From: "Sanders, Susan U." <[log in to unmask]>
Date: Sat, 15 Sep 2012 22:32:12 +0000

For clarification in speaking about searching PubMed.gov, and mapping
to MeSH, try this:

Use the words heart attack (without quotes).  I retrieved 186,831
citations. Look in the Search Details at your query translation and
you'll see the way that the PubMed system ran the search, like this:

"myocardial infarction"[MeSH Terms] OR ("myocardial"[All Fields] AND
"infarction"[All Fields]) OR "myocardial infarction"[All Fields] OR
("heart"[All Fields] AND "attack"[All Fields]) OR "heart attack"[All
Fields]

Try using the MeSH subject heading Myocardial Infarction.  I retrieved
136,867 citations.

PubMed uses "Automatic Term Mapping," see here for explanation:
http://www.nlm.nih.gov/bsd/disted/pubmedtutorial/020_040.html

If you use heart attack, PubMed "maps" the term to MeSH and also
searches to retrieve phrases and individual terms that are in All
Fields, hence more citations than if you used a MeSH term.  What MeSH
gets you is a more specific set of results.  What keywords gets you
(all fields) is a more sensitive retrieval, but with more irrelevant
citations.

Best,
Susan

Susan Sanders
Clinical Medical Librarian
University of Missouri Kansas City


________________________________________

From: Linda Schwartz <[log in to unmask]>
Date: Fri, 14 Sep 2012 07:19:17 -0400

I find one point in Cynthia's posting misleading. PubMed does not
convert keywords like "heart attack" to MESH headings. MESH headings
are a more powerful way to search and can be searched through PubMed.
Try both searches in the basic search box and you will get at least
5,000 hits difference in retrieval. (Using quotes around each phrase
makes the retrieval difference even more extensive!)

I think Paul's statement is spot on. GScholar content is less reliable
- not unreliable but less reliable - requiring MORE analysis that I
personally do not hear users doing. Also important is the algorithm
upon which retrieval in Google Scholar is based rewards popularity -
even if that popularity is based on a bad work that has been cited
repeatedly because it is so poorly done!

Educating the user is important. GScholar is just one tool in the box
and you need to know how each tool works, its strengths, its
weaknesses, and its coverage and apply time for analysis of results.
Unfortunately, users do not educate themselves but rely on the tool to
translate their words into an effective search - without bothering to
find out if it really does so.

Linda Matula Schwartz, MDE, AHIP
Director, Library Services
Lehigh Valley Health Network
Allentown, PA 18105

-----Original Message-----

From: Cynthia Porter <[log in to unmask]>
Date: Wed, 12 Sep 2012 12:44:39 -0700

Thank you Ken, I'm glad you asked this question.  I've been thinking
about Google Scholar lately.

I work in a medical library and I like searching in PubMed because it
converts keywords to Medical Subject Headings (MeSH), so it goes
beyond a keyword search (e.g., I search "heart attack" but I get all
the articles with the MeSH "Myocardial infarction").  Doesn't Google
Scholar depend on matching keywords?

Several times I have encountered full text links to journal articles
in Google Scholar for items that are not available online from our
library.  I think that some of these resources were posted without
publisher consent.  Is it wrong to refer a student to this online
resource?  I've sent the link to students, but not a copy of the
document.  This practice just doesn't feel right to me.  I guess I
could tell the student to search it themselves in Google Scholar, but
I don't like it when people tell me to "Google it."

Cynthia

Cynthia Porter
[log in to unmask]
Distance Support Librarian
A.T. Still Memorial Library, Arizona
Mesa, AZ 85206