In yesterday’s posting I discussed the enduring ambiguity around the concept of Open Access “mandates,” many of which have nothing mandatory about them. One prominent example of how this ambiguity plays out and creates confusion is the ROARMAP registry. ROARMAP stands for Registry of Open Access Repositories Mandatory Archiving Policies — this despite the fact that a very large percentage of the policies it collects are not, in any meaningful sense of the term, mandates. In fact, a considerable number of them are not even policies — some constitute mere guidelines and others are only proposals (some of which are identified as such in the registry and some not).

In this posting I’ll provide examples of the pervasive errors and misrepresentations in ROARMAP, and then briefly discuss why I think they matter.

ROARMAP provides a list of several hundred policies or proposals of various kinds; at the top of the site’s main page these are broken down by rough category, all of which are characterized as “mandates”: “Institutional Mandates,” “Thesis Mandates,” “Proposed Sub-Institutional Mandates,” etc. The list itself is organized alphabetically by country, and for each entry links are provided to the relevant repository and to a page that offers “policy details.”

Click through to the “Policy Details” page for a few of these entries, however, and one of the problems with the registry quickly becomes apparent. Those who maintain it have not made fact-checking easy; in the case of a great many entries there is in fact no link at all to the text of the policy summarized on the page, and in many cases where a link is provided, the link is incorrect or dead. Some of these mistakes are certainly unintentional errors, and are neither surprising nor particularly troubling; the database appears to be a labor of love carried out by busy people with demanding jobs. Inadvertent errors and out-of-date content are both to be expected and can easily be fixed.

But ROARMAP is characterized by another form of pervasive error, and this one offers a bit more cause for concern, since it always has the same effect: to exaggerate the “mandatory” nature of the policy in question. This error consists in the mischaracterization of institutional OA statements by selective quotation.

For example, consider the case of Oregon State University. The ROARMAP entry quotes its OA “mandate” as follows:

An Open Access policy was unanimously approved by the Oregon State University Faculty Senate at the June 2013 meeting. The policy grants Oregon State University a non-exclusive license to exercise any and all rights under copyright relating to its faculty’s scholarly articles, in any medium, and to authorize others to do the same, provided that the articles are not sold for a profit. The policy further directs faculty to submit an electronic copy of the accepted (post-peer review, pre-typeset) manuscript of their articles to OSU Libraries for dissemination via its institutional repository.

That is the end of the extract provided by ROARMAP. But the policy itself goes on to say that “at the request of a Faculty member… (OSU) will waive application of the license for a particular article” — a very important qualifier that is left out of the ROARMAP summary. In practice, this means that an Oregon State author can opt out of the policy’s requirements at his or her discretion. In other words, the “mandate” is effectively optional, though one would not know that from reading the ROARMAP summary. A similar disconnect between the ROARMAP summary and the institution’s actual policy can be seen when comparing summaries with policy texts in the cases of the University of Hong Kong (summary vs. text), Connecticut College (summary vs. text), Trinity University (summary vs. text), Rollins College (summary vs. text), and the California Institute of Technology (summary vs. text). These are only a few of many such examples.

Interestingly, the most egregious disconnects between ROARMAP summaries and policy texts seem to occur when the summary was created by someone from ROARMAP; where the summary language or extract is credited to a representative of the institution in question, the text generally (though not always) includes either the entirety of the policy statement or a more accurate abstract of it.

To be sure, there are some entries in ROARMAP that do accurately reflect the non-mandatory nature of the policy in question. But this brings us back to the nagging question: why is the systematic exaggeration necessary? Where the policies (or guidelines) make participation in the program optional, why the insistence on referring to them as “mandates”?

I posed that question last week, by email, to Tim Brody at the University of Southampton, the contact person given on the ROARMAP “About” page. As of this writing, I’ve received no response — if an answer does eventually come, I’ll update this posting. (And I have no doubt that discussion of this question will come up in the comments.)

It does seem to be in the nature of OA advocacy to emphasize constantly the movement’s “dramatic growth” and its “inevitable” dominance of the scholarly communication sphere. Some go so far as to say that we should no longer discuss OA as something that may eventually “become the norm,” but rather should “begin acting in ways that acknowledge that Open Access is the norm.” ROARMAP is a product of the OA advocacy community, so it’s not especially surprising that its language would be constructed in such a way as to shape the discourse along similar lines: portraying guidelines and optional OA programs as “mandates” helps to build the impression of pervasiveness and inevitability that typifies OA advocacy rhetoric. It is less obvious, however, that this approach contributes to the scholarly community’s ability to discuss OA in a rational way.