Steve Ventura, Associate Professor, Environmental Studies and Soil Science
 Ben Niemann, Professor, Urban and Regional Planning
 Todd Sutphin, Information Processing Consultant
 Rick Chenoweth, Professor, Urban and Regional Planning

Land Information and Computer Graphics Facility
 B102 Steenbock
 University of Wisconsin-Madison
 Madison, WI 53706

The role of spatial information technologies in decision-making has been debated almost since the inception of their use in local government land information systems. In perhaps over-simplified terms, one question is whether land information is used to help make decisions, or is it used to justify decisions made for many other reasons?' Niemann (1987) and Zwart (1988) epitomized the debate in their point and counter-point conference articles about "better information" resulting in "better decisions" through modernized land information systems.

At the local level, 'just getting the job done' public agency practitioners generally have not been concerned about the role or impacts of spatial information. If they take time to consider these issues, it is likely that most would follow disciplinary training and assume that improvement in quality and availability of land information benefits the citizens and organizations they serve, and that improvement particularly in form and access extend availability of information to audiences that otherwise may be excluded from decision-making processes.

Little empirical evidence has been reported to support or deny this belief in a positive role for land information in land use decision-making in the contemporary US local government context. A theoretical model of the role of information in local land use planning was developed by Knapp et al. (1995). They explicitly looked at the effect of information about local government infrastructure investments and land use regulations on the timing of development decisions. This study did show that information modifies development decisions. However, the models were limited to interaction between local government and developers; it did not account for all the other actors in development decisions, particularly actors that could be more influential if empowered by information.

In the land use debate, the difficulty of ascertaining who may be affected by land use decisions (and how this might change with differences in access to information) is exacerbated by the diffuse nature of the decision-making process. Many citizens are affected by land use decisions but may not be directly involved in the decision-making process. The optimistic view is that of a Jeffersonian democracy, where well informed citizens exert an indirect influence on the process, through elections, meetings, and even through consumer choices. A cynical view suggests that various elite groups control the decision-making process, for their own benefit. From an information standpoint, a question critical to understanding which view prevails may be 'whose information?'... whose worldviews are represented in a data base and in analytic tools to understand the data, and do these representations exclude the views of segments of society?

Questions about the role of land information in local government decision-making have been difficult to resolve because characterizing the decision-making process has been and continues to be difficult to ascertain. In particular, it is difficult to determine what role land information plays in local land use decisions because the process is influenced by so many other factors, including political, economic, legal, bureaucratic, personal and social pressures. And, the actors involved may not always be entirely open, knowledgeable or forthright about what has influenced decisions. Moreover, research must be done in situ; we don't have the luxury of controlled experiments, where we can suffuse a jurisdiction with information to observe the result while controlling or accounting for this host of other factors.

Both NCGIA (1998) and UCGIS (1997) have recognized the need to understand the role of GIS in society, including how public agencies and their constituents understand and use analyses and products from land information systems. Harris and Weiner (1998) and others also raise questions about differences in access to spatial information technologies and how this impacts different segments of society. However, before we can assess which groups may be empowered and which may be marginalized in land use decisions, we need to understand if and how land information is used and by whom. This requires clarification of some terms that express, in essence, why data are collected and disseminated.

The recent interest in GIS and society builds from work going back to at least the 1970s. During this period, the discussion has evolved from a focus on spatial information technologies exclusively within a single organization to a view of these affecting all segments of society. The germane questions throughout this evolution are which group(s) are generating and which are using spatial data, and thus where do the costs and benefits accrue.

'Land records modernization', a term possibly first used by Larsen et al. (1978), focused attention on data automation and the role of spatial information technologies within a single organization such as the departments comprising a local government. The primary beneficiary was the organization, through more efficient data retrieval.

The terms 'multipurpose cadastre' (NRC 1980, 1983) and 'multipurpose land information system' (Niemann et al., 1987) started a focus on the ability to move data between multiple organizations. The concept was encapsulated in the term data sharing, epitomized by NCGIA's Initiative 9 and resulting book with a similar title "Institutions Sharing Geographic Data" (Onsrud and Rushton, 1995). An underlying assumption was that data sharing was good public policy, reducing duplicative activities, ensuring that citizens didn't pay twice for the same data, allowing agencies to work from a common set of facts, and so forth. Still, the focus was primarily public agency to public agency.

The explosion of GIS in the 1990s has been accompanied by increased focus on issues of data access. This may be attributable to many factors, including much greater private sector use of GIS, more facile user interfaces, and proliferation of local land information systems with detailed information of interest to citizens and businesses. The issues that characterize data access are primarily the result of public-private interaction -- a balancing act that public agencies must sort through between open records requirements and protection of rights to privacy. 'Good public policy' arguments underlie both points. Privacy is constitutionally protected. The motivation for the Freedom of Information Act and parallel state open records laws are generally quite clearly described as providing citizens access to the means of government decision-making. The term 'democratizing' has been used to describe the culmination of GIS maturation wherein the system provides facile access to public data, increasing the ability of citizens to participate in decision-making processes (e.g., Tulloch, 1997).

Recent discussions and papers (e.g., the special issue of Cartography and Land Information Systems (volume 25, no. 2) on public participation GIS suggest that there is a step beyond access to data, a step requiring systems designed explicitly for representing worldviews of traditionally under-represented groups and providing them with meaningful and culturally relevant information products. In keeping with the evolution described herein (data automation à data sharing à data access), this might be described as data advocacy, though clearly the underlying concept goes well beyond just considerations of data.

On the public side, data advocacy involves creating GIS-supported decision-making systems that include processes open to all affected groups. On the private side, it involves eliminating real or perceived barriers to participation in decision-making, including technological barriers such as representation, access, training, even spatial literacy. The evolution could be described as moving from focus on a single public agency to multiple public agencies to public-private cooperation (primarily one-way, and primarily to those already embracing spatial information technologies) to public-private collaboration (with two-way data flows). Data advocacy is still a perspective that reflects a public agency's role in building and maintaining information systems, as opposed to other forms of public participation GIS that may involve implementation by non-governmental organizations involved in public decision-making processes.

Our project contributes to the discussion about the role of data and land information in land use decision-making by purposefully improving the type, quality, and availability of land information and analysis in a jurisdiction with an on-going and highly charged land use decision-making scene. It is, in essence, an attempt to move toward data advocacy and perhaps a higher degree of "democratization" that this implies. We will attempt to gauge the influence and impact this has on land use decision-making processes and outcomes through first-hand observation, post-decision reconstruction, surveys, and other methods. Key questions include:

• is new information being used?
  - in what form?
  - in what parts of the land use process?
  - how is it used (to help make or to justify decisions)?
  - does it or can it represent groups not traditionally empowered in decision-making

Dane County, Wisconsin (the County) is one of the fastest growing counties in the Midwest. It also continues to be one of the most productive agricultural counties in the state, typically ranking first in the state and in the top fifty nationwide in gross agricultural sales. For at least a couple of decades, the conversion of farmland to residential and commercial purposes has been contentious. The County Board frequently splits along rural/urban lines on land use issues, with pivotal votes coming from fringe suburban areas. More effective land use planning was a major theme in the campaign of the current County executive. She has followed with an effort to bowdlerize a regional planning commission that is widely regarded as ineffective (in both technological and political terms) and incorporating this function in the County's more technologically sophisticated planning department, and by embracing geospatial information and visualization technologies as part of a suite of tools to more effectively involve the public in land use planning and management.

A cooperative relation has existed between the County and the University of Wisconsin-Madison Land Information and Computer Graphics Facility (LICGF) for almost two decades. LICGF has conducted research and development on land information technologies and applications in this "real world" context with the County, essentially reducing the County's risk in adopting innovative technologies. As a result, the County has a sophisticated automated land information system used primarily for real property listing, tax assessment, deeds recording, and soil and water conservation. They have recently begun to use it for land use planning as well.

A recent rejuvenation of the LICGF-Dane County relation represents initial evidence that the "whose reality is represented" question must be considered in understanding how information is used. In essence, the County Executive's interests were piqued when she was shown a very different picture of how much land could be considered 'open space and farmland' than that shown in a 25 year land use plan done by the autonomous Dane County Regional Planning Commission (RPC). In essence, we provided evidence from the County's own databases that countered the PRC's suggestion of almost completely unencumbered open space beyond city boundaries. Our GIS-based analysis of land use classification from tax assessment roles indicated less than 50% of the county was developable farm and open space; in contrast, RPC used data for their land use plan that showed 85% of the county in this category, based on air photo interpretation. The tax assessors view is arguably closer to the land owners view of what its use is or could be.

Since the Spring of 1996, when the current County Executive first took office, we have assisted the County in developing and disseminating land use information. We have guided the County Executive's staff in the analysis, display and dissemination of their own geospatial data, particularly information related to land use, ownership, assessment, and resources. We have attempted to make high quality geospatial data and information readily accessible to anyone interested in using it in local land use planning, in several forms and through several venues.

To begin answering some of our questions posed earlier, we have simultaneously been observing how decision-makers and interested citizens and organizations react to and use geospatial technologies and information products. These can be thought of as experiments about the form and access to information. Activities include:

• Presentations with feedback. Overviews of County conditions, resources and trends based on geospatial analysis were presented in several venues, including a County-sponsored "land use forum" series, civic groups meetings, meetings with key staff at many of the involved agencies and organizations, and a University-based seminar series on the role of geospatial technologies in land use planning in which many of the key decision-makers were invited to speak about their views on this topic. The County land use forum series included "listening sessions," which provided an opportunity to observe directly how participants were thinking about and using geospatial information.

• "What do you like" experiment. In the largest County land use forum (over 300 participants), about 25 large format (48 X 36 inches) maps were presented. Participants were given a brief survey after an "open-house" period of observation, with questions about which products they found useful, what other products might be useful for land use planning, and how they saw themselves interacting with geospatial technologies. As an incentive to complete the survey, participants were promised a copy of the map of their choice. This experiment provided information both directly from the survey, and indirectly from tabulations of which maps participants selected as their reward for survey completion.

• Allocation experiment. In another County land use forum session, participants were divided into small groups and asked to place dots on large format maps to designate areas for future residential development. The maps portrayed factors related to growth opportunities and constraints. Different dots represented different numbers of residences (by dot color) and development density (by dot size).

• Web-based feedback. We are developing a series of planning guides describing how to use geospatial technologies and the County's data for several aspects of land use planning. These will be available in both hard-copy and Web-based formats. We will monitor the distribution of the bulletins and record IP domains of computers accessing the Web-based material. We will also ask viewers to provide feedback on the material and comments on the general approach with an open-ended mailto and a fill-in-the-blank survey.

• Training experiment. We provided free training in commercial GIS and visualization software, and their use in planning using the County data sets. We offered five 1-day sessions in five different aspects of land use planning. The sessions were widely advertised and open to anyone. The course included a general introduction to spatial information technologies and then described their use in planning applications in detail. This experiment is providing us with data on who in the community is interested in improving their aptitude for GIS-based planning and which aspects of planning they consider GIS as an appropriate tool. The courses were so popular that we have continued to offer them on a cost-recovery basis.

• Software module development. We are building five software modules for land use planning (Exploration, Analysis, Allocation, Impact, Public Access). These will be enhancements or extensions of existing GIS and visualization software. Selection of existing packages and development of new software will be based on citizen and County staff participation during module development, providing feedback on the usefulness, ease of use, and ability of modules to engage and involve citizens in land use issues.

• Public access terminal experiment. With the County's cooperation, we will develop a "public kiosk" interface to the County's data and analyses. The intent is to provide a kiosk that can supports "what if" analysis, not just information retrieval and display. We will monitor who deploys this interface, who accesses data through it, and solicit feedback from users through open-ended mailto and a fill-in-the-blank survey.

The overt goal for our activities is to make geospatial data and information readily accessible to anyone interested in using it for local land use planning and related applications. This will be accomplished through the litany of activities listed above. Emphasis is on trying to de-mystify and simplify user interfaces and other aspects of geospatial technologies that have hindered access to GIS databases and analysis using currently available commercial tools. We are also attempting to record participants ideas and feedback on data and particularly, analyses, so that the process enables them to guide what kind of information and information products are generated.

The covert goal of our project is to determine if this unprecedented access to and education about geospatial technologies and products makes any difference in the planning process. We intend to survey and interview participants to directly find out how individuals and factions/organizations perceive the impact and utility of geospatial technologies. We will also be gathering indirect evidence such as where and how GIS-derived products and facts are used in documents and meetings, whether these are used to help make decisions or used to justify decisions, and which factions or organizations seem to be able to make most effective use of them.

Two general strategies are being used to answer the three key questions raised in the introduction.

First, behavioral measures are used to directly observe land use decisions and the products that were used in the process. Behavioral measures can be built into software programs to track individual decisions in a manner similar to computer tutorials used in advanced training programs in a wide variety of professions. Behavioral measures can also be obtained from direct observations of group meetings that are later scored using content analysis procedures developed in the social sciences.

Second, self-reported behavior using questionnaires can be obtained where direct measures of behavior are not possible or inefficient. Questionnaires have the advantage of being able to explore the beliefs and attitudes that underlie the behaviors in question and, using attitude/behavior theory, can be used to design strategies that are most likely to be effective in efforts to transfer technology and disseminate its use.

While true experiments comparing the influence and impact of land information, analysis and visualization between different communities of users may not always be possible, the careful use of quasi-experimental designs such as time-series or non-equivalent control groups should be useful in isolating those factors that most influence the successful adoption and use of information technologies.

Preliminary evidence indicates a high degree of interest in improved land information analysis and visualization from County staff and other actors typically involved in land use decision-making. Requests for additional information products from County staff continue. The requests are increasingly specific about the type and form of analyses and products, indicating a more sophisticated understanding of the spatial data base and what is feasible in analyses and products. The county continues to invest more than 1/2 million dollars annually in land information systems and staff. The use of spatial information technologies is touted by County staff and officials as a key component in resolving some of the County's vexing land use issues.

Observations of participants in the land use forums indicates that citizens motivated to become involved in county-wide land use issues come into the process with a relatively high level of computer acumen. For example, we asked questions about computer literacy and access, and learned that 92% of the participants had access to computers and 70% were on the Internet. The majority of participants selected a complicated composite map depicting several factors related to growth management (as opposed to simpler single-theme products) as their reward for participating in our survey.

The allocation experiment provided explicit evidence about preferred development patterns and strategies of different societal factions (e.g., developers, farmers, rural or urban elected officials, environmentalists, etc., at least to the extent that the small groups were identifiable as particular factions). Preferred strategies conformed with expectations. For example, one group placed almost all their dots in and near existing urban areas, a "compact growth" strategy favored by environmentalists; another group scattered dots through a more rugged section of the county, with none in a region of highly productive farms, suggesting a farmland preservation strategy. All groups understood the spatial analysis involved in the exercise and how the underlying data contributed to their ability to make allocation decisions.

Almost all workshop participants found the land use planning training very useful. Over 280 people applied for 100 available slots, indicating a high level of interest in becoming better able to use spatial technologies and data for land use planning.

It is too early in our experiments and observation to say whether our infusion of better land information has engaged more people in the decision-making process or influenced land use decisions. Clearly though, it has been an important component of the County's process. Land information and spatial analysis were prominent components of the public land use forums. A land use vision statement by the County -- Design Dane (Falk, 1998) -- included many maps that were clearly the products of a GIS. A content analysis of that document revealed at least 38 different calls for information products or spatial analyses to support the County's proposed land use goals. In it, the County Executive called for "improving the way we do business by developing new information technologies to make more informed growth decisions."

The extent to which other actors in land use decision-making adopt and use the products and the technologies remains to be determined. At this point, participation seems to be primarily by technologically inclined. It seems thus far that this comprises most of the community directly involved in land use issues. This is not to suggest that other groups who may be less technologically sophisticated are not affected by land use decisions (and thus by GIS if it plays a significant role in decision-making). We will need to surmount the difficult methodological hurdle of determining who wins and who loses in the overall land use process, and then whether GIS was among the causal factors in these outcomes.

At this point, practitioners promoting more accessible land records can be comforted that we have no evidence to suggest that this has disadvantaged any groups or individuals. Their influence appears to be positive, though of course this is an ongoing drama with many layers and perspectives that await more comprehensive evaluation.

Falk, K.M., 1998. Design Dane! Creating a Diverse Environment Through Sensible, Intelligent Growth Now. Dane County Executive's Office, 47 pages.

Harris, T. and D. Weiner, 1998. Empowerment, marginalization and 'community-integrated' GIS. Cartography and Geographic Information Systems, 25 (2), 67-76.

Knaap, G.J., L.D. Hopkins, and K.P. Donaghy, 1998. "Do Plans Matter? A Game-Theoretic Model for Examining the Logic and Effects of Land Use Planning", Journal of Planning Education and Research, forthcoming

Larsen, B., J. Clapp, A. Miller, B. Niemann and A. Ziegler. 1978. Land Records: The Cost to the Citizens to Maintain the Present Land Information Base, A Case Study of Wisconsin. Madison, WI: Department of Administration, State of Wisconsin.

NCGIA (National Center for Geographic Information and Analysis), 1998. Panel on Geographies of the Information Society, Eric Shephard, Chair. http://www.ncgia.ucsb.edu/varenius/initiatives/ncgia.html

NRC (National Research Council), 1980. Need for a Multi-purpose Cadastre. National Academy Press, Washington DC

NRC (National Research Council), 1983. Procedures and Standards for a Multipurpose Cadastre. National Academy Press, Washington DC

Niemann, B.J., Jr., 1987. "Better Information for Better Decisions: No Question About It." Proceedings, URPIS 15 pp. 187-194.

Niemann, B.J., Jr., J.G. Sullivan, N.R. Chrisman, S.J. Ventura, A.P. Vonderohe, D.F. Mezera, and D.D. Moyer, 1987. "Results of the Dane County Land Records Project," Photogrammetric Engineering and Remote Sensing, 53(10)1371-1378.

Onsrud, H. and G. Rushton, ed., 1995. Institutions Sharing Geographic Data, Center for Urban Policy Research, New Brunswick, NJ

Tulloch, D., 1997. A Theoretical Model of the Life Cycle of Community Multipurpose Land Information Systems Development. Ph.D. Dissertation in Land Resources, University of Wisconsin-Madison, May 1997.

UCGIS (University Consortium for Geospatial Information Science), 1997. Research Priorities: GIS and Society. http://www.ncgia.ucsb.edu/other/ucgis/research_priorities/paper10.html

Zwart, P.R., 1988. "Some Observations on the Real Impact of Integrated Land Information Systems upon Public Decision Making in Australia." Papers from the 1988 Annual Conference of the Urban and Regional Information Systems Association, 1988, Vol. 1, pp. 68-79.