Comments on DOAJ proposed new selection criteria

The Directory of Open Access Journals has posted a draft of their new selection criteria and is requesting comments by July 15, 2013. Kudos to DOAJ for taking this open, transparent approach to their process. Following are my comments.

DOAJ proposed criteria and my comments in bold

We have tried to construct objective criteria that can facilitate compliance verification easily. In order to be listed in the DOAJ, a journal must meet the following criteria:

  • Journal will be asked to provide basic information (title, ISSN, etc.), contact information, and information about journal policies

Comment: make sense!

  • Journal is registered with SHERPA/RoMEO

Comment: this should be a recommended good practice, not a criteria for inclusion in DOAJ. The reason is that registering with SHERPA/RoMEO is controlled by SHERPA/RoMEO, not the journal.  SHERPA/RoMEO is an important service, but it is UK-based and focuses on UK priorities. Few Canadian scholarly journals are listed, for example. Before considering such a suggestion, someone should ask SHERPA/RoMEO if they are prepared to take on the task of including all of the world’s journals on request, including journals from every language used by scholars around the world.

  • Journal has an editorial board with clearly identifiable members (including affiliation information)

Comment: not every high-quality journal has an editorial board; some small journals are managed very well by a single managing editor.  Including a criterion like this may have the unfortunate effect of changing the way that journals are published, and not necessarily for the best. Suggested change to: Journal has a transparent and academically appropriate editorial practice.

  • Journal publishes a minimum of five articles per year (does not apply for new journals)

Comment: this should be struck, for several reasons. This criteria would discriminate against smaller journals. For example, last night in my scholarly communication class a student raised a point about a journal produced by a small community that one year declined to publish any articles, as no submitted articles were seen as meeting the journal’s standards. As open access becomes the default and inclusion in DOAJ essential to marketing, this criteria could end up defining what constitutes a journal. Houghton and colleagues found that the most cost-effective means of providing open access in the long term would involve a peer review overlay over articles deposited in repositories. To facilitate this development we need to allow creativity in how this work is done. Eliminating small journals would not be a good idea!

Another important point is that journals that have ceased to publish should still be made available. DOAJ should work towards noting that the journals are inactive, rather than eliminating them from DOAJ. Otherwise, authors who choose to publish in a journal in part because it is listed in DOAJ may find their work eliminated from DOAJ simply because the journal ceased to exist – a common occurrence even in the print / subscription world. Also, libraries use the DOAJ list to include open access works in library catalogues and serials lists, and dropping ceased journals is a loss of valuable content.

Finally, as may be obvious from the example above of the journal that refused to publish one year, a requirement of a minimum of 5 articles per year may drive journals to publish articles that they would otherwise decline. In other words, this will sometimes be an incentive to publish lower quality articles.

  • Allows use and reuse at least at the following levels (as specified in the Open Access Spectrum, ): 

Comment: please strike any reference to the Open Access Spectrum. This is a conceptual framework for open access that is not shared by the whole open access community. For example, the website points to PLOS journals using the OAS grid and Assess a Publisher or Publication with the OAS grid. This very much reflects a gold or open access publishing perspective which does not entirely leave out green or open access archives, but places it at a much lower priority level. Another consideration is that is open access definitions are opened up to a spectrum approach, there is no reason why others could not propose alternate spectra. For example, in recent discussions in Canada it appears that there are those who confuse national free access with open access. If we entertain a spectrum approach, why not a continuum from free access to a few people to global free access? Not all alternative possible spectra are negative examples like this. For example, a scholar-centered open access spectrum (which would appeal much more to me, as a scholar) might focus on the continuum of time of sharing, from when a research idea first occurs to you through to publication. Similarly, why not a spectrum from immediate open access to perpetual copyright with extremely limited rights – something that we should all remember is the real closed access.

  • Full text, metadata, and citations of articles can be crawled and accessed with permission (Machine Readability Level 4)
  • Provides free readership rights to all articles immediately upon publication (Reader Rights Level 1)
  • Reuse is subject to certain restrictions; no remixing (Reuse Rights Level 3)
  • Allow authors to retain copyright in their article with no restrictions (Copyrights Level 1)
  • Author can post the final, peer-reviewed manuscript version (postprint) to any repository or website (Author Posting Rights Level 2)


Machine readability is another example of a good practice to encourage which should not be required for inclusion in DOAJ. There will be variations in the ease with which different journals can achieve machine readability. Even PLoS uses locked-down PDFs, for example. More research is needed to determine whether machine readability of journal articles is always desirable. For example, if pictures of people are included, does the researcher have rights to permit facial recognition software? With the PLoS locked-down PDFs, do we really want the PDFs unlocked to facilitate data mining – wouldn’t it be much more useful to work towards having scholars share the data as open data, preferably linked to from the journal but housed elsewhere? Sometimes machine readability does make sense and is highly desirable – for example, I’d like to see the default for electronic works in general to be works that can be instanteously translated into the format of the reader’s choice, whether PDF, html, daisy or braille. Here, what is needed is not refusal to include journals in DOAJ if they are not at this standard, but rather education and support to help journals develop this capacity.
Provides free readership rights to all articles immediately upon publication is very basic to the definition of open access; this makes sense. I suggest adding the word “global” to avoid confusion with regionally limited free access, to: “Provides free global readership rights to all articles immediately upon publication”

Reuse is subject to certain restrictions; no remixing. It is good to see that journals that prefer to include some restrictions can be included in the Directory of Open Access Journals, but this statement is confusing and counter-productive. For example, as stated any journal that does allow re-use should rejected, so good-bye to the likes of PLoS and BMC!

Allow authors to retain copyright in their article with no restrictions (Copyrights Level 1)  Comment: it may be useful to encourage author rather than journal copyright retention, however this is not an essential part of open access and may not always be possible or desirable. For example, in the case of works-for-hire, some authors will not be able to claim copyright ownership. Another example came up at a recent conference, where scholars working with First Nations peoples are granting copyright in research articles to the First Nations peoples. A narrow requirement of author copyright retention would tend to prevent innovations in scholarly copyright at a period in time when I would argue that encouraging experimentation (articulating the commons) is optimal. Plus if a journal retains copyright but is clearly open access, the journal should be included in DOAJ.

Author can post the final, peer-reviewed manuscript version (postprint) to any repository or website (Author Posting Rights Level 2)

Comment: suggest add “at minimum” to encourage the common practice of allowing deposit of any version including the final version. Finally, thanks very much to DOAJ, PLoS and everyone else involved in this initiative and the Open Access Spectrum. While as this post likely makes clear I strongly disagree with many of the specifics, I do greatly appreciate all the work that the people involved in these initiatives have contributed towards open access. Update June 13: PLoS participating in the selection criteria team is a conflict of interest, because PLoS is one open access publisher and what they are attempting to do here is to control the definition of open access – if this is accepted, this will give them a competitive advantage over other open access publishers. Reader comments that meet the standard for commenting on IJPE are welcome, i.e. no anonymous comments and if you work for or are affiliated with a journal, publisher, or other initiative with an interest in these questions this affiliation must be stated in the comment.

Other posts on IJPE on related topics include the Creative Commons and Open Access critique series and through the open access definition label.