On Wed, 02 Jun 2010 14:50:28 -0700, Christopher D. Green wrpte:
>I just did my own little test. I searched both Google Scholar and
>PsycInfo for the phrase "implicit learning" (your results may vary).
>There was absolutely nothing dodgey about the results Google Scholar
>gave me (I checked the first 50 of several hundred thousand).
Now this is odd. First, it matters whether one uses "implicit learning"
(i.e., with quotes) or implicit learning.
For "implicit learning" (with quotes), google scholar returns:
|Results 1 - 100 of about 16,100. (0.11 sec)
Removing the quotes makes a difference.
NOTES:
(1) Although 16,100 hits are indicated, with 100 references per
page, the list of references ends at page 10 with the last citation being:
[CITATION] Implizites Lernen: Probleme und Perspektiven [Implicit learning:
Problems and perspectives]
A Buchner - Weinheim, Germany: Beltz Psychologie Verlags Union, 1993
When [CITATION} is listed, I believe that the hit is for a document
that has the cites that article in its reference list, that is, it is a
duplicate
count of the document.
(2) Google Scholar does not handle duplicate refernces well. Searching
for publications with Palij in Google Scholar produces these duplicates;
Principles of spatial problem solving.
M Levine, IN Jankovic, M Palij - Journal of Experimental …, 1982 -
psycnet.apa.org
SUMMARY Blindfolded college students learned simple paths either by moving
their
fingers over the successive points of a map of the path, walking through the
path laid
out on the floor, or (with the blindfold temporarily removed) viewing a map of
the path.
They were ...
[CITATION] IN and Palij, M.(1982). Principles of spatial problem solving
MJ Levine, IN Jankovic - Journal of Experimental Psychology: General
[CITATION] Principles of spatial problem &Avin%
M Levine, I Jankovic, M Palij - Journal of Experimental Psychology: General,
1982
[CITATION] Principles of spatial problem solving
M Levin, I Jankovic, M Palij - Journal of Experimental Psychology: General, 1982
This represent 3 duplicates of the first article, all within the first page of
100 hits. I did not check to see if there were additional duplicates on
subsequent pages.
(3) Using the CSA Illumina interface to PsycInfo (NOTE: PsycInfo
has different interfaces, the CSA interface is available at NYU
while, say, the EBSCOhost interface is used at CUNY -- they
do not return the same results on the basis of my prior experience),
searching for "implicit learning" (with quotes) returns:
|1054 results found for: KW=("implicit learning") in PsycINFO
And in the broader area of "Social Sciences:
|1896 results found for: KW=("implicit learning") in Social Sciences Subject
Area
It has been my experience that PsycInfo rarely produces duplicate
references (not counting articles that are republished in books of
reading). In searching the Social Sciences Subject Area, I found
duplicates. So, one needs to know how one's database handles
duplicate entries.
>[snip]
>PsycInfo returned far fewer (just several thousand) items, but the top
>50 were mostly from highly quality journals. The one really noticeable
>difference, however, was that it returned far, far more irrelevant items
>than Google Scholar did. I'm not sure how some of these got caught by
>the search engine. At first I thought I had simply made the mistake of
>not putting "implicit learning" in quotations marks, but when I went
>back and did that, I still got a raft of irrelevant citations.
I'm note sure what criteria Chris is using to define "irrelevant items"
and I'm sure reasonable people might disagree as what is and is not
relevant.
>I also took ISI's "Web of Knowledge" for a spin on the same search. I
>got less than a thousand (885) returns, Again, I seemed to get a lot
>that was irrelevant to the topic I was interested in. The journals
>looked solid, though it seemed that there were more applied journals
>than with the other two engines.
According to info on the NYU library website, "Web of Knowledge"
is described as the following:
Web of Knowledge
Allows you to cross-search ISI databases: Web of Science, Biological
Abstracts, BIOSIS Previews, Food Science and Technology Abstracts,
INSPEC, Medline, and Journal Citation Reports.
My first reaction is my would Chirs search Web of Knowledge when
the Web of Science would have been more appropriate:
Web of Science (ISI)
Web of Science provides seamless access to the Science Citation Expanded®,
Social Sciences Citation Index®, and Arts & Humanities Citation Index™.
A search of Web of Science for "implicit learning" (without quotes) produces:
Results
Topic=(implicit learning)
Timespan=All Years. Databases=SCI-EXPANDED, SSCI, A&HCI.
Results: 2,929
With quote marks:
Results: 846
It is unclear why such differences would occur.
>My conclusion: if I had to go for one, it would be Google Scholar. If I
>were not constrained to one, I'd check PsycInfo too. WoK is best if what
>you want is a list of article that have cited a particular article of
>the past.
Some final points:
(1) Google Scholar does not seem to handle duplicate cases well. If an
articles cited within an article, it appear to turn up as hits on GS. One
would then have to ignore the hits which begin with [CITATION] -- there
does not appear to be anyway to exclude these from the list (if someone
knows how to exclude them, please let us know).
(2) Citation analysis is a tricky business but Web of Science seems to be
the best at it. Consider the following number of citations for the
Levein, Janovic, and Palij (1984) JEPG
Google Scholar: Cited by 197 - Related articles - All 5 versions
(why GS thinks there are 5 versions of this article is puzzling)
Web of Science: 162 citations
PscyInfo (CSA Illumina): 71 citations
Now, as much as I would like to go with the Google Scholar citations,
I'm pretty sure that some of them are duplicates while I'm equally confident
thet Web of Science citation do not contain duplicates.
I have NO idea what how PsycInfo tracks citations but clearly it
doesn't do such a good job.
-Mike Palij
New York University
[email protected]
---
You are currently subscribed to tips as: [email protected].
To unsubscribe click here:
http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=2875
or send a blank email to
leave-2875-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
