“Is a digital era government more responsive to citizens? What model of democracy does it best support?” Discuss

Who would reject the idea of a democratic and responsive government? Digital era technologies have promised both to make government more responsive to citizen needs, and to revitalise flagging democratic structures. However, just as liberal and democracy were not always terms spoken in the same breath (Dryzek and Dunleavy, 2009), caution should be exercised before running together ‘responsive’ and ‘democratic’ government without critically assessing exactly what forms of responsiveness, and what forms of democracy, are being advocated. In this paper I set out to identity how two particular descriptions of government in the digital era seek to make government more responsive, and to identify the underlying models of democracy they are compatible and coherent with.

Government and Politics: many models of democracy

Three distinctions will prove crucial in the following discussions: between government and politics; between ideas of digital era government and ideas of open government; and between different levels of analysis at which we can consider democracy. I will address each of these in turn, making preliminary remarks about their relationship to each other.

The term government refers broadly to the entirety of the administrative and service provision functions of a state. Health services, tax offices, judicial systems and so-forth are all part of ‘government’. Reference to ‘the government’ picks out the executive of the state who are empowered, subject to constitutional constraints, to direct the management of government services. Government can thus be defined as the institutions involved in the provision, administration and regulation of public services and public resources, in the furtherance of certain goals. Politics is the contestation of those goals (often led by ‘the government’): both through debate over policy, and through competition over the power to legitimately determine the nature and ends of government services and regulations. Government can ‘respond’ to citizens on many different levels: through individual services being customised at the point-of-use; by executives adapting the structure and management of services to bring them into line with general patterns of need and desire; and by political discourse and changes of power that lead to alternative ends being specified for state services and institutions. Increased responsiveness of specific services may also be achieved (or at least, hoped for) by moving them outside the governmental system into private and voluntary sectors. It is possible to have responsive government without responsive politics, and to have responsive politics that struggle to make government services more responsive.

There are many theoretical positions that claim to show how digital technology impacts the potential responsiveness of government, politics and democratic systems. In this paper I address two key positions. The first, Digital Era Governance (DEG), focuses on trends in government administration emerging in part as a reaction to the limits and failures of New Public Management (NPM), and in part resulting from the potential of new technologies that are transferred from the commercial to the government sector (Dunleavy et. al. 2006). DEG outlines linked trends in the re-integration of government services; digitization of services; and the development of ‘needs-based holism’ that claims to re-engineer the business processes of the state around citizens. The second position I explore lacks the definitional clarity of DEG, but has achieved greater popular appeal, going under the somewhat rhetorical title of Open Government (OG). Where DEG draws on ideas from transactional services and the business sector, advocates of OG focus on the transfer of ideas from Internet start-ups, social networks and open-source projects. The OG movement draws together many disparate groups from across the political spectrum, and it is not clear that they share much more than a common dissatisfaction with current political processes and conduct, and a belief that the Internet is a key part of responding to their concerns. However, listening for core terminology of transparency, collaboration and participation (Lathrop & Ruma, 2010), alongside discussion of crowd-sourcing, social media, open source, open data, APIs and coding new tools to improve government demarcates a sizeable community of actors. I will comment later on variations within the broad approach of OG.

If DEG can be understood as an empirical account of developing trends within government, where increases to responsiveness proceed from digitally enabled internal reforms to processes, contemporary OG can be understood as an attempt to remake the government and political systems from the outside. In both cases I contest that these movements need to articulate clear accounts of the form of democratic governments that are created by the changes they identify. For any such account to be comprehensive it must include three elements: a theory of the democratic state; a theory of democratic process; and a normative theory of democracy.

Dryzek and Dunleavy (2009) articulate four classical models of the democratic state: Pluralism, Marxism, Elite Theory and Market Liberalism). They argue that, in contemporary contexts, pluralist ideas (that describe the state as consisting of multiple centres of power and groups engaged in moderate competition) play a key role in US and European political analysis, but are also joined by a de-economised Marxist and elite theory critique of capitalist polyarchies, and a remnant, or perhaps resurgent, new-right market liberalism. These models of the state specify the boundaries of government and outline the relationship between government and other institutions in society: including economic institutions, markets and interest groups. Whilst it might be argued that the digital era creates it’s own novel models of the state, seemingly similar digital era accounts of government can have very different lineages. For example, the idea of crowd-sourcing policy (using the Internet to intentionally gather inputs from beyond the government machine) can be linked to an idea of pluralism and the desire to include multiple, diverse voices in policy making (Wilson, 2008); from a belief that there are ‘elite’ experts best able to direct policy on a given matter outside the core of day-to-day political institutions, and that crowd-sourcing will locate them, allowing them to be co-opted to help government on specific issues (Sunstein, 2006); or from a belief in information markets, and the idea that from multiple inputs an ‘invisible hand’ will support the identification of optimal outcomes (Robinson et. al., 2008). The form that a democratic state takes affects not only how it is responsive to citizens, but also which citizens it is responsive to.

Political processes described as ‘democratic’ vary dramatically in the form they take. In direct democracy, citizens vote on every issue and express direct policy preferences. In a representative democracy, voters delegate responsibility to make decisions on specific issues to representatives, who, through electoral pressures, are supposed to take account of their constituent’s interests and preferences. Deliberative democracy emphasises the role of discourse in bringing about greater understanding and shared (or at least non-polarised) preferences across a population (Fishkin and Laslet, 2003). Whilst early Internet utopians focussed on it’s potential to support direct democracy (Morris, 1999), political scientists have since focussed on the role of the Internet in representative and party-politics, with utopians turning to the potential of the Internet to support improved deliberation (Schuler, 2010).

Keneth Arrow (1950) has shown that no democratic voting system with more than three options can be set-up to satisfy even basic desideratum for fully fair and efficient outcomes. As such, all democratic processes are either an attempt at approximating some democratic ideal, or are adopted pragmatically on the basis of some other higher-order values. It is necessary therefore to specify in any democratic theory not only the processes that are preferred, but the normative basis on which they are selected, and thus to identify when deviations from perfect democratic participation procedures may be acceptable.

Democracy as a set of institutional arrangements may be normatively valued: because it logically proceeds from some inherent values such as the equality or rights of all peoples; because it contingently turns out to be the best system to preserves inherent values such as individual liberty; or because it leads to a set of contingently desirable outcomes such as prosperity, peace and development. It is not uncommon to hear democracy justified, with words attributed to Churchill, as: “the worst option – except for all the rest”. However, even when such arguments are advanced, a positive account of the values or outcomes democracy preserves, and how it interacts with other fundamental values is necessary.

The liberal tradition, articulated progressively by thinkers such as Locke (1689) and Mill (1859), seeks a balance between the role of government and individual freedom – taking the two as in tension and accepting some limits on individual freedom whilst specifying safeguards to ensure the state, and in Mill’s case, the potential ‘tyrany’ of the democratic majority, does not unduly limit individual autonomy. By contrast Rousseau (1762) articulates a social contract whereby individual freedom is increased when individuals live by laws that they ‘given unto themselves’ in collective democratic decision making. For Rousseau, freedom is living in community governed by the prescriptions of the democratically discerned ‘general will’, not being free from state imposition. This positive freedom has been frequently criticised (Berlin, 1969; Popper, 1945), and has deep theoretical problems. However, it does highlight ideas of democracy as about promoting collective goods, rather than preserving private goods (e.g. individual liberty). The distinct, but related, idea that participation in democracy is constitutive of freedom (although in the interests of pluralism rather than a general will) is developed in Hannah Arendt’s work (1958) and finds some echoes, albeit with a very shallow interpretation of democratic participation, in Noveck’s digital era advocacy for ‘Collaborative Democracy’ (2009).

Dahl (1989; 2000) contends democracy is fundamentally founded on ideas of political equality. The more equal access is to political resources (money, power, information), the more democratic a society can be. The idea of democracy founded on an inherent property of human beings (equality in this case) parallels ideas of universal human rights[1]. Dahl admits that all democracies in practice fall short of the egalitarian ideal – but it would not make sense on Dahls account to promote democracy at the expense of equality. Justice is a further principle that may be related to democracy, both as a principle grounding democratic process, and as a contingent outcome of democratic decision-making (Rawls, 1971; Sen, 2009).

The complex constellation of different theories of democratic states, democratic processes and democratic ideals (normative theories) precludes asking any simple questions about whether particular developments in digital era government are democratic or anti-democratic. Rather, understanding these multiple layers of analysis facilitates attentiveness to the assumptions and implications of particular approaches to government, and can show how certain institutional and technological arrangements are sketched with particular notions of democracy implicitly or explicitly in mind. The following two sections will address DEG and OG conceptions of responsiveness and democracy, before we turn to some outline conclusions on government in the digital era.

Digital Era Governance

The responsiveness of governmental institutions, whether to bottom-up pressures from citizens, or top-down pressure from political authorities, is affected by their organisational structure. Organisation structures are intimately tied to, although not determined by, the nature and availability of communication technologies (Shirky, 2008). Changes in organisational structure are dynamic and complex, rather than uni-directional – with centralization and decentralization often taking place in parallel. At the same time that DEG identifies a re-integration of fragmented administrative, regulatory and commissioning functions of the state, the availability of technologies to support communication with, and performance management of, suppliers at a local level is playing a part in the extensive commissioning out of front-line services.

Local authority Children’s Services in the UK provide one such example of information systems affecting organisational reform. Since The Children Act (2004) the multiplicity of services for young people, from education and social care, to youth offending teams and teenage pregnancy services, have all been brought within Children’s Services directorates from separate organisations and departments. At the same time that a holism in policy making and commissioning has been sought through integration of central teams encouragement of joint commissioning, front-line provision has organisationally fragmented, with open competition and tendering leading to a diversity of service providers related to government by contract rather than through organisational hierarchy. However, commissioned services are intended to remain client centred through use of a Common Assessment Framework (CAF) and online client management systems for managing the provision of services to specific young people (Pithouse et. al., 2009). Child protection concerns are managed through the ‘Contact Point’ database designed to keep records on every single child in England and to flag any professional concerns about a child’s welfare. Shared management information systems further link commissioned services to commissioning hubs, where specialised procurement staff manage the contracts.

This is one, frontline, instantiation of government in the digital era. Whilst it doesn’t represent all aspects of DEG, we can explore ways in which such structures, coupled with Internet technologies, increase responsiveness to citizens[2]. Services can become more responsive to a direct request; more responsive to unarticulated needs; more responsive to trends in local need; and more responsive to policy changes at local and national levels. In an environment of fragmented services, a service-user may have to identify, from amongst many services, those appropriate to their need; those that they are eligible for; and those with capacity to support them. The need to share the same data with multiple services, often via form-filling, is removed with the introduction of lead-professionals able to access information from across a wide range of services, and to capture needs-assessment information from a service-user which can then be shared with other services (the ‘ask-once’ processes of DEG). In theory, this supports a service to be both more responsive to the needs that individuals articulate, and, through centralising information from difference ICT systems, to identify the full set of services an individual may be entitled to, without the individual needing to themselves be aware of, or apply for, those services. Dunleavy et al., (2006: 235) give one example of how the re-engineering of internal processes allowed a Canadian social security office to integrate information from existing IT systems and reduce a 30-page pensions application form into a welcome letter and statement of entitlement. However, whilst such integration of citizen information appears benevolent when notifying people of pension or benefit entitlements, when information in government databases is used to target services that individuals have neither requested, nor desire, the picture is more complicated. For example, many of the local Targeted Youth Support (TYS) programmes supported by the UK government (DCSF, 2010) encourage the sharing of data on ‘risk-factors’ between services to identify individuals who are statistically more likely to be involved in offending or truancy behaviour, and then to target services at these individuals. Whilst this may be seen as responsive government, it is certainly not responding to the desires or wishes of the individuals who are targeted: rather, it is responding to general policy agendas by treating individuals not as citizens, but as objects of interventions (Garret, 2004).

Responsiveness of frontline services to policy priorities is mediated through use of indicators (including National Indicator sets) and data derived from management information systems. Online ‘data observatories’ can support local areas to pool data from different domains, bringing together, for example, health, education and labour market statistics, and enabling improved commissioning and management responsive to local trends. However, the ability of local and national data collection to impact service responsiveness is constrained in two key ways. Firstly, it requires good quality data. Secondly, the structure and nature of data collected affects the available policy levers, and ultimately, the viable policy options for government.

Whilst some areas of government activity, such as traffic management, can take advantaged of ‘zero-touch technology’ to collect information on patterns of behaviour and service use, many frontline public services rely for their key detector mechanism (Hood and Margetts, 2007) on staff entering data into computer terminals. Studies of frontline staff IT skills, particularly staff located in community rather than office settings, are conspicuously lacking from the research literature on e-government, even though simple skills such as typing speed and literacy using online tools can dramatically impact both the extent to which staff fill in digital forms in accurate, timely and efficient ways; and the relationship of a professional to the information system and their client.

Even when good quality data is collected, that data is necessarily a reductive account of reality (Bowker, 2000). Scott (1998) notes that in attempts by governments to render their fields of operation ‘legible’ to their data collection processes, governments seek to impose an order on the diversity of local practices, and that imposition of order occurs not only in the way data about situations is represented, but in the way reality is modified to accord to the structures through which the state seeks to record it. For example, the articulation of UK policy and targets related to young people not in education, employed or training (NEET) has spurred the creation of local services that bring together diverse groups of young people whose only common attribute is their employment status (Yates and Payne, 2006). Once particular categories and structures of data collection are established, changes of policy that need to be managed through new metrics could require to complex and expensive changes to IT systems. Whilst DEG identifies the promise of modularised ‘utility computing’ for increasing the flexibility of government IT infrastructure, the realisation of that promise remains a long way off (Dunleavy et. al., 2006), and it does not resolve the tension of abandoning existing data series when they do not fit with new policy options.

To summarise the exploration of DEG so far: digital technology supports an increase in the potential responsiveness of front-line services, but such increase are contingent on local systems and ICT skills of staff; increases in the data available to commissioners through the Internet can support more responsive service development; however, responsiveness of local services to central policy is increasingly contingent upon, and intertwined with, capabilities and structures of management information systems.

DEG, as a comparative framework for looking at different states, is relatively silent on the political dimensions of governing. However, a number of remarks can be made on the relationship between DEG trends and democratic government. Firstly, in frequently framing the citizen-state relationship as customer-supplier, and adopting images and practices from the commercial sector, digital era governments risk setting up a significant expectations gap (Flinders, 2009). Whilst commercial sector entities can ignore provision of service where it is not cost effective, governments, with politically important universal service obligations, cannot. Furthermore, government is not building institutions from scratch, but operates with many legacy structures and relationships. The apparently inefficiency of a given government service when compared to a comparable commercial service may proceed from the complexity of delivering that service at a national scale to all people who need it, without price-discrimination, and working with legacy contracts and relationships. For example, public health services have to balance provision of services to individual patients with the careful use of central resources, and an awareness of health inequalities requiring some redistribution of resources to areas most in need.

In the digital era, government’s failings are not expressed solely in complaint letters, or occasional letters to local newspapers, but are shared and discussed through online social media and, increasingly, on dedicated websites set up for rating public services such as PatientOpinion.com. PatientOpinion.com CEO Paul Hodgkin has written about the creation of a ‘new economics of voice’ capable of driving changes in health service provision (2009). The terminology of ‘economics’ draws attention to market-liberal models of the state, in which equipping ‘customers’ of government with a choice of services and good information about those services is argued to drive improvement and close the expectations gap. However, Hodgkin’s own argument focuses on pluralist ideas of citizens actively engaging with health services to reshape them through dialogue rather than through consumer choice. Notably absent from such discussions are ideas that political representatives may play a role in reshaping services. Instead, representational politics is sidelined for digitally mediated citizen participation as the driver of change. The expectations gap does, however, impact on electoral politics. Whilst blame for disaffection in politics and the current crisis of legitimacy in UK politics has many roots, not least in a mainstream media focussed on scandal (Castells, 2009) and some very real recent scandals, bi-partisan adversarial politics of the UK and US does not benefit from contexts in which implausible forms of citizen-centred government are promised to the electorate, and then only delivered part-completed, late, and over-budget. Prime Minister Brown’s recent announcement of a MyGov personalised portal for every citizen is the latest in a series of bold technological claims creating another likely expectations gap.

Secondly, although linked to the idea of the ‘customer citizen’, is the perceived necessity in DEG to identify citizens in order to regulate entitlement to services, and the desire to make that identification digital. Although decentralised models of secure entitlement identification do exist (Cameron, 2005), governments have tended to push for centralised identity databases. For political theories that take state authority as contingently delegated from citizens (such as John Locke), the explicit but opaque (Edwardes et. al. 2007) implementation by the central state of databases that becomes the authoritative source of information on individuals’ rights is an anti-democratic step.

Thirdly, DEG affects the complexity of government, and, in open-book government trends, the public availability of information on governmental and political activities. At the central level, in reversing NPM fragmentation of the state, DEG arguably simplifies state structures, and brings many more services back under direct political control. Bimber (2003) discusses the relationship between the complexity and availability of democratic information, and the nature of those who participate in decision-making. Increased availability of information, and a simpler state, facilitates the involvement of a wider number of people in democratic decision-making, and challenges the dominance of bureaucratic and lobbying elites. This does not, in Bimber’s view, lead to mass engagement and the direct involvement of large numbers in public policy, but it does drive a quiet ‘revolution in the middle’ whereby new groups, formerly excluded by lack of access and limited organisational capacity, can mobilise and join discussions in the political arena: a progressive increase in pluralism as the informational costs of participation fall.

When it comes, however, to significantly opening up democratic and policy participation to new groups, it is not DEG framings of the state with the most to say, but the bloggers and essayists of the developing open government (OG) movement.

Open Government

The terminology of OG has a history going back to Enlightenment opposition to government secrecy and has more recently been used in discussing the Freedom of Information laws that many states implemented from the 1970s onwards (Chapman and Hunt, 2007) However, the current movement has two distinct points of focus: open data; and open policy making processes. Whilst much writing on OG draws on case studies of existing projects, it would be wrong to understand OG as an empirically grounded theory equivalent to DEG. Rather, it is a set of both ideological and practical prescriptions for changing government and politics, based on insights and ideas from open source and online communities.

On open data, the OG movement argues that all government data (or at least, non personally identifying data) should be made available online in machine-readable formats by default, with some exceptions for sensitive security data. Open data is seen as important for a number of reasons. Firstly, it increases the transparency and accountability of government. Secondly, it allows third-parties to provide innovative interfaces onto, or services with, government data (Mayo and Steinberg, 2007). Thirdly, given government is theoretically paid for by citizens then the products of government in data, which can be shared with near to zero marginal cost, should be made openly available to citizens (Pollock, 2009). Whilst some open data advocates focus on Public Sector Information (Aichholzer and Burkert, 2004), much attention in practice has been given to political data such as voting records, financial contributions to political parties and debate records.

Developers have used open data to build a wide range of web applications, ranging from tools presenting parliamentary or congressional voting records to help citizens discover their elected representatives legislative activities (Tauberer, 2010; Edwards, 2006), through to feedback tools that collect views from service users on government provision and feed those back to the government institutions in question (Hodgkin, 2009; King and Brown, 2007). Lawrence Lessig (2009) has suggested that much surfacing of government data, whilst desirable from a ‘naked transparency’ perspective, can have a negative impact on the overall responsiveness of political and governmental elements of the system. More information doesn’t always lead to greater efficiency or responsiveness. Advocates of targeted transparency note that a focus on online openness from government alone can disempower government vis-à-vis other non-open institutions such as the media and corporations (Fung and Weil, 2010).

Drawing on a perception of government’s limited or ineffective exploitation of technologies, OG activists have sought to “code a better country” – suggesting (and implementing) new tools and technologies based on open data. The modularisation of ‘government as a platform’ (O’Reilly, 2010) essentially bypasses the need to make government more responsive, as it allows a large community of technically capable citizens, alongside private and third-sector organisations, to meet specific needs through developing software on top of government services. As yet, the evidence base to show that open data initiatives do lead to widespread development of new applications responsive to citizen needs is not available. However indications from ‘Apps for Government’ competitions such as ShowUsABetterWay[3], and hack-day events such as Rewired State[4] suggest that, while many ideas have been explored, only a limited number have developed into completed ‘products’, few with widespread use. A number of the more successful application have focussed on moving dialogue with government into the open: allowing citizens to host peer-to-peer discussions around their interaction with governmnt. Websites such as the aforementioned PatientOpinion.com and the much-cited FixMyStreet.com are run interpedently of government and allow individuals to submit public feedback on government services. Whilst this can arguably increase the responsiveness of services to public pressure, King and Brown (2007) note a varied pattern, with many of the reports to FixMyStreet resolved quickly, but large numbers also remaining unresolved for long periods.

Opening up data is often seen as part of opening up the policy-making processes – allowing citizens to scrutinise government activity and to make informed contributions to discussions on issues they care about. However, open policy making goes further – describing ways in which government can use Internet-based tools to incorporate a diverse range of actors into policy making, through mass consultations (Sifry, 2010); design competitions; crowd-sourcing exercises (Sunstein, 2006) and other digitally mediated processes. Ellen Miller (2010: 193) articulates a vision of OG as challenging the role of insider policy-making elites, claiming: “the outsiders are becoming ‘insiders’”. However, Miller’s claims raise questions about whether OG constitutes an elite-theory critique of the state, or simply a challenge to established elites that their time is up, and the technological elites want their turn to run government.

The broad label of open government masks many different views on the democratic state. An emphasis on political transparency generally involves an acceptance of representative government, but a critique of the elite-domination of contemporary politics and a desire to increase the citizen control over elected representatives. By contrast, whilst Burton (2010) has argued that developers from the private sector should work for short periods within government, as a sort of civic minded ‘Developer Corps’ to support the solution of complex problems through the state, Bryant (2009) talks about technology allowing a smaller state and increasing individual freedom through limiting the domain of government. Many such ‘small state’ ideas draw upon notions from Benkler (2007) suggesting that digital technology facilitates collective action without coercive force, and allows individuals increased choice, thus increased individual freedom. Crenson and Ginsberg (2003) note that such ideas of ‘personal democracy’, whereby participation in the public sphere is based on individualised interaction rather than collective participation, only works “for those in a position to take advantage of its possibilities”. It is notable that whilst conceptions of ‘collaborative democracy’ advanced by current deputy US Government CIO Beth Noceck (2010) avoid individualist notions of participation, they split civic engagement into small tasks – similarly rejecting broad collective democratic participation in place of frequent, but narrow, state-society collaborations around specific issues.

Conclusion

Government in the digital era cannot be disentangled from government in an era of capitalist markets. Both DEG and OG are prone to frame citizens as consumers, and, whilst making viable claims to increase the responsiveness of some services to some individuals, fail to address distributional aspects of increased responsiveness. For normative conceptions of democracy based on equality (E.g. Dahl’s), this is a serious issue. For libertarian normative conceptions of democracy, prioritising individual freedom, it is a lesser concern. It is not the case, however, that digital era states are necessarily market-liberal states. Increases in information availability and openness of government processes support a form of ‘revolution in the middle’ that Bimber discusses (2004), increasing space for formal and informal plural interest groups to engage in public discourse. Digital era government can support many models of democracy: individualistic and collective. However, when DEG of OG are discussed solely in terms of ‘technical’ interventions, their wider impacts on democracy and social justice are often masked, and so constructing the discourse within which all citizens, not just the technically adept, can debate DEG and OG changes is a core task for any democratic theory that rejects the idea of blindly accepting new technological elites.

Bibliography

Children Act 2004 (c. 31), Available at: http://www.opsi.gov.uk/acts/acts2004/ukpga_20040031_en_1 [Accessed April 18, 2010].

2009. Rewired State. Available at: http://rewiredstate.org/ [Accessed April 21, 2010].

Aichholzer, G. & Burkert, H., 2004. Public sector information in the digital age: between markets, public management and citizens’ rights, Edward Elgar Publishing.

Archard, D., 2004. Children: rights and childhood, Theatre Arts Books.

Arendt, H., 1958. The human condition, Chicago.

Arrow, K.J., 1950. A difficulty in the concept of social welfare. The Journal of Political Economy, 58(4), 328–346.

Benkler, Y., 2007. The wealth of networks: How social production transforms markets and freedom, Yale University Press.

Berlin, I., 1969. Two concepts of liberty.

Bimber, B.A., 2003. Information and American democracy: Technology in the evolution of political power, Cambridge University Press.

Bowker, G.C., 2000. Biodiversity datadiversity. Social Studies of Science, 643–683.

Bryant, L., 2009. People power can reboot Britain. In Reboot Britain Essays. NESTA.

Cameron, K., 2005. The laws of identity. Microsoft Corporation

Castells, M., 2009. Communication power, Oxford University Press.

Chapman, R.A. & Hunt, M., 2006. Open government in a theoretical and practical context, Aldershot: Ashgate.

Crenson, M.A. & Ginsberg, B., 2003. From popular to personal democracy. National Civic Review, 92(2), 173–189.

Dahl, R.A., 1989. Democracy and its Critics, Yale University Press.

Dahl, R.A., 2000. On democracy, Yale University Press.

DCSF, 2009. Targeted youth support – Every Child Matters. Available at: http://www.dcsf.gov.uk/everychildmatters/Youth/targetedyouth/targetedyouthsupport/ [Accessed April 18, 2010].

Dryzek, J. & Dunleavy, P., 2009. Theories of the Democratic State, Palgrave Macmillan.

Dunleavy, P. et al., 2006. Digital era governance: IT corporations, the state, and E-government, Oxford University Press, USA.

Dunleavy, P. et al., 2006. New Public Management Is Dead–Long Live Digital-Era Governance. J Public Adm Res Theory, 16(3), 467-494.

Edwardes, C.A., Hosein, I. & Whitley, E.A., 2007. Balance, scrutiny and identity cards in the UK. Criminal Justice Matters, 68(1), 29–30.

Edwards, A.R., 2006. Facilitating the monitorial voter: retrospective voter information websites in the USA, UK and The Netherlands. Electronic Government, and International Journal, 3(3), 289–305.

Eggers, W.D., 2007. Government 2.0, Rowman & Littlefield.

Fine, A. et al., 2008. Rebooting America; Ideas for Redesigning American Democracy for the Internet Age,

Fine, A. et al. eds., 2008. The Power of Inclusion. In Rebooting America; Ideas for Redesigning American Democracy for the Internet Age.

Fishkin, J.S. & Laslett, P., 2003. Debating deliberative democracy, Wiley-Blackwell.

Flinders, M., 2009. BRIDGING – THE – GAP: REVITALISING – POLITICS – AND – THE – POLITICS – OF – PUBLIC – EXPECTATIONS – PB – Routledge, Representation, 45(3), 337.

Freiere, P., 1975. Pedagogy of the Oppressed, Penguin Education.

Garrett, P.M., 2004. The electronic eye: emerging surveillant practices in social work with children and families. European Journal of Social Work, 7(1), 57–71.

HMRC, 2009. HM Revenue & Customs: Joint HM Revenue & Customs and Companies House statement. Available at: http://www.hmrc.gov.uk/ct/ct-online/joint-filing-statement.htm [Accessed April 18, 2010].

Hodgkin, P., 2009. How the new economics of voice will change the NHS. In Reboot Britain Essays. Reboot Britain. NESTA.

Jim, K., 2009. House of Commons Hansard Written Answers for 27 Mar 2009 (pt 0017), Available at: http://www.publications.parliament.uk/pa/cm200809/cmhansrd/cm090327/text/90327w0017.htm#09032765004280 [Accessed April 21, 2010].

King, S.F. & Brown, P., 2007. Fix my street or else: using the internet to voice local public service concerns. In Proceedings of the 1st international conference on Theory and practice of electronic governance. Macao, China: ACM, pp. 72-80.

Lathrop, D. & Ruma, L., 2010. Open Government: Collaboration, Transparency, and Participation in Practice 1st ed., O’Reilly Media.

Lessig, L., 2009. Against Transparency. The New Republic. Available at: http://www.tnr.com/article/books-and-arts/against-transparency [Accessed April 21, 2010].

Locke, J., 1689. Two treatises on civil government, G. Routledge and sons, limited.

Mayo, E. & Steinberg, T., 2007. Power of Information Taskforce Report, Available at: http://www.cabinetoffice.gov.uk/reports/power_of_information.aspx [Accessed December 1, 2009].

Mill, J.S., 1859. On liberty,

Miller, E.S., 2010. Disrupting Washington’s Golden Rule. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Miller, P., 2009. Weary giants and new technology. In Reboot Britain Essays. NESTA.

Morris, D., 1999. Vote.com, Renaissance Books.

Noveck, B.S., 2009. Wiki government: how technology can make government better, democracy stronger, and citizens more powerful, Brookings Institute Press.

O’Reilly, T., 2010. Government as a platform. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Pithouse, A. et al., 2009. A Tale of Two CAFs: The Impact of the Electronic Common Assessment Framework. Br J Soc Work, 39(4), 599-612.

Pollock, R., 2009. The Economics of Public Sector Information. Available at: http://econpapers.repec.org/paper/camcamdae/0920.htm [Accessed March 4, 2010].

Popper, K.R., 1945. The open society and its enemies, Routledge.

Rawls, J., 1971. A Theory of justice. Harvard University.

Rawls, J. & Kelly, E., 2001. Justice as fairness: A restatement, Belknap Press.

Robinson, D. et al., 2008. Government data and the invisible hand. Yale Journal of Law and Technology, 11, 160.

Robinson, D.G., Yu, H. & Felten, E.W., 2010. Enabling Innovation for Civic Engagement. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Rousseau, J.J., 1762. The Social Contract. The classics of Western philosophy, 316.

Schuler, D., 2010. Online Deliberation and Civic Intelligence. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Sen, A., 2009. The idea of justice, Belknap Press.

Shirky, C., 2008. Here comes everybody: the power of organizing without organizations, Penguin Press.

Sifry, M.L., 2010. “You Can Be the Eyes and Years”: Barak Obama and the Wisdom of Crowds. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Sunstein, C.R., 2006. Infotopia, Oxford University Press US.

Tauberer, J., 2010. Case Study: Govtrack.us. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

UN, 1989. Convention on the Rights of the Child, Available at: http://www2.ohchr.org/english/law/crc.htm [Accessed April 18, 2010].

Yates, S. & Payne, M., 2006. Not so NEET? A Critique of the Use of ‘NEET’ in Setting Targets for Interventions with Young People. Journal of Youth Studies, 9(3), 329.

Footnotes

[1] Although even Dahl restricts democratic equality to ‘adults’, following a long line of thinkers failing to justify disenfranchisement of under 18s. See Archard (2004) for justifications of extending democratic rights to young people.

[2] I understand citizens to include children and young people, in line with the UN Convention on the Rights of the Child (UN, 1989).

[3] http://www.showusabetterway.co.uk/ – Accessed 17/04/2010

[4] http://rewiredstate.org/projects – Accessed 17/04/2010

The current push to make public sector information available as raw data in open digital online forms will radically transform research. Discuss, with reference to a number of case studies

Interest in ‘opening up’ public sector information (PSI) on the Internet has been growing in recent years (Aichholzer & Burkert, 2004), gaining a significant boost in Europe with the 2003 PSI Directive (European Commission), and gathering pace in the last twelve months, particularly in the United States and United Kingdom with high-profile central government open data initiatives (e.g. the launch of Data.gov.uk and Data.gov). However, whilst much has been written about the economic impacts of opening access to PSI (Pollock, 2009; Uhlir, 2009), and strong claims have been made about democratic potentials of open data (Mayo and Steinberg, 2007;), consideration of how open data impacts on academia have been limited, with benefits to researchers being little more than a footnote in studies, if noted at all (Sharif, 2009). This paper explores the extent to which the ‘revolutionary’ potential of open raw PSI could lead to radical transformations in how social scientific research is conducted and communicated, with a specific focus on the UK context.

Given the relative novelty of large quantities of PSI being available as open raw data (Data.gov.uk launched with just 1126 datasets in September 2009, and many of those datasets contained only aggregate statistics rather than raw data[1]), this paper necessarily takes a theoretical and speculative approach, looking at prior studies and papers in neighboring domains, and drawing upon emerging trends from primary exploratory research and from online resources (grey literature). The paper starts by clarifying definitions of PSI as open, raw data, before exploring relationships between data and research. After considering a number of case studies relating to research with PSI, open data and linked data, preliminary conclusions about the likelihood of radical change in research practice are advanced.

PSI

The OECD define PSI as information“dynamic and continually generated, directly generated by the public sector, associated with the functioning of the public sector…” (Vickery & Wunsch-Vincent, 2006), and the European Commission PSI Directive (2003) notes this includes: “social, economic, geographical, weather, tourist, business, patent and educational information” (§4). In the EU context PSI has historically been kept solely for use by government and it’s agencies (PIRA, 2000; Weiss, 2004), or has been available only in costly printed or proprietary formats, licensed to re-users under restrictive terms via contractual relationships and trading funds (De Saulles, 2005; Pollock, 2009).  The MEPSIR study (Deckers et. al., 2006) identifies six key domains of PSI: Business Information; Geographic Information; Legal Information; Meteorological Information; Social Data; and Transport Information. Undoubtedly much of this information is of interest and use to researchers. PSI, defined to exclude personally identifying information, should be distinguished from administrative data and micro-data, which may include disclosive information (Elias & Jones, 2006).

Whilst open data initiatives mean that many PSI datasets are available to the public for the first time, researchers in recognized institutions have long had access to PSI datasets, including through the Administrative Liason Data Service (ADLS) and the UK Data Archive (UKDA). However, the push towards open data introduces two key differences: opening up PSI removes the privileged research access that academics formerly had to this data; much PSI is being placed online in new raw and linked data formats, affecting the research potential of that data. We now turn to an account of what is meant by open data, before looking at different data formats being adopted for sharing open PSI.

Open

Just as advocates of free software have to distinguish between different models of free, often done using Stallman’s famous ‘Free as in beer vs. Free as in speech’ dichotomy, (Stallman quoted in Williams, 2002), so too those exploring ‘open data’ need to distinguish different forms of openness. Understandings of Open Science, the dominant, though not only, paradigm in the generation of public knowledge since the mid-1660s (David, 2004) stress the communal nature of scientific enquiry, with results shared ‘universally’ to all peers regardless of their personal attributes, and with all findings and methods disclosed regardless of whether a hypothesis was proven or disproven. This is a very specific form of openness between intellectual and organizational peers, developed for the purpose of allowing trust in scientific findings, by facilitating ‘organized skepticism’ (Merton, 1973 in David, 2004). Open Science has generally not been seen as undermined by boundaries that have the practical impact of restricting access to findings or methods only to those within a given domain of science. By contrast, advocates of Open Access (OA) argue that openness involves the removal of boundaries (e.g. payment of subscriptions, or requirements of institutional affiliation) that prevent any individual from accessing content. The OA movement is integrally an Internet based movement, evident in the Bethesda Statement on Open Access Publishing (2003[2]) that talks of publishing through ‘digital media’, and the Budapest Open Access Initiative (2002[3]) which discusses convergence between Open Science principles and the affordances of digital networked technologies. However, OA principles such as the Bethesda Statement often impose limitation on re-use of published works, not least restricting production of ‘hard copies’ to personal use only – a concession of the economic interests of print-based journal publishers. Some OA licenses take a different track, pushing against the interests of publishers, and incorporating non-commercial terms in their licenses. OA advocate John Willinsky seeks to provide a progressive compromise position in his ‘access principle’ (2006) replacing a descriptive definition of OA with an ideological commitment to “…extend the circulation of [a] work as far as possible and ideally to all who are interested in it and may profit by it.” Whilst more conservative than some OA principles (with the cautious caveat ‘as far as possible’), Willinsky’s principle introduces active encouragement to disseminate open works, rather than simply allowing that the works may be shared.

Open Knowledge arrangements (OKF, 2006), such as those compliant with the Open Knowledge Definition (OKD), move far closer to the ‘Free as in speech’ model, where open is taken to equate to placing content into, or close to, the public domain. The OKD (Version 1.0) is summarized by the statement “A piece of knowledge is open if you are free to use, reuse, and redistribute it”[4] and includes terms concerning access to works, royalty-free redistribution, derivative works, open technical standards without digital rights management, and non-discrimination in licensing of a work.

Talk of open data needs to be clear whether it is dealing with open ‘as in Science’, ‘as in Access’, or ‘as in Knowledge’. Both open science, and OA environments have struggled with sharing data (as opposed to published analysis of data), for reasons including cultures of data ownership, technology (e.g. the difficulty of archiving persistent data alongside journal articles), and legal restrictions. A number of the advisory council members overseeing the OKD have been involved in articulating explicit ‘open data’ principles for ‘open science’ arguing in the Panton Principles (Murray-Rust et. al, 2009) that “data related to published science should be explicitly placed in the public domain[5]”. Science Commons also encourage a ‘public domain’, rather than intellectual property based model, for open data.

When PSI is provided as a research input under OKD terms, it is necessarily open access, and, providing published research is similarly open access and open about it’s methodology, open science is supported and extended beyond restricted academic domains. Whilst licensing terms of PSI data projects such as data.gov.uk are still under development, and the current ‘Crown Database Rights’ and ‘Crown Copyright’ models are somewhere between OA and OKD approaches, for the purposes of research (where misrepresentation of data, prohibited by Crown Copyright, is ruled out by the commitments of scientific practice) we can consider a majority of open PSI as Open Knowledge.

Raw Data

Data is commonly processed between collection and release, and in the case of PSI, data is often released only after considerable manipulation. Some data processing aims correct errors in the dataset and control for measurement errors; other processing prepares data for specific uses: for example, removing disclosive information and converting individual observations into aggregate figures. In such aggregations, information from the original dataset is lost, and certain future uses of the data are precluded. Much of the data released through data.gov.uk is aggregate data, but a number of ‘raw’ datasets are also available – providing data points for each observation originally taken, rather than aggregates.

The call for “Raw Data” is one that has been popular with technology developers (Robinson et. al. 2010), but which researchers may have a more cautious approach to. Raw data is not always the most useful data (Gray, 2009). Whilst web developers building mapping mash-ups with PSI may not be concerned with occasional outliers and erroneous points in a dataset, social scientific researchers are likely to be concerned to clean and check raw data carefully. However, both developer and researcher will be interested in the data structure – as how data is structured can affect ways it can easily be used. In the case of data.gov.uk, particular effort is taking place, with the involvement in the project of linked data advocates Nigel Shadbolt and Tim Berners-Lee , to publish RDF linked data (Berners-Lee, 2009; Shadbolt et. al, 2006; Alani et. al., 2007). Unlike data tables that use data rows and named columns, RDF models data through ‘subject, predicate, object’ triples, where any element of the triple may be either a literal string, or a URI. Linked data URIs are dereferenceable. A HTTP request to a URI should return useable data about the subject of that URI: machine readable data when requested. Given URIs could exist anywhere in the Internet, this allows the distributed linking of datasets. When common schema and ontologies are used to model data, and common URIs used to refer the same subjects in multiple datasets, machine assisted information retrieval and processing, and in some cases, logical inference, becomes possible across datasets without requiring that they be integrated first.

The use of raw linked data for the provision of open PSI has interesting implications for researchers. Firstly, there is limited familiarity amongst researchers, particularly social science researchers, with the processes involved in using RDF data and querying RDF data stores. This may be expected to change over time, but as Dutton and Meyer (2009) have shown, the development and distribution of digital research skills throughout the social scientific research community can be a slow process, as generations of researchers trained before the advent of certain technologies move through their careers. Secondly, and at a more theoretical level, the choice of data structures and ontologies for modeling raw data can have a profound impact on the questions that can be asked of it (Bowker, 2000). Government data already has certain biases  present by virtue of having been collected for governmental, rather than research purposes. These biases can be compounded by choices about semantic structures applied in RDF models, and choices over which fields are treated as mandatory or optional. The third issue for research use of raw linked data is trust. Whilst meteorological and traffic data may be released entirely ‘raw’, or with minimal transformations applied, much ‘social’ data will have been heavily processed. In a web of linked data, knowing who has been involved in prior manipulation of data, and the extent to which their processing can be trusted as supporting, rather than undermining, effective social scientific research, is key. However, the trust layer of the semantic web stack (Berners-Lee, 1998) is one it it’s most underdeveloped (though heavily researched) aspects, and at present, no clear trust, audit or provenance model is applied to UK open PSI data.

Data and Research

Having access to data is fundamental to carrying out research (Elias, 2007), but, as Cole et. al. note (2008), social scientific data is often expensive or difficult to generate. Secondary research allows analysis of data created by other projects (Carmichael, 2008) and open raw PSI can provide a potentially rich resource for secondary research.

The relationship between researchers and their data varies between fields and disciplines. In some fields, small qualitative datasets gathered and analyzed by a single researcher predominate, whereas in others, vast shared datasets, generated through research technologies and instrumentation, are available for analysis and as inputs for computer aided research by globally distributed research teams. Sawyer (2008) articulates a distinction between ‘data-rich fields’, such as astrophysics, biology and ecology, and ‘data-poor fields’, including humanities and social science. He argues that in data-rich fields: “pooling and sharing of data is expected” resulting in scholars developing common understandings of the datasets; “form(s) drive methods” and the easy availability of data reduces the likelihood of additional data collection during research; and there are “few(er) theoretical groups”, as shared data forces out empirically unsupported positions, but at the same time, different positions become more entrenched as their conflicts originate in differential interpretation of shared data, rather than disagreement over the data itself. By contrast, in data-poor fields Sawyer argues that “data is a prized possession”;  “access to data drives methods”; and there are “many theoretical camps”. Whilst we might question Sawyer’s precise claims, for which he provides little empirical justification, there is clearly a relationship between data availability and research practices. Findings from research into digitization projects underline this. Meyer et al. (2009) report in the evaluation of five digization projects that, as digitized resources became available, a noticeable shift towards quantitative methodologies could be observed in proposed research papers.  However, the amount of data availability is not the only thing which affects it’s use.  The ‘contents’ of data matters as well – a point missing from Sawyers analysis. Where physical sciences are generally dealing with data about physical phenomena, social scientists are often dealing with data about people, and potentially sensitive data if it’s use could facilitate discrimination against groups, or harm to individuals (Solove, 2005). These ethical differences between uses of data in different fields can have significant impacts on how data is handled and used.

Whilst in an ideal process of social scientific work, facts are produced based on rigorous analysis of data, in ways that can be repeated and are open to Popperian falsification, when data is not published along with findings, open science cannot easily be practiced. Latour and Woolgar (1986) have suggested that ideal processes are rarely followed in practice, and thus increasing transparency of research may either require changes to research, or chances to claims about what makes research valid (Busek, 2008).

We now turn to three case studies relating open, linked and raw data to research.

Case Study 1: Traffic Data

One early UK government experiment with publishing open raw PSI was the release, through the Directgov Innovate website, of a single dataset containing locations of cycle accidents across the country between 2005 and 2007 (Clarke, 2009). The data was quickly used to produce mash-up maps showing cycle accident hotspots and to suggest ‘safer cycle routes’, and has since been cited extensively in discussions of government open data. However, the accuracy of the mash-ups can be questioned. Simply placing markers on a map where accidents have occurred without taking into account the severity of those accidents or drawing on background information about normal traffic levels at accident spots to show statistically significant high levels of accidents, can give a distorted account of where the safer cycle routes are. Whilst datasets providing traffic counts for road across the UK have since been released through Data.gov.uk[6], such data has not yet been combined with the accident counts in any publically available papers or mash-ups.

The STATS19 dataset from which cycle accident data was derived was already available to researchers through the UKDA, so was not new PSI, although it’s open release allowing for it, and derived data from it, to be freely redistributed was novel. Past use of STATS19 has found its accuracy to be questionable (Gill et. al., 2006), a finding echoed by cyclists discussing the released dataset in online fora (Cycle Chat, 2009). The dataset relies on reports from police forces, using a three-category classification of accidents into fatal, serious and slight injuries, where the last two categories are assigned according to subjective police officer judgments with no test of coder reliability. Whilst no social scientific data is perfect, and data errors are inevitable in any large dataset, the contrast between the standards applied to data gathered as PSI, and then released for research, as opposed to data gathered directly for research must be noted. For example, the DVLA estimate that 11.4% of their vehicle records, and 26.18% of their driver records contain errors (Watson, 2010). Although this is population data, as opposed to sample data, an error rate of over 25% may present problems to many research applications.

Lyons et. al. (2008) have found that more rigorous understandings of traffic accidents can be obtained through the comparison of multiple datasets including hospital admission records. They describe future analysis which draws upon “pseudonymised data linkage” and suggest that “increasing availability of the casualty postcode in the STATS19 data allied to increasing availability of electronic emergency department data should facilitate more accurate large scale police–health linkage studies.” However, in drawing upon potentially disclosive micro-data such as emergency department statistics Lyons et. al. have quickly moved beyond the straightforward use of PSI, and any datasets that result from such work are likely to again be subject to restrictions on distribution rather than to exist as open datasets.

Case Study 2: Clear Climate Code

Our second case studies comes from outside social sciences, and outside academia to explore how advocates of open science and accessible climate science have responded to a long-available data series – the GISTEMP (Goddard Institute for Space Studies Surface Temperature Analysis) analysis from NASA[7]. GISTEMP uses public domain PSI detailing observations from climate stations across the world, and generates monthly climate analyses of global temperature using a complex series of FORTRAN, C and Python programmes.  In 2009 a small group of programmers decided to re-implement the GISTEMP code purely in Python, with the goal of making it easier to understand: increasing the transparency of, and consequently confidence in, the GISTEMP results. This Clear Climate Code project[8] has successfully generated accessible code that, drawing on the same data as GISTEMP generates equivalent analysis. The project has also led to changes in the NASA GISTEMP code, after the re-implementation identified bugs in the original[9]. The existence of both open data, and open source code, allowed actors outside a narrow community of scientists to engage in the practice of open science, and to address wider relationships between science and society.

Case Study 3: Edubase and Educational Research

PSI regarding schools and educational attainment provide another example of data previously available under restricted conditions to researchers, but now widely available through open data initiatives. In particular, the EduBase dataset detailing schools across England has been converted into RDF linked data for Data.gov.uk and efforts are underway to link this to a range of further datasets including OFSTED school inspection data and pupil attainment data[10]. This involves creation of, and agreement on, unique identifiers for specific schools, and their reconciliation across government datasets. The modeling of Edubase in RDF also makes linkages with recently released geodata (Ordnance Survey, 2010) that can be linked to statistical information on Wards and other geographical areas. Whilst such linkages may have been achieved in past research, they were only achieved using proprietary tools and datasets. The ability of query across these linked datasets and, using semantic web agents (Brent, 2008) to draw in information from the wider linked web of data (where government minted URIs may become key nodes), introduces new possibilities for social scientific research.

Discussion

Does the release of open raw PSI herald the shift of social science from data-poor, to data-rich, with the resulting changes in disciplinary practice that Sawyer’s (2008) theory would predict? And can the creation of large linked semantic social datasets provide the forms of research technology that Collins (1994) suggests could enable social science to become a ‘high-consensus, rapid-discovery’ science? Revolutions are easy to predict in a synthetic analysis, but a grounded analysis must be more cautious.

It has already been noted that few open PSI materials were unavailable to researchers in the past, or at least, to researchers interested enough to seek them out. However, this does not mean that providing PSI as open raw data has no impact on the conduct of research. Firstly, as datasets formerly governed by restrictive agreements come to fall under Open Knowledge compliant licenses, any open access research based on this data is now potentially subject to review from a far wider community of peers. As the GISS-TEMP case shows, ‘amateurs’ from outside the academic field may take advantage of open data to explore the findings of academic study, to identify errors, and to extend, complement or disseminate academic findings. Research into physical science research practice and commercial problem solving finds that openness of both problem statements and data can support discovery of innovative solutions to entrenched problems, with many of the solutions coming from outside the ‘home’ field of the data or problem (Lakhani et. al., 2007; Lakhani, 2009). Not only may the same phenomena of a ‘computational turn’ that Meyer et. al (2009) identify from humanities digitization projects occur to some extent within social scientists research based on PSI, but actors from outside the social research fields may become involved in the analysis of social data, bringing different paradigms and approaches to it’s analysis. Open PSI repositories are not organized according to academic silos. That geodata or weather data of interest to climate researchers is available for discovery in Data.gov.uk next to schools and education data and local health statistics has the potential to challenge established research and data silos, much as Meyer and Schroeder (2009) suggest silos of disciplinary knowledge have been challenged by the growth of search based information retrieval across a web of documents. Open PSI does not make social sciences newly data-rich creating data-sharing via Sawyers suggested logic. However, as government is forcing certain datasets to be open it is setting new norms that, alongside requirements by funding bodies for publically funded data to be deposited in data repositories, point towards a more open future for social scientific data.  A more open future is not, however, a foregone conclusion. It has already been noted in the first case study that when research draws upon PSI alongside other data sources, it may end up dealing with personal data. Ohm (2009) has shown that even apparently anonymised data can, when combined with other data, lead to the exposure of private information. Against a pressure towards more open data, we may then expect both normative and legal frameworks around privacy to exert an opposite pressure to keep research datasets derived from PSI private in future, for fear of unintended privacy violations.

A second impact of open PSI on research concerns use of secondary data. Cole et. al (2008) note that even minor restrictions, such as delays between registering for data and gaining access can be “a significant barrier to use”, and thus we may expect the increasing availability of open raw PSI to increase researchers uptake of data for secondary analysis. Sawyer’s (2008) argument would suggest this may result in less primary data collection being carried out, with research drawing on pre-existing resources. Whilst Heaton (in Carmichael, 2008) suggests the secondary analysis is often useful not as the totality of a project, but as a complement to original qualitative and quantitative research. A number of advocates of data repositories cite their role in teaching: giving students direct experience of real data for secondary analysis (Corti & Bishop, 2005; Cole et. al. 2008), but if training students with skills for secondary analysis displaces the development of skills for primary data collection then a potential knock on impact onto the collection of primary data may be observed. Whilst a full normative analysis of the impact of increased provision of open raw PSI is beyond the scope of this paper, should research become more reliant on government collected information to the detriment of primary research (a possibility compounded by the status of government as both data provider and funding provider in many cases) then the increased nodality (Hood and Margetts, 2007) this would give government in the production of social knowledge does give cause for concern, particularly when one key functional role of social research must be to support public discourse around the success,  failure and future direction of government programmes.

Actual changes in the level of the use and citation of open raw PSI are something which will need to be tracked over coming years, an issue complicated in the bibliographic record by the lack of clear and widely agreed standards for data citation (Green, 2009), albeit with some efforts taking place to address this gap (Altman and King, 2007). However, it should be noted that Altman and King’s proposed data citation standard relies upon a computer fingerprint of a dataset, assuming it to be a static collection of quantitative information – rather than a dynamically developing set of RDF statements spread across multiple locations, or a real-time PSI data series. The development then of research technologies for working with open PSI may require some further attention.

For some, mention of RDF linked data encoding social facts holds out the possibility of new computational research technologies, enabling the forms of high-consensus, rapid-discovery science that Collin’s discussed (2004). Whilst exploration of RDF and semantic data has been widely explored in academic communities, from biology (e.g. OpenFlyData, Zhao et. al., 2009) to archaeology (e.g. CLAROS, Kurtz et. al., 2009), widespread uptake of linked data has not yet been seen in the social sciences, and so the forms of open raw PSI effort described in the Edubase case study point to new potential here. However, the potential of linked data is distinct from the potential of a semantic web – a potential that is widely dismissed in any case by social scientists who reject the idea of uniform ontologies, and argue that multiple paradigms (Kuhn, 1962) are used to describe the same data in different research contexts (Brent, 2008), frustrating semantic computation across datasets. Within a linked data approach, linkages are made not for semantic reasons, but for functional use – and the articulation of relationships between entities in PSI datasets is driven not by research ontologies, but by day-to-day concerns of government administration. Statistical computation across open PSI may allow more rapid and high-consensus discovery of more social facts, but, as Hollingsworth (2008) notes, theory is required before data can become knowledge, and open PSI promises little in the way of social theory-generating research technologies.

The considerations so far have focused on how the practice of research within the academic community may be transformed by open raw PSI. However, the most radical transformations for research may come about through the changing structure of communities of knowledge production brought about by widespread access to both data and tools of data analysis. The growth of open access data visualization tools such as IBM Many Eyes (Viégas et. al., 2007) which allow sophisticated exploration of public datasets increases non-academic engagement with data, as do mash-up projects such as the cycle accident maps of the first case study. Rather than being established through academia controlled peer-reviewed journal articles, widely held social ‘facts’ can be established through direct appeals to ‘transparent’ and auditable analysis of data. Of course, without critical appraisal of the data, what appears to be transparent analysis may be anything but – and this highlights a potential epistemic role for the academic researcher – contributing analysis and annotation of open PSI to both directories of data, and to communities where data is being used and manipulated. The cloud of online knowledge (Meyer and Schroeder, 2009) potentially becomes less the result of scholarly work, and more the result of agile and independent data analysis publishing works online. Researcher responses to such shifts are likely to be complex: whilst one possibility is increased researcher involvement in curating open datasets – for example, applying critical statistical analysis skills to convert a dataset from cycle accident incidents, into a measure of accident liklihood given usually traffic flows – if such activities are not rewarded by academic incentive structures then their occurrence is not likely to be widespread. And as Meyer and Schroeder have found (ibid.), change in research practices generally takes considerable time to diffuse across different disciplines.

Conclusions: norms, technologies, practices and relationships

The current push towards open PSI must be understood in the wider context of open data movements. An academic ‘movement’ (Frickel and Gross, 2005) for open data, following on from the OA movement, has benefited from open PSI. Government open data initiatives have arguably helped shift norms towards opening access to data, and the normative sustainability of closed-data open-science is being undermined. However, research technologies and academic practices are likely to lag far behind the cutting edge of open, raw, PSI use. Skills and practices take time to diffuse within and across disciplines, acting as a constraint on innovative developments – particular in the absence of funding arrangements or incentive structures to change behavior.  The real revolutions are arguably not in research technology or practice, but in the relationship between social science and wider society and the role researchers will need to play to remain relevant as open tools and open data allow a far greater range of actors to offer evidence-based answers to key social questions.

[1] Based on analysis of September 25^th 2009 database dump of the CKAN platform hosting dataset metadata: http://ckan.net/dump/hmg.ckan.net-20091125.csv.gz (Accessed 12^th April 2010)

[2] http://www.earlham.edu/~peters/fos/bethesda.htm#participants – accessed 2^nd March 2010

[3] http://www.soros.org/openaccess/read.shtml – accessed 2nd March 2010

[4] http://opendefinition.org – accessed 2^nd March 2010.

[5] http://pantonprinciples.org/ accessed 2^nd March 2010.

[6] See http://data.gov.uk/dataset/gb-road-traffic-counts (Accessed 12th April 2010)

[7] http://data.giss.nasa.gov/gistemp/ – Accessed 12^th April 2010

[8] http://clearclimatecode.org/ - Accessed 3^rd March 2010

[9] http://clearclimatecode.org/finding-bugs-in-gistemp/ Accessed 3^rd March 2010

[10] Authors notes from presentation by Jeni Tennison, Linked Data advisor to Data.gov.uk at OUCS Linked Data and Practical Semantic Web Workshop, March 2010. http://www.oucs.ox.ac.uk/rts/events/linked-data.xml Accessed 29th March 2010.

Bibliography

2009. How safe is your route? – Cycle Chat. Available at: http://www.cycle-cafe.net/forums/showthread.php?t=29659&page=4 [Accessed April 12, 2010].

2010. The Data Deluge. The Economist.

Aichholzer, G. & Burkert, H., 2004. Public sector information in the digital age : between markets, public management and citizens’ rights, Cheltenham: Edward Elgar.

Alani, H. et al., 2007. Unlocking the potential of public sector information with semantic web technology. Lecture Notes in Computer Science, 4825, 708.

Altman, M. & King, G., 2007. A proposed standard for the scholarly citation of quantitative data. D-lib Magazine, 13(3/4), 1082–9873.

Arzberger, P. et al., 2004. Promoting access to public research data for scientific, economic, and social development. Data Science Journal, 3(0), 135–152.

Berners-Lee, T., 1998. Semantic web road map.

Berners-Lee, T., 2009. Tim Berners-Lee on the next Web, Available at: http://www.ted.com/talks/tim_berners_lee_on_the_next_web.html [Accessed November 27, 2009].

Berners-Lee, T., 2009. Putting Government Data online – Design Issues. Available at: http://www.w3.org/DesignIssues/GovData.html [Accessed March 4, 2010].

Bowker, G.C., 2000. Biodiversity datadiversity. Social Studies of Science, 643–683.

Brent, E., 2008. Artificial Intelligence and the Internet. In N. Fielding, R. M. Lee, & G. Blank, eds. The SAGE handbook of online research methods. SagePublications Ltd.

Busek, E., 2008. Knowledge and Democracy: Is Freedom a Daughter of Knowledge? In N. Stehr, ed. Knowledge and democracy : a 21st-century perspective. Somerset, N.J.: Transaction ; London.

Carmichael, P., 2008. Secondary Qualitative Analysis Using Internet Resources. In N. Fielding, R. M. Lee, & G. Blank, eds. The SAGE handbook of online research methods. Sage Publications Ltd.

Clarke, P., 2009. Pedalling some raw data… |. Available at: http://innovate.direct.gov.uk/2009/03/10/pedalling-some-raw-data/ [Accessed April 12, 2010].

Cole, K., Wathan, J. & Corti, L., 2008. The Provision of Access to Quantitative Data for Secondary Analysis. In N. Fielding, R. M. Lee, & G. Blank, eds. The SAGE handbook of online research methods. Sage Publications Ltd.

Collins, R., 1994. Why the social sciences won’t become high-consensus, rapid-discovery science. In Sociological Forum. pp. 155–177.

Corti, L. & Bishop, L., 2005. Strategies in teaching secondary analysis of qualitative data. In Forum: Qualitative Social Research.

Crabtree, J. & Chatfield, T., 2010. Mash the State. Prospect, (167).

Cummings, J.N. & Kiesler, S., 2005. Collaborative research across disciplinary and organizational boundaries. Social Studies of Science, 35(5), 703.

De Saulles, M., 2005. e-Government and the Re-use of Public Sector Information. In 5th European conference on e-government: University of Antwerp, Belgium 16-17 June 2005. p. 121.

Dekkers, M. et al., 2006. MEPSIR Study – 2006. Available at: http://ec.europa.eu/information_society/policy/psi/actions_eu/policy_actions/mepsir/index_en.htm [Accessed December 21, 2009].

Department for Transport, H., 2008. Reported Road Casualties Great Britain: 2008 – Annual Report. Available at: http://www.dft.gov.uk/pgr/statistics/datatablespublications/accidents/casualtiesgbar/rrcgb2008 [Accessed April 12, 2010].

Dutton, W.H. & Meyer, E.T., 2009. Experience with New Tools and Infrastructures of Research: An Exploratory Study of Distance From, and Attitudes Toward, e-Research. Prometheus: Critical Studies in Innovation, 27(3), 223.

Eccles, K., Meyer, E.T. & Madsen, C., 2009. Digitisation as e-Research infrastructure: Access to materials and research capabilities in the Humanities. In 5th International Conference on e-Social Science. Cologne.

Eisen, M. & Salzberg, S., 2009. Open Access: The Sooner the Better. Science, 325(5938), 266-266.

Elias, P., 2007. The National Strategy for Data Resources for
Research in the Social Sciences. Available at: http://eprints.ncrm.ac.uk/449/ [Accessed April 24, 2010].

European Commission, 2003. Directive 2003/9 8/EC of Parliament and Council on the re-use of public sector information, Available at: http://ec.europa.eu/information_society/policy/psi/docs/pdfs/directive/psi_directive_en.pdf.

Evans, J.A. & Reimer, J., 2009. Open Access and Global Participation in Science. Science, 323(5917), 1025-1025.

Fensel, D., 2005. Spinning the semantic Web: bringing the World Wide Web to its full potential, The MIT Press.

Frickel, S. & Gross, N., 2005. A general theory of scientific/intellectual movements. American Sociological Review, 70(2), 204.

Gill, M., Goldacre, M.J. & Yeates, D.G., 2006. Changes in safety on England’s roads: analysis of hospital statistics. British Medical Journal, 333(7558), 73.

Gray, D.E., 2009. Doing research in the real world, Sage Publications Ltd.

Green, T., 2009. We Need Publishing Standards for Datasets and Data Tables. , 22(4), 325-327.

Hollingsworth, J.R., 2008. Introduction to Part 2. In N. Stehr, ed. Knowledge and democracy : a 21st-century perspective. Somerset, N.J.: Transaction ; London.

Hood, C.C. & Margetts, H.Z., 2007. The Tools of Government in the Digital Age 2nd ed., Palgrave Macmillan.

Jones, P. & Elias, P., 2006. Administrative data as a research resource: a selected audit. Economic & Social Research Council Regional Review Board Report 43, 6.

Kuhn, T.S., 1962. The structure of scientific revolutions, Chicago ; London: University of Chicago Press.

Kurtz, D. et al., 2009. CLAROS-Bringing Classical Art to a Global Public. In 2009 Fifth IEEE International Conference on e-Science. pp. 20–27.

Lakhani, K.B., 2009. Innocentive.com (A),

Lakhani, K.R. et al., 2007. The value of openness in scientific problem solving. Boston: Harvard University.

Latour, B. & Woolgar, S., 1986. Laboratory life: The construction of scientific facts, Princeton Univ Pr.

Lyons, R.A. et al., 2008. Using multiple datasets to understand trends in serious road traffic casualties. Accident Analysis & Prevention, 40(4), 1406–1410.

Mayo, E. & Steinberg, T., 2007. Power of Information Taskforce Report, Available at: http://www.cabinetoffice.gov.uk/reports/power_of_information.aspx [Accessed December 1, 2009].

Meyer, E.T. & Schroeder, R., 2009. Untangling the web of e-Research: Towards a sociology of online knowledge. Journal of Informetrics, 3(3), 246-260.

Murray-Rust, P. et al., Panton Principles. Available at: http://pantonprinciples.org/ [Accessed March 4, 2010].

Newbery, D., Bently, L. & Pollock, R., 2008. Models of public sector information provision via trading funds.

Obama, B., 2010. Memo from President Obama on Transparency and Open Government. In D. Lathrop & L. Ruma, eds. Open Government. Available at: http://www.whitehouse.gov/open/documents/open-government-directive [Accessed April 18, 2010].

Ohm, P., 2009. Broken Promises of Privacy: Responding to the Surprising Failure of Anonymization. SSRN eLibrary. Available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1450006 [Accessed March 4, 2010].

OKF – Open Knowledge Foundation, 2006. Open Knowledge Definition. Available at: http://www.opendefinition.org/ [Accessed March 4, 2010].

Olson, G.M. & Olson, J.S., 2000. Distance matters. Human-computer interaction, 15(2), 139–178.

Ordnance Survey, H.O., 2010. News Release: Ordnance Survey launches OS OpenData in groundbreaking national initiative – 01 April 2010. Available at: http://www.ordnancesurvey.co.uk/oswebsite/media/news/2010/April/OpenData.html [Accessed April 12, 2010].

PIRA International Ltd, 2000. Commercial exploitation of Europe’s Public Sector Information,

Pollock, R., 2009. The Economics of Public Sector Information. Available at: http://econpapers.repec.org/paper/camcamdae/0920.htm [Accessed March 4, 2010].

Robinson, D.G., Yu, H. & Felten, E.W., 2010. Enabling Innovation for Civic Engagement. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Sawyer, S., 2008. Data Wealth, Data Poverty, Science and Cyberinfrastructure. Prometheus, 26(4), 355–371.

Scott, J.C., 1998. Seeing like a state, Yale University Press New Haven, CT.

Shadbolt, N., Hall, W. & Berners-Lee, T., 2006. The semantic web revisited. IEEE intelligent systems, 21(3), 96–101.

Sharif, R.M., 2009. Maximizing the Value of Public Sector Information for Scientific and Socioeconomic Development in Africa | www.KMAfrica.com – Knowledge Management Africa KnowledgeHub. In KMAfrica. Available at: http://www.isivivane.com/kmafrica/?q=group.governance.maximizing.the.value.of.public.sector.information.for.scientific.and.socioeconomic.development.in.Africa [Accessed April 13, 2010].

Solove, D., 2006. A taxonomy of privacy. University of Pennsylvania Law Review, 154(3), 477–560.

Stehr, N., 2008. Knowledge and democracy : a 21st-century perspective, Somerset, N.J.: Transaction ; London.

Uhlir, P., 2009. The Socioeconomic Effects of Public Sector Information on Digital Networks: Toward a Better Understanding of Different Access and Reuse Policies: Workshop Summary. Available at: http://www.nap.edu/catalog.php?record_id=12687 [Accessed September 30, 2009].

UK Data Forum, 2009. UK Strategy for Data Resources for Social and Economic Research 2009-2012, Available at: http://www.esrc.ac.uk/ESRCInfoCentre/NDS/.

Vickery, G. & Wunsch-Vincent, S., 2006. Digital broadband content: public sector information and content, Organisation for Economic Co-operation and Development.

Viégas, F.B. et al., 2007. Manyeyes: a site for visualization at internet scale. IEEE Transactions on Visualization and Computer Graphics, 13(6), 1121–1128.

Watson, A., 2010. Statistics on the accuracy of DVLA records – WhatDoTheyKnow.com. Available at: http://www.whatdotheyknow.com/request/statistics_on_the_accuracy_of_dv [Accessed April 1, 2010].

Williams, S., 2002. Free as in freedom, O’Reilly Media, Inc.

Willinsky, J., 2006. The access principle: the case for open access to research and scholarship, MIT Press Cambridge, MA.

Wuchty, S., Jones, B.F. & Uzzi, B., 2007. The increasing dominance of teams in production of knowledge. Science, 316(5827), 1036.

Zhao, J. et al., 2009. OpenFlyData: The Way to Go for Biological Data Integration. In Data Integration in the Life Sciences. pp. 54, 47. Available at: http://dx.doi.org/10.1007/978-3-642-02879-3_5 [Accessed April 12, 2010].

Aichholzer, G. & Burkert, H., 2004. Public sector information in the digital age: between markets, public management and citizens’ rights, Edward Elgar Publishing.

Alani, H. et al., 2007. Unlocking the potential of public sector information with semantic web technology. Lecture Notes in Computer Science, 4825, 708.

Allan, R., 2009. The Power of Government Information. In J. Gøtze & C. B. Pedersen, eds. State of the eUnion: Government 2.0 and Onwards. AuthorHouse.

Bartholomew, D.J. et al., 2008. Analysis of Multivariate Social Science Data, Second Edition (Chapman & Hall/Crc Statistics in the Social and Behavioral Scie) 2nd ed., Chapman and Hall/CRC.

Becker, R.A., Chambers, J.M. & Wilks, A.R., 1988. The new S language: a programming environment for data analysis and graphics.

Beetham, D., 1994. Defining and measuring democracy, Sage Publications Ltd.

Beniger, J.R., 1986. The control revolution: Technological and economic origins of the information society, Harvard Univ Pr.

Benkler, Y., 2007. The wealth of networks: How social production transforms markets and freedom, Yale Univ Pr.

Bennett, C., 1985. From the Dark to the Light: The Open Government Debate in Britain. Journal of Public Policy, 5(02), 187-213.

Berners-Lee, T., 2009. Putting Government Data online – Design Issues. Available at: http://www.w3.org/DesignIssues/GovData.html [Accessed May 19, 2010].

Bilgin, A. et al., 2009. Graphviz-Graph Visualization Software, Graphviz.

Bimber, B.A., 2003. Information and American democracy: Technology in the evolution of political power, Cambridge Univ Pr.

Bizer, C., Heath, T. & Berners-Lee, T., 2009. Linked data–the story so far. International Journal on Semantic Web and Information Systems, 5(3), 1–22.

Boland, L. & Coleman, E., 2008. New Development: What Lies Beyond Service Delivery? Leadership Behaviours for Place Shaping in Local Government. Public Money & Management, 28(5), 313–318.

Boland, L. & Colman, E., Can Leviathan stop gobbling people up and start really listening to them?

Borg, S., 2001. The research journal: a tool for promoting and understanding researcher development. Language Teaching Research, 156 -177.

Bowker, G.C., 2000. Biodiversity datadiversity. Social Studies of Science, 643–683.

Brewer, J. & Hunter, A., 2006. Foundations of multimethod research, SAGE.

Brown, G., 2009. Speech on Building Britain’s Digital Future. Available at: http://www.number10.gov.uk/Page22897 [Accessed April 13, 2010].

Brudney, J.L. & England, R.E., 1983. Toward a Definition of the Coproduction Concept. Public Administration Review, 43(1), 59-65.

Bryman, A., 2008. Social Research Methods 3rd ed., OUP Oxford.

Burkert, H., 2004. The Mechanics of Public Sector Information. In G. Aichholzer & H. Burkert, eds. Public sector information in the digital age: between markets, public management and citizens’ rights. Edward Elgar Publishing.

Butler, D., 2007. Data sharing: the next generation. Nature, 446(7131), 10-11.

Cabinet Office, 2009. Calling Open Data Developers: We need your help. Cabinet Office: Digital Engagement Blog. Available at: http://blogs.cabinetoffice.gov.uk/digitalengagement/post/2009/09/30/Calling-Open-Data-Developers-We-need-your-help.aspx [Accessed November 27, 2009].

Cameron, D., 2010. Letter to Government departments on opening up data | Number10.gov.uk. Available at: http://www.number10.gov.uk/news/statements-and-articles/2010/05/letter-to-government-departments-on-opening-up-data-51204 [Accessed July 9, 2010].

Card, S., 2009. Information Visualization. In J. A. Jacko, ed. Human-Computer Interaction. CRC Press.

Castells, M., 2009. Communication power, Oxford ;;New York: Oxford University Press.

Castells, M., 2001. The Internet Galaxy: Reflections on the Internet, Business, and Society 1st ed., Oxford University Press, USA.

Chi, E.H., 2000. A taxonomy of visualization techniques using the data state reference model. In Proceedings of the IEEE Symposium on Information Visualization. pp. 69–75.

Cho, H. & LaRose, R., 1999. Privacy issues in Internet surveys. Social Science Computer Review, 17(4), 421.

Coleman, E. & Boland, L., 2010. Can Leviathan stop gobbling people up and start really listening to them? A call for a new form of public sector governance.

Costello, A.B. & Osborne, J.W., 2005. Best practices in exploratory factor analysis: four recommendations for getting the most from your analysis. Practical Assessment, Research & Evaluation, 10(7), 1–9.

Couper, M.P., 2000. Web surveys – a review of issues and approaches. Public opinion quarterly, 64(4), 464–494.

Crenson, M.A. & Ginsberg, B., 2003. From popular to personal democracy. National Civic Review, 92(2), 173–189.

Dahl, R.A., 1989. Democracy and its Critics, Yale Univ Pr.

Dahl, R.A., 2000. On democracy, Yale Univ Pr.

Daniel, W.W., 1993. Strategic research: Political convenience or practical reality? Policy Studies, 14(3), 4–17.

Davies, A. & Lithwick, D., 2010. Government 2.0 and Access to Information: 2. Recent Developments in Proactive Disclosure and Open Data in the United States and Other Countries (2010-15E), Library of Parliament, Canada. Available at: http://www2.parl.gc.ca/Content/LOP/ResearchPublications/2010-15-e.htm#a4 [Accessed May 24, 2010].

De Saulles, M., 2005. e-Government and the Re-use of Public Sector Information. In 5th European conference on e-government: University of Antwerp, Belgium 16-17 June 2005. p. 121.

Dekkers, M. et al., 2006. MEPSIR – Measuring European Public Sector Information Resources, Available at: http://ec.europa.eu/information_society/policy/psi/actions_eu/policy_actions/mepsir/index_en.htm [Accessed December 21, 2009].

Dervin, B., 1994. Information <-> democracy: An examination of underlying assumptions. Journal of the American Society for Information Science, 45(6), 369-385.

Dewey, J., 1938. Logic-the theory of inquiry, Henry Holt and Company.

Dierking, H., 2010. Engineering Good Government. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Diesing, P., 1971. Patterns of Discovery in the Social Sciences Reprint., AldineTransaction.

Ding, L. et al., 2010. Making Sense of Open Government Data.

Downs, A., 1957a. An Economic Theory of Democracy, Addison Wesley.

Downs, A., 1957b. An Economic Theory of Political Action in a Democracy. The Journal of Political Economy, 65(2), 135-150.

Dryzek, J. & Dunleavy, P., 2009. Theories of the Democratic State, Palgrave Macmillan.

Dunleavy, P. et al., 2006. Digital era governance: IT corporations, the state, and E-government, Oxford University Press, USA.

Dunlop, C. & Kling, R., 1991. Computerization and controversy: value conflicts and social choices, Academic Press Professional, Inc. San Diego, CA, USA.

Eaves, D., 2009. Three Laws of Open Data (International Edition. Eaves.ca. Available at: http://eaves.ca/2009/11/29/three-laws-of-open-data-international-edition/ [Accessed June 1, 2010].

European Commission, 2003a. Directive 2003/9 8/EC of Parliament and Council on the re-use of public sector information, Available at: http://ec.europa.eu/information_society/policy/psi/docs/pdfs/directive/psi_directive_en.pdf.

European Commission, 2003b. Public Sector Information Directive, Available at: http://ec.europa.eu/information_society/policy/psi/actions_eu/policy_actions/index_en.htm [Accessed December 21, 2009].

Eynon, R., Fry, J. & Schroeder, R., 2008. The ethics of internet research. The SAGE handbook of online research methods, 23.

Fine, A. et al., 2008. Rebooting America; Ideas for Redesigning American Democracy for the Internet Age,

Fink, A., 2006. How to conduct surveys : a step-by-step guide 3rd ed., Thousand Oaks ; London: Sage.

Floridi, L., 2004. Information. In L. Floridi, ed. The Blackwell Guide to the Philosophy of Computing and Information. Wiley-Blackwell.

Fung, A. & Weil, D., 2010. Open Government and Open Society. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Gray, D.E., 2009. Doing research in the real world, Sage Publications Ltd.

Hague, B.N. & Loader, B., 1999. Digital democracy, Routledge.

Hatcher, J. & Waelde, C., 2007. Open Data Commons » Licenses. Available at: http://www.opendatacommons.org/licenses/ [Accessed June 1, 2010].

Hayek, F.A., 1944. The road to serfdom,

Heald, D., 2006. Transparency as an instrumental value. In Proceedings of the British Academy. p. 59.

Held, D., 2006. Models of democracy, Polity Pr.

Hindman, M.S., 2009. The myth of digital democracy, Princeton Univ Pr.

HMG, 2009. Putting the Frontline First: Smarter Government. Available at: http://www.hmg.gov.uk/frontlinefirst/action1/transparency.aspx [Accessed March 7, 2010].

Hogge, B., 2010. Open Data Study, Open Society Institute. Available at: http://www.soros.org/initiatives/information/focus/communication/articles_publications/publications/open-data-study-20100519.

Hood, C.C. & Margetts, H.Z., 2007. The Tools of Government in the Digital Age 2nd ed., Palgrave Macmillan.

Hookway, C., 2008. Pragmatism. Available at: http://plato.stanford.edu/entries/pragmatism/ [Accessed June 23, 2010].

Huang, J., Thornton, K.M. & Efthimiadis, E.N., 2010. Conversational Tagging in Twitter.

Janesick, V.J., 1999. A Journal About Journal Writing as a Qualitative Research Technique: History, Issues, and Reflections. Qualitative Inquiry, 505 -524.

Kamarck, E.C., Nye, J.S. & (Program), V.O.G.I.T.2.C., 2002. Governance.com, Brookings Institution Press.

Kish, L., 1965. Survey sampling., New York ; Toronto: J. Wiley.

Kling, R. & Iacono, S., 1995. Computerization movements and the mobilization of support for computerization. In S. L. Star, ed. Ecologies of knowledge: Work and politics in science and technology. pp. 119–153.

Kling, R. & Iacono, S., 1988a. Mobilization of Support for Computerization: The Role of Computerization Movements, The. Soc. Probs., 35, 226.

Kling, R. & Iacono, S., 1988b. The Mobilization of Support for Computerization: The Role of Computerization Movements. Social Problems, 35, 226.

Kling, R., McKim, G. & King, A., 2003. A bit more to it: Scholarly communication forums as socio-technical interaction networks. Journal of the American Society for Information Science and Technology, 54(1), 47–67.

Lakhani, K.B., 2009. Innocentive.com (A),

Lathrop, D. & Ruma, L., 2010. Open Government: Collaboration, Transparency, and Participation in Practice 1st ed., O’Reilly Media.

Lawley, D.N. & Maxwell, A.E., 1962. Factor Analysis as a Statistical Method. Journal of the Royal Statistical Society. Series D (The Statistician), 12(3), 209-229.

Le Grand, J., 2007. The other invisible hand: delivering public services through choice and competition, Princeton Univ Pr.

Lievrouw, L.A., 1994. Information Resources and Democracy: Understanding the Paradox. Journal of the American Society for Information Science, 45(6), 350–57.

Malmud, C. et al., 2007. 8 Principles of Open Government Data. Available at: http://resource.org/8_principles.html [Accessed June 1, 2010].

Mayo, E. & Steinberg, T., 2007. Power of Information Taskforce Report, Available at: http://www.cabinetoffice.gov.uk/reports/power_of_information.aspx [Accessed December 1, 2009].

Meyer, E., 2010. Socio-technical Interaction Networks: A discussion of the strengths, weaknesses and future of Kling’s STIN model. Social Informatics: An Information Society for all? In Remembrance of Rob Kling, 37–48.

Miles, M.B. & Huberman, A.M., 1994. Qualitative data analysis: An expanded sourcebook, SAGE publications, Inc.

Murray-Rust, P. et al., 2009. Panton Principles. Available at: http://pantonprinciples.org/ [Accessed March 4, 2010].

Needham, C., 2008. Realising the Potential of Co-Production: Negotiating Improvements in Public Services. Social Policy and Society, 7(02), 221-231.

Newbery, D., Bently, L. & Pollock, R., 2008. Models of public sector information provision via trading funds.

Nissenbaum, H., 1998. Protecting privacy in an information age: The problem of privacy in public. Law and Philosophy, 17(5), 559–596.

Noveck, B.S., 2006. Peer to Patent: Collective Intelligence, Open Review, and Patent Reform. Harv. JL & Tech., 20, 123.

Noveck, B.S., 2010. The Single Point of Failure. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Noveck, B.S., 2009. Wiki government: how technology can make government better, democracy stronger, and citizens more powerful, Brookings Inst Pr.

Obama, B., 2010. Memo from President Obama on Transparency and Open Government. In D. Lathrop & L. Ruma, eds. Open Government. Available at: http://www.whitehouse.gov/open/documents/open-government-directive [Accessed April 18, 2010].

OED Online, 2000. datum. In Oxford English Dictionary. Oxford University Press. Available at: http://dictionary.oed.com/cgi/entry/50057802/ [Accessed May 22, 2010].

OKF – Open Knowledge Foundation, 2006. Open Knowledge Definition. Available at: http://www.opendefinition.org/ [Accessed March 4, 2010].

Parycek, P. & Sachs, M., 2010. Open Government – Information Flow in Web 2.0. European Journal of ePractice, (9). Available at: http://www.epractice.eu/en/document/313345 [Accessed May 23, 2010].

Pellegrini, T., 2009. Semantic Web Awareness 2009, The Semantic Web Company. Available at: http://www.semantic-web.at/file_upload/1_tmpphpvuVU1T.pdf [Accessed May 23, 2010].

PIRA, 2000. Commercial exploitation of Europe’s Public Sector Information,

Pizzicannella, R., 2010. Co-production and open data: the right mix for public service effectiveness? In Draft papers. 10th European Conference on eGovernment. Limerick, Ireland. Available at: http://pizzican.wordpress.com/2010/05/23/co-production-and-open-data-in-public-services/#comment-3 [Accessed May 24, 2010].

Pollock, R., 2009. The Economics of Public Sector Information. Available at: http://econpapers.repec.org/paper/camcamdae/0920.htm [Accessed March 4, 2010].

Prat, A., 2006. The more closely we are watched, the better we behave? In PROCEEDINGS-BRITISH ACADEMY. p. 91.

Putnam, H., 1995. Pragmatism: an open question, Wiley-Blackwell.

Putnam, H. & Conant, J., 1995. Words and life, Harvard Univ Pr.

Rivadeneira, A.W. et al., 2007. Getting our head in the clouds: toward evaluation studies of tagclouds. In Proceedings of the SIGCHI conference on Human factors in computing systems. San Jose, California, USA: ACM, pp. 995-998. Available at: http://portal.acm.org/citation.cfm?id=1240624.1240775 [Accessed July 17, 2010].

Robinson, D. et al., 2008. Government data and the invisible hand. Yale Journal of Law and Technology, 11, 160.

Rosling, H., 2006. Hans Rosling shows the best stats you’ve ever seen, Available at: http://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen.html [Accessed May 23, 2010].

Rosling, H., 2010. The Changing World — A New Mindset for a Multipolar World, The World Bank, New York. Available at: http://www.youtube.com/watch?v=cfCdZv1KiSU [Accessed May 23, 2010].

Rosling, H.A., 2007. Visual technology unveils the beauty of statistics and swaps policy from dissemination to access. Statistical Journal of the IAOS: Journal of the International Association for Official Statistics, 24(1), 103-104.

Rowley, J.E., 2007. The wisdom hierarchy: representations of the DIKW hierarchy. Journal of Information Science, 0165551506070706v1.

Scott, J.C., 1998. Seeing like a state, Yale University Press New Haven, CT.

Sifry, M.L., 2010. “You Can Be the Eyes and Years”: Barak Obama and the Wisdom of Crowds. In D. Lathrop & L. Ruma, eds. Open Government: Collaboration, Transparency, and Participation in Practice. O’Reilly Media.

Skolnick, B.P., 1995. The uses of social science research: a report of a workshop organised by the ESRC Programme on Information and Communication Technologies,

Smith, A., 2010. Government Online, Pew Research Centre.

Snowdon, D., 2010. Its information to data we need, not DIKW. Cognitive Edge. Available at: http://www.cognitive-edge.com/blogs/dave/2010/05/its_information_to_data_we_nee.php [Accessed May 19, 2010].

Stasavage, D., 2006. Does Transparency Make a Difference? The Example of the European Council of Ministers. In PROCEEDINGS-BRITISH ACADEMY. p. 165.

Stephenson, D.W., 2009. Data Triggers Transformation. In J. Gøtze & C. B. Pedersen, eds. State of the eUnion: Government 2.0 and Onwards. AuthorHouse.

Stiglitz, J.E., 2000. The contributions of the economics of information to twentieth century economics. Technology, 1441.

Sunstein, C.R., 2007. Republic. com 2.0, Princeton Univ Pr.

Suzuki, R. & Shimodaira, H., 2006. Pvclust: an R package for assessing the uncertainty in hierarchical clustering. Bioinformatics, 22(12), 1540.

Tashakkori, A. & Teddlie, C., 1998. Mixed methodology: Combining qualitative and quantitative approaches, Sage Publications, Inc.

Tennison, J. & Sheridan, J., 2010. Linking UK Government Data. In Linked Data on the Web (LDOW2010). WWW2010, Raleigh, North Carolina. Available at: http://events.linkeddata.org/ldow2010/ [Accessed May 19, 2010].

Torkington, N., 2010. Rethinking Open Data – O’Reilly Radar. O’Reilly Radar. Available at: http://radar.oreilly.com/2010/02/rethinking-open-data.html [Accessed July 18, 2010].

Uhlir, P., 2009. The Socioeconomic Effects of Public Sector Information on Digital Networks: Toward a Better Understanding of Different Access and Reuse Policies: Workshop Summary, Available at: http://www.nap.edu/catalog.php?record_id=12687 [Accessed September 30, 2009].

Varnhagen, C.K. et al., 2005. How Informed Is Online Informed Consent? Ethics & Behavior, 15(1), 37.

Viégas, F.B. et al., 2007. Manyeyes: a site for visualization at internet scale. IEEE Transactions on Visualization and Computer Graphics, 13(6), 1121–1128.

Von Hippel, E., 2005. Democratizing innovation, the MIT Press.

Weinstein, M., 2008. TAMS Analyser for Macintosh OS X, Available at: http://tamsys.sourceforge.net/.

Weiss, P., 2004. Borders in Cyberspace: Conflicting Public Sector Information Policies and Their Economic Impacts. In G. Aichholzer & H. Burkert, eds. Public sector information in the digital age. Edward Elgar Publishing.

Wiber, C.K., 1999. Holistic Method. In P. A. O’Hara, ed. Encyclopedia of political economy. Routledge, p. 451.

Worthy, B., 2008. The Future of Freedom of Information in the United Kingdom. Political Quarterly, 79(1), 100-108.

Yin, R.K., 2008. Case study research: Design and methods, Sage Publications, Inc.

This exploratory study has undertaken wide-reaching preliminary work exploring OGD use. The typologies it offers require further work to generate operational definitions for each, and to further test their adequacy on a larger sample of OGD use-cases. Whilst this study inevitably touched on OGD supply, further work looking directly at relationships between supply and use of OGD would be valuable. Two strategies are relevant for this wider work. Firstly, a Socio-Technical Interaction Networks approach (Meyer 2010) would build on identification of a possible OGD Computerization Movement (§4.1), and would support more comprehensive analysis of ideologies driving OGD use, and of how actors and technologies are interacting, whilst also drawing attention to the actors missing from, or peripheral to, networks around OGD. Secondly, work to explore downstream uses of OGD (e.g. use of interfaces built with OGD) would provide illumination beyond the direct OGD use explored in this study, allowing deeper evaluation of OGD value-chains. Understanding the full civic value-chain of OGD use is important to inform future investment decisions concerning OGD supply.

6.2. Contributions and Conclusion

This paper makes three contributions. Firstly, it outlines a model for thinking about different processes of OGD use, based on careful analysis of real-world use cases. Secondly, it sketches connections that can be drawn between OGD and models of democratic engagement and public sector reform. This lays the foundations for future studies, including those oriented towards normative assessment of the relationship between OGD and civic change. Thirdly, in focusing on OGD from the end-use perspective, rather than more common perspectives of data and tool providers, and in addressing the democratic and public service arguments for OGD distinct from economic benefit arguments, it contributes to a rebalancing of the OGD debate towards civic over technological or economic concerns. By way of conclusion concerning the implications of OGD use, it notes that whilst unambiguous (though not always realized) pathways for OGD to impact on reforming public services are apparent, the relationship between data and democracy, (democracy understood as something other than ‘personal democracy’) is harder to define, more equivocal, and worthy of much further theoretical and practical attention.

• Promoting transparency and accountability
• Empowering citizens to drive public sector reform
• Releasing the economic and social value of information
• Putting Britain at the leading edge of semantic web developments.

Assessing the magnitude of impact that OGD can have in these respects is beyond the scope of this study. However, a number of modest remarks can be made from the evidence-base presented in §4 regarding democratic and public sector reform implications of OGD use.

5.1. Changing gatekeeping

OGD users have been exploring how scrutiny can be crowd-sourced, and have been using data to check state information and policies. The availability of OGD removes one aspect of both the state (and mainstream media’s) gatekeeping power (Castells 2009), although other socio-economic structures may constrain the extent to which new actors can widely advance their alternative interpretations of data. With both high-profile (COINS) and niche (School Revenue Balance) datasets this study has seen individuals able to create their own interpretations of government data, and to contribute those interpretations to spaces of online debate.

However, where OGD has allowed a ‘revolution in the middle’ (Bimber 2003), with new actors using data to enter public debate, those actors are, at present, technically skilled, as well as issue-interested, and thus far the engagement to be observed is data-aware individuals engaging with issues, rather than conventional issue-based civil society institutions engaging with OGD. Notions of communicative governance (§2.4.2) suggest OGD should allow citizens to debate the meaning of data with the state. Within the core network of actors around data.gov.uk debate was occurring, but the debate is frequently dominated by technical and data-standard type concerns. Widespread capacity within the state to engage with debate around meanings of data, and to use data as the starting point for democratic dialogue is yet to develop.

5.2. Empowering individuals

Exploring OGD can give individuals greater understanding of the state, and direct access to facts and information can empower individuals in their interactions with the state – addressing the information inequalities that concerned Dahl (2000). There is a risk that a focus on data-for-developers, and expectation that all data will be accessed through online interfaces can lead to a neglect of mechanisms for direct OGD access, although infrastructures and principles of data publishing could be developed to ensure all citizens have good opportunities to go direct to a trusted data source to get the detailed and granular information they need.

In-so-far-as new mediators of government information, such as UKSchoolMap and Schooloscope give information on topics over which citizen-consumers have genuine choice, then these can also empower individuals through market-mechanisms. It is notable though that both market and non-market models of empowerment described in this section are individual centred: and how OGD feeds into models of collective empowerment is unclear from the instances considered in this study. Attentiveness to Bowker (2000) and Scott’s (1998) concern that schemas of data collection can impact its future use, is warranted in this context, particularly if geographies of data collection (wards, super-output-areas etc.) do not match geographies of civic mobilization.

5.3. Innovation and reform

The OGD related drivers for public sector reform evident so far appear to come not from citizen engagement with the planning process, but from OGD-enabled innovation and the use by third-parties (often commercial or social enterprises) of OGD to provide planning support and new services to the state. OGD can create a space for innovation outside the bureaucracy, and at present, many OGD users are motivated to pursue innovative development on a speculative basis – either for personal learning, or potential reward in terms of recognition, profits or patronage. How far this innovation can move from dealing directly in predominantly digital and informational services to impact upon wider areas of public sector reform is yet to be seen, and certain issues may arise about the accountability gains or losses that result if OGD drives a transfer of public functions to private enterprise. However, particularly given the economic focus of EU PSI agendas, it is within the entrepreneurial use of state-centric OGD that some of the most significant implications for public sector reform are likely to be.

5.4. Digitized government: equitable architectures?

In section 4.1 it was suggested that the motivational driver shared by many OGD users of ‘digitizing government’, in part a response to a critical assessment of past government ICT initiatives (Dunleavy et al. 2006), has many of the hallmarks of a ‘Computerization Movement’, focused on problem-solving via application of technology to a situation, and noting no limitations of appropriate digitization. Via the same routes noted for general innovation above, OGD can support the development of new modular digital infrastructures around government data, and many of the processes of OGD use noted in §4.2 do adopt a very modular approach, providing components of an OGD infrastructure for others to build on. However, Hood and Margetts (2007), drawing on Dunlop and Kling (1991), note that visions of technology-driven change often suppose situations in which “conflict, politics and adversarial legalism are unknown or at least unimportant” (p. 179). There is a risk that many claims about the potential implications of digitized government via OGD ignore, and thus do not address and seek to tackle, key constraints to the equitable distribution of benefits (and power) that can be created/shared through OGD use. Unequal distributions of skills, social capital, and time have a big impact on who is making use of OGD, and the reach they can have with OGD use. Similarly, some problems remain difficult to solve, even with (and sometimes because of) government datasets, and the appropriate response to such problems may be social rather than technological.

Although most embedded cases cannot identify their direct social impact, they do posses implicit or explicit connections with different models of public sector reform/democratic change. Across the full 55 instances of OGD use, a total of 9 distinct types of civic OGD-use can be identified, set out in Table 5 and framed by the typology outlined in §2.4.

Many OGD uses fit into multiple categories. For example, Schooloscope is both oriented towards improving the quality of demand by giving better information to children and parents as ‘consumers’ of education, and could constitute a co-produced information resource. The following sections draw on instance and embedded case examples to elaborate on specific connections from Table 4.

4.3.1. Politics

Looking at the use of COINS we see conventional forms of scrutiny (C4, newspaper led) complemented by crowd-sourcing initiatives (C3-C4) and independent individuals (C1, C5) and companies (C2) providing their own platforms for members of the public to explore government spending, and, in theory, to scrutinize that spending. The availability of multiple education datasets also allowed an independent school governor (E5) to scrutinize a specific government policy. However, as noted in 4.2.6 much scrutiny can fall short: unless data is also used to lobby for change.

Scrutiny-type processes may however feed into individuals voting choices, or direct interaction with elected representatives and senior officials. A number of instances refer to the use of political data to explore electoral histories, perhaps for strategic voting. Data-journalist Simon Rogers notes that during the 2010 UK Election “numbers were debated every single day”, suggesting that data is coming to play an increasingly fundamental role in political debate. That role can be direct, when OGD allows citizens or media outlets to check the claims made by politicians, or may be more subtle, though no less political.

For example, the ‘Facebook Crime Quiz’ takes data on crime rates in a local area and makes a game out of checking whether a user’s perception of crime matches the reality. The ‘Awsometer’ mobile-phone application takes a user’s location and checks a variety of open datasets for positive public provision nearby, giving an ‘Awsomeness score’ to that location. The application’s developer was inspired to create it as a response to the ASBOrometer iPhone application which displays anti-social behavior statistics for a locality, describing it as a form of “digital pamphleteering”.

4.3.2. Co-production

It was difficult from the data available to identify instances in which OGD supported significant engagement between state and citizens, or state and service providers that could be clearly described as “collaborative” or “community based” where community is understood as a community of interest or local community. Rather, loose linkages between public services and independent actors facilitated by the flow of OGD support a range of models of ‘co-production’ (Brudney & England 1983; Needham 2008; Pizzicannella 2010). Co-production of information was most common, although a number of survey responses also described the co-production of planning: small enterprises using OGD resources to create environmental, education or housing reports subsequently sold to or shared with government to inform strategic planning.

The Lichfield My Area Map (E3) illustrates that informational (and other forms of) co-production may also occur between layers of government: allowing a council to seamlessly provide information on services provided by another tier of government. Few examples of co-produced services were available in the sample, but, in reducing ‘friction’ in information exchange, there is clearly potential for OGD to play a role in the functioning of commissioned-out services.

4.3.3. Markets

UKSchoolsMap and Schooloscope (E1, E4) are illustrative of OGD interfaces that support ‘consumers’ of services to choose provision based on their specific preferences, although the impact these can have depends on how underlying markets are functioning (and how effective market forces are in specific public service domains). Whilst demand-side market-based OGD use is more visible, a number of data-use instances, survey responses and interviews suggested OGD has a key role in stimulating a competitive marketplace for public services.

80% of all survey respondents agreed with the statement “Innovators from outside government will use OGD to build better online services than government can”. The perceived incapacity of government to develop in-house, or to procure small-scale, modular and agile services (not only online services) was a recurring theme in discussions with participants at hack-day events. As one interviewee noted, when government’s effective monopoly in specific areas of data collection is converted into monopoly control of that data it “actually drives others out of the marketplace” and crowds out innovation. When data is open then access to the data does not provide a competitive advantage to firms with exclusive data-access agreements. Competitive advantage has to come from offering innovative value-added services on top of the data.

50% of survey respondents indicated providing a platform for others to build upon was amongst their motivations for OGD use, but also noted their OGD use was experimental and without long terms sustainability plans. No conventions were evident in any of the cases explored for indicating to end-users when the bulk-data in an application was last updated, nor for indicating transparently how the data had been processed before it was presented.

Social networks and support from peers were important in inspiring or enabling many OGD uses. Whether it was the training provided by a Local Information Service (LIS) manager to non-technical users seeking facts for their work, or the supportive events and mailing-lists around data.gov.uk, many OGD re-users noted the importance of communities of practice for enabling their OGD use.

“When I saw the EduBase dataset was out there I didn’t know anything about RDF or linked data at the time, but luckily…there’s a good community of people around data.gov.uk…”
Local authority web-manager

Attentiveness to those excluded from, or peripheral to core socio-technical networks around OGD is important in future analysis of the democratic potentials of OGD. The close connections between technically-oriented OGD re-users, and data.gov.uk OGD providers is notable, and potentially reinforces the ‘data for developers’ focus of the platform. One interviewee, responsible for a LIS noted that training and capacity building were as important as tool building in getting data used effectively:

“It’s a long process – it has taken several years really to build people’s confidence [in using data in their work].”
LIS Manager

Creating interfaces onto data, and informational representations of datasets, involves making value judgments, both about what to present, and how to present it. Reflecting on the changing role of newspapers in relation to data, data-journalist Simon Rogers explained:

“I think our role as gatekeepers has changed to become roles as interpreters – to help people interpret data and use it and get the most out of it, which is a tricky thing.”

Schooloscope (E4) is similarly aware of its role as an interpretation layer on top of raw-data. Rather than present statistics as UKSchoolMap (E1) does, Schooloscope provides a visual and textual narrative about the quality of each school, applying an algorithmic interpretation of the data on each school to make qualitative statements such as “Pupils at The Cherwell School are more or less content”. It is both a search interface and an information source: articulating a value-laden interpretation of each schools’ performance.

Economic forces (Schoolscope has had over £100,000 investment; the unfunded UKSchoolMap suffered downtime and accuracy issues during the data collection period), the link structure of the web (Hindman 2009; Guardian school appeals data appears high in Google results than data.gov.uk’s raw-data), and wider structures of media networks (Castells 2009), are likely to continue to impact on the practical availability of different interpretations of data to end-users.

Non-obvious interfaces onto data (i.e. doing more than putting geo-data on a map, or transactional data in a table) can be significantly more complex to create than straightforward representations that follow the existing structure of data.