I think the Poynter studies actually confirm my own philosophies. My research and observations confirm the results of the Poynter studies so they were not a surprise.

The eye is drawn to stimulating information, to be sure, but the overall cognitive system is not so simple. Even the most green of novices on the web recognize the importance of hyperlinks and headings. It is rare that a web user is looking for an image. Instead, the quest is usually for textual information. Thus, the our cognitive economy has been wired to search for action items (e.g., web applications and links) and rapid navigation aids (e.g., headings, navigational elements, groups of text). Most images only tell a story at the end of the search, when the user has found other (text) content.

There are several caveats of course. For example, images are very important for game and pornography web sites. Thus, they are valuable to those crowds. However, I wouldn't be suprised to find that even users of these image dependent sites would be drawn to links and headings (e.g., HOT DOWNLOADS).

In case you did not read enough of the various article on this interesting topic, here are some more:

Give Them Words, Not Pictures http://www.clickz.com/cgi-bin/gt/cz/perm/perm.html?article=1823

Poynter-Stanford Project (Source Article) http://www.poynter.org/eyetrack2000/index.htm

Eyetrack Online News Study May Surprise You http://www.editorandpublisher.com/ephome/news/newshtm/stop/st050300.htm

I'm curious, what do other folks think?

-- John S. Rhodes (john@webword.com), June 21, 2000

Answers

John,

I'm curious to see the results for specific sites and then against analysis of those sites capabilities. For example, the BBC site, news.bbc.co.uk, where they have experienced photo editors who know how to crop images down, and they also make use of graphics for commentary and feature articles in the RHS column of the front page.

Personally I find images and the RHS graphics on the BBC site very compelling. I know that they catch my attention regularly. I also know that this result is not generally true and sites like CNet I mostly ignore images. Why is that? Can the Poynter study answer that?

I would also like to see breakdown of the data against personal interest. For example, as a Briton living in the US, I visit several news sites regularly besides BBC these include theherald.co.uk, scotlandonsunday.com and scotsman.com. I read these for specific types of news. For example I may wish to know if Celtic won on Saturday. To answer that question, a picture of Henrik Larsson and a headline saying something like "Celtic triumph over Hearts" is enough. I don't need to read the story. On the other hand, if I want to know about the club politics and how the new team coach is settling in then I look for stories about Kenny Dalglish and Martin O'Niell. To answer my curiousity, I need to read the story.

Therefore, I want to see results correlated by Lakers fans against basketball stories and text versus pictures tracked etc. Golfers against pictures of Tiger Woods and so on.

I don't think it is reasonable to draw design conclusions from the averages. Before we can truly draw conclusions we need to know what the goals were for the viewer/reader.

David

-- David Anderson (david@uidesign.net), June 21, 2000.