University policies may be voluntary (i.e. they request that researchers make their work Open Access) or mandatory (i.e. they require that researchers make their work Open Access. The evidence shows that only mandatory policies produce the level of self-archiving from researchers that fill repositories.
Issues to take into account
The main issues to take into account in developing an Open Access policy are:
Evidence suggests that the earlier an article becomes available to others the greater its eventual impact. Institutional policies should threefore require deposit of articles at the earliest possible time. The optimal moment to specify is after peer review, when the final corections have been made and the paper is being sent to the publisher for the last time. The author should deposit the article at this point, in the format in which it was prepared (e.g. Word).
Even where publishers permit self-archiving of the author’s final version of an article, they may still wish to impose an embargo period during which Open Access is not permitted. This is to protect their sales. The period varies considerably from publisher to publisher and discipline to discipline. It may be as short as 3 months for some science journals or as long as 24 months in the humanities and social sciences.
Traditionally, in most cases authors are required to relinquish copyright to journal publishers in return for having an article published. Increasingly, though, authors are retaining copyright, often encouraged to do so by their institution. The recent mandate from Harvard University’s Faculty of Arts and Sciences requires authors to retain their own copyright, for example.
In these circumstances the author usually assigns the publisher a ‘licence to publish’. This is normally done through adding an ‘author addendum’ to the publisher's copyright transfer agreement (CTA) which the publisher requires the author to sign (and which normally transfers copyright to the publisher). Specimen author addenda and licences have been drawn up by various organisations and authors can choose which one suits them best. The two most commonly used ones are those from SPARC/Science Commons and from SURF/JISC (the national ICT organisations in the Netherlands and the UK).
Although increasing numbers of authors feel strongly that they wish to retain copyright in their own work, there are also many who do not relish the procedure of organising this with their publisher. Open Access can be achieved satisfactorily without this negotiation if authors are reluctant to engage in it (see types of policy wording).
Choice of journal in which to publish
Authors generally have strong opinions on the best journals in which to publish their work. Open Access policies should always ideally leave that choice with them, though some (big) funder mandates have closed the door on journals where the publisher does not permit the conditions that the funder requires. The biggest of these is the US National Institutes of Health which, along with the Wellcome Trust, mandates Open Access for all research it funds. If journals do not permit Open Access within 6 months of publication then the author is required to publish in a journal that does. The result of this tough policy is that journals are changing their terms to comply with the funder mandates: very few journals wish to deny themselves the chance to publish articles from work funded by these prestigious funders.
Individual institutions do not have the same clout as the NIH, though, so it is best for institutional policies to be crafted in such a way that authors are free to publish in whichever outlet they choose. It is possible to word a policy that gives authors this freedom yet still effectively provides Open Access even if publisher policies seem to preclude this (see below).
Types of institutional Open Access policy
There are three basic types of policy:
Type 1: Immediate deposit with immediate Open Access
This type of policy requires authors to deposit their articles upon acceptance for publication, once the final corrections have been made, and make them immediately available in Open Access through the repository. The outcome of this type of policy is immediate Open Access; the disadvantage is that some journals do not permit immediate Open Access and so this type of policy constrains the choice of journals in which an author can publish.
Type 2: Later deposit, after the embargo period
This type of policy requires authors to deposit their articles after publication, and at the end of the publisher’s embargo period. The advantage is that this complies with publisher requirements but the disadvantage is that it delays Open Access and runs the risk of the author forgetting to deposit the article so long after publication.
Type 3: Immediate deposit with optional later access
This type of policy requires immediate deposit but if it is submitted to a journal with an embargo then the policy permits access to be opened only at the end of the embargo period. During that period, however, the article’s metadata (title, authors, affiliation, abstract, references) should be fully Open Access at all times (publisher embargoes cannot be applied to metadata, which are not copyrighted).
The advantage here is that this policy complies with publisher embargoes but concomitantly ensures that all the required research outputs are gathered into the repository at the earliest possible point, that is, when the article has been accepted for publication and is in its final form.
If there is a publisher embargo to be accommodated, the deposit process allows the author to enter the date at which the embargo is to be lifted and the full-text of the paper be made Open Access. During the embargo period the metadata (title, author names, affiliation, abstract) of the article are Open Access (only the full-text is closed) so that indexing services like Google can find and index the article. Its existence is thus known, even if the full-text is not Open Access until the end of the embargo period. Would-be users can email the author for a copy of the article or, if they have access to the journal through a subscription, they can find the published version. Either way, the article is being used and its impact is starting to grow.
The EPrints and DSpace repository softwares have an ‘eprint request’ button for each article, so that when the full-text is under an embargo, would-be readers simply click on this button and an email request is sent to the corresponding author to ask for a copy by email. This is a ‘fair use’ or ‘fair dealing’ fulfilment of a personal request from an interested would-be user.
Institutions are therefore designing Type 3 policies. Most new institutional and funder policies are of this type. The policy from the Istituto Superiore Sanita (National Institute of Health) in Italy is an example. An optimal wording for this type of policy, which can be used as the basis for any institutional policy, is here.