Correction appended below(Sara): Good morning. I'm Sara Quinn, visual journalism faculty member of The Poynter Institute and director of the Poynter EyeTrack07 project.
(Pegie): And I'm Pegie Stark Adam, affiliate faculty member at Poynter and co-director of the study.
(Sara): We're here to share the results of Poynter's major new eye tracking study for print and online news.
This is the first time these results have been seen anywhere. We consider ASNE to be the very best forum for this important project.
In our study, we've chosen to focus on a comparison of print and online news reading -- looking for similarities and differences in reading behavior. We've also drawn a comparison between broadsheet and tabloid formats.
As you will see, we've tested nearly 600 regular readers in four U.S. markets. That's a very large sample. We've tested readers of the
Rocky Mountain News in Denver,
The St. Petersburg (Fla.) Times, the
Star Tribune of Minneapolis and the
Philadelphia Daily News.
Test subjects were asked to read that day's edition in either print or online. Our tests were conducted late last year and encompassed thirty days of publication.
(Pegie): The newspapers recruited 100 readers for each of the six test sites by calling subscriber lists, contacting the general population and by running ads in their publications. Here's what our readers looked like:
Subjects were broken into two age groups:
- 56 percent of our sample were age 18 to 41
So, the average age was 39.
There were equal numbers of men and women:
We also looked at the number of times they read the paper or news Web site each week:
- 29 percent read print or online one to three times per week.
- The rest of the sample -- a big 71 percent -- read print or online four or more times a week.
Although education and employment were not recruiting criteria for the subjects, our sample was well-educated and made good incomes:
- 87 percent had some college. And online subjects had the highest education level.
- 75 percent were employed.
(Sara): To be clear, subjects read the paper for a total of 15 minutes while they wore eye tracking glasses that recorded their gaze. Not the MOST normal of situations -- certainly not the same as if we'd sat with them at their kitchen table -- but subjects reported that the experience was comfortable and they were able to read as they would normally.
These were regular readers of the publications.
Our goal has been to create a baseline assessment of average news reading behavior. Again, it's the size of our sample -- 600 readers -- that allows us to generate this assessment.
We're delighted to be able to share our findings with you. And Pegie and I look forward to talking with you throughout the rest of this conference.
(Pegie): To give you a short overview of the project, we'd like to show you a video, created by Al Tompkins of The Poynter Institute.
(SHOW THE VIDEO)
(Sara): Now, on to the major findings.
Our first finding has to do with the amount of story text read.
One. People in our study selected what they wanted to read, of course, and then ... they read a lot! We were amazed by these numbers.
A big surprise was that a much larger percentage of story text was read online than in print.
(Sara): To look at a comparison, on average, online readers read 77 percent of what they chose to read. Broadsheet readers read an average of 62 percent. Tabloid readers read an average of 57 percent...
Tabloid readers, on average, read a higher percentage of jumped story text than broadsheet readers. Sixty-eight percent of the jump text read by tabloid readers and 59 percent of the jump read by broadsheet readers.
(Pegie): As you can see in our study, one of the most important findings had to do with reading. And, we found that there were two different styles of reading.
(Pegie): One was methodical, the second is scanning.
Here's how we defined these:
Methodical readers tend to:
- Read from top to bottom, without much scanning around the page.
- Read in a full, two-page view when reading in print.
- Online, use drop-down menus and navigation bars to locate stories.
Scanning readers tend to:
- Scan pages, headlines and other display elements.
- Read part of a story, then jump to photographs or other elements without going back to the same place in the text.
- Online scanners jumped around the screen, looking at different elements, then they would eventually click on a story and read.
(Pegie): Two. So, our second finding is that, about 75 percent of the print readers tended to be methodical in their reading.
About half of the online readers were scanners, while the other half were methodical in their behavior.
Online, there was very little difference in the amount of text read between methodical readers and scanners.
In print, methodical readers read a higher percentage of text than scanners.
And this was comparable in broadsheet and tabloid.
Tabloid scanners read a considerably smaller percentage of text.
(Sara): Three. The prototype portion of our study showed the value of alternative story forms as they related to comprehension and retention of information. By alternative, we mean things like a Q&A, a timeline, a short sidebar or a list.
Subjects were asked to read one of six different versions of a story about the spread of bird flu.
Three versions were print and three were online. What you see here are the print versions.
Each included identical information -- fact for fact -- but the story structure differed.
At the end of the test, subjects were quizzed about the story in an exit interview.
In both the print and online, subjects who answered the most questions correctly had read the version of the story with the most alternative structure -- no traditional narrative.
Again ... this was in the prototype portion of our study.
In the eye tracking part of the daily print publications, we also found that this type of story form drew a great deal amount of visual attention, compared to regular text.
On average, we saw 15 percent more visual attention to alternative or "unconventional" text than to regular text. In broadsheet, this number rises to 34 percent.
(Pegie): When we say that an element drew more attention, or was seen "more than expected" we are saying that that element drew "above average" attention, proportionally.
For example, let's assume that there were 20 large photos available to be seen and 100 small photos. Now, let's say that there were 10 "eye-stops" on those large photos and 20 on small photos. (Let me define "eye-stop" for you: Each time someone's eye stopped on a particular element, it was counted as an "eye-stop.")
If you do a calculation based on the average, proportionally you will see that large photos generated more attention, compared to small photos.
Four. In the first eye track study in 1991, we found that larger headlines and photos drew the most attention.
This finding was supported in our study. And larger headlines and photos got dramatically more attention than smaller headlines and photos. In fact, they were the first points of entry in print.
But online is different. It's navigation bars, teasers and story lists that get primary attention.
Five. Here are some things we found out about photos.
Color photos draw dramatic attention in broadsheet, compared to black and white photos.
Live, documentary news photos -- photos of real people doing things in real time -- got more attention than staged photos. Studio or staged photos received little attention.
And mug shots got relatively little attention in any format.
(Sara): These key findings just scratch the surface of the full EyeTrack07 study. There's much more to come about how readers looked at both editorial and advertising material.
You'll find more detail and a broader discussion about what these findings mean at our Poynter conference April 10 to 12 in St. Petersburg, Fla.
The EyeTrack07 report itself, will be released in June.
Go to
eyetrack.poynter.org for more information.
(Pegie): We see so much more out there that needs to be tested in news in print, broadcast and online.
We are interested in future studies using EyeTrack and in partnerships with you and other news organizations.
What we have gathered for this study is a very rich data set. We have almost 600 videotapes of readers reading in print and online that we can go back into to mine even more information.
More and more things come to mind every day.
We're also interested in taking a closer look at the way headlines are written in an engaging way.
Here are some other things we're considering:
- News delivery on large format screens, in high definition, on telephones and even smaller screens.
- Elements of television news, like text that moves across the bottom of the screen, or animated graphics.
- And major innovation in searchability online.
And, of course, we can only imagine what's possible in the future.
We're already talking about hosting a seminar early next year that would create models of news delivery for the future. We'll bring industry visionaries together to brainstorm about what might be ahead. And we plan to use EyeTrack to test those models.
Our April EyeTrack07 conference has already been filled.
But we're taking applications now for a hands-on EyeTrack seminar in August that will further explore how these findings can be used to engage the reader.
Pegie and I look forward to talking with all of you about the study -- we'll be out in the halls and in the lounge. Please feel free to contact us.
We're interested in hearing your ideas.
Thank you so much for your time.
EDITOR'S NOTE: This text has been edited to reflect changes in the findings. Similar changes have been made to the presentation video.
CORRECTION: In the section above regarding alternative or unconventional story forms, the percentage increase in visual attention for broadsheet was incorrect.
