When APA first designed its Web site, it organized the wealth of information available on the site according to APA's own organizational structure. Click on the "Practice" section, and you would find information of interest to clinical psychologists. Click on "Public Interest," and up came information on APA's advocacy efforts. And so forth.

The only problem? The site's users--who range from high school students and other members of the general public all the way up to clinical and academic psychologists--couldn't find what they needed.

That's quite typical, says psychologist Dick Horst, PhD, president of a Silver Spring, Md., company called UserWorks, which tested the site for APA last spring.

"Designers and the in-house people who contribute the content to a site are well acquainted with the structure of their organization, but users may not be and don't care to be," Horst says. "They don't care what department they get their information from. They just want their information."

At the company's recommendation, APA reorganized its home page into areas specifically targeting psychologists, students and the public. The newly streamlined home page that debuted last summer is just one example of how human factors experts like Horst are improving the usability of the World Wide Web.

As the number of Web sites skyrockets, psychologists are co nducting research, designing sites and testing sites that are already up. No matter when they get involved in the process, their goal is always the same: making sure users can get what they need without being distracted by a site's bells and whistles.

A user-centered approach

Web sites have traditionally been designed by graphic designers, computer programmers and marketing departments who are often interested in getting the sites up as quickly as possible.

That's a problem, says Donald A. Norman, PhD.

"There's a real need for psychology here, for an understanding of human behavior," says Norman, a professor emeritus of cognitive science at the University of California at San Diego and a former executive at Apple Computer and Hewlett Packard. He is now president of learning systems for a Web-based learning company called UNext.com.

Without psychology's involvement, says Norman, Web sites often ignore the very people they ostensibly target. Their structure is often unclear. Fancy graphics slow the user down. And information is hard to find, particularly for those interested in comparison shopping.

Now that organizations are realizing that Web users have millions of other sites to choose from, they are starting to embrace what Norman calls human-centered design--an approach that puts people first and technology second.

That can make a big difference, especially since psychologists' recommendations often counter the conventional wisdom. Take a phenomenon called "banner blindness," for example. Although Web site designers have long assumed that the most important element on a page should stand out in some way, psychologist David M. Lane, PhD, of Houston's Rice University, has shown that that approach actually backfires.

Lane and co-author Jan Panero Benway, PhD, found that users looking for a specific piece of information tended to overlook a large button at the top of the page offering just that information. Instead, they searched the options available in the standard navigation menu at the center of the page. Although some designers speculate that users have simply learned to ignore elements that resemble ads, Lane points out that the original finding occurred among Internet novices using an ad-free company intranet.

The real issue is perceptual grouping, says Lane, who suggests that designers incorporate critical functions like "Make a reservation" into the main menu rather than physically isolating them.

"Psychologists often have a difficult time convincing software engineers that something doesn't work, until they take them behind one-way windows or show them videotapes of people trying to use the site," says Lane, an associate professor of psychology and statistics at Rice. "Psychologists have a good feel for how hard human behavior is to predict."

Hands-on work

Psychologists are also getting involved in actually building sites, although it's their expertise rather than their degree that attracts customers.

"Usability is becoming the big buzzword these days," says Catherine D. Gaddy, PhD, a project director at Human Factors International, Inc., a Baltimore-area company that offers Web site design, testing and training. "We don't say, 'Hi, we're psychologists.' We say, 'Hi, we're usability experts who know human performance.'"

The design process begins with what Gaddy calls "Human Factors 101," an analysis of the site's proposed users and the tasks they will be coming to the site to accomplish. Drawing on her training in experimental psychology, Gaddy then creates task-flow diagrams. Building out from there, she designs a black-and-white, text-only site, then gradually adds color and graphics to make the site attractive.

A big part of Gaddy's job is wrestling with graphic designers, who are sometimes more excited about multimedia film clips and other technological innovations than usability. When designers lobby for the use of animation, for instance, Gaddy has to point out the human performance data proving that there will always be some part of the eye and brain paying attention to these elements rather than the site's message.

"Glitz is fun," she says, "but it can be distracting."

Of course, psychologists sometimes get called in after the fact.

"Often the first phone call comes when an organization is just weeks away from a launch," says Dick Horst of User- Works. "They're getting nasty comments back after releasing a beta version or their help desk is being flooded with calls."

Like design, testing begins with an analysis of a site's users and tasks. Testers sometimes watch study participants tackle assigned tasks from behind a one-way window. Or they may engage in dialogue, asking participants to think out loud as they attempt to use various site functions.

Although the process obviously isn't as rigorously controlled as a scientific experiment, the testers rely on their background in experimental psychology to interview participants, collect data and avoid biasing results. Most sessions are videotaped.

"People enjoy participating in these studies," says psychologist Marguerite W. Autry, PhD, the human factorsproject engineer who tested APA's site at the UserWorks lab. "For once in their lives, someone's listening to what they have to say."

Rebecca A. Clay is a writer in Washington, D.C.