Computing a curriculum: descriptor-based domain analysis for educators

摘要

College educators need objective ways of assessing coverage, overlaps, and gaps in courses in their curricula, of validating their present offerings, and of monitoring subject-matter trends. This article presents a new methodology for attaining these goals through the use of descriptors from commercial bibliographic databases. A small set of master terms is chosen to model a college, department, or academic degree program, and then large numbers of descriptors that co-occur with the master terms are downloaded in online retrievals. The number of master terms with which descriptors intersect, and the number of documents these intersections produce, yield weights by which the descriptors’ relevance to the curriculum can be prioritized. Curricula are thus grounded in the subject indexing of evolving literatures. Suitably arranged, the descriptors form a rich outline of the subject matter, both central and peripheral, that coursework in a field might cover. From this outline, the descriptors with the highest weights are extracted as a “Virtual Curriculum,” against which the subject-matter of existing courses can be validated. If individual courses are assigned duplicate high-weight terms, overlaps in course content become visible. High-weight terms that cannot be matched to any existing courses reveal possible curricular gaps. Because online bibliographic databases are dynamic, domain analyses such as this can be repeated periodically to monitor trends and update judgments. A single analyst can carry out all or much of the work; the main costs are for online searching and the analyst’s time. The results are comparable to those produced by a national committee of experts. The study reported here used nine master terms to model the curricula for Drexel University’s graduate programs in information systems and library and information science. Descriptors from the INSPEC and ERIC databases were processed with Dialog search software (principally the RANK command) and SPSS.