Competency E: design, query and evaluate information retrieval systems

“Core Competency E — design, query and evaluate information retrieval systems.”

Introduction

“Information retrieval” means just what it says, retrieving information. This means that an information retrieval (IR) system could be an internet search engine such as Google or AskJeeves or a library’s OPAC. Even a book can be an IR, since when it is opened, one can retrieve information that is stored therein. The definitive IR for me was the old card catalog of my childhood library. The cards were color coded – subject, author, title; white, canary yellow, and mint green, all typed on a typewriter or hand printed, and each card represented a book or other piece of material in the collection. When I was a child, a trip to the library was really a treasure hunt. The clues were in the card catalog. The hints were the signs posted to the ends of the shelves which listed what was located on those shelves, which Dewey call numbers or the first two letters of the authors’ last names for fiction.  The Wizard of Oz, for example, could be found on the shelf labelled, “BA-BE” in the children’s section. One had to have an idea of what to look for in the catalog – the white subject cards were probably the most perused, but sometimes, one would be looking for another book by a favorite author (those were mint green cards). The OPACs used in place of the card catalogs create faster and perhaps more relevant results (though maybe not as creative or as fun ones), and they work on the same reference and cross reference principle. Keywords, Boolean, or specific searches can all result in cross references and there is less need to be exact in one’s query.

Design

I had to opportunity to learn about designing an IR when I created a database to catalog my mother’s collection of American Girl™ dolls (see E_Evidence_1). I used InMagic’s DBTextworks to create a catalog of the collection. Each field had to be consistent in each entry and each field also needed to be searchable. For example, if my mother wanted to know which dolls she had acquired in a certain year, I would have to make those terms searchable. I chose what I knew made each part of the collection, each doll, unique from the others, but at the same time, I wanted to be able to search by similarities as well. I created a spreadsheet from the database for my mother so that she could see the all that data at a glance without turning on her computer.

Query

When the County library system where I work changed from ACS to SirsiDynex and iBistro for searching the catalog,  that was when I first learned about the query. We attended training in the new systems before it went live so that we could assist our patrons with information retrieval.  The mote specific the query, the more relevant the outcome. A “keyword” only search might display 1000s of results, but the advanced search option allows the querent to use the Boolean search terms and/or/not, as well as specifying in which category to search (i.e., title, author, type of material, language, etc.) Having this system available is useful when one has a patron looking for a title that might be used by several authors for different books. For example, a patron might be looking for  Daughters of Artemis  by L. M. Townsend. If that author is not specified, they may get materials with that title by J.D. Ironmonger, Lauren Wright Douglas, and at least two titles by other authors.

Evaluate

The measure of an effective IR system is user satisfaction. I had the opportunity to evaluate several IR systems. In the first piece of evidence I am presenting here, I compared RefWorks and Google Scholar in the context of discovering which system would be more useful for academic work (see E_Evidence_2). In the next piece of evidence (see E_Evidence_3), I evaluated a user’s ability to retrieve information in a virtual world setting. I based my evaluation on ease of use, relevancy of results, and whether results could be repeated consistently. I found Google Scholar to be the most user friendly. It allows both keyword and Boolean search terms and produced consistently relevant results. RefWorks was more complex and has more features, but with assistance and practice, the features and links are very useful for students. I found the virtual world scenario the most enjoyable IR experience, however it was that very experience that was a drawback in the system because it was distracting from my search. There was really no way to query for a specific result and because some of the sites are created and run by non-professionals, the information presented may be inaccurate, albeit enjoyable.

Conclusion

Librarianship is a multifaceted profession. At the heart of this profession, however, is service. By being fluent in the languages of several IRs, a librarian can better serve those seeking information. In the end, the librarian is her or himself the best IR there is because he or she can access, analyze, and evaluate the accuracy and relevance of the information  that is retrieved through the systems discussed here and others.  I feel that I have the experience and knowledge to use information retrieval systems to assist information seeking constituents in a plethora of settings.