1. The Concept Map
Few searches are for single, simple topics that can be expressed as one word; most topics can be broken down into subtopics, or concepts, that are used in setting up the search. An easy way to identify the concepts is:
- Write out the title of what would be the ideal article answering the question being asked.
DISPOSAL OF HAZARDOUS WASTES IN CANADA
- Look at the title and separate the concepts.
DISPOSAL |
OF HAZARDOUS WASTES |
IN CANADA |
- Beside each concept write a few commonly used synonymous terms or phrases. The computer will only retrieve what you ask it to; it is necessary to search for variations of the words used to express the concepts.
DISPOSAL |
cleanup, dumping, recycling, incinerate |
OF HAZARDOUS WASTES |
toxic wastes, industrial wastes, PCB, dioxins |
IN CANADA |
Ontario, Quebec, Alberta, Canadian |
Turning this into a search, the terms listed horizontally will be connected by logical
ORs (broadening the search) and those for each concept will be connected by logical
ANDs (narrowing the search). These operators will be discussed below.
- While few searches get this long, the query in the concept map would develop into a search like:
(disposal or cleanup or dumping or recycling or incinerat*) and (hazardous wastes or toxic wastes or industrial wastes or pcbs or dioxin*) and (canada or ontario or quebec or alberta or canadian)
The need to use such synonyms varies from database to database. If the database only has free text terms, you must attempt to come up with as many variations as the authors might have used. Some databases, such as Medline or PsycINFO, have predefined subject headings which remove much of the need for long lists of synonyms. Don't be afraid to leave out redundant concepts; they can hamper complete retrieval because you are requiring that the words be there. Effective searching is a balancing act between finding too much or too little.
HINT:
A couple of rules of thumb for most bibliographic information systems are:
- If everything retrieved is exactly on topic, you are likely to have missed some references.
- The closer you get to retrieving everything on a topic, the more irrelevant material you will retrieve along with it.
Some people ask about how extensive a list of keywords they should have.
- DO list mandatory terms; What makes your topic unique?
- DON'T list redundant terms as separate topics.
- DO list components of topics. (e.g., fruits or apples or pears or peaches or . . . Most databases don't have anything indicating that an apple is a fruit).
- DON'T list anything that doesn't describe your topic in some way.
- DO make use of features specific to the database such as hierarchical subject codes or weighted subject headings.
- DON'T list all possible synonyms. If you try to get every last reference you are liable to retrieve large amounts of irrelevant material.
- DO restrict some searches to specific fields where appropriate (publication year, author, cited reference).
- DON'T enumerate all aspects of a topic if you want them all. Use WELLS instead of DEEP WELLS or SHALLOW WELLS.
- DO watch for words that could also be phrases. e.g., DATABASE is also DATA-BASE or DATA BASE (a hyphen sometimes counts as a space but not always).
- REMEMBER to include singulars and plurals.
- REMEMBER to include British, American, and other variant spellings.
- Abbreviations should also be spelled out, UK or UNITED KINGDOM.
- Some concepts, such as HIGH TEMPERATURE, are relative and not always well described. Search for all other aspects first, then see if there is some standard indexing that can be incorporated if necessary.
June 20, 2005
Previous Chapter Next Chapter
Comments and Questions are welcome!
Librarian, Information Services and Resources
Last Updated: October 7, 2004
