S. adults) appear to orient their social networking use as a political echo chamber.

"The echo chamber trope is familiar; people freak out about it," said Rainie. "What's interesting in our data, is that 'empty chambers' are a far bigger deal: the majority of Americans who are wholly removed from hyper-partisanship and hyper-engagement.

"This presents a big strategic decision for foundations and their grantees: Which group is foremost in your mind? Do you seek to activate the people who are most politically engaged, or those who are fairly disengaged?"

Some organizations may prefer to seek the attention of highly engaged political influencers on civic issues. Pew's data implies that midterm elections, which tend to be more state-focused, may present a stronger opportunity than national elections for political engagement that is less polarized (and thus perhaps less alienating to the majority).

For national politics and elections, Pew's data visualizations of social media conversations show that the most conservative or liberal people (activists) tend to talk mainly to themselves, not to each other, especially on Twitter. This creates a visual "barbell" effect: two heavily self-referential clusters of connections connected by a thinner bar of cross-communication.

In contrast, social media conversations about state and local politics tend to show smaller clusters of polarization connected by thicker lines of communication. That is, people tend to talk to each other more across political divides -- implying engagement that midterm elections on state issues might prove more constructive.

For organizations that seek to engage people who are politically disengaged, Rainie noted that the 'empty chambers' are mostly full of people who are younger, with smaller personal networks (online and otherwise) and who may face economic disadvantages. These people are generally less engaged in their communities -- but they do tend to be very interested in solving problems DIY-style.

"They believe big institutions, including government, have utterly failed them," said Rainie. This might sound like it would be harder for organizations to engage people constructively on issues, but he sees it as an opportunity.

"When people's networks shift from reflecting hierarchy to DIY, that's a real opportunity for institutions to reclaim a place in those personal networks," Rainie said. "If you think like a friend rather than a broadcaster, you can deliver all sorts of important information in the wrapping of social connection. Think of yourself as a node in people’s networks. Many nonprofits are doing that right now. When information comes from a friend, it has more meaning."

Part of being able to act like a friend on social media involves understanding the patterns of how people interact with social media. Rainie's presentation identified five "media zones" characterized by different patterns of seeking information and interaction. Typically, an individual enters each of these zones at different times over the course of a day or week. For most of these zones, social media and the cachet of friendship attracts attention and engagement, sometimes moreso than established news brands.

When trying to increase civic engagement among people who are diengaged from civics and politics, it might be most effective to focus on getting to know the most-connected people in the social networks of your community. Then, don't lead with civics or politics; rather, interact with them to demonstrate how the issues they care about relate to civic matters.

"There is very little pure play political conversation in Twitter or elsewhere in social media. People are talking about other stuff," said Rainie. "Infiltrating the networks in your community allows you to find connectors -- the people with broad, diverse networks that contain people on both ends of the political spectrum. When connectors speak, lots of people from different perspectives pay attention to them."

Consequently, a useful engagement goal to pursue during this election season might be to identify connectors and seek to have them share your information with their networks. This is more nuanced to attempt, and measure, than strategies to simply increase pageviews on your website. But it could prove more effective in increasing and sustaining overall civic engagement.