"Search Engines" and the Databases You Tell Them to Search (Click this hyperlink to see it graphically explained.)

        Search engines are programs that sort "text strings" looking for matches with the text strings you put in the search windows.  "Text strings" are not "words" or "phrases" to a search engine because it does not parse them like the human mind does.  For that reason, they do not distinguish between "root" as something belonging to a plant and "root" as the product of dividing a number by itself (i.e., "square root," "cube root").  Search engines look only at the pattern of letters and spaces.

        Databases of scholarly articles are compiled by human beings who have gathered them from peer-reviewed publications in specific disciplines or to cover a range of related disciplines.  Database owners usually sell their "product" through search engine companies with whom they sign exclusive contracts, such as EbscoHost's deal with the MLA International Bibliography or JSTOR's "Language and Literature" or "History" journal collections.  You can ask the search engine what journals it covers, and experienced users will make sure that they are using the right search engine based on that coverage.  For instance, though Ebsco, Wilsonweb, and Infotrac all cover language and literature, only EbscoHost covers the MLA Bibliography, which is the standard reference for the field.  Ask your instructors which search engines and databases they use before you start your research, and you can eliminate a lot of uncertainty and wasted effort!

        Because most Internet users are used to searching for information using commercial search engines like Google or Yahoo, they do not differentiate clearly between a search engine and a database that the search engine searches.  For instance, though most users do not know it, Google silently searches the contents of WorldCat, the OCLC catalogue of the world's library catalogues, for every search they perform.  They did not ask Google to do this.  Nor did they tell Google to search the Russian pornographer's site, or the French medievalist's site--Google's search algorithm is largely mathematical and does not discriminate among sources for source quality.  Users must do that.

        Academic search engines, however, usually have to be told by users which databases to search.  This is because they are primarily designed for users who know exactly what disciplines (viz., "majors") they want to search in and do not want other disciplines' databases included because that will produce false hits (e.g., the biology researcher who accidentally includes a mathematics database in a search for the term "roots").  Our library currently (2007) subscribes to Ebsco, WilsonWeb, and Infotrac, Lexis-Nexis, and WorldCat.  Unfortunately for us, even "academic" search engines have included non-academic, non-peer-reviewed databases in their search capacity to help sell their product, but all serious academic search engines will include in their search windows a check box or pull-down menu that enables you to restrict the search to "peer-reviewed only" sources.  That is how you keep the trash out of your search.  The exception is Lexis-Nexis, which serves journalism and legal scholars, because they routinely need to know what is in the public press as part of their jobs.  For peer-reviewed journalism or law opinions, though, they turn to JSTOR's "Journalism" or "Law" sources, or to the "peer-reviewed only" searches in the other three commercial engines.  WorldCat searches library catalogues, so its results usually used to be used to locate print resources, though all libraries are increasingly stocking their collections with "electronic editions" of books and digital images of archival materials.