ER&L 2009: Google Scholar

The Seismology of Google Scholar: Does Google Scholar rock your world?
Pam Howard
Ya Wang
San Francisco State University

From proxy data:
Use of Metalib decreased markedly when they proxied Google Scholar on their Databases list, set it up as an SFX source, and started teaching it in instruction classes.

In the "what a good idea" department:
SFSU proxied their Google Scholar link from their databases page so that Google would always recognize the user's IP address, even if they switched computers or were off campus.

They also have SFX data for Google Scholar as a source from 2006-2008. They show comparison tables for requests and clickthroughs:
2006 - EBSCO is #1, GS is #11
2007 - EBSCO is #1, GS is #3
2008 - EBSCO is #1, GS is #2

They used the Google Visualization Motion chart to display the change of 2008 monthly SFX request/clickthrough for the above two databases. Dumped their requests/clickthroughs into a Google Docs spreadsheet; the motion chart takes that obscure data and puts it in motion over time so you can see the changes. [JBK comment: Awesome.... What a great way to tell a data story.]
http://code.google.com/apis/ajax/playground

MetaLib - theirs is hosted by the CalState Xserver. They have only entered 4 DBs into MetaLib (Academic Search Premier, ERIC, JSTOR, ProQuest Newspapers). The real-time search appears as a delay to users - when they show it live, it's long (~40 sec for only 4 databases). The same search in Google Scholar takes only 0.09 seconds. Which tool is the user going to prefer?

They can't capture all proxy counts for Google Scholar, but do capture them all through their Fed search interface. In the time/use comparison they show, when they proxied Google Scholar and stopped teaching Metalib. They can't count all access to Google Scholar, but they know that use of Google Scholar is increasing, and use of Metalib is decreasing.

ILLiad and Google Scholar:
Once they added ILLiad as an SFX source, use of ILLiad via SFX jumped. Shortly after, they proxied GS and started teaching it. The percent of ILLiad requests coming from Google Scholar through SFX jumped from ~1% in 2006 to over 5% in 2008.

Conclusions:
-SFSU treats GS as a regular database. Usage has increased and GS has become the most popular database in the library.
-Instruction has an impact on GS - Instructors feel it is a good resource discovery tool and have started teaching it.
-No dramatic decrease in the use of other DBs - GS co-exists with other DBs like Web of Science and Academic Search Premier
-Next steps might include replacing the Metalib link on their beginning search page with a link to GS.

Comments

[Dale A]

Shall we proxy our Google Scholar link?

[Jamene]

I think so. The convenience to our users is incredible - never having to tell someone how to set their Scholar Preferences as long as they access it via a proxied link. The perq for us is having the ability to count hits to GS via the EZproxy logs. We would then have two sets of data for GS - an "incoming" count via EZproxy logs, and an "outgoing" count via SFX source data. We could do the same comparison for our SFX source databases. Perhaps we could use all this data to compare usage of GS with usage of subscription databases, but that's debatable. At the very least we'd be placing GS on the same level as our paid databases when it comes to linking. Whether we follow up with instruction of GS similar to SFSU's is another discussion.