Photo Slideshow
See all of the photos from this story
Download PDF
Download a printable PDF version of this article

Story Index

Eyewitness News

By Boaz Dvir

UF communication researchers help newspapers track where their readers look, in print and online

Think quickly: Where did you look first when you turned to this article? Was it the large, catchy headline? Was it the nice photo? Was it the graphic box?

You’ve probably paid little attention to how you read a newspaper or news Web site. But in an ever-more-competitive business, newspaper publishers and editors need to know what atracts readers to their products so they can deliver news and advertising the way readers like it and allocate their scarce resources effectively.

So last year the Poynter Institute — a St. Petersburg, Fla.-based journalism think tank — asked 600 print and Web news readers from around the country to let it watch how they got their news.

Dubbed EyeTrack07, Poynter researchers fitted these study participants with special glasses that tracked where they looked as they read. Poynter videotaped the subjects reading news and looking at ads in broadsheets, tabloids and Web sites at the St. Petersburg Times (print and online), the Denver Rocky Mountain News, the Philadelphia Daily News and the Minneapolis Star Tribune (print and online).

To help design the research in a scientifically valid fashion and make sense of the raw footage, Poynter turned to University of Florida public relations Professor Mary Ann Ferguson, an expert in research methodologies and data analysis.

“We couldn’t have done it without the college,” says Pegie Stark Adam, co-director of EyeTrack07.

Because Poynter wanted to track regular people reading real news in real time, it was vitally important to plan the research program meticulously, from data gathering to data analysis, Ferguson says.

“Unlike science that’s done in the lab, where you can continually be doing correction, client-based science is very different. It requires that there be a lot of work done beforehand,” says Ferguson, who’s done hundreds of such projects during her career. “Once the subjects started reading the papers, it was too late.

“We had a long, long list of things at the beginning that would have been nice to do, but we had to continually get real, to ask ourselves, ‘What can we really do here?’”

Poynter has been using eye-tracking technology since the early 1980s, when publishers’ greatest concern was whether retooling their presses to print color was worthwhile. Fast-forward to 2007, when newspapers are having to reinvent themselves in the Internet age.

EyeTrack07 is the largest and most ambitious of Poynter’s studies and the first to contrast and compare print and Web news sources.

Each news organization recruited a diverse group of 100 research participants who were regular readers of news. Each looked at a newspaper or a news Web site at least once a week and seven out of 10 read news four or more times a week.

Poynter conducted the research over a consecutive five-day period at each location from July through November 2006.

Ferguson says it was important to provide some framework for various news elements in the papers or on the screen, Ferguson says.

“We had to have a universe of what they read before we could code it,” Ferguson says. “And we had to have some agreement about what that universe was.”

“Mary Ann really did a lot to help us set up the research protocol so that the data analysis would be foolproof,” says EyeTrack07 co-director Sara Quinn.

Each morning before the test subjects arrived, Quinn and Adam scoured the papers, tagging master copies with stickers representing more than 300 elements, from headlines to photos to advertisements. They did the electronic equivalent with the Web pages.

To try to mimic the laboratory environment as much as possible, Poynter used identical chairs and tables at each of the sites. When the participants arrived, they donned the eye-tracking glasses and settled in for 15 minutes with their own un-tagged copy of the newspaper or Web site.

The glasses feature two cameras above the right eye — one records the reflection of the eye in a monocle, the other captures the target, in this case, news stories, headlines, photographs, ads, links and related elements. Although it appears cumbersome, participants told Poynter the device had little effect on their long-acquired reading habits.

Poynter also exposed the study subjects to three “prototype” newspapers that contained exactly the same information presented in different ways. The goal of this portion of the project was to evaluate how readers learn and recall information.

By the time Poynter completed the information-gathering stage, it had videotaped 582 useable reading sessions.

After each day’s testing at each site, it shipped hard drives of footage to Gainesville, where Ferguson and her student-researchers went to work.

“It was a huge dataset, because for each subject we had to record who it was, what paper, what date,” Ferguson says. “Then every time the eye stopped, a number needed to be entered.”

Eye-Opening Results

Two-person student teams spent more than 2,000 hours over six months extracting detailed data from more than 102,000 eye stops. While one researcher ran the video and identified the coded element as a reader’s eyes locked onto an item, the other researcher tracked the subject’s reading progression, for instance when they flipped pages.

After New York-based Mediamark Research Inc. analyzed the data, Ferguson and her project manager, recently graduated doctoral student Christine Popescu, worked with Poynter to make sense of the findings and turn it into a book called Eyetracking The News: A Study Of Print And Online Reading.

Poynter released the lead findings of EyeTrack07 at a meeting in March of the American Society of Newspaper Editors in Washington. Then in April, it hosted a three-day conference in St. Petersburg with about 50 journalists from The New York Times, Atlanta Journal-Constitution, Chicago Tribune and other newspapers from as far away as Denmark and India.

Ferguson says she hopes this study gives guidance to an industry trying to figure out how to keep its printed product relevant while also being a player in the online world.

“The thing that I love about applied research is that we’re dealing with very real problems in very real time and so we have to be willing to find the quick answers, the ‘good enough’ answers,” she says. “From a scientist’s point of view, we might want to know why did that happen, how can we make it happen again? From the applied perspective, we just want to know if it’s happening. Is anyone reading my newspaper?”

Newspapers have been losing subscribers and advertising dollars for a long time, but in recent years their biggest challenge has come from the Web, which can offer many of the same services faster and cheaper.

Journalists realize news travels differently these days, they’ve just had no data-based idea how. So they’ve been making assumptions — for instance, that online readers suffer from some sort of a collective attention deficit disorder.

EyeTrack07 put many of those assumptions to the test. Several results have surprised even the most plugged-in journalists. The biggest one, according to Ferguson, is that online readers stayed with a given story longer than broadsheet and tabloid readers.

Extra! Extra! Read all about it: Online readers finished an average of 77 percent of what they started!

Journalists have long assumed readers would stick longer with print stories. But EyeTrack07 participants made it through only 62 percent of a story text in broadsheet and 57 percent in tabloid. Furthermore, about two-thirds of Web stories started were read in their entirety.

“One of the most valuable things that came out of this study is that people are reading more online,” Ferguson says. “Regardless of the length of the story, online people are reading more than they are reading in print. We did not know that before this study.”

EyeTrack07 busted another long-held belief among journalists — that readers rarely stay with a story once it “jumps,” or continues on another page. The study shows that, in fact, most of them do: 59 percent of the time in broadsheets and 68 percent in tabloids.

Both results spoke to John Temple, editor and publisher of the Rocky Mountain News.

“If you have a good story, they will read it, and they will go very deep into the story,” Temple said at the April conference.

“We still have to be editors,” added Philadelphia Daily News Editor Michael Days at the conference. “I’ve read 80-inch stories that I could have read another 20 inches on. And I’ve read 10-inch stories that were too long. So, it all depends on ... how was it done? What’s the reporting? How’s the writing? Is it a topic that really engages and captivates?”

But the study did reinforce some truisms of journalism: large, color photos attract much more attention than small, black-and-white photos. And newspaper readers are attracted to big headlines.

“Bold energetic headlines and energetic photos matter, and I think we need to give that energy to our readers,” St. Petersburg Times Editor and CEO Paul Tash says in EyeTracking The News.

The study served up surprises on the advertising side, as well. Ad buyers have long assumed that the most-expensive full-page ads deliver the best results. But the study shows that, in broadsheet, three-quarter page ads, which are usually surrounded by news copy, elicit just as strong a response.

Turns out, readers’ eyes follow the news and tend to notice what’s adjacent to it. Without any news, they may simply skip over full-page ads. This could lead some newspapers to start charging just as much, or maybe more, for three-quarter-page ads. And advertisers may want to take advantage of this little-known three-quarter-page bargain, as long as it lasts.

Among the biggest differences between newspapers and Web sites are the entry points for stories. In newspapers, headlines and photos continue to be the point of entry, much as they did in the first EyeTrack study. Online, teasers, story lists and navigation bars serve a similar purpose.

EyeTrack07 also provided editors and publishers insights into what kinds of content online readers look for. Nearly half (44 percent) of online eye stops were on content available only on the Web.

“Increasingly, I think we need to think about how short videos, audio files, interactive graphics, slide shows and a host of other media can be used to tell the same story,” New York Times reporter Sewell Chan says in EyeTracking The News.

“EyeTrack07 teaches universal lessons,” Ferguson says. “It shows that beliefs are often just assumptions, and reality sometimes smacks in the face of expert logic. It’s safe to say that, before EyeTrack07, few, if any, journalists would have guessed that readers go deeper into online stories than they do in print. Now, they know better.”

Ferguson says this study still contains “a tremendous amount of unmined gold,” that she, her students and journalism researchers elsewhere should be able to evaluate for years to come.

Mary Ann Ferguson
Professor, Department of Public Relations
(352) 392-6660
Related Web site: eyetrack.poynter.org

Probing Prototypes

Hundreds of hours analyzing videotapes from the EyeTrack07 sessions just weren’t enough for UF doctoral student David Stanton, so he’s going back for more.

Stanton is doing his dissertation on a part of the study that Poynter has only partially explored — the prototypes.

All three prototypes contained the same information about bird flu, including background on the disease’s origins and the possibility of an epidemic. The only difference was the way in which the information was edited and packaged, both in print and online.

One prototype consisted of a conventional headline, narrative and photograph. Another contained a narrative story with more factual information in a map and box. The third prototype contained little narrative information, relying almost exclusively on graphics.

After reading one of the prototypes, the subjects were asked nine questions about bird flu designed to measure their recall and what they learned.

From the questions, Poynter learned that the most graphical prototype lead to the greatest learning and recall.

But Stanton wants to take that analysis a step further, using the EyeTrack07 videos to compare how print and Web readers looked at the different prototypes.

“My teaching areas are computer and online storytelling,” he says. “Online is the future. We need new, clear guidelines and measurements. When we publish a story online, we need to make sure we’re doing it in a form that’s appropriate for that story.”

Stanton says he is motivated by the knowledge that his research is vital to the newspaper industry, which is struggling for survival and desperately seeking new ways to attract and keep readers.

“I’m passionate about this study,” Stanton says. “It’s a useful, untapped area.”