Children's Issues in Online Catalog Design

Elisabeth A. Riba
Simmons College

Introduction

Over the last decade, libraries have made a massive shift from card catalog indices to online public access catalogs (OPACs). The trend has been so pronounced that in 1997 the Library of Congress discontinued its MARC catalog card service.

In this rush to computerize the catalogs, few OPACs adequately take into account the different demands of young users. Elementary school children make heavy use of the library, both for pleasure reading and to complete assignments, yet most OPACs appear designed for adult or academic users. How can we expect youngsters to use the same tools as adults? The gap between first graders' abilities and those of fifth graders is so great that even these age groups may be better served with different systems.

For children to use OPACs as effectively as adults, we need to change many of the basic assumptions.

Problems

OPACs today are primarily text based. This makes a certain amount of sense, since the information is also primarily text. However, elementary school children are new readers. The standard font on VT100 terminals is smaller than what's used in children's books, making it harder to read. Use of long words and complex phrases can also confuse users.

For example, The DRA Web of Library and Information Services, available at http://www.micds.pvt.k12.mo.us:8000/, includes the collection of three libraries, grade school through high school. The first page of this OPAC says "To activate any function in DRA Web, select the picture of what you want to do. For detailed instructions concerning this page, choose Navigating This Page." I analyzed the text from "Navigating This Page" with a grammar checker That screen received a Flesch grade level of 14.5, Coleman-Liau grade level of 14.4, and a Bormuth grade level of 10.9. This is hardly accessible to elementary school students who need the most assistance. Children benefit from good online help, but most OPACs are deficient in this area.

Once past the obstacle in the display and language, children face problems with input. Keyboard entry can be a barrier for people who haven't yet learned to type. Hunt and peck is an inefficient process which slows users down. Moreover, keyboards are sized for adult hands, so even those who know how to type properly may not be able to reach all the keys.

To get around the difficulties with typing, some OPACs have implemented graphical interfaces that use mouse or touch screen navigation. But children also have less manual dexterity than adults, so systems requiring very precise selections can be just as frustrating as keyboards. Plus, this format requires children to identify icons and recognize what screen elements can and cannot be selected.

Usually, these graphical interfaces are hierarchical browsers, using multiple screens to narrow down search terms. This kind of structure makes it easy to get lost. Solomon (1993) described one poorly designed OPAC by noting "a failure in alphabetic selection along the way would add additional levels to the trail of screens before a child could reach a satisfactory end" (p. 247)

These trees follow the assumption that all users will divide the information in the same way. Such structured searches may work when children already have some guesses where to find the material, but can be much more difficult with new information. Project Seed, which formed the basis of the oft cited Science Library Catalog, examined children's ability to categorize material. Borgman, Chignell and Valdez (1989) found even older children did not match the exact groups devised by the researchers. Most children classified whales with fish, rather than with mammals. If mammals is the only path to whales, then children will probably not find those materials. When asked to label their groups, children devised some very different terms. Air, carbon dioxide, oxygen and shade were classified parts of nature that are invisible while cactus, rain, snow and water were things you can drink. When dealing with unfamiliar terms, the children made even stranger groupings. Fifth graders placed herring with birds, possibly confusing the fish with herons. Likewise, stamen and stomata -- plant terms -- were clearly mistaken for the word stamina, when children grouped them under things that give you feelings or a lot of energy.

In these cases, children would have more luck if they typed the word in a search, rather than browsing for it. However, keyword searches have their own complications. Today's keyword searches require exact matches with rigid subject expressions that are not always intuitive. Solomon (1993) reported "the gap between user-generated search terms and the indexer-generated subject descriptors is aggravated by the variety of terms that searchers use to explore the information structure of the database" (p. 246). The articles describe dozens of searches that appear logical, but return nothing. Fasick (1994) noted that searching for poems, Colonial clothes, or famous Indians will all fail, because LCSH uses poetry and clothing, and does not use the term famous (p. 57).

Even when children know what to search for, misspelled words will also lead to incorrect results. Borgman, Hirsh and Walter (1989) observed that children's spelling skills do not begin to improve until fifth grade (p. 664). Even words the children already know can be invalidated by poor typing skills. Thus, the very method of inputting data increases the chance for error.

In comparison, card catalogs are much more forgiving. Users can often find relevant results, even from incorrect searches, by looking at adjacent entries. When poems is unsuccessful, it's easy to flip a few more cards and discover poetry. However, when OPACs can't make a match, they generally return no results or give error messages. This can be very discouraging for users. Fasick recounted "in one example a girl was trying to find out more about where foxes lived. After a while she found two hits. . . . This particular student's reaction to this dilemma was to switch her topic" (p. 57).

These issues are part of a larger problem, that affects both hierarchical and keyword searching. Although OPACs can be a useful tool, they were designed with the assumption that the users know how to search. However, these are skills learned through experience, something young children often don't yet have. OPACs expect the searcher to understand how to phrase their questions. Not only must users know which terms to use and how to spell them, but also how to combine them. Entering a natural language query in a Boolean search engine will not work.

Most search engines utilize Boolean searches, but these are not intuitive. In conversation, the word and is used like the Boolean term or. This leads to a lot of confusion. "The idea that using 'and' with two keywords results in a broader search, while using 'or' results in a narrower one seemed to be accepted by many children in Irving's study, whereas the opposite is actually true" (Fasick, p. 58). If a student searches for cats and dogs, they will only find works containing both concepts.

But once the OPAC returns results, the user then must decide what to do with them. If an experienced searcher receives too many results, there are ways of narrowing them down to find the most appropriate. Children don't necessarily have those skills. Determining which sources are relevant can save a lot of time, but is a judgment call.

Over time, users can learn effective strategies for OPAC searches. However, there's one final hurdle. There is no industry standard for OPACs. Card catalogs function the same way everywhere. Users only had to learn the skills once, and could then successfully search any library. Not so with OPACs. The secrets to search at the school library may not apply to any other library the person visits. WebCats (http://library.usask.ca/hywebcat/) is an online directory of OPACs. It lists over sixty vendors, and I have found significant variation even within different OPACs from the same company.

Current solutions

The issues children face when using OPACs are well known. All the articles said basically the same thing. Busey and Doerr (1993) summarized the difficulties best: "Information access breakdowns occurred due to problems with spelling, typing, spacing, punctuation, search syntax, subject headings, search strategies, classification schemes, and shelving arrangements" (p. 78).

However, surprisingly few OPACs have specifically addressed these issues. Most web-based OPACs in use by K-12 schools look the same as OPACs for adults. There are a few OPACs specifically designed for children, among them the Kid's Catalog of the Denver Public Library and the Science Library Catalog.

These OPACs were specifically designed to enable children to search the library unaided, and the end results were surprisingly similar. Both OPACs make heavy use of graphics, and prominently offer help on every screen. Both use a point-and-click hierarchical browsing for users to narrow down their selection. The Kid's Catalog has a keyboard for keyword searching, but early versions of the Science Library only allowed mouse navigation. The most recent version of Science Library has added typed keyword searches, because research showed that older children preferred that method. Once a book is selected, both OPACs offer maps pointing to the location on the shelves. The interfaces look quite friendly and accessible, and both libraries have made frequent revisions in response to patron comments.

As good as these systems are, they are only two OPACs. Most OPACs remain adult-oriented.

Proposal

So, what can be done about this situation? Many of the technological innovations of the last five years, especially the explosion in search engines, can be used to improve OPACs. After reading the research, I have a good idea what is needed.

I propose a system that would allow users to easily switch font size to ensure legibility. Input would use both mouse and keyboard; and designers might try to adapt some old laptop keyboards for smaller hands. For children who cannot type, text entry could also be accomplished by clicking on the desired letters onscreen.

Rather than limit users to either browsing or keyword searches, I recommend a Yahoo-like interface, which allows for both. Users could navigate down the hierarchy as far as they are able, and perform keyword searches of the entire database or just that portion of the hierarchy.

Every screen should have some help text, with a button for more detailed assistance. Online tutorials could ease novices over the learning curve, with scavenger hunt games to help hone searching skills.

Spelling checker technology can minimize the problems of misspelled words by suggesting alternatives. Online thesauri can take the user's query to suggest related LCSH terms and offer broader, narrower and related terms to help find the exact match. Linking the OPAC to circulation could let users restrict their searches to what's currently in the library. Relevancy ranking can sort the results based on which books are checked out the most, assuming the best books are generally the most used.

We should be able to adapt the technology to suit the users, rather than requiring the users conform to the technology's expectations.

References

Borgman, C. L., Bower, J., Auth, M. J., & Krieger, D. (1989, October 30 - November 2). From hands-on science to hands-on information retrieval. Proceedings of the 52nd Annual Meeting of the American Society for Information Science, 26, 96-103.

Borgman, C. L., Chignell, M. H., & Valdez, F. (1989, October 30 - November 2). Designing an information retrieval interface based on children's categorization of knowledge: a pilot study. Proceedings of the 52nd Annual Meeting of the American Society for Information Science, 26, 81-95.

Borgman, C. L., Hirsh, S. G., & Walter, V. A. (1995, October). Children's searching behavior on browsing and keyword online catalogs: the Science Library Catalog project. Journal of the American Society for Information Science 46(9), 663-684.

Busey, P., & Doerr, T. (1993, Fall). Kid's Catalog: an information retrieval system for children. Youth Services in Libraries, 77-84.

Edmonds, L., Moore, P., & Balcom, K. M. (1990, October). The effectiveness of an online catalog. School Library Journal 36(10), 28-33.

Fasick, A. M. (1994). Children's use of information technology. In Encyclopedia of Library and Information Science (Vol. 55, pp. 51-69). New York City : Marcel Dekker.

Hirsh, S. G. (1997, Spring). How do children find information on different types of tasks? Children's use of the Science Library Catalog. Library Trends, 725-745.

Hooten, P. A. (1989, Spring). Online catalogs: will they improve children's access? Journal of Youth Services in Libraries, 2, 267-272.

Solomon, P. (1993, June). Children's information retrieval behavior: a case analysis of an OPAC. Journal of the American Society for Information Science, 44(5), 245-264.

Solomon, P. (1994, Fall). Children, technology, and instruction: a case study of elementary school children using an online public access catalog (OPAC). School Library Media Quarterly, 43-51.