Observations on the website search behaviour of children

Summary

This paper offers some observations on the ways 9 to 12 year children search for information on websites and how this may differ from the search behaviour of adults.

First, it considers how young children might navigate the graphic interface of computer games and describes observations from a usability study undertaken last year involving children and adults using the Australian Museum site. The paper then outlines a more recent study of how twelve children undertook information-seeking tasks using the site of Commonwealth Parliament House.

It appears that many children do “read” and search websites in ways that are different to adults. Children appear to take a whole-screen approach when searching for specific information on a screen rather than the left-to-right, top-to-bottom scanning techniques more often used by adults.

When it comes to searching a website for information however, children seem less capable than adults at making judgments about the value and effectiveness of the particular search paths they have chosen.

This paper does not attempt to provide an answer to the question of how children use websites, rather the aim is to raise several issues for website designers and managers to consider when developing sites for young children.

Game World

An increasing number of Australian children are brought up with computers, using games and educational software from a very early age, and virtually all will have acquired basic computer skills by the time they leave primary school.

The Australian Bureau of statistics has found that over 90% of all children aged between 5 and 14 have used the internet. And, contrary to popular belief the majority of them say that they use it for educational purposes. A study by the Annenburg Public Policy Center at the University of Pennsylvania in 1999 produced a similar result with a finding that for teens aged 13 – 17 schoolwork had surpassed games as the most frequent online activity (1).

I have an eight year old daughter who enjoys playing computer games. For the last three years I have watched the way my daughter and a number of her young friends approach these games and use the various graphic interfaces they contain when undertaking the required tasks.

Many children come to a computer game (and the web for that matter) with an intuitive knowledge of the basic building block of the interface, the hypertext link. They know that some areas of the screen will be ‘hot’, allowing them to move to another screen with the simple click of the mouse. What appears interesting to me is how quickly they are often able to determine where these hot areas are when playing a game.

I have noticed that when my daughter and her friends are presented for the first time with a relatively complex computer-screen graphic containing many visual elements they often just look at it for a while before leaping in. They appear to be taking it in as a whole picture, before they start exploring it with the mouse. The mouse exploration of the screen elements often appears largely random and rarely starts at the top left and works it way down the screen to the bottom right. In contrast, I suspect that many adults when faced with the same situation will start the process of actively searching the screen earlier (either with the eye or mouse), employing the skimming and scanning techniques they have developed over the years for coping with print publications

It appears to me that children (at least those who are either preliterate or just starting to read) are attracted first to those particular areas of the screen that look interesting or catch their attention.

I have also noticed that the children who are more experienced with computer games seem to have an almost intuitive idea of which screen areas are likely to produce results and preferentially explore these first. It often surprises me how quickly they are able to find the productive links within a complex image. These children also seem to remember the many links they encounter during a game with relative ease. With websites for children however, this is often not the case and I suspect the reason for this difference lies in the greater care and testing that occurs during with the development of most computer games.

I believe educationalists and website developers can learn a lot by observing the way children select and use computer games.

“What the students’ responses and game preferences suggest is that the phenomenon of computer games and their meanings is by no means stable. Nor have we yet identified ways in which teachers and students might work together on such texts without falling into the traps of appropriation and approval/disapproval which bedevil classroom studies of popular culture.”
Computer games, culture and curriculum, Catherine Beavis (2)

Jared Spool from User Interface Engineering has spent many years studying how people use websites to obtain information. Spool found that image maps could cause unforeseen navigational difficulties for adults, leading him to describe image maps as a major potential usability problem (3). No such problems I am sure for a child who has conquered the dark with “Pajama Sam” or joined Madeline on her “European Adventure”. By the time they start surfing the web they are already very familiar with exploring image maps and following the links they may contain.

Australian Museum Site

In June 2001 the Australian Museum launched a revised version of its web site. Prior to the launch, Museum staff extensively tested the user-focus of the new site in-house. The Museum also wished to independently evaluate the site’s effectiveness in the immediate post launch period. I undertook a usability and accessibility evaluation of the website during July and August of that year.

In the past Museum staff had noticed that children and adults often have different objectives or reasons for visiting the site. The Museum was keen to assess how effectively the revised site was able to meet the needs of these two audience groups. As a result, we conducted two sets of evaluations, one using children in school years 6 & 7, the other using adults (primarily teachers and/or parents).

In addition, the Museum had identified several key services they wish to deliver to these two target groups through the site. A series of evaluation tasks, which involve these service areas and are reflective typical site use, were developed in conjunction with Museum staff.

Findings

I found the usability and accessibility of the Museum site to be of a high standard. While the study generated a number of findings and suggestions specific to the Museum site, there were two more general observations that I think are pertinent to this discussion.

First

Children appear to be more willing than adults to explore the different link options on any one page. However when it comes to primary navigational choices, children seem far more reluctant to abandon their first decision and make a new choice.

The Australian Museum site has four prominent primary navigational entry points to the site; About the Museum, Research & Collection, Features and Explore. On the home page, each of these choices has an image with rollover text that briefly describes the content of that section of the site.

Analysis of the behaviour of the child evaluators suggests that after making their first navigational choice they are likely to devote considerable energy to searching options within this navigational stream before reconsidering their primary choice and changing stream. This behaviour could be likened to working within a trench and digging it deeper and deeper in the search for an answer.

For example, when looking for information on holiday activities child LT choose “Features“. She made 6 unsuccessful interactions in this area of the site before asking for help. With the third evaluation task (finding information for a school project) her first choice again was “Features” followed by 15 unsuccessful interactions before being advised to look in the “Explore” area of the site. Similarly, JT made 11 unsuccessful attempts to find information about food in the “About the Museum” area of the site before abandoning the task without investigating another area of the site.

A study by Terry Sullivan et al involving the retrieval of information from two web sites by a group of 12 year olds and a group of 16 olds reported similar behaviour.

“We also took note of an apparent anomaly in our young subjects information foraging behaviour. About half of our subjects persisted in following a known unproductive path three or more times prior to abandoning it. The verbal protocols suggest that a least some subjects were aware of the futility of these repetitive efforts. ‘I know this won’t work, but I don’t know what else to try’ one subject commented.” (4)

These observations suggests that when a site provides a primary (top-level) navigation option devoted to children, which contains leads to all areas of the site that are relevant to them, then children are likely to dig deep in this area when searching for information.

It also highlights the need to provide clearly labeled cross-linking between the different sections of the site at the second and third page levels allowing children to reconsider their primary navigation choices at different stages in the information search and retrieval process.

Second

Care needs to be taken when choosing navigational icons. The Museum site uses a strip of six icons under the search box on virtually all higher-level pages of the site as links to important information or services. One was in the shape of a wine glass.

The Museum was interested in seeing how effectively people were able to find out about the availability of food at the Museum from the site and so one of the tasks asked participants to use the site for this purpose.

The wine glass icon was a link to information about the Australian Museum Society. Needless to say, when it came to looking for information about availability of food at the Museum all the evaluators carefully considered this icon. Even though the rollover indicated the icon was a link to the Australian Museum Society, 75% of the children clicked this icon in their search for information about food. None of the children were able to find the information on the site.

Children like adults see icons being used in the real world for a wide range of functions, but probably rely mainly on Resemblance and Exemplar icons (as described by Jenny Preece, Human-Computer Interactions) (5) and are less capable than adults at decoding Symbolic and Arbitrary icons. Websites also make extensive use of icons and this use should not come into conflict with preconceived notions as to what particular icons represent in the real world.

Australian Parliament House

In order to gain further insights into how children might navigate websites I asked twelve Year 6 children from Stanmore Public School to undertake two information retrieval tasks using the Australian Parliament House site. The tasks were in keeping with the learning focus of their current class work and were developed in conjunction with their teacher, Mr Gordon Parrish. Six girls and six boys participated in this study and I would like to thank them and Mr Parrish for their help.

The participants were given about 6 minutes to do the first task and up to 14 minutes for the second task.

Each participant worked on their own and started at the site homepage for each task. The first two pages for each task were recorded along with the number of clicks (or interactions when using the search facility) and the number of times they returned to the homepage.

Task 1: Who is the Shadow Minister for Reconciliation and the Arts? And, what state is that person from?

Nine out of 12 participants achieved this task, taking between 3 and 11 clicks.

None of the participants who completed the task exhibited the trench digging behaviour I observed with the Australian Museum task. However, two of those who were unsuccessful clearly demonstrated this behaviour making 10 and 14 clicks respectively in the attempt. In each case most of their search effort was confined to a single area of the site.

Two children however, appeared to be reluctant to move very far from the home page returning to it repeatedly. One child, who was able to find the answer, took 11 clicks and visited the homepage 5 times in the process. Another child, who was unsuccessful, took 7 clicks and made 4 visits to the homepage.

Task 1 Clicks/Interaction Record

Number of Clicks	Number of Subjects	Comments
3	3
5	3
6	1
7	1	Unsuccessful. 4 homepage visits
8	1
10	1	Unsuccessful. Trench digger
11	1	5 homepage visits
14	1	Unsuccessful. Trench digger

Returning to the homepage and selecting a new search pathway can be a good search strategy. This sort of behaviour is similar to the “hub-and-spoke” navigation reported by Tauscher and Greenberg (6) and others. The strategy can only succeed however, when it is not done at the expense of adequately exploring the options provided by the initial path. A reluctance to move more than two clicks away from the homepage I characterise as Home Stayer behaviour.

Task 2: There were three Australian Prime Ministers during World War One (1914-1918). William (Billy) Hughes is probably the best known. Who were the other two Prime Ministers during the war?

This task is considerably more difficult. Only one child was able to complete the task unassisted, taking only 3 clicks and less than 4 minutes. This child’s web skills are quite remarkable.

The other eleven participants were provided with a clue about 5 minutes before the end of the task time and nine found the answer with this assistance. In most cases it took between 3 and 5 clicks for them to find the answer after being told which primary choice on the homepage would lead to it.

The twelve participants spent between 3 and 22 clicks (or interactions) on this task.

Five of the participants exhibited the trench digging behaviour described earlier. For example, before being provided with a clue about which area of the site would lead to the answer, one child expended 8 out of 11 clicks on search engine interactions. Another, found themselves at the War Memorial site for 9 out of 18 clicks and another spent 8 out of 12 clicks in the Parliamentary Education Office site.

Two children exhibited the home stayer behaviour described earlier. Before receiving assistance, one child returned to the homepage 5 times in 11 clicks and another 7 times in 12 clicks.

Task 2 Clicks/Interaction Record

Number of Clicks	Number of Subjects	Comments
3	1
10	1
11	2	One trenchdigger
13	1	Trench digger
14	1	With 5 homepage visits
15	2	One trench digger. One – 7 homepage visits
16	1	Trench digger
19	1
20	1
22	1	Trench digger

Other observations:

1. Site Map and Search

None of the participants used the site map for either task. Also, no subject initially used the search facility although two did try it at some stage when attempting the second task.

The site map and the search facility cannot be directly accessed from the homepage, but via a link labeled “find”. This situation undoubtedly made it more difficult for the subjects to locate the search input, raising two possible explanations for their apparent reluctance to use it. First, maybe they were just unable to find search, or second, perhaps children prefer not to use search engines.

I suspect that the second proposition might be the most likely. The reason for this is that many search engines are poorly designed and maintained and very few if any have been designed with an eye to how children might interact with such a facility. The search engine on the Parliament House site is neither easy to use nor particularly effective.

Allison Druin and others from the Human-computer Interaction Lab at University of Maryland have looked at how children understand and use hierarchical domain structures and whether they can conduct complex searches if sufficiently supported visually and conceptually (7). Part of this work involved studying how boys and girls search for information about animals contained in hierarchically nested envelopes with categories of information. They observed that boys seemed to tip everything out and go through the pile on the floor whereas the girls were quite careful in their search style, but at times seemed more interested in browsing rather than finding the answer.

Although when it came to computer searches this apparent difference between boys and girls disappeared, the researchers concluded:

“This leads to the notion that the application should fully support both structured searching and browsing as equally valid and efficient methods of obtaining information.” (8)

The study by Terry Sullivan et al involving the retrieval of information from two web sites by children mentioned earlier, also highlighted the importance of having a well-designed search facility on sites for children.

“Children’s success on the simple-fact questions leads us to conclude that youngsters may know how to look for information in web-based systems, but that they may experience substantial difficulty in knowing where to look. In this context, we posit that orientation and navigation aids are especially important when designing sites for use by children. More concretely, we suggest that topical sites intended for younger audiences would benefit greatly from a simple sitemap or other site-wide navigation and orientation aid.” (4)

2. Above the fold

The first task, which required finding the Shadow Minister for Reconciliation and the Arts, highlighted the importance of keeping information above the fold.

In newspapers the headline stories are always presented ‘above the fold’, that is in the top half of the front page. The same principal applies to web pages, although the position of the fold varies according to the screen resolution and browser used to view the page.

The children at Stanmore School use Apple computers with relatively small screen displays. When undertaking the first task many of the participants quickly found the Parliament of Australia “Who’s Who”web page. However, on their screens the links to the Shadow Ministers were below the fold and most of the children did not scroll and so did not find the link.

Know Your Audience

In conclusion, I would like to comment briefly on the importance of clearly identifying the target audience for a site early in the design process, bearing in mind that the prime users of a site will not always be the same as the intended audience for the site’s content. This is particularly the case with sites that are designed to appeal to very young children where the main site navigator may be a parent or teacher.

For example, the subsite “Kids Korner”, which is part of “Family.com” a site emanating from the US, is not likely to appeal greatly to the under twelves, even though a large proportion of its content is for them. The design of the site is aimed directly at their parents, providing quick access to a range of age-appropriate activities.

On the other hand, ABC TV children’s Playground is designed to appeal to young children between 2 and 16, according to information presented in the parent area of the site. The content of the site however, seems to be heavily skewed towards the under tens.

The extensive use of Shockwave for games on the site slows down performance leading I suspect to an imbalance in the time versus reward ratio for many young users. My young daughter certainly fits into this category.

Two years ago, when she was six years old, I watched as she visited the Playground site following a promo on television. She clicked on a jigsaw puzzle in the games area and waited. The Shockwave logo came up. “What’s this dad?” I told her and said it wouldn’t take too long. Twenty seconds later a 6 piece jigsaw puzzle appeared which she did in a couple of seconds.

She tried another jigsaw. Click. “Oh, it’s that shockwave again.” Another 20 second wait for a two second game at the end of which she abandoned the site. My daughter, who will happily spend an hour or so playing a computer game for the umpteenth time, did not revisit the Playground site for more than a year. And when she did, the games were still much the same, as was her reaction to using the site.

References

1. Turow, J. “The Internet and the Family”. Annenburg Public Policy Center, University of Pennsylvania, 1999.
2. Beavis, C. “Computer games, culture and curriculum”. Page to Screen ed. Ilana Snyder, 1997
3. Spool, J. “Web Usability: A Designer’s Guide”, 1998.
4. Sullivan T., Norris C. et al. “When Kids use the Web: A Naturalist Comparison of Children’s Navigation Behaviour and Subjective Preferences on two WWW Sites”. www.pantos.org/ts/papers/wkutw/
5. Preece, J. “Human-Computer Interactions”. Addison Wesley, 1994.
6. Tauscher, T and Greenberg, S. “How People Revisit Web Pages: Empirical Findings and Implications for the Design of History Systems”. International Journal of Human Computer Studies, 1997.
7. Revelle, G., Druin, A. et al “Young Children’s Search Strategies and Construction of Search Queries”. Human Computer Interaction Lab, University of Maryland. www.cs.usmd.edu/hcil/querykids
8. Druin, A., Bederson, B. et al. “Designing a Digital Library for Young Children: An Intergenerational Partnership”.  Human Computer Interaction Lab, University of Maryland. www.cs.usmd.edu/hcil/kiddiglib