Nieman Foundation at Harvard
Earlier this year, when the U.S. Chamber of Commerce launched a multimillion-dollar television ad blitz for (and against) candidates running for federal office, I wanted to know how much it spent to broadcast ads supporting a U.S. Senate candidate in Hawaii.
A spokesman for the Chamber wouldn’t provide the figure directly, but it’s the kind of information that’s public two times over: You can get it from the Federal Election Commission or the Federal Communications Commission. My options:
1. Wait months for the next round of campaign finance reports to be filed, and get the data from the FEC then; or
2. Fly 5,000 miles Hawaii to physically walk into local TV stations and request the “public file,” which includes FCC-mandated information about the campaign ads that candidates and groups buy.
Maybe you’re thinking there has to be a better way.
The FCC thought so, too, when it proposed requiring stations to post their public files online. But the National Association of Broadcasters pushed back, saying the new rules would be “very burdensome” to local stations, especially during election season when such files are frequently updated. The NAB went so far as to say it doubted “many if any” members of the public would choose to access such information online. (There’s also a potential financial interest for stations; in comments to the FCC, the association expressed concerns about requiring the television industry to upload “potentially hundreds of thousands of pages containing commercially sensitive information, such as rate information.”)
It’s still up in the air. All this as the 2012 election season rolls on, and into the uncharted super PAC-fueled reality wrought by Citizens United.
So ProPublica decided that reporters and citizens should take matters into their own hands. Instead of waiting for broadcasters and the FCC to figure things out, ProPublica is crowdsourcing a nationwide effort to collect the public files that trace the money spent on televised campaign ads: “If TV Stations Won’t Post Their Data on Political Ads, We Will,” the headline read.
“We,” in this case, means anybody who’s willing to help. As of 9:30 a.m. on Thursday, 74 people from 49 television markets — in 32 states and the District of Columbia — had volunteered to join the project.
“While [the information] is public, it’s not really accessible in an easy way for people, so we saw a crowdsourcing opportunity,” ProPublica social media editor Daniel Victor told me. “The only way to get this information is to crowdsource it.”
But just because crowdsourcing could work when traditional reporting isn’t a viable option doesn’t mean that it will. In 2009, for example, ProPublica acknowledged how difficult its Stimulus Spot Check public reporting assignment was. When Amanda Michel, then ProPublica’s director of distributed reporting, introduced that project — which was designed to track state-by-state stimulus spending — she called it ambitious but “doable.”
“All you need to do is rummage around on the state’s Department of Transportation Web site and make several follow up-calls over the next week,” she wrote at the time.
That proved to be an understatement. Tracking contracts between local governments and contractors can be a difficult task even for seasoned reporters. In a report released earlier this month, the consumer group U.S. PIRG found that only one state — Illinois — keeps online information about the projected and actual benefits created from economic development subsidies. Only six state government websites post contracts between vendors and the state, the group said.
“We want to make sure we’re not asking people to assemble mountains and mountains of papers.”
In a 2009 interview with Poynter, Michel said that ProPublica ended up learning just as much about citizens’ access to information in different states as it learned about stimulus spending.
ProPublica was still able to collect data, and to provide a snapshot of various stimulus projects in states across the country. But it was a lesson on what Jeff Howe, ex-Nieman Fellow and author of Crowdsourcing, called the sixth rule of successful crowdsourcing: “Keep it simple, break it down — give the crowd something each individual can work on, yet can aggregate into something great.” Some of the site’s other crowdsourcing collaborations have been more straightforward (like calling a member of Congress to get her or his position on a piece of legislation), and ProPublica has earned acclaim for the distributed reporting projects it conducts under the ProPublica Reporting Network banner.
One recent example was the site’s project to determine where members of Congress stood on the controversial SOPA and PIPA legislation that aimed to curb Internet piracy but caused a public outcry over how the bills might stymie free speech. These kinds of undertakings are valuable not just to the public but to reporters in regional newsrooms who can mine state-specific data to localize a story. (ProPublica took another step in that direction with the Wednesday announcement of a partnership with Digital First Media that will give MediaNews Group and Journal Register newspapers pre-publication access to ProPublica news apps so that reporters can use it to write local stories.)
Similarly, Victor says he hopes the latest crowdsourcing project will produce enough good data for a collection that local reporters can tap for regional coverage. The tricky thing about journalistic collaborations with the public is that you never know what you’re going to get. That’s true, too, in professional reporting, but with crowdsourcing you’re potentially dealing with data on a larger scale.
To determine whether crowdsourcing was the way to go this time around, ProPublica first conducted a small experiment. The site enlisted a team of graduate students from Northwestern’s Medill journalism school to request public files from Chicago-area stations. Here’s how Victor explains it on the site:
For our experiment, we asked our Chicago volunteers to check on spending by five super PACs that individually support Mitt Romney, Rick Santorum, Ron Paul, Newt Gingrich and Barack Obama. There were no records of spending in Chicago by four of them, but Restore Our Future, a pro-Romney super PAC, advertised on all five stations. The super PAC paid the five stations about $800,000 in the past month.
Now, ProPublica wants help getting the same kind of data from TV stations across the country. For this to work they’ll need help from more people — journalism students, professional reporters and anyone else who’s willing. For now, the focus will remain on money paying for ads related to the presidential race rather than state or local races. (“I wouldn’t say ‘no’ to anything, but initially presidential, because it would get out of hand fast,” Victor said. “We want to make sure we’re not asking people to assemble mountains and mountains of papers.” Keep is simple, break it down.)
At one point in the project planning, Victor says ProPublica considered having people text message line-item data from public files to the site. But that seemed both unnecessarily complicated and risky, introducing the possibility of errors while also eliminating the possibility of sharing the reports themselves. Properly framing a project — from the scope of data involved to the steps necessary to obtain and share it — is key to an effective crowdsourcing effort. Ultimately, Victor says volunteers need to feel like they are actually accomplishing something.
“There is very, very much value added. If we’re getting all this great data, they know we have this great reporter who can write great stories.”
“I really focus on the personal incentive,” Victor said. “My big concern with this particular project is will this be too difficult? Are they going to get to the office and be totally frustrated?”
So before even asking Medill students for help, Victor personally went to five TV stations in three states. He says he found slight variations in forms that stations used, but getting the data was straightforward and required little time. A small sample size, sure, but Victor’s personal experience was encouraging enough to email Medill to ask for help. (Victor says he also experimented with posting a Facebook message only visible in Illinois to see how many volunteers the request would yield, but that only a handful of people responded.) When the test run with grad students went “swimmingly,” ProPublica decided that it was time to make the open call to the public at large.
“It has to be presented as: ‘You’re not doing our work for us — you’re doing work that will enable us to do great work,'” Victor said. “There is very, very much value added. If we’re getting all this great data, they know we have this great reporter who can write great stories.” ProPublica reporter Kim Barker, news apps developer Al Shaw, and Victor will be among those working on the data and the stories that result from them.
Some elements of the project are still up in the air. Will volunteers all be able to scan-and-email the documents, or will some have to resort to other means of transmission? How often will volunteers be expected or asked to check in at stations? (Victor: “If we build up an energized group that wants to do weekly checks in every market in the U.S., wonderful. If targeted efforts in specific markets at specific times works better, that’s fine too. No way to know yet.”)
The kind of scoops on Victor’s wish list would require comparing information from public files with FEC reports to see if there are advertisers who aren’t complying with campaign finance disclosures, for example.
“There are certain tasks that you do need a reporter for and there are certain tasks that you do not need a professional reporter for,” Victor said. “Sometimes you’re talking to a relatively smaller community that is already passionate about that issue. Sometimes that community doesn’t exist and you need to create it, which is something I think we can do.”
(If you’d like to help out at your local TV station, click here to sign up for the ProPublica Political Advertising Project.)
