Digital Government Session
Organizer & Chair: Joseph Hardin, University of Michigan

Panelists:	Noshir Contractor, University of Illinois at Urbana-Champaign
	Tracy Westen, Center for Government Studies
	Russ Neuman, University of Pennsylvania
	Larry Brandt, National Science Foundation

The section on Digital Government contained presentations reflecting two facets of representative government that confront researchers and developers in the human sciences: the selection of members of government through the electoral process, the interaction of candidates and voters; and the operation of government through the construction and reconstruction of the organizations that provide the services of government, the government agencies. Contractor and Neuman discussed the challenge of new information technologies for agencies that necessarily have legacy systems and are composed of complex human organizations. Both focused on the unique status of government and the ways in which they differ from for-profit organizations. Brandt then described an initiative to connect researchers and agencies that are interested in using emerging information technologies. Finally, Westen looked at the use of digital resources in support of elections, bringing a case study of an actual system that has been in use and is under continuing development.

Contractor's presentation extended the issues discussed and reported at the panel on Digital Government at an earlier NSF-sponsored Workshop on Human Centered Systems. That report underscored the recognition that from an organizational standpoint federal agencies differ from private sector businesses in terms of scale, the nature of the service-profit relationship, heterogeneity of participants, as well as concerns about access, privacy, and security. The report argued persuasively that these differences present a national challenge for the technical design of new information technologies as well as the migration of legacy systems. Contractor argued that the organizational and social issues involved in the design and deployment of these technologies pose an additional dimension to that national challenge. He identified some of the lessons learned from past evaluations of federal information systems, suggested two meta-theoretical perspectives, those of public goods and self-organizing systems, that are relevant to the design and deployment of successful federal information systems, and proposed collaboratories as an infrastructure to design and deploy National Challenge testbeds for digital governance. (from Contractor)

The idea of collaboratories, which combine the expertise and perspectives of domain experts, social scientists, computer scientists and users in the construction of testbeds that are then operated with a high degree of evaluation, was well received. Examples of analysis and lessons learned from ongoing research emphasized the importance of user support when agencies are developing tools and, specifically, the need to support agency staff in helping users gain access to information and services - good interfaces and interactive help systems are not enough for many users. This ability to look outside a purely technical approach and utilize all the available resources of an agency, including but not limited to technical fixes, is a core component of a comprehensive human centered approach to technology adoption and adaptation.

Brandt pointed out that most Federal agencies are struggling with the creation of a strategic vision and an operational philosophy for information technologies. The goal of NSF's new Digital Government initiative is to build a research domain of problems that intersect the traditional CISE research communities with the mid to long-term research, development, and experimental deployment needs of the Federal information service communities, by supporting projects that innovatively, effectively, and broadly address potential improvement of agency, interagency, and intergovernmental operations and/or government/citizen interaction. The Initiative arises from the recommendations of a 1996-97 Workshop on R&D Opportunities in Federal Information Services, chaired by Herb Schorr, Director of USC's Information Sciences Institute, with Sal Stolfo of Columbia's Computer Science department serving as co-chair. The Workshop final report can be found at http://www.isi.edu/nsf/ (from Brandt) The necessity of human-centered design to efforts of the type described by Brandt was emphasized in the discussion. Especially important was building in from the start in-depth evaluation and analysis of the activities and responses of all the participants. The development of motivated responses to calls for innovation should be built on the understanding of not only successful, but also unsucessful efforts to apply information technolgies to government agencies and their services. The cross discipline approach and the proposed integration of the research community, including but not limited to the computer science community, with the activities of federal agencies was very well received.

Neumann identified what he termed an old and familiar story in the sociology of complex organizations and in public administration research. He pointed out that self-interested demands of both agency personnel as well as private vendors and consultants is unlikely to move us very far outside of the box in thinking about inventing a new approach to the networked provision of public service, stating that the Reinventing Government and National Performance Review initiatives have run out of steam and neither drew heavily on the insights of researchers studying the impact of information technology on organizational structure and productivity. All this in the face of a quantum change in scale and function of the federal government as we move from the industrial to the information age. The problem, as we stand on the cusp of this transition, is that connecting stovepipe agencies and offices to the Internet may not be the right idea at all. It may prove to more like giving cellular phones to our pony express riders -- not a bad idea in itself, but perhaps we should be rethinking our equine assumptions about postal connectivity. (from Neumann) The presentation of a matrix of IT impacts and where these impacts might affect public functions was meant to be a starting point for discussion that was not constrained by old assumptions. The discussion surrounding the choice of row and column elements for the matrix highlighted the differences in government functions, such as collective defense vs social services, and the varying impacts that, say, lower transaction costs or disintermediation or increased speed of communication might have on the various sectors. The broad view gained by abstracting to this level was recognized as particularly beneficial to those trying to initiate change across the various functional units, such as Brandt and the initiative he described.

Westen presented the Center for Governmental Studies' new working system of interactive digital political communication called The Democracy Network. The Center first tested The Democracy Network in two "laboratory settings" during the 1996 Presidential and selected local elections. This system is an attempt to determine whether innovative uses of interactive digital communications technologies can stimulate candidates to provide voters with a broader range of electoral information. Preliminary testing indicates The Democracy Network has the capacity to improve the quality of voter information, as well as reconfigure the parameters of electoral debates, in a number of ways. Increases in candidate participation, the addressing of more specific and substantive issues in the campaign, encouragement of issue centered debate among candidates, and the use of the system by voters to analyze candidate positions on issues and to volunteer to work in campaigns were all discussed. (from Westen) Not surprisingly, some skepticism was expressed concerning the impact of such systems on larger elections, ones where large amounts were spent on advertising, but there was considerable interest in the actual results that were reported on the use of the system in the campaigns to date. The detailed approach to design of the system, emphasizing use from both the candidates' and voters' perspectives, reflected evident concern on the part of the developers in making the system approachable, overcoming first generation hudles to use, and providing built-in incentives for both sets of users to participate. These discussions highlighted issues in the Human Centered Design arena.

Presenter Synopses

Organizational and Social Issues in the Design and Deployment of Human Centered Systems for Digital Governance

Noshir Contractor
University of Illinois, Urbana-Champaign
nosh@uiuc.edu

Overview

This presentation extends the issues discussed and reported at the panel on Digital Government at an earlier NSF-sponsored Workshop on Human Centered Systems. That report underscored the recognition that from an organizational standpoint federal agencies differ from private sector businesses in terms of scale, the nature of the service profit relationship, heterogeneity of participants, as well as concerns about access, privacy, and security. The report argued persuasively that these differences present an national challenge for the technical design of new information technologies as well as the migration of legacy systems. This presentation argues that the organizational and social issues involved in the design and deployment of these technologies pose an additional dimension to that national challenge. It makes this argument in three steps. First, it identifies some of the lessons learned from past evaluations of federal information systems. Second, it identifies two meta-theoretical perspectives from the social and organizational sciences that are relevant to the design and deployment of successful federal information systems. Third, it proposes Collaboratories as an infrastructure to design and deploy National Challenge "Test-beds" for digital governance.

Past studies of the use of information technologies in government agencies have debunked several of our conventional myths and tacit assumptions, by emphasizing the social and organizational character of the deployment process. For instance, they have found that: * Government agencies are more likely to deploy an information technology if the individuals in charge of making the decision about the technology believe that the technology will reinforce their political power within the agency (Political reinforcement theory). Government agencies are more likely to use a new technology if they are given the flexibility of "re-inventing" various features and aspects of the technology. (Unintended uses of technologies). * Government agencies will often ignore (or sidestep) requirements of entering data into large integrated information systems if the individuals charged with the responsibility of entering the data are not the ones who get any direct benefit from the databases. Government agencies are more likely to collect more complete data if these are automatically collected as part of the workflow, rather than requiring individuals to perform special data entry activities. * Government agencies are not always committed to the collection of accurate data in their databases, especially if these data are likely to hinder management's ability to negotiate future budget requests. * The support for improved boundary spanning activities (with customers) by government agencies does not necessarily entail the design of better kiosks and front-end interfaces for customers directly. This is because customers with the greatest need in terms of help and advice on education, health, welfare, etc. are often the least able to take advantage of these interactive services. Instead, government agencies need to design technologies that support their employees in assisting such disadvantaged customers. Substantive contributions from the social and organizational sciences Findings such as those reported above point to the need for integrating social and organizational issues into the design and deployment of National Challenge Info structures for Digital Governance. More specifically there are at least two meta theoretical frameworks in the social and organizational sciences that directly speak to the issue at hand. First, theories of collective action and public goods offer important insights into the social and organizational factors that influence the actual use of a technologically well designed information and communication system. These theories begin with the premise that public goods, such as federal information systems, have two distinctive characteristics: the impossibility of exclusion, meaning that members of the collective cannot be excluded from using the information even if they did not contribute to it, and jointness of supply, meaning that one person's use of the information does not necessarily diminish the level of information for other users. Within this framework, these theories examine individuals' investments in, and benefits from, participating in the production (or creation) as well as the distribution (or maintenance) of public goods such as information systems. Second, theories of self-organization describe the mechanisms by which the norms surrounding the use of information systems evolve over time. It is widely recognized and accepted that the mechanisms influential in the design and initial use of an information system are not the same as those influential in the sustained use, diffusion, and growth of such systems. Recent developments in computational modeling and simulation techniques have the potential of providing us with the ability to assess and visualize the impacts of initial design and deployment decisions on the long term use of these information systems by individuals. Specifically, they can offer insights into how the creation, maintenance, and dissolution of communication, trust, resource, and knowledge networks influence (and are influenced by) the adoption and use of collective goods such as information systems.

The lessons learned from past studies as well as the recent substantive and methodological developments in the social and organizational sciences suggest the need for incorporating these components into the design of National Challenge Test-beds for Digital Governance.

Historically there have been very few examples of teams that were successful in strategically structuring multidisciplinary teams in the design and deployment of large scale information systems. Instead, it was more likely for a team of computer scientists and engineers to design systems, and for organizational scientists to come in after the event to evaluate these systems. The critical importance of the National Challenge "Test-beds" make this loosely-coupled relationship a potentially costly error.

Recent experiences of several NSF initiatives, including our own Team Engineering Collaboratory (the infrastructure used by physically distributed researchers on the NSF-funded Project CITY - Civil Info-structure TechnologY), suggest that a "Collaboratory" model is more likely to successfully and concurrently dovetail the expertise and efforts of computer scientists, domain experts, users, human-computer interface specialists, and organizational scientists. These Collaboratories, with their triadic focus on people, instruments, and data, are especially well suited to design federal systems that specifically address National Challenge questions such as:

• How do we structure federal information systems that provide incentives and reduce costs for all individuals who are responsible for contributing to, and obtaining information from, the systems?

• How do we structure sustainable information systems that minimize the cost of capturing and disseminating the migratory (or the "what?") knowledge and the embedded (or the "why?/how?") knowledge of an individual worker within such systems?

• How do we design systems that recognize and respond to the varying needs (in terms of data aggregation, and routine versus one-shot queries) and capabilities (in terms of technological access) of multiple external "customers" of the government , including private citizens, businesses, and the media?

• To what extent can the government offer incentives and safeguards for various external customers to provide authentic data TO the government via these information systems?

• How can agencies use these information systems as knowledge networks for the creation, maintenance, and dissolution of adhoc inter-agency initiatives?

The opportunity of using a new generation of technologies to usher the era of digital governance is flush with possibilities that were until recently unimaginable. It would be frustrating to see this cycle of progress delayed by our inability to recognize and incorporate the important social and organizational dimensions that influence the design and deployment of these new systems.

The NSF Digital Government Initiative

Lawrence E. Brandt
Program Director for Digital Government
Division of Experimental and Integrative Activities
National Science Foundation

The Federal government is a major user of information technologies, a collector and maintainer of very large data sets, and a provider of critical and often unique information services to individuals, states, businesses, and other customers. Yet, most Federal agencies are struggling with the creation of a strategic vision and an operational philosophy for information technologies. The goal of NSF's new Digital Government initiative is to build a research domain of problems that intersect the traditional CISE research communities with the mid to long-term research, development, and experimental deployment needs of the Federal information service communities, by supporting projects that innovatively, effectively, and broadly address potential improvement of agency, interagency, and intergovernmental operations and/or government/citizen interaction.

The Initiative arises from the recommendations of a 1996-97 Workshop on R&D Opportunities in Federal Information Services, chaired by Herb Schorr, Director of USC's Information Sciences Institute, with Sal Stolfo of Columbia's Computer Science department serving as co-chair. The Workshop final report can be found at http://www.isi.edu/nsf/

NSF is working toward the release of an initial Program announcement in mid-April 1998, with proposals due in mid-July for 1998; the deadline date for future years will be mid-March. Other related announcements may be issued jointly with other Federal agencies later. At present, the mandatory requirements for proposals are:

• Significant participation by university-based researchers in domains funded by NSF/CISE.

• Significant participation in the definition and execution of the work proposed by at least one Federal agency.

• A domain area that is primarily if not uniquely governmental in nature, or where Federal government requirements are unique.

• Integration of domain experts (users, customers) during the project life cycle.

• A plan for evaluation, during the project and at its conclusion, by the proposer to determine the usefulness and usability of its research products in a Government context.

• Proposals for domain-specific cross-agency pilots or testbeds (see Categories of Proposals below) must describe a framework or architecture to encourage the sharing of data, development of application program interfaces, software objects, software re-use, modularization, and importation of new functionality from other projects.

Questions about the Digital Government Initiative may be directed to:

Lawrence E. Brandt
Program Director for Digital Government
Division of Experimental and Integrative Activities, Suite 1160
National Science Foundation
4201 Wilson Blvd.
Arlington VA 22230

Phone - 703/306-1963
Fax - 703/306-0632
Internet - lbrandt@nsf.gov
Home page - http://www.cise.nsf.gov/eia/staff/lbrandt/index.html

The Public Sector and the Private Sector in the Digital Age:
New Questions for Research

W. Russell Neuman
Annenberg Public Policy Center
University of Pennsylvania

This is a call for thinking outside of the box. The need to upgrade legacy information systems and human interface designs in government agencies is important, urgent, difficult work. It deserves the sustained and energetic support of the research community in computer science, statistics and the social sciences. In my view, however, there is a gap in the research agenda that moves considerably beyond the question of efficient conversion from legacy to modern information systems. If I am correct, such a gap might well be addressed with the support of the National Science Foundation. I'd term it - the ultimate national challenge.

The function and scale of the federal government changed dramatically as we moved from the agricultural to the industrial age. By the third administration at the end of 18^th century, the American federal bureaucracy had grown to over 600 full time-employees. The federal banking system had three employees. Not much of a bureaucracy yet, although some at the time argued we were already overstaffed. At the turn of this century, we face a very different scale and complexity of governance. I need not cite the statistics.

The function and scale of the federal government will change yet again as we move from the industrial to the information age. The problem, as we stand on the cusp of this transition, is that connecting stovepipe agencies and offices to the Internet may not be the right idea at all. It may prove to more like giving cellular phones to our pony express riders - not a bad idea in itself, but perhaps we should be rethinking our equine assumptions about postal connectivity.

This is an old and familiar story in the sociology of complex organizations and in public administration research. The self-interested demands of both agency personnel as well as private vendors and consultants is unlikely to move us very far outside of the box in thinking about inventing a new approach to the networked provision of public service. The Reinventing Government and National Performance Review initiatives have run out of steam and neither drew heavily on the insights of researchers studying the impact of information technology on organizational structure and productivity. It may be that the fields of political science, public administration and applied MIS are paradigmatically challenged.

The social, economic and behavioral sciences are just beginning to utilize their scientific resources to address questions of functionality, efficiency and equity at the borders of the public and private spheres. Rather than lag behind, there might be an important opportunity for the NSF to draw out and nurture new lines of interdisciplinary inquiry applied to national challenges at the highest level.

There is a useful matrix that contrasts a list of governmental functions and a set of generic properties characteristic of integrated digital systems. This is not a solution, but I hope a provocative starting point for fresh thinking about what digital governance could really mean.

Matrix of IT Impact and Public Functions

Tracy Westen
President, Center for Governmental Studies and The Democracy Network
Adj.Prof. of Law, USC Annenberg School for Communication

Background: The Center for Governmental Studies has created a new working system of interactive digital political communication called The Democracy Network. The Center first tested The Democracy Network in two "laboratory settings" during the 1996 Presidential and selected local elections, using two different sets of candidate-supplied materials:

• Digitized, on-demand video position statements supplied by ten Presidential and several state legislative candidates on up to ten issues, carried over Time Warner's Full Service Network's digital News Channel in Orlando, Florida; and

• Textual position statements (of up to 1,000 words each) supplied by dozens of Presidential, state and local candidates on over 100 issues, carried over the Internet during the Santa Monica, California election.

The Center then refined and further tested The Democracy Network's Internet features in two local elections: Los Angeles (Spring 1997) and Seattle (Fall 1997). Based on these further tests, the Center will make The Democracy Network available for use in up to six states during the 1998 Congressional and State Elections and then make it more widely available to all states for the Presidential, national and state elections of 2000.

Features: The Democracy Network is an attempt to determine whether innovative uses of interactive digital communications technologies can stimulate candidates to provide voters with a broader range of electoral information. It provides participating candidates with IDs and Passwords, allowing them to log onto The Democracy Network from remote locations, select any substantive issue of their choosing, prepare a textual (and in some instances video) statement of up to 1,000 words, remotely insert that statement into an "Issue Grid" for viewing by the public and then rebut their opponents’ statements. Opponents’ positions are labeled "No Comment" until filled. The "Issues Grid" thus allows voters to watch candidates engage in a "digital debate" as their positions on specific issues evolve over time.

The Democracy Network also allows voters to watch debates between ballot measure committees, review candidate biographies, volunteer for or make campaign contributions to candidates, communicate with candidates and other voters about specific issues, review the coverage of participating media and utilize a number of other features. The Democracy Network is supported by national foundations, contains no advertising and is free to users and all candidates on the ballot.

Preliminary Results: Preliminary testing indicates The Democracy Network has the capacity to improve the quality of voter information, as well as reconfigure the parameters of electoral debates, in a number of important ways.

• Candidates will participate if The Democracy Network is available and explained to them-with participation rates as high as 91% in some instances.
• Candidates are more likely to use the system to address specific issues than espouse generalities.
• Candidates are more likely to make factual and substantive comments in the system than generalized emotional appeals.
• Candidates and voters will use the system to "push" opposing candidates to address issues they might otherwise avoid.
• Candidates will participate more actively when the are able to do so remotely through the system's Remote Updating System (RUS).
• Candidates believe the system will ultimately reduce their campaign costs.
• Voters will significantly use the system if it is publicized and known to them.
• Voters principally use the system to review candidates' substantive positions on issues.
• Voters will use the system to volunteer for candidate campaigns.

Further Potential Impacts: Initial research on The Democracy Network also suggests the system has the capacity to improve the electoral process in a number of additional ways:

• By reducing the financial disparities in candidates' spending.
• By increasing voter participation.
• By exposing voters to contrasting viewpoints in a convenient format.
• By allowing voters to control the time of their exposure to the candidates' messages.
• By allowing both horizontal comparisons of competing candidates and vertical comparisons of the same candidate.
• By giving voters the choice of information in a multimedia format.
• By letting voters review the statements and questions of other voters.
• By letting voters review their own decisions over time.
• By giving minority candidates greater visibility.