Librarians are sometimes asked to identify high-impact open access journals. That’s a hard question, in part because there’s not much consensus about how to measure journal impact. But we’ll try in this and some following posts.
One approach is to look at statistics that measure the average number of citations that have occurred to articles published in the journal, though there are lots of variants on that general theme.
OA ranking using Scopus and SJR
One site that ranks journals is SCImago, which uses citation data from SCIverse Scopus (an Elsevier database).
If we look just at open access journals (which for convenience I’ll define as journals that appear in the Directory of Open Access Journals), then we find that there are about 7300 open access journals. By comparison, Ulrich’s lists about 65,000 peer reviewed scholarly journals of any kind, so about 10% of all journals appear to be open access. A small subset of these journals have been indexed with all of their citations noted. For instance, SCIverse indexes about 18,000 journals, mostly in the sciences and engineering, but doesn’t keep track of whether particular journals are open access. For these 18,000 journals, SCImago computes something they call “SCImago Journal Rankings” (SJRs) based on the number of times citations to the journal have appeared in other indexed journals.
If we combine the two sources of information (OA status from DOAJ, and SJR from SCImago), we find that about 2000 of those 18,000 journals are open access, and hence have SJR numbers. I’m attaching a table listing the 50 open access journals with the highest SJR values as of 1 March 2012. Some observations
- The highest ranked journal (Cancer Journal for Clinicians) has an SJR value of 9.895 and a Journal Impact Factor of 94.333. According to SCImago, it is the 4th most heavily cited journal among ALL journals indexed. With 62 citations per article based on SCImago calculations, its articles on average appear to be the most highly cited of any scholarly journal of any kind.
- The vast majority of highly ranked OA journals are in biomedicine, though there are a few exceptions (Optics Express, PLoS One, New Journal of Physics).
- Looking at the whole sample of 2014, the vast majority of titles are still in STEM disciplines, presumably because Sciverse mostly indexes science journals. So using this data might be useful to identify high-impact journals in Chemistry, but not in Political Science.
| Title | ISSN | SJR rank | SJR |
| Cancer Journal for Clinicians | 1542-4863 | 4 | 9.895 |
| Molecular Systems Biology | 1744-4292 | 70 | 2.349 |
| MMWR Recommendations and Reports | 1545-8601 | 80 | 2.031 |
| Frontiers in Cellular Neuroscience | 1662-5102 | 84 | 1.974 |
| Optics Express | 1094-4087 | 86 | 1.956 |
| . . . |
Ranking based on Article Impact
SJR is only one of many citation-based metrics. eigenfactor.org has computed a similar metric, Article Impact, based on citation data from ISI Web of Science. The list of highest-impact OA journals overlaps but is somewhat different.
More Information
Resources downloadable from my website:
- Tables of the 50 highest ranked OA journals, ranked by SJR and by Article Impact.
- A spreadsheet containing all 2014 OA journals rated by SCImago, with additional information such as the subject area of the journal.
- A spreadsheet containing 625 journals rated by eigenfactor.org
I’ll follow up with other ways to identify high-impact journals in other posts.
My estimate of active, scholarly peer-reviewed journals based on Ulrich’s is just over 26,000. The calculations are a bit tricky as you need to deduplicate, i.e. in Ulrich’s print and electronic journal forms of the same journal are two records. Details can be found in this appendix of my draft thesis: http://pages.cmns.sfu.ca/heather-morrison/appendix-c-how-many-active-scholarly-peer-reviewed-journals/
Interesting. Only 26,000? In that case, the scimago ratings based on scopus citation data cover a larger percentage of journals than I had expected — 19,000 out of 26,000. It’s also a bit puzzling that in that data set only about 2000 (under 11%) are also listed in DOAJ. If both SCImago and DOAJ were random selections of all journals then we’d expect DOAJ to represent 11% of all journals as well, which would predict that there were a total of perhaps 7500*19000/2000 ~ 70,000 journals total.
I’ve noticed in Ulrich’s that their flags for “active”, “peer reviewed” and “scholarly” don’t always quite match my own assessment of journals, so that may be where the differences lie. Possibly DOAJ is much more lenient in its categorization. It would be quite interesting if we were to discover that there aren’t as many OA peer reviewed journals as we had thought.
Do you have any data on what percentage of your 26,000 journals appear in DOAJ, and conversely what percentage of DOAJ appear in your list? That would be a good consistency check.
I agree that deduping ISSNs is a tough problem. The SCImago data includes both ISSN and EISSN when the same journal has both, so it’s already deduped for that.
Using Heather’s methodology today, I am seeing 27,397 active, academic, refereed journals (23,921 print + 3,476 e-only). Of these, 6709 are called open access by Ulrich’s. Of these, 4208 are available online (and 1845 are e-only). These numbers give a somewhat higher percentage of journals that are OA.
I wonder what an OA journal that is not online is like. I think we need to conclude that both DOAJ and Ulrich data probably have fairly high error rates, as well as coding criteria that may not always match what we think they mean. That’s not too much of a problem for some types of question. If we want a sample of the top OA journals in Chemistry then missing a few isn’t a disaster. But if we want to know the 10 top OA journals (perhaps because we’re giving a prize or something), then missing a few is worse than including too many.
Wendy, have you found any legal way to export that list of 6709 journals so that we could look them up in bulk on a citation database like Web Of Science or Sciverse?
As I wrote earlier, if you trust citation metrics then you might also be interested in a bridge I have available.
Open Access is the practice of providing unrestricted access via the Internet to peer-reviewed scholarly journal articles.
Open Access is also increasingly being provided to theses, scholarly monographs and book chapters.
Open Access comes in two degrees: Gratis Open Access is no-cost online access, while Libre Open Access is Gratis Open Access
plus some additional usage rights.
Open content is similar to Open Acces, but usually includes the right to modify the work, where as in scholarly publishing it is
usual to keep an article’s content intact and to associate it with a fixed author or fixed group of authors. Creative Commons
licenses can be used to specify usage rights. The Open Access idea can also be extended to the learning objects and resources
provided in e-learning.
OMICS Group Inc. is one of the Open aceess publisher which provides journals in the form of Open Access.
Hi, thanks for very interesting analysis. Two remarks:
(1) can you provide the doaj-scimago.xlsx file also in CSV, as openoffice and gdocs have problems handling that large spreadsheet;
(2) the last 2 comments – “open access …” – seem to be spam by one of predatory OA publishers, I think you’d like to remove them.
thx
I am trying to access the three documents cited in your original post of 3-2-2012 – with citation statistics for OA journals. The links do not work. We are preparing a presentation for faculty tomorrow, and it would be so helpful to have access to these documents. Can you email them to me or give me URL’s that work? Thanks, Bev Allen, Colorado State University-Pueblo Library
I second Marcin, these .xlsx files are a bit pesky to work with. An open format would be useful like CSV or ODF.