Right now I'm using Web of Knowledge, ScienceDirect, etc. to find papers to read and get some metadata. My university has a subscription, so this is nice. As far as I know these databases simply have a selection of journals, download all of the articles from those journals, and display them for me, updating every time the journal publishes. My question is: what percentage of journals are actually represented here? Some people say there are between 10,000,000 and 100,000,000 published academic documents in the sciences. Web of Science has about 50,000,000, and it's not clear how many ScienceDirect has.

A good comparison would be: if you take everything that Elsevier, Wiley, Springer, and Nature have, combined, and subtract everything that Web of Knowledge has, how many things are left?

Another way to phrase it: Of the major publishers (like the ones I listed above), how many are represented in these databases?

  • I am not sure what your question is and I doubt there is a quantitative way to answer your question. I would say that they are quite complete in general. It is well known that Bradford's Law and Lotka's Law applies quite cleanly for most of these academic databases. Could you rephrase your question perhaps?
    – Shion
    May 15, 2013 at 20:33
  • @JeffE Does that include the four journals I listed above?
    – jclancy
    May 15, 2013 at 21:08
  • 8
    if you take everything that Elsevier, Wiley, Springer, and Nature have, combined — then you're still missing everything published by ACM, IEEE, SIAM, AMS, MIT Press, USENIX, ACL, ArXiv, personal web pages, ....
    – JeffE
    May 15, 2013 at 21:11
  • google seems a good database. if the first page does not satisfy you then google scholar is better
    – seteropere
    Dec 2, 2013 at 6:29

4 Answers 4


Previous research has shown strong evidence that there is a language bias in academic databases in the social sciences and humanities[1]. However, evidence for the sciences is mixed: Some cannot find an US bias using the Science Citation Index (SCI)[2], others actually identify also a strong language bias in the SCI[3].

Furthermore, you have to consider differences in publication cultures between disciplines: The humanities and social sciences still approve monographs and edited books as publications which are seldom listed in theses databases.[4]

[1] Bookstein, A./Yitzhaki, M. (1999): Own-language preference: A new measure of “relative language self-citation”, in: Scientometrics, 46 (2), 337-348.
Hicks, D. (1999): The difficulty of achieving full coverage of international social science literature and the bibliometric consequences, in: Scientometrics (1999), 44 (2), 193-215.
Hicks, D./Wang, J. (2011): Coverage and overlap of the new social sciences and humanities journal lists, in: Journal of the American Society for Information Science and Technology, 62 (2), 284–294
Yitzhaki, M. (1998): The ‘language preference’ in sociology: Measures of ‘language self-citation’, ‘relative own-language preference indicator’, and ‘mutual use of languages’, in: Scientometrics, 41 (1-2), 243-254.
[2] Van Leeuwen, T. N. et al. (2001): Language biases in the coverage of the Science Citation Index and its consequencesfor international comparisons of national research performance, in: Scientometrics, 51 (1), 335-346.
[3] Luwel, M. (1999): Is the science citation index US-biased? in: Scientometrics, 46 (3), 549-562.
King, D. A. (2004): The scientific impact of nations, in: Nature, 430, 311-316.
[4] Nederhof, A. J. (2006): Bibliometric monitoring of research performance in the Social Sciences and the Humanities: A Review, in: Scientometrics, 66 (1), 81-100.


The precise extent of scholarly publishing is unknown, difficult to estimate, and the subject of some debate. That said, there is lots of evidence that Web of Knowledge and ScienceDirect are likely only the tip of the iceberg.

That is sort of the takeaway from an article I found in from Library Journal on Online Databases-Online Scholarly Journals: How Many? (Google WebCache because the LJ copy seems gone) by Carol Tenopir and published in February 2004.

Apparently, one of the best sources online is Ulrichsweb which is a serials tracking website that is part of ProQuest. I just did a search for "Academic / Scholarly" and "Journal" and came up with more than 107,000 different results.

  • 1
    Your link to google's webcache seems defect, how about a link to the original article? Dec 2, 2013 at 10:01
  • I've made the change. The article had not been reposted when I wrote this (AFAICT) so I used the cache version instead. This is correct. Thanks!
    – mako
    Dec 23, 2013 at 23:37

Not very. For example, Elsevier is a large publisher, but only still publishes a fraction of the total number of journals and articles in the world. By my count, they publish at least 2,797 journals. They also market Scopus as

the world's largest abstract and citation database of peer-reviewed literature and quality web sources

By my count again, and you may need a subscription to access this information, there are 3230 journals indexed by Scopus. So unless Elsevier publishes 87% of all the journals in the world, Scopus is missing some. In fact we know they are since Pubmed indexes 5,064 journals.

That said, most "reputable" publishers spend the money and/or produce quality products that they get included in the major indices.


An answer for you depends on your purposes and standards.

If you're hoping to rule out the possibility that any published prior art is relevant to a patent claim, no database will be anywhere near the degree of completeness that you would prefer.

On the other hand, every database is complete for some purpose. For example, my understanding is that Web of Knowledge has at least a mere mention (a journal title, volume number, and page number) for every item that has been cited in the core literature since the end of the nineteenth century.

That presumes a definition of core literature that is decided by Thomson Reuters editors in the form of the product's "master journal list" on the basis of Bradford's law. Those seeking an introduction to Bradford's law could start with Sturgeon's law, which apologized for the low quality of much science fiction by observing that "90% of everything is crap." Bradford's law is the more specific, better studied converse in regard to scientific journals, that if we can examine a small but highly valuable fraction of them all, then for many purposes we can ignore the rest.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .