There are several different types of bibliographic database available. The following are a few examples.
Humans read every article and assign MeSH classification terms.
In Ovid: From 1966, ~3900 journals.
Also available from NLM's PubMed.
Indexed by machine. Wide coverage, fast processing, reduced quality of indexing.
In Ovid: From 1996, ~7500 journals:
Processed by machine. Search for all papers which cite a given paper.
Available to McGill community at http://isiknowledge.com. Under Searchable Database Products select Web of Science, then Cited Reference Search.
Provides electronic delivery of full-text articles.
Available with username and password at http://source.cisti.nrc.ca
Indexed by machine, publicly available, ranking by ‘relevance’.
Includes ‘articles from a wide variety of academic publishers, professional societies, preprint repositories and universities, as well as scholarly articles available across the web’.
Coverage patchy, may improve as they work with publishers.
Includes searching of non-free sources, but with abstracts only.
Nature of ‘scholarly’ is not defined by Google.
Importance of using refereed sources.
Dubious nature of some refereeing, especially conferences.
Available on the Web at scholar.google.com (launched 2004 Nov, still ‘beta’ as of 2007 Mar).
Cochrane Library: a collection of databases containing systematic, up-to-date reviews of randomized controlled trials (RCTs). Main component is Cochrane Database of Systematic Reviews (CDSR). ~660 full-text, regularly updated, systematic reviews of the effects of health care, plus the 'protocols' of ~620 reviews currently in preparation.
Best Evidence: full-text access to reviews from top clinical journals on a regular basis, selected as being methodologically sound and clinically relevant, with enhanced abstracts and commentaries.
Medline search results in Ovid can be restricted to the Best Evidence database by applying the "limit" Article Reviews - Best Evidence to the search strategy.
Medline searches restricted to the broader EBM Reviews limit will provide links to records in both the Best Evidence ("Article Reviews") and CDSR ("Topic Reviews") databases.
Further resources from McGill's Health Sciences Library:
Some databases, including Medline, use a highly structured set of subject headings to classify their contents. If such subject headings are available it's a good idea to start your search by using them, rather than just searching for text words.
When you type in a search word or words describing a single concept, the computer will attempt to match what you've typed to one or more of its standard headings. If it's not sure which one is appropriate, it will present you with a list of possibilities.
Make sure the Map Term to Subject Heading
box has a check mark in it, then click on Perform Search
.
You can click on one of the offered subject headings to see it in the context of the subject-heading tree. You can then click on the i icon to see a Scope Note describing what that heading includes.
Once a particular subject heading has been chosen, you will be shown a hierarchical list, showing a parent (more general) subject heading and any child (more specific) subject headings.
Beside each subject heading will be the number of corresponding entries in the database. You can choose a more specific heading if you like.
Note that if you choose a subject
heading which has children, all of the children may or may not be included
along with it. Inclusion of the children is called exploding
the term. Normally it happens by default. If you go into the tree display
then it doesn't happen by default, but you can
check on the Explode
box if desired.
You can either
include All Documents or
Restrict to Focus. The former will give you more
hits, maybe more than you want; the latter will give a more
focussed search, but may miss too much.
You can restrict to focus by checking on the Focus
box.
Introduction to Information Literacy (by McGill Health Sciences Library) includes