Presentation on theme: "Pay It Forward: Investigating a Sustainable Model of Open Access Article Processing Charges for Large North American Research Institutions Ivy Anderson."— Presentation transcript:
1Pay It Forward:Investigating a Sustainable Model of Open Access ArticleProcessing Charges for Large North American Research InstitutionsIvy AndersonCalifornia Digital LibraryCNI 2015 Spring MeetingSeattle, WAThanks Ralf – I’m going to talk about a project that’s intended to be an exploration of the ideas that Stuart described, and in many respects a response to the European perspective that Ralf just shared with us.
2What is the Pay It Forward project? A grant from the Andrew W. Mellon Foundationto investigate the impact on North American research-intensive universities of a largescale conversion to gold Open Access funded via Article Publication Charges (APCs)18-month project, January 2015 – June 2016Pay It Forward is the title we’ve given to a grant that UC Davis and CDL have received from the Mellon Foundation to investigate the impact of a largescale conversion to gold OA funded via Article Publication Charges on large N Am research universities. UC Davis and CDL were fortunate to have received a planning grant from Mellon in 2014 to plan this project, and to have been awarded the full grant this year; we’re now in the first phase of the grant, which goes through June 2016.And no, Pay It Forward has nothing to do with a project of the same name launched by UC Press this year in conjunction with their new OA initiatives. Even though we work closely with the Press on many projects, we obviously did not consult on our project names!
3Pay It Forward will attempt to answer a single question: Can a large-scale conversion to open access scholarly journal publishing funded via APCs be viable and financially sustainable for large North American research-intensive institutions?“Viable” = author willingness to publish OA under the right conditions, and a scholarly publishing infrastructure that will allow high quality research and scholarship to flourish“Sustainable” = costing no more than, and ideally considerably less than, current subscription costs for comparable content today, with a rate of growth that is supportable over timeThe goal of this project is to answer one primary question – can such a shift to gold OA be viable and financially sustainable for the institutions that publish the lion’s share of research in the US and across the globe. There have been any number of studies that look at the economics of OA from a global perspective, but very few – Ralf’s analysis notwithstanding – that have looked at the problem from a local institutional perspective. One early study that did that was at Cornell, about a decade ago, which suggested that gold OA would be significantly more expensive for ARL member institutions than the current subscription system. No matter how attractive the economics of OA might look from the perspective of the scholarly system as a whole, no institution will be incentivized to move in that direction if it isn’t sustainable on a local level. However, we’d had some reason to question whether those findings were still relevant. As one of the largest public research institutions in the world, with a significant publishing profile – UC publishes something like 2% of the world’s research literature – we felt it was imperative to look at the question of gold OA from the perspective of the finances of our own institution.In addition to financial sustainability, we wanted to also consider the viability of this potential shift – by which we mean, is it good for scholarship as a whole, and will academic authors embrace it?
4Why this project, why now? Increasing disconnect between European and North American approaches to open accessNorth America (Green)Europe / UK (Gold)Finch ReportHorizon 20/20APC “Total Cost of Ownership” Agreements (UK, Austria, Netherlands…)Tri-Agency Open Access PolicyNIH Open Access PolicyOSTP DirectiveFaculty OA PoliciesFASTROK, but why study this right now? Ralf’s paper really gives us the answer. While N. America is primarily moving in a green OA policy direction, policy developments in Europe and the UK are driving a conversion to gold OA. In other words, two of the largest research publishing economies in the world are working at cross-purposes when it comes to open access developments. The US and Canada together comprise 31% of worldwide output, while Europe and the UK comprise (34%) – and these developments are heading on a collision course.
5Why this project, why now? Increasing disconnect between European and North American approaches to open accessNorth America (Green)Europe / UK (Gold)Confusing economic situationInstitutional license fees and APC revenue are both increasingPotential for publisher double-dippingLack of transparencyConfusion aboutWhat content one has access toWhat content one is paying forThis is setting up an inreasingly confused economic situation. License fee and APC revenue are both increasing, double-dipping opportunities abound. Gold OA is now something like 10-13% of all publishing worlwide, and projected to continue to accelerate rapidly over the next 5 years. We in the US need to better understand the options and implications for us.
6UC Faculty Open Access Policy Local drivers:UC Faculty Open Access PolicyCampus Open Access Fund PilotsFaculty began asking: “Does this mean I have to pay to publish?” and “Will the library pay?”Our Libraries wanted to understand how gold OA would impact our budgets if we were to subsidize publicationAnother driver for UC was faculty adoption of an OA policy in Even though this is a green policy, many faculty are confused about the differences between green and gold OA, are they were starting to ask if this meant they would have to pay to publish. At the same time, the UC libraries launched a coordinated set of open access funds across the campuses, and so everyone - the libraries, faculty, and administrators – began asking about the financial implications.
7In 2013, CDL did some preliminary modeling of APC costs and possible scenarios at the request of our University LibrariansSo in 2013, at the request of our University Librarians, CDL did some preliminary modeling of what the impact of a conversion to gold OA might look like. CDL licenses most of the journal content available systemwide at UC, and we’d also been purchasing customized reports from Thomson Reuters about UC publishing rates in the journal packages that we license to inform our journal negotiations, so we had a good base of both financial and authorship data to work from.
8UC 2010 publishing breakdown, by publisher We took a representative sample of nine journal packages, including both our largest packages and a handful of smaller ones – and we also looked for some diversity, including a university press publisher and some smaller more specialized packages. It turned out that half of UC’s annual publishing output (as recorded in WOS) was concentrated in these nine publishers, about 18,500 articles out of a total of nearly 37,000.We then applied a variety of APC scenarios to our authorship data – including various levels of APC, different levels of grant funding, and finally, estimates of UC corresponding authorship – and compared the results to the license fees we paid to those publishers.Data from Web of Science or publisher reports
9UC costs per publisher under different APC scenarios compared to licensing costs Formula: APC * % Grant funded * No. of 2010 UC articles per publisherAPC$3,000$2,000$1,500$1,000% Grant funded0%50%25%Publisher APublisher BPublisher CPublisher DPublisher EPublisher FPublisher GPublisher HPublisher ITOTALSLegend:> 15% Above License Fee>15% Below License Fee< 15% Above License Fee< 15% Below License FeeThis produced a kind of a heat map, showing which scenarios produced an OA cost that was higher than our current license fees, and which would be lower. Obviously in this particular model we used somewhat arbitrary figures – in this example, APCs running from the $3,000 that is typical of many hybrid journals, to $1,000 at the low end. We assumed that grant funding would be used to pay APCs in many cases, so we applied some gross percentages to the model, ranging from 0% to 25, 50, or 75%. The intent of this project wasn’t to be precise – we didn’t really have the resources for that – but to do some modeling that was indicative and reasonably well grounded in what we know.The red areas show where OA would be more expensive, and the green areas where it would be cheaper than our licenses. The pale red and pale green show the same thing, but within what might be considered a tolerable variance of less than 15% in either direction. You can see from this that for some publishers, our(Based on 2010 publishing output and 2011 licensing costs)
10UC Publishing breakdown, by discipline We did the same kind of modeling by discipline. In this case, we already had a database that broke down our journal expenditures by discipline based on LC Class, and we had a rough idea of how UC publishing mapped to major categories LHS, PSE, and SSH from a annual accountability report that UC’s office of research prepares, so we felt we had enough tools to produce a general idea of what this kind of modeling would look like from a disciplinary perspective. This was an important dimension, because some of the characteristics that we wanted to model, for example levels of grant funding and volume of output, vary considerably by discipline – with most grant funding in biomedical research, and very little in SSH.(Disciplinary breakdown is an estimate based on2012 UC Acountability Report graphs)
11Modeling gold OA by discipline Current Licensing CostsImpact of Different APC and Grant Funding Scenarios, By DisciplineAPC$1,500$2,500$1,000$500% Grant funded50%75%25%0%Life and Health Sciences$9,376,466$14,931,540$12,442,950Physical Sciences$11,363,274$8,018,790Social Sciences and Humanities$4,502,907$3,133,780All Disciplines$25,242,647$26,084,110This produced an interesting overall result, showing that even with very aggressive levels of grant funding and relatively low APCs in the life sciences – for example an average APC of $2500 and 75% of these funded via grants - OA might be more expensive in those disciplines, but that this could potentially be made up with lower APCs in other disciplines. Here again, this was purely theoretical, although some of the assumptions were grounded – for example, Berkeley’s experience with its open access fund in 2009 found that 75% of the requests were able to be supported from grant funds rather than requiring library funding.
12Factoring in corresponding authorship The most interesting result was when we factored corresponding authorship estimates into the mix. We did this on the basis of a single sample – we had earlier hand-sampled UC publications in Nature journals, and found that 43% of the time, the corresponding author was from an institution other than UC. So on the basis of that admittedly limited data point, we estimated a 60% corresponding authorship rate for LHS and PSE journals. We kept SSH at 110%, since co-authorship rates are much lower in those journals anyway – another data point informing this decision was a study by the London School of Economics suggesting that SS journals had an average co-authorship rate of 4. So we chose a fairly conservative estimate in this domain.
13Co-authorship patterns may make gold OA more sustainable In this scenario, using a life sciences APC of $2500 w/ 75% grant funding, an avg APC of $2,000 in PSE w/25% grant funding, and a low APC of $500 with no grant funding and no sharing of corresponding authorship, the model was about 20% less expensive than our current licenses. A low APC in SSH seems reasonable given the levels of APC currently being charged by publishers like SAGE, for example, which has a socsci OA megajournal.As I said, these models were meant to be indicative rather than precise – they used a limited number of data points to illustrate how the interaction of different variables would produce very different costs. What we really need to know is which of these assumptions and variables are the right ones.
14Pay It Forward will attempt to answer a single question: Can a large-scale conversion to open access scholarly journal publishing funded via APCs be viable and financially sustainable for large North American research-intensive institutions?“Viable” = author willingness to publish OA under the right conditions, and a scholarly publishing infrastructure that will allow high quality research and scholarship to flourish“Sustainable” = costing no more than, and ideally considerably less than, current subscription costs for comparable content today, with a rate of growth that is supportable over timeSo this brings us back to Pay It Forward. We agreed that if we really wanted to understand the cost and sustainability of gold OA, we needed to dig into the details and create a realistic, data-driven model.
15Project Components: Qualitative and Quantitative Users: the value of different publishing components to academic authors and editorsFaculty and student focus groups / surveysPublishers: Publisher survey on open access perspectives and plansQuantitativeReasonable publishing financial models that cover basic costs and ongoing innovationAnalyze the cost of publishing - both traditional and non-traditional publisher use casesCurrent and potential costs to the partner universities under a variety of scenariosFive-Year period of study,What our partner universities spent on scholarly journals fromAnalyze partners’ publishing output across a range of disciplines, including multi-institutional authorship patterns, research funding, and growth over timeThe project has both qualitative and quantitative components, to ensure that we’re not driven purely by dollars but are also taking into account social and behavioral dynamics and values.Qualitative – fac, grad students, post-docs, focus groups to inform surveys; publisher…Quantitative – cost of publishing – both trad and non-trad to develop a realistic, ‘ground-up’ cost model- and then a much richer and more detailed elaboration of the kind of modeling we’d done earlier, examining publishing output and licensing costs under a variety of scenarios that, in this phase, will be informed by detailed research and analysis.
16Co-Led by UC Davis and CDL Core Project TeamMacKenzie Smith, UC Davis (Co-PI)Laine Farley, CDL (Co-PI)Greg Tananbaum, ScholarNext (Project Manager)Ivy Anderson, CDL (Quantitative Lead)
17Project Partners Four research-intensive universities Harvard, Ohio State, University of British Columbia, University of California (10 campuses)Three industry partnersBibliometric: Scopus, Web of SciencePublishing: Association of Learned Society and Professional Publishers (ALPSP)We didn’t want this to be UC-centric, so have engaged a set of like partners, both public and private, all of whom share our characteristics of high publication output, but possibly with different emphases.Elsevier and TR will be helping us directly with bibliometric analysis – we’re acquiring raw data from WOS that can be analyzed by our modeling team, and reports from both WOS and Scopus to give us concrete data points by discipline about both worldwide and institution-specific publishing outputALPSP, a society with some 300+ member organiziations including both the large commercial publishers and society publishers, will be assisting us with the publisher survey to gain a better understanding of publisher attitudes and strategic directions.
18Project Consultants Greg Tananbaum, ScholarNext Project manager, Publisher surveys and costsCarol Tenopir, University of TennesseeUser studiesDavid Solomon, Michigan State UniversityBo-Christer Bjork, Hanken School of EconomicsAPC research, Scenario modelingMark McCabe, Boston University & University of MichiganScenario modeling, economic analysisAnd we’ve assembled an impressive array of consultants to do this work and bring all of the pieces together -
19Project Timeline Phase 1: January-March Phase 2: April-June Finalize data specifications, begin data gatheringConduct focus groups Develop publisher surveyPhase 2: April-JuneCollect and refine dataConduct user surveysConduct publisher surveyPerform publishing cost analysisPhase 3: July-DecemberComplete survey analysisComplete financial and bibliometric data analysisBuild and refine modelsPhase 4: January-JuneReview and refine modelPrepare documentationWrite up findingsTimeline – we’re still in the initial phases of the project, busy finalizing data specifications, conducting focus groups, and developing survey questions. We’ll be heads-down in the project work for the remainder of the calendar year, and expect to have results to share sometime after December.
20DeliverableThe main outcome of the project will be a detailed, flexible, and publicly accessible financial model that depicts what the emerging APC model would cost large research institutions under a variety of rigorously-modeled scenariosWhat level of APC is realistic and sustainable in a given discipline?How might costs be distributed among institutions, research funders, and other players?Our goal is to develop a replicable methodology that that others can apply in a local context
21Work to Date: Focus groups 8 focus groups conducted so far: 60 participants4 locations: UBC, Harvard, UC Irvine, UC DavisOSU to come in AprilDiversity of disciplines, ages, and perspectivesTranscription has begun, with analysis and questionnaire drafts to follow
22Early focus group impressions Range of perspectives:True believers, skeptics, most people somewhere in the middleMany senior faculty already post green versions in a repository or personal websiteSupport for OA as readers and as a moral good, but most have access to what they need nowConcernsWhere funding will come fromRicher nations may dominate publicationPotential for APC price increasesPredatory / vanity publishingLack of transparency – ‘publishing is broken’APCs are too high – publishers charge what the market will bearLibrary roleCoordinating/administrativeInstitutional publishing licenses
23Looking toward the future: Global Outlook Global R&D Share2015 STM ReportS&E Global Article ShareNSF 2014“The cumulative effect of sustained above-global-average growth in R&D spending in emerging economies has been a profound shift in the global make-up of research.”(STM Report, 2015)