Why has Poynter done this study now?

We wanted to take a scientific look at how people navigate through news in various story forms — and how these forms differ in broadsheet, tabloid and online. Editors and publishers have asked for specific information about this as they make decisions about where to put their resources and how to tell compelling stories most effectively.

How does this study differ from previous Poynter studies?

This is the first study to compare both print and online. With more than 600 test subjects, EyeTrack07 is the largest study Poynter has undertaken. Subjects came in for one test and were asked to read that day's edition of the publication.

Who paid for this study?

Poynter has funded more than two-thirds of the study; the participating news organizations paid the balance.

Which news organizations were studied?

Two tabloids:
Rocky Mountain News
Philadelphia Daily News

Two broadsheets:
St. Petersburg Times
Star Tribune of Minneapolis

Two news Web sites:
St. Petersburg Times
Star Tribune of Minneapolis

How does the equipment work?

Two small cameras are mounted above the subject’s right eye. One records the position of the eye as it is reflected in a small monocle. The other camera records what the subject is viewing. These two images are married to create a video that superimposes a crosshair over the newspaper or monitor. The crosshair is a mathematical representation of the position of the cornea. As the subject reads, the crosshair follows his or her gaze.

Doesn’t the eyewear get in the way of a normal reading experience?

Certainly, the reading experience isn’t the same as if the person were sitting in his or her living room or at a coffee shop. But because this was a scientific study, there were certain controls that we had to have in place. Each subject was asked to sit comfortably and to read as he or she would normally. The mobility of this particular equipment allowed them to move freely, to sit back in the chair, to raise or lower the newspaper, etc.

Most subjects said that they were aware of the headgear, but that it was not a significant distraction from their normal reading experience.

Who was tested?

49% men
51% women
56% age 18-41
44% age 42-60
29% read print or online editions 1-3 times a week
71% read print or online editions 4 or more times a week

How long were people asked to read?

No time limit was specified when they began the test. Each subject knew he or she would spend as much as 90 minutes with us. Subjects were given identical instructions, in which we told them to read as long as they would normally, or until they were asked to stop – whichever came first.

Ultimately, we did stop them after 15 minutes of reading the live newspaper or news site. The time limit we chose was based on the volume of data that would have to be viewed and processed later. In general, each minute of recorded test required as many as ten minutes to process and code.

Did you tell them what to read?

No. They were allowed to read whatever they liked. We did ask them to refrain from reading the classifieds, as this was not our focus.

What elements were tracked, exactly?

The list of more than 350 specific elements included captions, headlines, photographs, graphics, briefs, stories, obituaries, sports and agate listings, blogs, podcasts and more. We also made note of whether or not the elements were packaged together.

How did you test comprehension?

An exit interview was used to help determine comprehension and retention using questions about a set of prototypes.

How were the prototypes used?

Six versions of a story presentation were carefully edited and designed so that they contained exactly the same information – fact for fact. Three prototypes were print; three were online.

The first version of both print and online was fairly simple: headline, written narrative and photograph. The second version was more graphic. The third version was very visual, with no traditional narrative.

Will Poynter do more testing?

Yes! We see a lot more out there that needs to be tested in print, broadcast and online. We are interested in future eye tracking studies, and we're looking for partners.

We'd like to take a look at:

• News delivery on large format screens, in high definition and smaller screens
• Moving text and animated graphics in broadcast news
• Major innovation in searchability and interactivity online

Previous Poynter studies include:

• “Eyes on the News, 1990-91”, conducted by Drs. Mario Garcia and Pegie Stark Adam
• Eye Track II: Stanford and Poynter Study, 1999-2000, conducted by Poynter Visiting Faculty Member Andrew Devigal and Stanford University’s Dr. Marion Lewenstein.
• Eyetrack III: Looking at online readers, 2003-04, conducted by Steve Outing of Poynter and Laura Ruel of the University of North Carolina