The Crowdwire


Social Seniors: Voting With Their Tweets

The AARP Age Brand List2senior vote matters, big-time. Why? Because there are lots of older Americans - more than 40 million in the 65-plus bracket - and they turn out in large numbers on Election Day.

Right now, the question is how they will vote. For some time, polls showed Romney decisively  leading among older Americans, but no more.

"Support for Romney among Americans age 60 and older has crumbled, from a 20-point lead over Democratic President Barack Obama to less than 4 points,” Reuters reports, and the implications are enormous. "If Romney loses seniors, he loses this election, period," one analyst told the news service.

What’s driving this shift, and can social media help us understand what older voters are thinking?

If that strikes you as an absurd question - seniors don’t tweet, do they? - think again. More than half of Americans 65 and older are on the internet, and 34% of those who are use social networks, according to a study by Pew Internet.

Seniors are still less engaged with social technologies than other age groups (86% of internet users 18-29 use social platforms). But when it comes to politics, their presence has an outsized significance, because they vote heavily. A social comment about the election is more meaningful if it comes from a likely voter.

Two recent social moments offer a revealing glimpse of what seniors are thinking. Bluefin Labs, home of The Crowdwire, tracks and analyzes the social conversation on about 600 major brands, and compiles a list of the most talked-about brands for a regular feature in Ad Age magazine. On the current list (see above), AARP comes in at Number 6. An advocacy organization for older people trending right up there with Starbucks and H & M - what gives?

It’s about politics. Speaking at an AARP conference in New Orleans, Romney running-mate Paul Ryan called for the repeal of the Obama health-care program, and was loudly booed. The social networks lit up, due partly to wide TV coverage of the story. On the political shows, liberal commentators portrayed it as a campaign stumble, noting that the crowd cheered for Obama’s remarks (delivered via video screen) on the same subject. However, CBS News and other outlets noted that, in addition to the boos, Ryan drew applause when he introduced his mother and when he talked about the need to cut the deficit.

To make sense of the social response, we went to the data. First, we looked at who made the comments, and found that among the groups commenting most frequently were people who watch network and cable news shows, and those who had posted previously about health-insurance - both pointing to an older demographic. We then isolated comments specifically about Ryan’s Obamacare remarks and found that 43% were anti-Ryan, 22% were pro-Ryan, and 35% were neutral or unclassifiable.*

The other striking moment was Sunday’s 60 Minutes broadcast, featuring interviews with both Obama and Romney, which drew an unusually large social response for that show. Over the last year, public social commentary about 60 Minutes has averaged about 2,700 comments per show. That’s low for a show that routinely draws more than 12 million viewers, and may be related to the fact that the median viewer age is 60. Though seniors are on social media in growing numbers, they don’t match the comment volume of younger people.

For this show, however, there were about 33,000 comments - more than ten times the norm. Who was making them? The average viewer age for this broadcast was 58, slightly younger than normal but still on the senior side. That doesn’t mean all the comments came from older people, but it suggests a fair number did. We confirmed this by examining Bluefin Labs data about those who made the remarks. Among the most frequent commenters were married people, grandparents, viewers of both network and cable news shows, and fans of the game show Jeopardy! Definitely an older crowd.

As for the content, isolating remarks that were about specific issues (as opposed to general mentions of the show), we found that the candidates’ views about health care and taxes drew the most response, accounting for 26% of all commentary on the show. As for the opinions expressed in the comments, 54% were anti-Romney, while just 4% were anti-Obama. Comments clearly supporting either candidate accounted for less than 10% of the total, and the remainder was neutral or unclassifiable.

In short, seniors may not dominate social media, but the social conversation mirrors what the pollsters are finding: Health care matters to these voters, and that could be bad news for Romney. “Among seniors,” The Washington Post reports, “the issue rivals the economy as a top voting issue, undercutting Romney’s appeal in Florida, Ohio and Virginia. Generally, the more voters focus on Medicare, the more likely they are to support the president’s bid for reelection.”

Seniors 60 Minutes Pie

*Figures reflect public comments on Twitter and Facebook.

Ready, Aim, Fact-Check: A Look at the New “Super PAC App”

The engine behind both presidential campaigns is fundraising, and we’re talking a lot of money. Fundraising in this election has already crossed the billion-dollar mark, and over half of that has been spent on TV advertising. Super PACS are behind many of the ads, but outside of political circles these organizations are dimly understood.

How can ordinary voters make sense of the campaign money game? Entrepreneurs Dan Siegel and Jennifer Hollett believe they have a solution – the Super PAC App, which they developed to bring transparency to the 2012 presidential campaign.

The idea is simple: aim your smartphone at any campaign commercial as you watch it, and the app pulls up key facts about who’s behind it, and even helps you fact-check its claims.

Available for iOS or on the web, the app uses audio-fingerprinting technology to identify any Romney or Obama commercial and provide helpful information.

We were impressed by the idea and decided to put it to the test:

Functionally speaking, the app was easy to use and worked with both live TV commercials and those on YouTube. We had some follow-up questions, so we talked with co-creator Dan Siegel, a recent graduate of MIT’s Sloan School of Management:

What motivated you to create this app?

Super PAC App is free, ad-free, and entirely focused on the user experience.

We were motivated to build something that connects users with important information in a fun and easy way. The official campaigns and outside groups have already raised over $1.4 billion, much of which is pouring into slickly-produced thirty-second TV ads. Who is behind these ads, what are they saying, and are the claims based on any facts? That’s important information that otherwise takes hours of homework to find. With Super PAC App, you can find out easily, as you’re watching the ad.

Why did you think it would appeal to voters?

With your TV screaming political messages at you during every commercial break, there’s a tendency to tune out. We were inspired to create something that cuts through the noise, because voters shouldn’t have that hopeless “forget it” feeling. You like your TV? Keep watching. You like your couch? Stay on it. You don’t understand what these political ads are all about? Use Super PAC App.

How does your audio-fingerprinting system work? When a new ad airs on TV, how long does it take to show up in the app?

The audio fingerprinting is powered by our partner, TuneSat. We are tracking groups that have run ads to date, or that have run presidential ads in prior elections. When these groups upload an ad online, it automatically is pulled into our database and fingerprinted. To date, this has been happening at the same time that an ad is aired on TV. So if it’s on your TV, chances are pretty good that it’s in Super PAC App.

Are the fact-check resources in the Claims section non-partisan, or at least ideologically balanced? How can a user be absolutely sure they are objective?

We point users to non-partisan, trusted news outlets that write specifically about the claim in question. This includes “fact checks” but can also include more context or information about the claim. We avoid articles that are stories about the ad itself, unless the “story” is checking the facts or claims of the ad. Super PAC App won’t tell you “true” or “false”, but it will connect you with information to make that assessment yourself.

Who’s downloaded the app already?

We were excited to see that we hit #1 in the App Store on our day of launch, within our category of free news apps, and cracked the top 100 for all free apps. It was definitely a moment when Jennifer (my co-founder) and I looked at each other and knew—we built something that people cared about. It was a rewarding feeling. We high-fived, did a dance, and then got back to work because with a user base, keeping the app fresh matters that much more.

The app has been downloaded in all fifty states and in fact across the world. We’re excited to see people using Super PAC App in swing states that are getting hit hardest with these ads. We’re also interested to see that users are checking us out in Saudi Arabia, Iran, China, Russia and other places we weren’t necessarily targeting.

Can you share some usage demographics with us?

Most of our users are repeat users, coming back to learn more. Our ad database is growing by the day, of course, and the avalanche of ads seems like it’s picking up. The average user is inside the app for nearly three minutes per session, so they’re spending time digging around, which is great.

What’s been the most surprising thing you’ve found in this experience?

Frankly it’s always a surprise to find out people want what you’re building. You can do market and customer research ahead of time, but until the product goes live, you’re basing a lot on assumptions and a gut feeling. We built this thinking it was needed and people would want it. But at the end of the day, we were three people in a small office getting no sleep. There were plenty of moments I looked up from my computer at 3 am, saw the team cranking away, and thought, I hope this matters.

Do you see a future in the app after the election?

We have built a tool that connects users with trustworthy information in real-time. Political commercials happen to be the “flavor of the moment” that can benefit from real-time fact checking, but we are just getting started. There are other ways we can apply the same concept to benefit users and we’re exploring that for after the election.

The Brown-Warren Debate in Word Trees

Perhaps you’ve read what the news reports and pundits had to say about the first debate between U.S. Senator Scott Brown and Democratic challenger Elizabeth Warren. But do you know what TV viewers were saying in social media as they watched the debate?

Here’s the conversation, visualized in two word trees, Brown to the left and Warren to the right. We explained word trees in an earlier post. The main thing to bear in mind is that they show the phrases most commonly used in response to the two candidates as they debated. We didn’t have room to show complete sentences, but these fragments give a sense of the social chatter.

Warren Word Tree

Charting the Unpredictable

Charting Unpredictable Gaphic

All those polls you’re seeing might give the impression that presidential elections have become predictable. Don’t believe it. Elections often hinge on completely unpredictable events and how the candidates respond to them.

We’ve just seen two such surprises: 1) The uprisings of Muslims around the world in response to the “Innocence of Muslims” film, and the related killing of U.S. Ambassador to Libya Christopher Stevens; 2) The video of Mitt Romney speaking at a Florida campaign fundraiser, characterizing 47% of Americans as "victims" dependent on government aid.

We drilled into the social-media response to these stories to detect their impact on the campaign.

As out-of-the-blue controversies go, these were initially quite different from each other. The Muslim protests began as a foreign policy challenge, while the “47%” story was all about politics. But as they unfolded, each broke out of its category. Soon after the death of Ambassador Stevens, Romney criticized the Obama administration’s response, turning a policy crisis into a political issue. Conversely, Romney’s comments at a political fundraiser gave rise to a vigorous policy debate about class, entitlement programs and the size of government. 

In short, they overlapped not just in time but in content. So we decided to look at them side-by-side. The chart above tracks the social response to both (the Y axis represents comments per hour). Given the importance of the female vote in this election, the gender disparity between the two conversations was striking. Women accounted for nearly half of the response to the “47%” story, but represented just 39% of the commentary about the Muslim uprisings. The graph also shows highly re-tweeted comments that helped drive the spikes in each conversation.

We wanted to learn more about the people who reacted to these stories. Given the wall-to-wall media attention, we expected to see heavy commentary from a wide range of social users, not just news and politics junkies. But, in fact, those people were by far the most energetic participants. The list below shows the audience segments commenting most frequently on each story. These categories are defined by users’ previous public comments in recent months, before these stories broke.

Leading the commentary on the Muslim protests were those with established Anti-Obama views, who were 2.5 times more likely to comment than Pro-Obama people. The “47%” story drew the strongest response from established Anti-Romney people, who commented more often than Pro-Romney users by a factor of 4.*

As is generally the case with political topics, the most frequent commenters included many people who follow the news on three cable channels: CNN, Fox and MSNBC.

Muslim and 47 Controversies

*The Pro-Romney and Anti-Obama populations are not identical, nor are the Pro-Obama and Anti-Romney groups. Some people comment only about the candidate they oppose, never mentioning the opponent; others only remark on the candidate they support.

Beyond the Bloc: African-Americans and Election 2012

Obama at Ebenezer

As the election nears, pollsters and news outlets are focusing on key demographic groups that could play a decisive role in the election. According to some reports, African-Americans may not turn out for Obama as strongly as they did in 2008. In fact, recently some black pastors have announced their opposition to both Obama and Romney - the former because of his gay-marriage stand, the latter because he’s a Mormon - and are urging their congregations not to vote at all on November 6th.

"Some black clergy see no good presidential choice between a Mormon candidate and one who supports gay marriage, so they are telling their flocks to stay home on Election Day,” the Associated Press reports.

In a recent Crowdwire analysis, we found that among those who watched and commented most on the Democratic National Convention, two groups predominated: political junkies and African-Americans. Both groups were identified by other TV fare we know they watch, based on their public social-media remarks. In the first case, it was political news shows; in the second, it was the BET Network, which has a large African-American audience.

That many African-Americans tuned in to the DNC and commented on it - and the sentiment of their comments was largely positive - suggests that at least part of the black population is still engaged by Obama’s message. So how do we square this with the news about African-Americans opposing both candidates?

We dug into the data and found some answers. First we looked at whether large numbers of BET viewers had remarked in social media that they were troubled by same-sex marriage or Mormonism. That would suggest that the trend identified in the story about black churches is a broader phenomenon, and black voters might not turn out this year as strongly as they did in 2008.

But our analysis did not show this. In fact, BET viewers have spoken up very sparsely on gay marriage and Mormonism, and even when they did, the commentary did not lean decidedly pro or con.

So what gives? Here’s a theory: While the headlines and political experts often speak about “black voters” collectively, it’s important to remember that the black population numbers in the tens of millions and is not monolithic.

BET’s target audience - African Americans and consumers of black culture in the 18–34 age range - is not identical to the population of black churchgoers, which includes many older people. And there’s no reason to assume the two groups have identical takes on gay marriage and Mormonism, or on the presidential candidates.

Indeed, viewing any racial or ethnic category as a single voting bloc is highly reductive. People are more complex and unpredictable than that, and social media are a new way of understanding that complexity.

[Photo: Obama speaks at Ebenezer Baptist Church, Atlanta, 2008]

Obama vs. Romney: Who’s Tuning In?

This election is drawing a big social response, and we’ve been giving you the numbers. But as Stephen Colbert points out, numbers alone don’t tell you much.

Five million social comments were posted by people watching the conventions live on TV, according to data gathered by Bluefin Labs. Who were those talkative TV viewers? If we knew more about them, we might get a better sense of where the election stands.

So we put the data from the two convention weeks under The Crowdwire microscope. With the right tools, it’s possible to discern the political leanings of social authors, based on their previous public posts. Earlier, we reported that during the Republican National Convention, people who oppose Romney commented more actively than any other group, and significantly more than the average social user. Viewers of the same political persuasion - anti-Romney/pro-Obama - also led the response to the DNC, posting about twice as frequently as the average user.

Another way to reveal who spoke up about the conventions is to examine the kinds of TV shows they watch. Political junkies figured prominently among those posting remarks on the conventions. The graphics below show the Top Ten shows watched by those posting often about each convention. In general, the shows on these lists are heavily political. In fact, for the viewers of each convention, among all other TV shows they’ve watched, the number-one most common show in both cases was the other convention. Junkies like politics, whatever side of the aisle it’s coming from. On the RNC list, every show is partly or entirely about politics.

RNC Show Affinity

Now look at the DNC list and you’ll notice a striking difference. Four of the Top Ten shows on this list - shown in blue - are not political, but more in the pop-culture category. And three have large African-American viewerships: “Love & Hip Hop: Atlanta,” “Sunday Best,” and “The BET Awards 2012.”

Maybe it’s not surprising that a convention re-nominating Obama for president would draw black viewers, and that they would post social comments. African-Americans use Twitter far more than whites do, according to a Pew Research study. And in 2008, Obama won strong support from black voters.

But media outlets have been questioning if that will hold this time around. Thus, any evidence that black voters might be engaging with the Obama campaign in large numbers is noteworthy. Also, in the last election, the President got a major lift from his pop-culture crossover appeal and support from celebrity entertainers, including Oprah Winfrey’s endorsement and the Yes We Can video. Perhaps he hasn’t lost his mojo.

DNC Show Affininty

All data based on public comments posted on Twitter and Facebook.

Colbert Does a Number on Social Politics

            You have to watch this Stephen Colbert clip, for two reasons:

            1) It’s a brilliant slam of the hype around social media and politics, and there is a lot of hype. 

            2) It points the way past the hype. Social metrics such as tweets-per-minute are being widely reported, including right here, because they’re interesting. When 4.5 hours of political speeches generate 5 million comments from the public, that’s significant. But numbers alone aren’t enough. Whether it’s a convention, a debate or a candidate’s gaffe, the point is not just how much was said on the social platforms, but what was said, and who said it.

            In short, you have to drill deeper. Which is exactly what we at The Crowdwire are trying to do. As the election unfolds, we hope you’ll come back often and drill with us.

Follow us on Twitter here and Facebook here.

The Election Goes Social (With a Bang)

The Democratic National Convention produced stunning social-media numbers. It was the most socially commented-on event so far in this election. The chart below tells the story, based on data gathered by our colleagues at Bluefin Labs.

Of the two nominating conventions, social-media users showed much greater interest in the DNC, by a margin of 2.5 to 1. We were especially struck by the gender reversal - more men commented on the RNC and more women talked about the DNC. Since women are considered the prize in this election, and the DNC was the first major  event of this campaign on which they led the social commentary, this could be significant.

The social population doesn’t mirror the general public, so these figures should not be viewed as a poll-like sample of overall voter sentiment. Nonetheless, a conversation this large  - 5 million remarks posted in just three days - is a treasure trove of information. More than anything, it shows that politics is getting very social very quickly.

As Election Day approaches and the campaign gets hotter, this trend will almost certainly continue, and The Crowdwire looks forward to being your guide.

Positively Clinton

People are still talking about Bill Clinton’s speech. The former President even got #arithmetic to trend on Twitter.

Talk is great, but did Clinton generate the good kind of talk that he was aiming for? Did his speech make a favorable impression with viewers, and were they moved to say positive things about it to others?

To answer this question, we analyzed the social response to the night’s three lead speakers: Sandra Fluke, Elizabeth Warren and Bill Clinton. Using a sentiment analyzer developed by Bluefin Labs, we tracked both the negative and the positive commentary across the 90-minute window in which they spoke. The results are shown on the graph above.*

The volume of positive commentary soared immediately after Clinton started speaking, and the negative dropped sharply. At its height, Clinton’s positive sentiment score exceeded 40 percent, which is unusually high. As the speech continued, the lines flattened out, then diverged again at the end when Obama came onstage and hugged Clinton. 

The speech certainly didn’t win over everyone in the social sphere. “Clinton has talked so long I’ve forgotten the lies he told at the start,” tweeted one person. But CNN host Piers Morgan was channeling many others with this remark, retweeted more than 3,000 times:

Piers Morgan tweet

To receive updates about our work, please follow The Crowdwire on Twitter or Facebook.

*We checked these results by directly examining samples of the conversation. Because language is ambiguous and some comments are neutral, not all posts are classified as negative or positive. Thus, at any given point, the two percentages will not add up to 100.

A Few Words About Michelle & Ann

In a previous post, we noted that Michelle Obama’s speech to the DNC elicited an enthusiastic response from social media. Though Ann Romney’s speech to the RNC drew its share of warm comments, the overall reaction was mixed.

We wondered if there might be a way to visualize the two responses side-by-side. Hundreds of thousands of comments are involved, so this is no simple task. Word clouds are good at showing the most commonly used terms in a body of data, but not how they fit together in expressions and sentences.

That’s where word trees can be useful. Bluefin Labs has a nifty new word-tree generator we’ve been wanting to try out. You enter any phrase and it calls up all the social comments (in a given data set) using that phrase, and displays them in a branching-tree format.

The above word trees diagram the results we obtained after entering “Michelle Obama is" and "Ann Romney is" into the tool. These are partial trees, representing just the start of the most common sentences that employed these phrases to describe each woman. The full trees, too large to show here, continue branching far out to the right, ultimately forming a veritable forest of complete sentences.

We’re working on better ways to display future word trees. Still, this truncated view provides a peek at the way people on social platforms are talking about Michelle Obama and Ann Romney.