CHI 97 Electronic Publications: Late-Breaking/Short Talks
CHI 97 Prev CHI 97 Electronic Publications: Late-Breaking/Short Talks Next

Age Group Differences in World Wide Web Navigation

Beth Meyer, Richard A. Sit, Victoria A. Spaulding, Sherry E. Mead, and Neff Walker
School of Psychology
Georgia Institute of Technology
Atlanta GA 30332-0170 USA
+1 404 894 2680
{ gt9020a, psg95rs, psg95vs, gt6575a, pswkrnw}


In this study, we examined the effects of age and training on efficiency and preferences in a World Wide Web search activity. Older participants were able to complete most of the tasks, but took more steps to find the information than did younger adults. Factors in this inefficiency were patterns of returning to the home page and revisiting pages that had been seen before during a search. Interactive training improved efficiency and altered preferences. We discuss implications for training and design.


World Wide Web, information navigation, usability, aging, training, older users.

© 1997 Copyright on this material is held by the authors.


As the World Wide Web (WWW) becomes more widely used, site designers must allow for a more diverse set of users. This includes an increasing number of older adults (over age 65). There is little existing research on older adults' Web use, particularly their ability to navigate complex sites. Research with other tasks has shown that older people are willing and able to learn to use computers, but that they consistently have more difficulty than younger people [1]. Therefore, to make WWW documents truly accessible to older users, we must understand their special needs in moving through information networks.

Several types of age-related declines could impede older WWW users. The literature describes a decline in working memory; it shows that systems that make users remember items from one task to another are less usable for older adults [1]. Thus, older WWW users could have trouble remembering which pages they have seen, or how they arrived at the current page. Further, increased age is associated with increased motor noise and slower movements [3]. This could affect the use of scroll bars or image maps. More positively, research has shown cognitive adaptability in older adults [4]. So, they may respond well to training in complex tasks. With proper design and training, older adults can learn to navigate the WWW.

We studied both younger and older adults, in two different training conditions, as they learned to find specific information in a web site. Our observations suggest that interventions, including navigation aids and training, could help both younger and older WWW users to find the information they need.


In this study, participants searched a complex web site for the unique answers to a set of nine questions. Thirteen older adults and seven younger adults have participated. Seven older participants and four younger ones received a "hands-on" navigation tutorial before starting the tasks, while the other participants only received a "hands-off" description of navigation methods. None of the participants had significant WWW experience, though all had computer and mouse experience.

To perform the search tasks, participants used Netscape Navigator (v. 2.0), running under Microsoft Windows (v. 3.1), to browse a set of 19 related web pages. The files were stored on the hard drive of a personal computer, to keep control over the material that participants could see. The content was taken with permission from the Scientific American WWW site [2].

The site included a home page, two articles with links to supplemental pages, several articles without such links, a page of article summaries and links, and two navigation aids that could be used to reach any page (a "site map" and a "table of contents"). Most search targets were within articles and their supplemental pages; one was on the home page. Some pages also included "index tabs," labeled buttons at the top of a page that linked directly to important pages (and were highlighted to indicate the current page).

After the first task, participants started each search from the end point of the previous search. We presented the nine tasks in the same order to each subject. This allowed us to compute the shortest possible path lengths for each task. Participants had 10 minutes to complete each task.

We logged all movements between and within pages. We also asked participants to "think aloud" while searching, conducted exit interviews, and audio recorded the session.


The two age groups differed in search efficiency. Across the nine tasks, older participants took an average of 9.7 steps to find a target, while younger participants took an average of only 6.4 steps. This age difference was most pronounced for the two tasks that required the most steps (four) for an optimal search. For these two tasks, older subjects needed an average of 19.5 and 13.4 steps, while younger subjects averaged 9.3 and 6.3 steps, respectively.

Also, in both age groups, participants with the "hands-on" tutorial seemed slightly more efficient than those without this training. Across age groups, subjects with the "hands-on" training took 7.8 steps on average to find a target, as opposed to 9.3 for those without this training.

Older participants were more likely than younger ones to return to the home page as the first step in a search. Younger participants did this in less than 15% of searches, but trained older subjects did in 23% of their searches, and the figure for older subjects without the training was a big 37%. Several targets were only a link or two away from the endpoint of the prior search, such that returning "home" was often a move away from the target. So this strategy of returning to the home page reduced the older subjects' search efficiency. However, it may also have avoided the need to remember precisely where they had gotten when they had finished the previous search.

Also, older participants were more likely than younger ones to return to pages that they had already visited within the same search task. The average younger participant did this about 6 times during the entire set of tasks, while the average older participant did this almost 19 times. The older users may have had problems remembering which pages they had visited and what was on those pages.

Training seemed to affect users' preferences. Overall, participants with the interactive training used the "Back," "Forward" and "Home" buttons on the Netscape browser for a greater percentage of actions (26% vs. 20%). For the younger users, the trained group used the scroll bar but not the keyboard to move within a page (42% of actions vs. 0%), while the group without the training used the keyboard for a greater proportion of actions than the scroll bar (27% vs. 20%). A similar interaction for the older groups was less pronounced. The trained older subjects did use the keyboard to move within a page for 8% of their actions, perhaps because the precise movements required by the small scroll bar controls were harder for them.

Older participants used "index tabs" more than younger ones (9% of actions vs. 3%), and those who had the training more than those who did not (26% vs. 6%). Also, the only group to significantly use the site map was the older trained group. The training may have encouraged the older adults to take advantage of the site map. However, the younger people (even with the training) may not have felt a need to consult the site map, given their efficiency.


It is encouraging to note that older adults with no WWW experience were usually able to find the information. However, some of the observations raise serious concerns about the ability of older people to use existing Web interfaces. For example, consider the single task that older users typically performed in nearly 20 steps. It was done on a constrained set of 19 pages without true Internet access. To attempt this task - and several did not complete it - older users on average followed roughly enough links to view every page on the system. In fact, when some users were not able to readily locate the information, they sometimes appeared to use the strategy of trying every link on a page. Of course, such long wanderings and random searches would be disastrous on the real Internet. The inability to recognize that a link had already been tried would also be disastrous in a complex search. And the pattern of starting searches at the home page is problematic on the full WWW, where no single home page can be any real index to all that is available (as this home page was).

There are some possibilities for addressing these problems. For example, interactive training on the navigation methods seems to reduce the need to start most searches from one home page. However, it also appears that older users need some better means of seeing where they are and where they have been, rather than constantly having to keep track of this themselves. In future work, we will discuss more specific design features to address these problems. Without them, information on the WWW could be effectively inaccessible to a major segment of the population.


This research was supported in part by NIH grant No. P50 AG11715 under the auspices of the Center for Applied Cognitive Research on Aging, one of the Edward R. Roybal Centers for Research on Applied Gerontology. We would also like to thank the editors of Scientific American magazine for their permission to use material from the November 1996 version of their WWW site.


1. Czaja, S.J. Aging and the acquisition of computer skills. In Aging and Skilled Performance, W.A. Rogers, A.D. Fisk, and N. Walker, Eds. Lawrence Erlbaum Associates, Mahwah, NJ, 1996, pp. 201-220.

2. Scientific American. Available online as

3. Walker, N., Philbin, D.A., and Spruell, C. The use of signal detection theory in research on age-related differences in movement control. In Aging and Skilled Performance, W.A. Rogers, A.D. Fisk, and N. Walker, Eds. Lawrence Erlbaum Associates, Mahwah, NJ, 1996, pp. 45-64.

4. Willis, S.L. Improvement with cognitive training: Which old dogs learn what tricks? In Everyday cognition in adulthood and late life, L.W. Poon, D.C. Rubin, and B.A. Wilson, Eds. Cambridge University Press, Cambridge, MA, 1989, pp. 545-572.

CHI 97 Prev CHI 97 Electronic Publications: Late-Breaking/Short Talks Next

CHI 97 Electronic Publications: Late-Breaking/Short Talks