As the NCeSS research programme unfolded within the changing technical environment, instead of focussing on grid computing, e-Social Science broadened out to include a diverse range of initiatives exploring how computer support and networking, as well as new sources of data including that harvested from the Web, could be used in new ways to capture people’s views and map their behaviours and their networks. These projects included an exploration of new forms of digital data, such as mobile phone logs and GPS to track people’s interactions (see Chapter 9); the creation and exploitation of metadata (that is, data about data, such as its provenance) to facilitate the sharing and reuse of research data (Edwards et al., 2011); linking data about individuals from different sources and the confidentiality and ethical issues that this raises (Duncan et al,. 2011); webometrics, that is, measuring the number, types and patterns of hyperlinks in the Web (Thelwall, 2009); creating maps of geo-referenced data to reveal patterns such as the location of crime hotspots (Hudson-Smith et al., 2009); large-scale social simulations of, for example, the demand for housing in a city and how it changes over time (Birkin et al., 2010); parallelization of statistical routines to make more efficient use of computing time (Das et al., 2010); enabling researchers to collaborate in marking up videos to highlight significant aspects of the social interactions they record (Fraser et al., 2006); mining large bodies of unstructured text for patterns (Ananiadou et al., 2009a; 2009b); and developing software for delivering behavioural interventions over the Internet (Webb et al., 2010). Many of these initiatives will be further described in the chapters that follow.
As these examples reveal, the e-Social Science programme became highly disparate, expanding to include an increasingly wide range of emerging digital technologies, and drawing on many of the new forms of digital data that were becoming increasingly accessible. The various projects demonstrated that a modest input of technical support could ease existing research processes. This proved particularly productive when there was very close engagement between computer scientists and social scientist users in order to track and respond to changing requirements so that research practices and computing tools could co-evolve. However, successful co-production requires that effective local support structures are established and delivered ‘at the elbow’ of the users (Procter et al., 2013a). This leads on to the wider issue of user adoption, and the barriers to and facilitators for this.
1.2.3 User Adoption
As we noted earlier, the adoption of innovations in research methods and tools has been on a smaller scale to date than the e-Research vision initially anticipated. e-Science’s radical ambitions for transforming everyday research have been tempered in the light of growing evidence about the very real barriers slowing widespread adoption of advanced tools and services across the science community. This extends to the social sciences too. We have already noted that computer packages to support most tasks in the social science research cycle were available before the e-science programme was launched. What many social scientists seek are more efficient or user-friendly versions of these existing digital tools rather than a transformation in their approach facilitated by novel e-Infrastructure, and they have often lacked the resources or incentives to take up the new methods that it offers.
Although a small cadre of ‘early adopters’ – mostly involved in the e-Social Science research programme – have been keen to experiment with innovations and to take risks, adoption of even the broader e-Infrastructure by the wider social science research community has been handicapped by a complex of factors (as has e-Research as a whole: see Voss et al., 2010; Procter et al., 2013a). These include a lack of awareness of the opportunities e-Infrastructure provides; problems in translating innovations in one field into benefits for one’s own research; risk aversion; and levels of IT support that are often dictated by institutional policies and priorities rather than individual researcher needs. Late adopters are often resistant to training and require shallow learning curves if they are to invest in new skills and adopt new ways of working. They may feel they can achieve their career goals – publications and promotions – using the tools with which they became familiar as graduate students. This environment is not conducive to the wide uptake of innovative tools and services or the pushing of boundaries.
Another factor hampering uptake is the uncertain path of technological innovation, which affected the whole of the UK e-Science programme from its launch in 2001. During the early stages of any innovation, the existence of competing technical solutions can be a disincentive to adoption. The emergence of alternatives to grid computing middleware, such as Web 2.0 tools as noted above, introduced uncertainty about the future direction of e-Infrastructure technology development. Studies of previous infrastructure innovations suggest that technological uncertainty may deter some potential users from engaging, at least until a clear technical winner has emerged (Edwards et al., 2007). This uncertainty has been amplified over the last decade as publicly funded research services have faced competition from commercial suppliers, for example, in the provision of cloud computing, with infrastructure, platforms and software all offered to users as subscription services. While this relieves users of the cost of support and maintenance, they lose control over the development path, which is driven by commercial priorities.
A further uncertainty in the future trajectory of emergent e-Infrastructure is its sustainability, that is, the resource-intensive path from research, through software development to delivery of services and support to users. To illustrate: even the more tractable new users will adopt new tools and services only when these are ‘hardened’ to production level, that is, become easy to use, stable, reliable, documented, maintained and fully supported. This requires that software development pathways be created that ensure that e-Infrastructure is able to move beyond the research stage, that is, beyond proofs of concept, demonstrators and prototypes, to production level tools and services. It is ease-of-use and the utility of e-Infrastructure, and its contribution to advancing social scientists’ own substantive research that would persuade them to adopt new ways of working.
The achievement of sustainability is adversely affected by several aspects of the current academic reward system. One is the distinction between ‘pure’ computational research and ‘applied’ software development, with the former bringing rewards for ‘proof of concept’ software innovations but the latter – involving re-building the software to make it robust and efficient – being little rewarded within academia, to the extent that there are few developers to be found even in computer science departments, let alone social science departments. Yet without significant development work most ‘proof of concept’ innovations – such as those emerging from the e-Science programme – are unusable except in the hardware and software context in which the researcher constructed them. Earlier in this chapter, the advantage of software co-production was noted, but this requires collaboration not just between computer scientists and social scientist users, but also the addition of developers to the team, who can re-build innovative tools so that they become project-independent.
There is a similar distinction between both research and development on the one hand and service delivery on the other. The latter requires documentation, online or face-to-face support, FAQs, software maintenance, bug fixes, distribution, porting to new operating systems and so on. Service delivery to support e-Infrastructure is essential for effective and widespread use of e-research resources, but has little place in academia except in a very few specialized units.
Given the co-ordinated efforts of computer scientists, developers and service providers needed to deliver e-Infrastructure that can be readily deployed by users, and the lack of such organizational and human resources in many academic departments, it is not surprising that researchers tend to restrict themselves to the sorts of social science that can be achieved through an unsystematic mix of existing technologies with which