BackTalk: Do We Want or Need Metasearching?

Metasearching (a.k.a. federated searching or broadcast searching) is considered by some to be the next evolutionary step of database searching. Proponents believe that novice users, such as undergraduates, are baffled by the number of databases they have to choose from and need one common interface to meet all their research needs. A common interface, they claim, could wean novice searchers from using the Internet as their primary source for research. Metasearching, then, could become the "new Google."

There is the speed factor to consider. Searching Google's index of six billion web pages only takes a second or two. Performing a metasearch of only 30 databases with a limit of 20 results each is painstakingly slow in comparison, and metasearching provides no relevance ranking. Wouldn't the results take less time if the user chose fewer databases to search? Certainly, but the argument for needing metasearching is based on the assumption that users don't know which databases to select.

If undergraduates who understood both the concept and the results were asked to choose between metasearching or more content, most would choose the latter.

Improving the beginning researcher's selection of research tools is part of the educational process. Students will only rely solely on free Internet sources if they are allowed to. Classroom teachers who desire authoritative references for student research are insisting that at least a portion of the periodicals cited in a paper be peer-reviewed. Most database interfaces provide a way to limit to articles from peer-reviewed journals. This feature is not available with metasearching.

Most undergraduate research can be performed in today's general-purpose databases, which are both robust and user-friendly. Many vendors market databases that index thousands of periodicals over a wide range of disciplines. Their default search is by keyword, which is, by the way, the only search metasearching can currently perform. Full-text content is obvious in the results (not always so in metasearching), and linking to other content providers is possible. Results are displayed in descending chronological order (again, not always so with metasearching). Doesn't sound too difficult, does it?

Usage statistics tell us that many thousands of searches are done each month on academic databases. Most students conduct research on their own. Others have been helped by librarians, classroom teachers, or other students.

We already have many ways to assist database users. Libraries in which students are struggling with online searching need to develop web sites that are well organized to guide users to their OPACs and databases; databases selected both for their content and ease of use; web and/or printed handouts to guide searchers; knowledgeable and friendly reference staff available most library hours; comprehensive instruction programs; and good relationships with classroom teachers.

Some claim that libraries should take advantage of metasearching to serve patrons content with a "good enough" answer. Good enough pretty much sums up metasearch quality, but should our schools and colleges be satisfied with mediocrity? Do we pride ourselves with producing good enough library researchers, good enough graduates?

Information literacy is a laudable goal, promoting a life skill. Selecting a research tool is one of the first concepts that should be learned. Library users who become information literate can take advantage of Boolean operators, thesauri, and more to provide an excellent reference list. Metasearching can, at most, use Boolean operators; it can search words in subject headings but not as subject headings. Metasearching, then, is a step backward, a way of avoiding the learning process.

Database vendors have exerted considerable effort to improve their interfaces and search capabilities over the past several years. The costs for those improvements must be passed on to the customer, the libraries. Should libraries now avoid using the improvements for which they have paid?

Metasearching is costly, with prices varying by vendor and level of capability. In addition to hardware, software, and annual maintenance fees, the reportedly considerable amount of staff time required to configure the software should be factored in by libraries investigating this technology. The choice is between spending the money and time on content and instruction or on a new, largely untried technology promising "good enough" results.