Chapter 4 – Developing the UK-based e-Social Science Research Program
In this chapter we review the development of the research program of the National Centre for e-Social Science (NCeSS) funded by the U.K. Economic and Social Research Council (ESRC). We begin by tracing the origins of the Centre, we outline the work-in-progress across the Centre’s emerging research program, and we consider how the Centre contributes to the needs of social science researchers and provides them with opportunities to address key substantive research challenges in new ways. We then turn to the continuing efforts of the NCeSS Hub to develop the Centre’s research agenda and to encourage the wider adoption of e-Social Science. We conclude by reviewing some of the challenges which NCeSS and the research community as a whole must address if the goals of e-Social Science and e-Science are to be realized fully.
The National Centre for e-Social Science (NCeSS) was established by the U.K. Economic and Social Research Council (ESRC) in 2004 as its contribution to the U.K. e-science program. NCeSS’s mission is to enable social scientists to exploit innovations in digital infrastructure so that they are able to address key challenges in their substantive research fields in new ways. This infrastructure, known as the ‘Grid’ or increasingly commonly as ‘e-Infrastructure’ (or ‘Cyberinfrastructure’ in the U.S.), comprises networked, interoperable, service-oriented, scalable computational tools and services.
The Centre forms part of the ESRC’s strategy to develop leading-edge methodological tools and techniques to enhance the U.K. social science research community’s capacity to collect, discover, access, manipulate, link, share, analyze and visualize both quantitative and qualitative data.[i] To achieve its aims, NCeSS coordinates a program of e-social science research and makes available information, training, advice and support. The Centre is leading the development of an e-infrastructure for the social sciences and is also responsible for providing advice to the ESRC on the future strategic direction of e-social science.
NCeSS has a distributed structure, with a coordinating Hub at the University of Manchester, and a set of major three-year research Nodes and smaller one-year projects distributed across the U.K. The Hub acts as the central resource base for e-social science issues and activities in the U.K., integrating them with ESRC research methods initiatives and the U.K. e-science core program.
In this chapter, we review the progress of the NCeSS program, report on its current impact on social science research, reflect on the development of the research roadmap and, taking into account factors likely to influence future adoption, consider its trajectory over the next five years.
First steps in U.K. e-Social Science
In late 2000, the U.K. government announced funding for a research initiative in e-science (Hey and Trefethen, 2004). This comprised a so-called ‘core program’ of research into e-science technologies and applications through a series of demonstrator projects, together with the commitment of the six individual research councils to fund programs specific to their disciplines and communities.[ii] The ESRC launched its contribution to the program by commissioning four scoping studies:
- Grid-enabling quantitative social science datasets (Cole, Schurer, Beedham, & Hewitt, 2003).
- Qualitative research and e-social science (Fielding, 2003).
- Human-Centred Design and Grid Technologies (Anderson, 2003).
- Social shaping perspectives on e-science and e-social science (Woolgar, 2003).
The focus for these studies reflected core social science research orientations which were likely to raise quite different requirements for applications; the existence of an already well-established and mature infrastructure for the curation of research data through the Economic and Social Data Service;[iii] and the understanding that social scientists had a distinctive contribution to make to the U.K. e-science program by applying social science research methods to investigate barriers to the adoption of e-science.
Following the recommendations of the scoping studies, the ESRC allocated £500k to fund Pilot Demonstrator Projects (PDPs): small scale projects to explore how e-infrastructure could be used to generate exemplars to showcase the potential to the wider social science research community, and test the level of interest in e-social science. This approach acknowledged that the numbers of social science researchers ready at this early stage to grasp the potential of e-social science was likely to be small. A phased approach was seen as essential to facilitate a bootstrapping process where the success of ‘innovators’ or ‘early adopters’ would trigger interest and, eventually, lead to adoption by the wider community.
Eleven PDPs were funded, four with a main focus on aspects of data infrastructure and integration, six on statistical analysis and modeling (see, e.g., Peters, Clark, Ekin, Le Blanc, & Pickles, 2007) and one on collaborative qualitative video data analysis. Two PDPs also explored how e-social science might break down the divide between quantitative and qualitative research methods. Social shaping issues, as represented by, for example, the usability of new technologies, figured in many of these projects.
The limited funding per PDP meant that deployment of usable research tools and services was unlikely. This would require resources on a scale that only a significantly bigger program could provide. Accordingly, in 2004, the ESRC announced the next step in its e-social science strategy: the formation of the National Centre for e-Social Science (NCeSS) with funding of £6M for a first phase over the period 2004-2007 and funding for a second phase contingent on the outcome of a review. The Centre was to be structured around a coordinating Hub, a small number of large (£500K) projects or ‘Nodes’ and a larger number of smaller (£50K) projects. It was designed to ensure sufficient resources were available in the Nodes to developed innovative tools and services, at least to the stage of delivering demonstrators or, better, usable prototypes to the wider community, while the smaller projects maintained the program’s capacity to explore more speculative ideas.
The National Centre for e-Social Science 2004-2007
Building on the pattern established by the scoping studies, the specification for the NCeSS research program in its 2004-2007 phase identified two principal strands of research.
The applications strand aimed, through substantive problem-driven research projects, to explore the use of e-infrastructure to make advances in quantitative, qualitative and mixed-methods economic and social research. Its long term aim was to build an e-infrastructure providing new and more powerful research resources to the wider social science research community. Such resources would include datasets, analysis tools and services, and virtual research environments providing integrated access to them.
The social shaping strand aimed to understand how e-infrastructure is being developed, how it is being used across the research community, and what its implications are for scientific practices and research. ‘Social shaping’ was defined broadly to include all social, economic and other influences on the genesis, implementation, use, usability, immediate effects and longer-term impacts of the new technologies.
This first phase of the NCeSS program began with a call for Hub applications. The University of Manchester’s proposal was selected from a shortlist of three and began work in May 2004. At the same time, the second-ranked Hub proposal was awarded a Node grant. There followed two calls for Node applications. The first was issued in May 2004 and resulted in three nodes being commissioned. The second was issued in November 2004 and resulted in a further three Nodes being commissioned. Significantly, four of the original eleven PDPs were successful in graduating to Node scale funding. The research objectives of the seven Nodes were as follows.
- The Collaboratory for Quantitative e-Social Science Node (CQeSS: Lancaster University and Daresbury Laboratory) aimed to develop tools and services to advance the state of the art in quantitative methods. It focused on developing middleware that would allow users to exploit distributed research resources such as datasets and more powerful computational facilities while continuing to be able to employ their favorite desktop analysis tools (Crouchley et al., 2005; Grose et al., 2006).
- The Mixed Media Grid Node (MiMeG: Bristol University and King’s College, London) focused on developing tools to support distributed, collaborative video analysis (Fraser et al., 2006; Tutt, Hindmarsh, Shaukat, Fraser, & McCarthy, 2007; Shaukat and Fraser, 2007). Digital video has become an invaluable tool for social scientists to capture and analyze a wide range of social action and interactions. Video-based research is increasingly undertaken by research teams distributed across institutions in the UK, Europe and worldwide but there was little existing technology to support collaborative analysis.
- The Modeling and Simulation for e-Social Science Node (MoSeS: Leeds University) aimed to develop a suite of modeling and simulation tools for application in policy making (Birkin et al, 2005; Birkin, Townhend, Turner, Wu, & Xu, 2007; Townend, Xu, Birkin, Turner, & Wu, 2007). The chosen policy applications drivers were healthcare, transport planning and public finance. Social science problems of this type are characterized by a requirement for extensive data integration and multiple iterations of computationally intensive scenarios.
- The Digital Records for e-Social Science Node (DReSS: Nottingham University) sought to develop new tools for capturing, replaying, and analyzing multi-modal digital records of people’s activities (Crabtree, French, Greenhalgh, Rodden, & Benford, 2006; Crabtree et al., 2006; Greenhalgh et al., 2007; Knight et al., 2006). Social scientists worked in close partnership with computer scientists on three substantive research driver projects in order to explore the salience of new forms of digital record for research and to determine requirements for tools.
- The Geographic Virtual Urban Environments Node (GeoVUE: University College, London) focused on developing geographical information systems (GIS) tools and research environments to enable users easily to map and visually explore spatially-coded socio-economic data (Batty, Steadman, & Xie, 2006; Milton and Steed, 2007). Driver applications included urban planning and design.
- The Semantic Grid Tools for Rural Policy Development and Appraisal Node (PolicyGrid: Aberdeen University) brought together social scientists with interests in rural policy development and appraisal with computer scientists with experience in Grid and Semantic Web technologies. The objective was to explore how Semantic Grid tools (Chorley, Edwards, & Preece, 2007; Heilkema, Edwards, Mellish, & Farrington, 2007; Pignotti, Edwards, & Preece, 2007) could be used to support social scientists and policy makers using mixed-methods research techniques (e.g., surveys and interviews, ethnographies, case studies, simulations).
- The Oxford e-Social Science Node (OeSS: Oxford University) addressed the inter-related social, institutional, ethical, legal and other issues surrounding e-infrastructures and research practices. The focus was on confidentiality, privacy, data protection, intellectual property rights, accountability and trust and risk in distributed collaborations (Axelson and Schroeder, 2007; Carusi and Jirotka, 2007; Dutton, 2007).
Under this first tranche of research funding, twelve small grant projects were also commissioned.
Node research themes and synergies
The Node awards can be broken down into six applications strand Nodes and one social shaping strand Node, although it is worth remarking that all applications strand Nodes addressed social shaping issues (e.g., usability) in their work plans. Of the applications strand Nodes, the majority (MoSeS, PolicyGrid, MiMeG and CQeSS) had data analysis tools as their main focus and one (DReSS) had data collection and management as its main focus. A more detailed analysis of Node research themes (as defined in the call for Node submissions) is shown in Table 4.1. Unsurprisingly, because of their established use of computing, quantitative methods (i.e., statistical analysis, simulation and high performance computing) figure quite prominently in the Node themes whereas qualitative methods figure less prominently. Table 4.1 also reveals an interesting gap in this first phase of the NCeSS program with the absence of any contribution to the text/data mining theme.
Table 4.1 suggests a significant number of potential synergies between the Nodes. For example, DReSS and PolicyGrid form a cluster around data description / discovery / management / re-use; CQeSS, DReSS, PolicyGrid and MoSeS form a cluster around quantitative methods and, within this cluster, CQeSS and MoSeS form a mini-cluster around high performance computing and MoSeS and PolicyGrid form a mini-cluster around simulation. Similarly, a cluster (MiMeG, DReSS, PolicyGrid, GeoVUE, OeSS) exists around the theme of collaboration.
Supporting emergence of program synergies and outreach
Theme clusters are important to the NCeSS program for strategic reasons. NCeSS is a managed program where the objective is not only to ensure that research projects individually achieve their potential but that they work effectively together, that collaborations and synergies flourish and that, collectively, the program is able to establish partnerships with the wider e-science community in the UK and internationally. A problem of research programs is that individual projects often find it hard to release the effort to take on the additional burden of collaboration. As we will consider later, communication appears to be a key issue for the success of interdisciplinary projects, but this has to be complemented with resources if collaborations between program members are to work in practice.
In NCeSS, alongside administrative arrangements introduced to foster synergies and cooperation between the Nodes, the most effective measure has been to initiate a joint project that provides motivation and a focus for collaborative activities by the Hub and Nodes and, most importantly, funding for them. This is the e-infrastructure for the Social Sciences Project (Daw et al., 2007). Its main objective is to deploy research resources and demonstrators selected from those being developed within the Nodes, Small Grant Projects and PDPs. In this way, the project aims to:
- provide a platform for disseminating the benefits of e-social science to the wider social science research community;
- enhance understanding of issues around resource discovery, data access, security and usability by providing a test bed for the development of metadata and service registries, tools for user authorization and authentication, and user portals;
- lay foundations for an integrated strategy for the future development, support and sustainability of e-social science infrastructure and services.
However, by providing the Nodes with the motivation and resources to work together, the e-infrastructure project has also proved an effective response to the challenges of promoting synergies across NCeSS and other ESRC investments, coordinating activities and identifying areas in which to promote the benefits of common policies and technology standards.
A key objective for e-science is to foster and further scientific research by collaboration across disciplinary and geographical boundaries. e-Social science requires international collaboration in order to integrate an expanded range of data and computing resources. Development of a comprehensive e-infrastructure supporting social science research in a wide range of contexts can only be achieved through the establishment of globally recognized standards and through exchange of experience on implementation and use.
NCeSS has devoted substantial effort to establishing these international links. Members of the Hub have been actively involved in ICEAGE (The International Collaboration to Extend and Advance Grid Education[iv]) whose major objective is to establish a worldwide initiative to inspire innovative and effective Grid education. The Hub also participated in the recently completed E.U.-funded AVROSS project studying the adoption and sustainability of e-infrastructures (Barjak et al., 2007).
The Hub has established a series of annual International Conferences on e-social science.[v] The 1st and 2nd conferences were held in Manchester in 2005 and 2006 respectively. With the support of the U.S. National Science Foundation (NSF), the 3rd international conference was co-organized with Michigan University in 2007. The conferences bring together international representatives of the social science and e-infrastructure research communities in order to create better mutual awareness, harmonize understanding, and instigate coordinated activities to accelerate research, development, and deployment of e-infrastructure to support the social science research community. NCeSS also hosts a program of visiting research fellowships welcoming scholars from outside the U.K. to make an extended visit to the Hub and Nodes. The EUAsiaGrid project[vi] in which NCeSS is a partner enables the Centre to extend these activities significantly by establishing strong links with key players within the E.U. and beyond. The partners in this project share experiences of building and using e-infrastructure in the social sciences and pursue a program of studies about the development of these technologies across the international e-science community.
Despite or perhaps because of this encouraging start, much remains to be done to foster international collaborations in e-social science. The way forward is to begin with commonality of more established e-science at a global level, build upon the NCeSS experience of studying how e-infrastructures and research practices are mutually shaped and then engage other interested partners in e-social science.
Developing the Research Roadmap
Since the launch of NCeSS, the Hub has been working to develop the research roadmap and to extend its engagement with the social science research community. A variety of specific activities have been used to facilitate this.
First, NCeSS has organized a rolling program of Agenda Setting Workshops (ASWs) to which social scientists are invited to hear about opportunities for using e-infrastructure and to reflect on how these technologies might address the obstacles they face in pursuing their research objectives. A total of thirteen ASWs have been held since November 2004. Topics have included: development and use of ontologies; combining and enhancing data; applications of simulation and modeling in policy making; trust and ethics; visualization techniques; using text mining to bridge quantitative and qualitative methods; collaboration in e-science; confidentiality issues and clinical data; quantitative methods; qualitative methods; and recording human activities.
The ASWs have helped to identify new areas for the application of e-infrastructure in the social sciences. For example, a theme emerging from several of the Workshops is that Grid-enabled datasets, services and tools are key enablers for the wider take-up of e-social science. The ASWs have also enabled NCeSS to begin to profile the social science research constituency in terms of its awareness and readiness to adopt new technologies and begin to map the membership of the ‘interested’ community, that is, those prepared to take up new tools and services if properly supported.
Second, as the outcomes of the ESRC’s own strategic deliberations crystallized in mid 2007, these were fed into the roadmap. The ESRC National Data Strategy[vii] identified the provision of a world class social science data infrastructure as essential to improving the re-use of existing data collections and to meeting the challenges of the ‘data deluge’ arising from a profusion of new, ‘naturally occurring’ sources of digital data. The ESRC Key Research Challenges[viii] identified succeeding in the global economy; international relations and security; understanding and shaping individual decisions; education and life chances; religion, ethnicity and society; population change; and environmental change as its strategic research priorities for 2007-2010.
Providing a world class data infrastructure will require implementation of a technology strategy that will make available to researchers better tools for describing, locating and accessing data, cleaning it, maintaining its confidentiality, combining datasets, and facilitating secondary analysis. Meeting the ESRC’s strategic research priorities will require more collaborative and inter-disciplinary approaches, and the infrastructure and tools to support them.
Third, there have been two external reviews of NCeSS, one focusing on the Hub and the second on the ESRC’s e-social science strategy as a whole. The latter concluded that the ESRC had made an extremely promising start in its e-social science investments. As a result of these external reviews, the NCeSS Hub was refunded in 2007 for an additional five years and the NCeSS research program received a new round of funding in mid 2007. A new phase of Node commissioning was initiated in late 2007, based on the evolving research roadmap, with the aim of funding eight three-year Nodes from 2008.
Research Roadmap 2007-2012
The 2007-2012 research roadmap retains the twin tracks of an applications strand and a social shaping strand. A key difference, however, is that the 2007-2012 roadmap has not only to identify new avenues for the application of e-social science but also to identify the best ways of building on the achievements of the first phase. Three major themes were identified in the applications strand: data infrastructure, data analysis and collaboration. Within these, several more specific issues were highlighted in the second phase Node call:
- Many of the new sources of social data are distinctive in that they are ‘born digital’ and continuously updated by people’s everyday activities. Research is needed to explore how to realize a ‘population observatory’ in which social scientists can discover, access and use these new forms of data, while remaining sensitive to ethical issues relating to privacy, confidentiality and access.
- As sources of social data grow and proliferate, the problem of resource description becomes ever more acute and yet remains critical to data discovery and use. Mechanisms are needed to automate the adding of metadata to datasets. Ways of harnessing community generated contributions (tagging or ‘folksonomies’) and combining them with more formal representations (ontologies) are also required.
- Secure procedures for accessing confidential data are an overarching requirement for increasing the research value of existing and new sources of social data. Current provision, based on physically secure locations, represents a major barrier to the effective exploitation of such data.
- Multi-scale modeling, including that which combines physical, biological and socio-economic phenomena, is becoming an increasingly important tool for exploring complex, inter-connected systems such as climate change and disease. Real-time modeling driven by data from sensor grids and from population observatories is increasingly possible.
- The extension of text mining techniques to social and economic data will be critical in exploiting new large-scale data resources, with a wide range of potential applications. For example, the capacity to gather qualitative data now vastly outstrips the capability to analyze it. The related technique of data mining has also yet to find wide application in academic social science research.
- Solving complex problems often involves multiple steps where the output of one step is used as the input in the next. Tools are needed to assist in the management of research processes that involve the synthesis of complex, multi-stage analyses.
- Virtual research environments, that is, persistent digital spaces where researchers can share data and tools, are vital for distributed collaboration. Examples are rapidly emerging, including a growing number taking inspiration from social networking sites,[ix] but work remains to be done to improve the usability and interoperability of tools and to extend support to cover the overall research ‘lifecycle’.
- Inter-disciplinary collaborations involving the social, medical and natural sciences will be essential for success in tackling many ESRC research priorities. New collaborative resources, such as ‘population laboratories’ providing secure access to medical and social data, and new powerful tools for data linking and analysis will be essential in order to support such collaborations.
Further social shaping research is critically important as e-science moves beyond its initial phase in the U.K. and elsewhere. e-Science is beginning to take shape in ways that need to be understood as both applications and technologies co-evolve. Studying patterns of adoption will help to reveal factors which enable or inhibit the diffusion of e-infrastructure-supported research methods and will facilitate interventions that can enhance the e-infrastructure’s effectiveness. The social shaping agenda addressed four themes:
- Genesis of e-infrastructure, including historical comparisons with the development of other communications technologies and a consideration of broader institutional and political contexts;
- Social, economic and other determinants of the design, uptake, use and sustainability of e-infrastructures;
- Implications for the nature and practice of science, including social science, and for the character and direction of knowledge production, validation and use;
- International comparisons, examining how different national science policies and legal frameworks influence the funding and organizations of these developments.
Measuring the impact of investment is vital both for the effective management of research programs and the shaping of future science policy. It is very challenging, however, to capture the full range of impacts, especially within short timescales.
The NCeSS program is ‘work in progress’ and e-social science is far from being a routine undertaking. It carries significant risk arising from such factors as the uncertain path of technological innovation; the lack of experience among potential users of the technologies involved; issues arising from the social organization of research including embedded research practices and the established reward structure; and the management of processes of change. We have outlined various initiatives that NCeSS has undertaken to counter these risks in the pursuit of its core aims, notably the Agenda Setting Workshop series and the e-infrastructure for the Social Sciences Project. In this section, we reflect on some of the most significant challenges that need to be tackled as NCeSS seeks to build upon its achievements to date.
The target audience for NCeSS dissemination and capacity-building activities can be segmented by academic discipline, methodological approach and career stage but the most useful profile is in terms of knowledge and skill. It is on this basis that the Hub’s strategic planning is organized around the tri-part division between the ‘early adopters’, the ‘interested’ and the ‘unengaged’.
The early adopters are largely already part of NCeSS (Hub, Nodes, Small Grant Projects, Pilot Demonstrator Projects, and Visiting Fellows) and are keen to push to the limit what e-social science makes possible. The Hub supports early adopters by facilitating networking and encouraging the sharing of technical expertise. The e-Infrastructure Project is playing an important role in this.
The interested form the test-bed for e-social science applications. In return for their assistance in helping to identify requirements, NCeSS offers them demonstrators of how e-infrastructure might aid their research. NCeSS also supports their adoption of new tools and services (produced by early adopters) and monitors their experiences to feed back into the development process. The Hub has used an action-oriented ethnographic approach in a series of small-scale case studies to understand the needs that arise in the everyday research practices of the interested and, in parallel, promoted e-infrastructure developments to address their needs.
Not all of the unengaged have the same view of e-infrastructure. Some simply do not know what is available (Lin, Procter, Halfpenny, Voss, & Baird, 2007), and they are the target of NCeSS’s awareness-raising efforts that again employ demonstrators to illustrate the potential of new tools and services. Others are aware but do not have the time or inclination to invest in new ways of working. Yet others are epistemological skeptics, believing that any use of information and computer technologies taints the resultant social science. These will be the hardest to win over.
Even the more tractable of the unengaged will adopt new tools and services only when these are ‘hardened’ to production level (that is, become easy to use, stable, documented and supported); when they offer immediate benefits that quickly outweigh the costs of learning to use them; and when they complement existing research practices. Only when NCeSS has moved beyond proofs of concept, demonstrators and prototypes to near production level tools and services can it begin to engage them. It is the ease-of-use and utility of e-infrastructure, and its contribution to advancing social scientists’ substantive research, that will persuade them to adopt new ways of working, not the provenance of the technology. In this respect, the major breakthrough in wider adoption of e-infrastructure will be when its e-science origins become invisible, just part of the normal landscape of research methods. In this way, even the epistemological skeptics can be enticed into the fold.
Understanding and tackling barriers to adoption
Complementing its small-scale, focused studies, the Hub has also been involved in mapping the adoption of e-infrastructure across disciplines at the U.K. and European levels. This mapping provides the groundwork to develop mechanisms to tackle barriers to adoption and improve outreach (Barjak et al., 2007; Voss et al., 2007b). The e-Infrastructure Use Cases and Service Usage Models (eIUS)[x] project aims to gather and document evidence of how e-infrastructure is currently being used to facilitate research processes. The Enabling Uptake of e-Infrastructure Services project (e-Uptake)[xi] aims to develop strategies for widening adoption of e-infrastructure. Both projects are funded under the Community Engagement strand of the U.K. Joint Information Services Committee’s e-infrastructure program.
What this work confirms is that, as with any innovation, barriers that potential users of e-infrastructure face are numerous and, singly or in combination, they can delay or even prevent adoption (Rogers, 1995; Molina, 1997). For e-infrastructure to be widely adopted, costs as perceived by users must be outweighed by the benefits. Potential users must be aware of e-infrastructure, must understand the advantages it can bring to their research, must be willing to invest in new skills, and must have access to the facilities and support they need for successful adoption. At the same time, e-infrastructure services must be reliable, robust and easily usable if researchers are to be persuaded to trust their mission critical work to them. Moreover, users must be confident that services will not only continue to be available but also improve in response to their needs so that the benefits increase over time.
Having mapped the adoption of e-infrastructure and identified how users respond to barriers, strategies then need to be devised that enable uptake, for example, by providing clear and well-supported migration routes. These routes are likely to be different for late adopters drawn from among the unengaged and early adopters at the forefront of e-social science. Late adopters, for example, may require direct and personalized support in the form of staff development courses (both face-to-face and self-paced on-line learning); specific consultancy to develop new applications to utilize services in novel ways; and a single well-curated source of information about technical components and services, along with exemplars of their use.
As users’ requirements mature, their support needs may also change. Moreover, different user communities will be at different phases of the adoption cycle at any one time and so support has to be provided for all phases simultaneously. Explicitly or implicitly, e-infrastructure users will go through cycles of evaluating their requirements and assessing the appropriateness of services while providers will similarly go through cycles of improving services and developing new ones to meet emerging needs. Accordingly, potential user communities – and their experiences of barriers and responses to them – are likely to be highly diverse. Extensive, flexible and varied resources will be needed to promote e-social science to them effectively. Furthermore, technical and social issues cannot be separated given the dynamic between them – which is the focus of NCeSS’s social shaping research agenda. Accordingly, e-infrastructure adoption will only reach full maturity when social, organizational, cultural, ethical and legal issues are resolved in tandem with the creation of technology-based tools and services. This is the main message of studies of previous efforts to develop large-scale infrastructures: success depends on aligning technical components and stakeholder interests (Edwards, Jackson, Bowker, & Knobel, 2007).
e-social science is underpinned by a vision of the transformation of research practice into collaborative activity that combines the abilities and resources of distributed groups of researchers in order to achieve research goals that individual researchers or local groups could not hope to accomplish (Voss, Procter, Budweg, & Prinz, 2007a). Also, e-social science is inherently cross-disciplinary, given that its development is dependent on close collaboration between social scientists and computer scientists. Beyond this, its full potential will be realized only when it spans further disciplines, including medical and environmental sciences, thereby enabling the investigation of the full complexity of determinants of individual action and social behavior.
Yet there are numerous institutional constraints on collaboration, most notably the academic reward system in which career advancement and scientific reputation depend heavily on individual achievements. The reality of individual competition over discovery claims, grants, promotion and space in top-ranking journals is far removed from the ideal of openness and sharing of data and other resources promoted by the e-science vision. While the current rapid spread of ‘Science 2.0’ (Waldrop, 2008) provides grounds for optimism that research cultures are indeed changing, continuing investigation of barriers to real-world interdisciplinary collaboration, and mechanisms to overcome them remains a critical element of the NCeSS social shaping research program.
A variety of factors conspire to make issues about the sustainability of e-infrastructure impossible to ignore. First, U.K. e-science has reached that point in the innovation lifecycle where it must seek engagement with users beyond the early adopters and it is important that the pace of adoption does not falter. To progress, the interested must be supplied with tools and services that meet their needs and they must be supported as they seek to embrace innovative practices, or their enthusiasm will wane. Second, the existence of competing technical solutions can be a disincentive to the adoption of innovations, and recent divergent developments are a cause of confusion within the e-infrastructure community. Five years ago, the e-infrastructure technical roadmap was indistinguishable from that of the ‘Grid’. Now, however, the rise of Web 2.0 has caused many to consider alternative, more lightweight approaches where sophisticated grid-based solutions are not required. While this may have the effect of promoting adoption in some quarters, it also carries the risk of deterring others from engaging, at least until a clear technical winner has emerged. Third, if the accumulation, sharing and re-use of resources called for by the e-research vision generates a substantial increase in both the numbers and the types of research resources available, then priorities will have to be established to determine which will receive continued support, given that funding is unlikely to keep pace.
This brings us to a fundamental question for sustainability, which is how research resources originating in time-limited projects can be re-built to production level quality, then curated, maintained and managed so that they remain viable for use by the whole research community in the long term. In particular, where, in a landscape of multiplying, diverse and distributed resources, will the necessary effort and expertise come from, and what funding models are most appropriate to pay for it (Voss et al., 2007b)? Funding bodies are concerned that sustaining the burgeoning body of research resources will consume an ever increasing proportion of their budgets.[xii] The existing institutional infrastructure as represented by current service providers and the funding models that support them are in tension with the opportunities that the new technical infrastructure affords. It is time to consider whether a blueprint for a new institutional infrastructure, possibly with a greater number and diversity of service providers, is necessary, but there are few signs as yet that the relevant stakeholders (existing or potential) are ready or able to explore and agree how to best exploit – and fund – the options available to them.
Impact and its measurement
Measuring the impact of the e-social science program is vital for planning its future strategic direction. However, evaluating the impact of innovation is notoriously difficult, especially in the short term since results take time to disseminate and their significance may only become clear after a considerable delay. Moreover, impacts do not flow solely from technical advances; these are mediated by a wide range of social, institutional, cultural and economic factors. Impact can manifest itself as improvements in research performance, for example, efficiencies in data integration, or in novel research outputs, such as new findings that could not have been achieved without innovative technologies. In addition, impact in terms of added value obtained from innovations can be economic, social, personal or more diffusely realized. For these reasons, conventional uni-dimensional impact measures (such as peer reviews, citation analyses and other bibliometric approaches) have severe limitations. Other approaches, such as impact modeling, which involves extrapolating from previous trends, are not suited to e-science because it is fluid and fast changing. The problem is not the lack of data; there exists a deluge of data from sources such as bibliometric databases, e-repositories and the Web. It is the techniques that would enable this abundant data to be transformed into evidence for impact assessments that are not equal to the task.
Measuring impact exemplifies the kinds of challenge that e-science was devised to address. In this sense, the solution to creating an evidence base for the e-(social) science roadmap lies in its own hands: the use of network analysis, data and text mining to extract and analyze information from extensive sources, within a research program that will foster the necessary interdisciplinary collaborations. The recently announced NSF Science of Science and Innovation Policy program represents a very interesting and timely step towards this.[xiii]
A feature of the current state of e-social science is that, despite the substantial investment by the ESRC, its adoption remains piecemeal. Although bids for Nodes and Small Grant projects obviously had to attend to the strategic direction set out in the specification of the calls if they were to be successful, nevertheless the proposals reflected the interests of the groups who authored them, as they slotted their e-social science projects into their ongoing, wider research programs. The upshot is that the outcome of a large-scale e-infrastructure program like NCeSS cannot be guaranteed through top down strategic planning. Instead, much of the Hub’s coordinating activities, and especially the e-infrastructure for the Social Sciences Project, aim to harness bottom up innovations that are driven by experimentation by the Nodes and Small Grant Projects as they respond to requirements proposed by their local, substantive social science driver projects. A future ‘Whig history’ of the emergence of e-infrastructure will no doubt identify a clear line of development but the reality along the way is of uncertainty and risk-taking, only some of which will issue in what are later recognized as successes.
In this environment, user capability becomes central, and this requires striving to provide e-social science tools and services that enable social scientists to improve their research practices and generate results that would not otherwise have been possible. NCeSS must therefore continue to work very closely with social science researchers to seek out their requirements and it must be catholic in the selection of technologies to meet them, broadening out from the original emphasis on Grid computing. At the same time, it must be remembered that the technical and the social are inextricable, requiring attention to awareness-raising, training, support and other activities designed to reduce the cost of adoption.
Anderson, A. (2003). Human-centred design and Grid technologies. Swindon: ESRC.
Axelson, A-S., & Schroeder, R. (2007). Making it open and keeping it safe: e-Enabled data sharing in Sweden and related issues. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper139.pdf
Barjak, F., Weigand, G., Lane, J., Kertcher, Z., Poschen, M., Procter, R., & Robinson, S. (2007). Accelerating transition to virtual research organization in social science (AVROSS): First results from a survey of e-infrastructure adopters. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper141.pdf
Batty, M., Steadman, P., & Xie, Y. (2006). Visualization in spatial modeling. In J. Portugali (Ed.), Complex artificial environments, pp. 49-70. Berlin: Springer.
Birkin, M., Clarke, M., Rees, P., Chen, H., Keen, J., & Xu, J. (2005). MoSeS: Modelling and Simulation for e-Social Science. In Proc 1st Int. Conf. on e-Social Science, Manchester, UK.
Birkin, M., Townhend, P., Turner, A., Wu, B., & Xu, J. (2007). An architecture for social simulation models to support spatial planning. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper214.pdf
Carusi, A-M., & Jirotka, M. (2007). From data archive to ethical labyrinth. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper171.pdf
Chorley, A., Edwards, P., & Preece, A. (2007). Tools for tracing evidence in social science. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper124.pdf
Cole, K., Schurer, K., Beedham, H., & Hewitt, T. (2003). Grid enabling quantitative social science datasets – A scoping study. Swindon: ESRC.
Crabtree, A. French, A., Greenhalgh, C., Rodden, T., & Benford, S. (2006). Working with digital records: developing tool support. Paper, International Conference on e-Social Science, Manchester, UK. Available at: http://www.mrl.nott.ac.uk/~axc/DReSS_Outputs/ICeSS_1_2006.pdf
Crabtree, A., French, A., Greenhalgh, C., Benford, S., Cheverst, K., Fitton, D., Rouncefield, M., & Graham, C. (2006). Developing digital records: Early experiences of record and replay. Journal of Computer Supported Cooperative Work, 15(4): 281-319.
Crouchley, R., van Ark, T., Pritchard, J., Grose, D., Kewley, J., Allan, R., Hayes, M., & Morris, L. (2005). Putting social science applications on the Grid. Paper, International Conference on e-Social Science, Manchester, UK. Available at: http://www.ncess.ac.uk/events/conference/2005/papers/papers/ncess2005_paper_Crouchley.pdf
Daw, M., Procter, R., Lin, Y-W., Hewitt, T., Jie, W., Voss, A., Baird, K., Turner, A., Birkin, M., Miller, K., Dutton, W., Jirotka, M., Schroeder, R., de la Flor, G., Edwards, P., Allan, R., Yang, X., & Crouchley, R. (2007). Developing an e-infrastructure for social science. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper127.pdf
Dutton, W. (2007). Reconfiguring access to information and expertise in the social sciences: The social shaping and implications of cyberinfrastructure. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper127.pdf
Edwards, P.N., Jackson, S.J., Bowker, G.C., & Knobel, C.P. (2007). Understanding infrastructures: Dynamics, tensions, and design. Final report of the workshop “History & Theory of Infrastructure: Lessons for New Scientific Cyberinfrastructures”. National Science Foundation. Available at: http://deepblue.lib.umich.edu/handle/2027.42/49353
Fielding, N. (2003). Qualitative research and e-social science: Appraising the potential. Swindon: ESRC.
Fraser, M., Hindmarsh, J., Best, K., Heath, C., Biegel, G., Greenhalgh, C., & Reeves, S. (2006). Remote collaboration over video data: Towards real-time e-social science. Journal of Computer Supported Cooperative Work, 15(4): 257-279.
Greenhalgh, C., French, A., Tennent, P., Humble, J., & Crabtree, A. (2007). From ReplayTool to digital replay system. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper161.pdf
Grose, D., Crouchley, R., van Ark, T., Allan, R., Kewley, J., Braimah, A., & Hayes, M. (2006). sabreR: Grid-enabling the analysis of multi-process random effect response data. Paper, International Conference on e-Social Science, Manchester, UK. Available at: http://redress.lancs.ac.uk/resources/Grose_Daniel/Grid-Enabling_Analysis_Multi-Process_Random_Effect_Response_Data_R/NCeSS06.pdf
Hey, T., & Trefethen, A. (2004). UK e-Science Programme: Next generation grid applications. International Journal of High Performance Computing Applications, 18(3): 285-291.
Heilkema, F., Edwards, P., Mellish, C., & Farrington, J. (2007). A flexible interface to community-driven metadata. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper144.pdf
Knight, D., Bayoumi, S., Mills, S., Crabtree, A., Adolphs, S., Pridmore, T., & Carter, R. (2006). Beyond the text: Construction and analysis of multi-modal linguistic corpora. Paper, International Conference on e-Social Science, Manchester, UK. Available at: http://www.ncess.ac.uk/research/sgp/headtalk/20060630_knight_BeyondTheText_paper.pdf
Lin, Y.-W., Procter, R., Halfpenny, P., Voss, A., & Baird, K. (2007). An action-oriented ethnography of interdisciplinary social scientific work. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper179.pdf
Milton, R., & Steed, A. (2007). Mapping carbon monoxide using GPS tracked sensors. Environmental Monitoring and Assessment, 124(1-3): 1-19.
Molina, A.H. (1997). Insights into the nature of technology diffusion and implementation: The perspective of sociotechnical alignment. Technovation, 17(11-12), 601-626.
Peters, S., Clark, K., Ekin, P., Le Blanc, A., & Pickles, S. (2007). Grid enabling empirical economics: A microdata application. Computational Economics, 30(4): 349-370.
Pignotti, E., Edwards, P., & Preece, A. (2007). Semantic workflow management for e-social science. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper174.pdf
Rogers, E. M. (1995). Diffusion of innovations. 4th ed. New York: Free Press.
Shaukat, M., & Fraser, M. (2007). A security framework for data distribution in qualitative analysis tools: Digital rights management in MiMeG. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper137.pdf
Townend, P., Xu, J., Birkin, M., Turner, A., & Wu, B. (2007). Modelling and simulation for e-social science: Current progress. Paper, e-Science All Hands Meeting, Nottingham, UK. Available at: http://www.allhands.org.uk/2007/proceedings/papers/889.pdf
Tutt, D., Hindmarsh, J., Shaukat, M., Fraser, M., & McCarthy, M. (2007). Can you see what I see? Developing coherent support for distributed collaborative video analysis with MiMeG. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper168.pdf
Voss, A., Procter, R., Budweg, S., & Prinz, W. (2007a). Collaboration in and for e-research: making the ‘O’ in virtual organisation work. Paper, German e-Science Conference, Baden-Baden. Available at: http://www.cwe-projects.eu/pub/bscw.cgi/d1248208/Collaboration%20in%20and%20for%20e-Research.pdf
Voss, A., Mascord, M., Arguello Castelleiro, M., Asgari-Targhi, M., Procter, R., Fraser, M., Jirotka, M., Halfpenny, P., Fergusson, D., Atkinson, M., Dunn, S., Blanke, T., Hughes, L., & Anderson, S. (2007b). e-Infrastructure development and community engagement. Paper, International Conference on e-Social Science, Ann Arbor, MI. Available at: http://ess.si.umich.edu/papers/paper170.pdf
Waldrop, M. M. (2008). Science 2.0: Great new tool, or great risk? Scientific American. Retrieved 13 June 2008 from http://www.sciam.com/article.cfm?id=science-2-point-0-great-new-tool-or-great-risk
Woolgar, S. (2003). Social shaping perspectives on e-science and e-Social Science. Swindon: ESRC.
[ix] For example, my experiment (http://www.myexperiment.org) and nature network (http://network.nature.com/).
[xii] A recent example is the decision by the UK Arts and Humanities Research Council to discontinue funding of the Arts and Humanities Data Service from April 2008.