The Specific Facet Search Strategy in ERIC/FirstSearch

This week I performed a specific facet search in the ERIC/FirstSearch database. With this strategy, you find your most specific facet first. Then, you combine facets, one at a time, until you reach your facet that produces the most results. Search results should be checked each time before adding another facet to make sure that the records retrieved make sense.

The topic of interest and naive question are the same as before.

Topic of Interest: Nonuser studies

Naïve Question: What are some of the characteristics of nonuser groups identified by public libraries?

The following table reflects the thesaurus terms gathered from ERIC/FirstSearch.

There were no thesaurus terms found for the first facet of nonuser groups, so I performed a keyword search here.

I then tried each of the following search strings using each facet individually to determine which facet is the most unique.

• SS1 keyword field search: nonuser* OR non-user*

RESULT: 442 records

• SS2 subject field search: public libraries

RESULT: 7,361 records

• SS3 subject field search: library research

RESULT: 6,552 records

SS1 is the most specific facet. I started with that and then added SS3 which yields the second smallest amount of results to get:

• SS4: (nonuser* [in keyword] OR non-user* [in keyword]) AND library research [in subject]

RESULT: 35 records

A final search was done using all three facets to see if anything can be found with the public libraries aspect added.

• SS5: (nonuser* [in keyword] OR non-user* [in keyword]) AND library research [in subject] AND public libraries [in subject]
  

Final result is 12 records.

Citation: Gallup Organization, Inc. (1975). The role of libraries in America. Princeton, NJ: Author.

This is a wonderful technique to find out which facet produces the least amount of results. Also, I was not expecting to see my results get limited so drastically once search statements started being combined because the facet with the most results is used last. However, any additional AND statement can bring the number of results down by any amount.