Saturday, November 19, 2011

Google Scholar Citations

Google Scholar has a new feature, Google Scholar Citations, that allows you to generate a profile page with your papers. It also lists citations and calculates the infamous h-index. Have a look at my profile here to see if it's an interesting feature for you.

Is it?

To set up a profile page, you first need a Google account. If you already have one, it takes like 2 minutes or so. If you set up your page and enter your name, you'll be offered a list of papers that might be yours, that you can then edit. Mine was pretty good, probably because my name is not very common. A few papers seem to be missing, some listed items weren't papers but deceased websites that I wrote a looong time ago, and my research statement also appeared, but by and large it worked well.

The citation count is not exactly the same as on inSPIRE. In some cases Google Scholar counts more, in other cases less. It's not clear to me what causes the difference.

And my dear husband is evidently author of a paper with 3990 citations. Yes, I am very proud of him :o)

11 comments:

Phil Warnell said...
This comment has been removed by the author.
Phil Warnell said...

Hi Bee,

Thanks now I can have a look to see if I’ve missed something you or Stefan wrote among all those listed here. One thing it had me to discover is that you co-authored a paper with Stefan back in 2003 which also happens to be one of most cited for either of you. This then further confirms for me that you formed to be a good combo right from the start.

Best,

Phil

Uncle Al said...

"my dear husband is evidently author of a paper with 3990 citations.

Not to mention being mighty good with the lab work, too. "8^>)

Bee said...

Hi Phil,

Right, Stefan and I wrote a paper together. We shared an office too at this time. Those were the days :o) The paper also taught me a lesson: The title doesn't contain any of the relevant keywords that the topic later became known under, which is the reason why it has happened like a dozen of times that I had to write an email to somebody saying, look, we did that already some years ago. A big mistake I've tried not to repeat. Best,

B.

Phillip Helbig said...

Check out the discussion at http://telescoper.wordpress.com/2011/11/18/google-citations/ .

Massimo said...

Thank you for this link, very useful.
So, regarding the infamous h-index, do you think that the inconsistency between the number quoted by Google and that produced by WebOfScience is due to citations to preprints ?

Bee said...

Hi Massimo,

It doesn't seem to me like preprints make the difference. Both Spires as well as Google scholar seem to list & count preprints without journal reference. Stefan's guess why his article has about twice as many citations on Google scholar as on Spires is that they may count internal reports which are online, but not on the preprint server. As to why most of the other papers seem to have slightly less citations I don't know. (My first guess was omitting self-citations, but that doesn't seem to be the case.) Best,

B.

stefan said...

Citation numbers (and the derived indices) depend strongly on the database you are checking out.

This is especially striking for the Geant4 paper, which is

S. Agostinelli et al.:
GEANT4: A Simulation toolkit.
Nucl. Instrum. Meth. A506 (2003) 250-303.

Citation numbers quoted at different openly available databases are as followed:

Science Direct: 2891

ADS: 1734

Spires: 1466

Google Scholar: 3990

I guess the list of publications evaluated to get the citation statistics is quite different for these four databases.

Spires, for example, has quite a strict focus on publications in high-energy physics and may miss some publications actually citing the Geant4 paper, for example in medicine.

From the list of publications Google Scholar is listing with me as an author, I can see that their database includes a lot of non-peer-reviewed publications. Besides preprints, this includes technical reports, contributions to annual reports, and even slides of presentations. All of this "grey literature" can contain references to the literature, hence contributes to the citations counts.

Spires and ADS also include preprints in their statistics, so preprints alone are not the decisive factor to explain the big difference in the citation number.

But I can imagine that the Geant4 paper has a lot of citations in unpublished technical reports and in areas not covered in Spires or ADS.

Cheers, Stefan

Kay zum Felde said...

Hi Bee,

thanks for the info, even if I have no paper published yet. A paper with 3990 citations is really great. Stefan, I am impressed.

Take care Kay

Massimo said...

I see.
Well, I have gone through a few profiles of friends and collaborators, and I have noticed that in most case the difference is of order 10-20% (higher number yielded by Google Scholar). In many of these cases I have seen instances of "double dipping", i.e., ArXiv preprint and subsequently published article both citing article A, and Google Scholar counting those as two separate citations.
I think this could be conceivably a large source of disagreement between the two, but it seems that it should be fixable (and the author herself has the chance to do it... whether they will or not, it's a different story).

rab said...

Bee: I forgot to thank you for this. Most useful, especially for keeping track of who is reading one's work.