Going Green: a year of supporting the HEFCE OA Policy

A year since HEFCE linked the mandatory deposit of accepted manuscripts to the allocation of QR funding, this post describes how the Library’s Scholarly Communications Team has helped academic colleagues successfully adapt to the new ruling which has resulted in an unprecedented level of Green OA deposits to our repository.

As we entered 2016, developing support services for the incoming policy was the team’s highest priority. The biggest obstacle was a very low Green OA baseline: from 2013 to early 2016, the number of Manchester-authored papers available through pure or hybrid Gold OA journals rose steadily; yet annual Green OA deposits to our repository stalled at ten percent of the institution’s journal article output over the same period.

During a pilot the previous year a variety of support models were tested with selected schools and we found that the highest compliance was achieved when the responsibility on authors was minimal and the Library offered to deposit and set embargoes on their behalf. So in February, we made a proposal to develop a fully mediated service which was accepted by the University’s senior research governance committee. The task then was to reorient our existing Gold APC service to encompass Green OA workflows which scaled to enable our team of six staff to support deposit of ~6000 papers a year.

To allow authors to send us their manuscripts we created an authenticated deposit form (branded as the Open Access Gateway) with just two input fields: acceptance date and journal/proceedings title, plus a drag and drop area to attach the file. Authors could also use the form to request Gold OA payment if they acknowledged a grant from a qualifying funder.

In the months before the policy launched we worked closely with Library marketing colleagues to deliver a communications campaign which raised awareness of the policy and built momentum behind the new service. Our message to academics was, ‘Just had a paper accepted? Now make sure it’s REF eligible.’

In April the policy launched, our deposit form went live, and the firehose to the University’s publications opened. Early signs were promising; in the first weeks, we received roughly ten manuscripts per working day which represented a significant increase against our baseline. However, more persuasion was needed for those authors across campus who weren’t sending us their manuscripts. We therefore began to chase those authors via email and sent a follow-up email, copying in a senior administrator, if we had no response.

We particularly focussed on non-compliant papers with the highest altmetric scores which had an increased likelihood of being selected for REF. The chaser emails were effective and many authors replied immediately with the manuscript attached. Of course, our emails also prompted some authors to ask questions or offer opinions about the policy which required additional resourcing. 

Sending chaser emails significantly raised the institution’s compliance rates, and it was clear that we would need to continue to do this systematically as awareness of the policy gradually spread across campus. This placed additional strain on the team as searching Scopus for papers and chasing authors proved an intensive process. We explored alternatives (e.g. using the ORCID API to identify our new papers) but no automated process was as effective as the painstakingly manual check of Scopus search results.

By August we’d refined our reporting to the University to include school/division level compliance against projections. To achieve this we recorded the compliance status and author affiliations of every single University paper falling within the scope of the policy in a masterfile spreadsheet. We then used SciVal to calculate the average output of the REF-eligible staff from each of the 33 schools/divisions over the past five years. This enabled us to project how many accepted papers we would expect each school/division to have per month during the first twelve months of the policy. Every month we produce a compliance report, like the one below, which supports standing agenda items at research governance committee meetings across the University.

 

 

As we moved into the new academic year, monthly deposits continued to rise. The team were at maximum capacity processing incoming manuscripts, so to speed up the flow of papers through our assembly line we purchased a license for the online forms service Typeform and developed a back office ‘If-This-Then-That’ style form. This allowed us to distil esoteric funder and publisher information into a simple workflow enabling papers to be processed with reduced input from higher grade staff.

Now, twelve months on from the introduction of probably the most swingeing OA policy ever introduced, we can take stock of how well we have adapted to the new ruling.  In the absence of a Snowball-type metric for measuring our compliance we currently track compliance three ways:

• % top papers compliant: In January, the University ran its annual Research Review Exercise, involving research staff proposing their best post-2014 papers for internal review. This provided the first opportunity to gauge compliance of those papers with the highest chance of being returned in the next REF. During the exercise, a total of 1360 papers were proposed for review which were within the scope of the policy, of these a very encouraging 95% were compliant with the new OA requirements. 
• % compliant against projections: Our chosen metric for reporting compliance to the University. We report the proportion of compliant papers with at least one REF-eligible author against the total number of papers we would have expected to have been accepted by all our REF eligible staff. Against this measure, 68% of the papers are compliant, 7% are non-compliant, and 25% are not currently recorded. Many of these unrecorded papers will not yet be published so we will chase them once they are indexed in Scopus. A large number of our papers arising from the ATLAS/LHCb collaborations are also available via Gold OA and are compliant but we have not yet recorded them in our masterfile.
• % compliant overall: To date, we’ve recorded 4656 papers of which 4031 (87%) are compliant, 459 (10%) are not compliant, and 166 (3%) are being actively chased by our team.

In total there’s a 55% Green OA/45% Gold OA split, and given that Green OA represents more inconvenience than most of our academic colleagues unfamiliar with arXiv have ever been willing to tolerate, it is very unlikely indeed that the University would have achieved such high compliance had the Library not provided a mediated Green OA deposit service. The data confirms our approach helped make Green Open Access an organisational habit practically overnight. 

The approach has come at a cost however; over the past year, supporting the HEFCE OA policy has taken up the majority of the team’s bandwidth with most of our 9am-5pm conversations being in some way related to a paper’s compliance with one or more funder OA policy.

Now that our current processes have bedded in, and in anticipation of the launch of the new UK Scholarly Communications License (UK-SCL) – for more on this read Chris Banks’s article or watch her UKSG presentation – and further developments from Jisc, we hope that over the next 12 months we can tilt the balance away from this reductionist approach to our scholarly output and focus on other elements of the scholarly communication ecosystem. For example, we are already in discussions with Altmetric about incorporating their tools into our OA workflows to help our academics build connections with audiences and are keen to roll this out soon – from early conversations with academics we think this is something they’re really going to like.

Whatever lies in store, it’s sure to be another busy year for the team.