Competency E

Competency E

Design, query, and evaluate information retrieval systems.

The design of databases should be user-oriented because easy retrieval of information is the goal (Weedman, 2019). Each item in a database is represented in a way that makes the records representing the item findable and usable. The representation of the item contains metadata that gives ways of identifying a record. One of the things that must be represented is the “aboutness” of the information or document that a user is trying to retrieve. This aboutness is isolated by an indexer who creates a list of unique attributes that describe the information being represented. Controlled vocabularies are standardized searchable terms. Vocabularies can also be hierarchical or faceted, meaning that they can be associated with another term or subject. Controlled vocabularies are in contrast to natural language systems that may contain ambiguities or inconsistencies but use language that a user is more likely to use as a query term. Vocabularies can be pre-coordinated or post coordinated. A pre-coordinated vocabulary is done by the indexer and was popular when patrons used physical card catalogs because they coordinate a subject heading with subheadings to narrow the subject down with more specificity. A post-coordinate vocabulary means that the user is coordinating terms at the time of the search. A controlled standardized vocabulary increases the precision of the records being retrieved from a database in part because it disambiguates terms so there is no overlap in what a term represents. This means that you are not retrieving irrelevant records, but only records that match the query. Precision is a way of evaluating the effectiveness of a database. Recall is also important in evaluating databases; it is the degree to which the database retrieves all the relevant data (Weedman, 2019).

Querying databases could require knowledge of filters and Boolean logic. Academic databases have many choices of filters to narrow or broaden results such as year of publication, author, and subjects. Boolean logic that combines the terms AND, OR, and NOT to broaden or narrow results can also be used in combination with the filters. If you want to limit the number of results, combine two search terms with AND. This means each hit must have both terms. To broaden results, combine search terms with OR. To eliminate certain terms from the search you can use NOT. Using quotation marks around search terms allows for querying an entire phrase in that order. Without quotes George Washington could reveal hits with George and Washington, but not necessarily about our first president.

Conclusion

Having the opportunity to design a basic database helped me to understand the theory behind various query methods. I feel confident that I will be able to explain to students why they need to query databases differently than they would Google.

Evidence 1: Creating a Controlled Vocabulary Assignment

I did this project as part of INFO 202, Information Retrieval Design. We were given records and had to create attributes for each record as an individual assignment. Then, I combined my list with my group member’s lists to create a unique set of attributes to represent each record.

Evidence 2: Search Strategies Assignment

I completed this assignment in INFO 210 Reference and Information Services. I used various search strategies in different databases and compared the results.

Evidence 3: Redesigning a Website for Usability

I completed this group project as part of INFO 202. We redesigned a pet store website to increase usability and findability with a clear hierarchical organization and consistent search terms. My role in the project was to create the graphics to represent the original and redesigned hierarchies. I also wrote most of the discussion of the redesign.

References

Weedman, J. (2019). Designing for search. In V.M. Tucker (Ed.). Information retrieval system design: Principles & practice (6^th ed., pp. 119-142). Academic Pub.