The Future of Public-Sector Internet Services: Pt. II

A national survey reveals the state of government participation on the Web.

by / October 1, 2001
This is Part II in a two-part series of articles reporting on the findings of a national survey of public managers on the Future of Public Sector Internet Services For Citizen Participation and Service Delivery. The study was conducted by John O'Looney of the Carl Vinson Institute of Government and was supported by the National Science Foundation. The respondents were drawn from state and federal government subscribers to Government Technology and from local government officials from communities with populations over 50,000.

One of the more important strategies emerging from the business side of the Internet is the idea of information syndication. Information syndication occurs whenever an information provider contracts with several producers of information in order to reshape (and possibly resell) the information in new, more digestible, compact or comprehensive formats.

Many of the more successful Web portals, or sites where a person can easily access any of the available services provided by an organization, as well as other related or value-added information, are based in part on a syndication model. The more commercial of these sites bring in local weather and sports or personalize stock or news information. The information itself may come from several different original sources. For example, with respect to financial information, the portal may bring in stock charts from Big Charts, research opinions from Merrill Lynch, and breaking news from Reuters or CNN.

If personalized government portals are to provide the same level of service that exists in many of the better private-sector Web portals, they will also likely need to follow a syndication model. The appropriateness of governments acting as syndicates of content generated by nongovernmental organizations or interest groups is a question that is rarely raised in the context of print media. As such, government policy makers and managers have little experience on which to develop policies for the Internet.

In addition, the experience of governments with the appropriateness question in the print world may not be particularly applicable to the Web environment. For example, in the print world, one can easily distinguish among three ways of including outside information:
1) government publishing information that comes from outside under its own imprimatur; 2) government acceptance of an advertisement from an outside party; and
3) the inclusion of a footnote reference to a publication of an outside group within an article written by a government official.

These cases also appear in the Web environment, but in addition, in a Web environment, one can provide a link to an outside resource (with or without identifying the link as going to a nongovernmental Web site) or, one can automatically incorporate the information from outside within the framework and brand of the government. Although there are parallels to print-world situations, there are also major differences. For example, incorporation of an article created by an outside source into a government publication typically involves some individual review of the article's content; links and automated incorporation can often occur without any review of the appropriateness of the information to which the article is being linked. Even if the information is reviewed, the outside content creator, without the government's knowledge or approval, may change it the next day.

In addition, while the linked material may be found acceptable, the links that go from this material to other resources may facilitate a reader accessing content that would be objectionable to the government. The answer to the question of appropriateness of syndication is obviously one that will involve more than simply following one or more paper-world metaphors.

The respondents were asked to assess the appropriateness of governments offering hyperlinks to particular types of information related to governmental issues but that are produced by organizations outside of government. The appropriateness level was measured in terms of the frequency at which respondents felt that government-sponsored personalization services should make available outside resources of particular types.

The pattern of responses to these questions suggests that respondents' concern regarding links to outside resources and groups was strongest related to the possibility that the government might link to commercial advertising. The high level of concern (56 percent of respondents thought this should never happen) is probably based on the worry that the government's brand would be seen as something that could be bought. Linking to political party organizations was also identified by a larger percentage of respondents as an activity that should probably be avoided or prohibited. Forty-five percent of respondents thought that such links should never be made available as part of a government personalization process.

Respondents expressed a similar, but slightly lower level concern for links to other election campaign-related resources, whether these are provided by interest groups (Mean: 3.55) or the candidates themselves (Mean: 3.67).

Respondents were substantially more likely to approve the appropriateness of providing links to public forums on policy issues (Mean: 2.77), to non-partisan groups that sponsor and report what is said at candidate forums (Mean: 3.0), and, to a somewhat lesser extent, to hypertext links to interest groups that provide policy (but not election-focused) information (Mean: 3.11).

Respondents were most likely to approve linking to material provided by other governments. Nevertheless, a few respondents thought it inappropriate for their government to create hypertext links even to other governments.

The overall pattern of responses suggests that governments are very concerned about the value of their brand. This desire to protect the government by limiting the scope of information that the government will provide is probably connected with two traditional public administration norms: first, that there should be a strict line of separation between policy and administration; and second, that publicly-oriented and commercially oriented activities need to be clearly delineated and kept separate. The majority of the respondents also indicated that they would be extremely reluctant to do in the Web world what they would not do in the physical world; i.e. disrupt the classic separation between politics and administration.

These findings suggest that public administrators are likely to be extremely cautious with regard to tapping into what is widely recognized as the primary feature of Internet-based resources -- their ability to connect to ever-widening and interwoven resources from across the nation and the globe. Essentially, such caution on the part of public managers may result in the public sector remaining a second-class player in the development of Web resources. Rather than accessing the comprehensive range of government-related Web resources from governments themselves, citizens may choose to access these resources from private-sector Internet content creators and syndicates. The danger here is that by protecting its brand and traditional norms too much, governments may cede control over their information and services to outside providers who are better able to provide comprehensive, one-stop services.

An Alternative Delivery Model
In its beginnings, the Internet promised an expansion of virtual public space in a time when physical public space was being depleted by the process of private malls replacing the more public town and downtown spaces. The increased presence of commercially-oriented Web sites does not by itself undermine the value of public sector Web resources to create new public spaces -- unless the commercial Web resources replace public Web resources (e.g. due to their capacity to better integrate the delivery of desired services and information). In this regard, there may need to be a reassessment of the appropriateness of government organizations being centers of information. Traditionally, there has been an effective model for such a role -- the public library. The ethics of public librarianship would likely provide public managers with more leeway to create links from government Web sites to outside resources. Even though libraries are funded by public tax dollars, their "freedom to read" ethics, though sometimes controversial, are generally accepted by the public as appropriate provided certain practices are followed.

Given this situation, governments may want to consider either following the public library model or perhaps even housing their Internet resources under an existing public library organization. Following the public library model, public managers could loosen their self constraints regarding linking to outside resources. However, such loosening of restraints would need to be accompanied by the adoption of practices that are not yet common among government Internet units. Specifically, governments that linked to non-public-sector outside resources would need to ensure that these links were explicitly marked and labeled in ways that would provide the user with clear signals that they were leaving the government's Web site; and provide certain metadata about the outside resource (e.g. sponsored by a partisan political organization, provided by a commercial firm, or approved by a registered non-profit organization, etc.).

Who Should Control Personalization or Customization?
Personalization and customization in general refers to the ability of an Internet Web site or service to be shaped or re-shaped so as to better meet the individual needs or wants of a user.

As the terms are often used interchangeably, it is possible, and perhaps important, to make a distinction between instances where citizens, customers or clients make choices about what they will be exposed to and in cases where a government agency or service provider selects or prioritizes the information to be delivered to a user of its online information services. Some have suggested that the term "personalization" should refer to the former cases where personal choice is involved, and the term "customization" should be used to refer to agency efforts to make information or service features personally relevant to a citizen or consumer. As with most other aspects of digital technology, the choice between the two types of personalization need not be an either/or one. Rather, because these technologies are infinitely mutable, it is possible to fashion Internet services that embody some degree of both types of personalization (or in the more precise terminology, both personalization and customization).

The Appropriate Locus of Control Over Personalization
In a non-Internet world, the control over a personalized service tends to reside by default with the owners of the personalization technology. For example, one of the goals of the data-matching studies of social-service clients (i.e. studies that attempted to identify when two or more agencies were serving the same clients) was to enable the personalization of service delivery by government human service agencies. However, all of the data used for personalization and the personalization choices that were made in the social services areas where these studies were conducted were entirely under the control of the agencies themselves. In an Internet-based personalization system, governments may still own the server side of the system, but it becomes possible for citizens or clients (in both the programmatic and technical senses) to access and potentially control certain system parameters.

Allowing citizens to fully control the content of their personalized government Web portal (or one-stop government information and service delivery Web site) is an attractive option from the point of view of likely service satisfaction. Some cyberspace theorists (see: Shapiro, 1999; Lessig, 1999), however, have expressed concern about the potential for citizens making personalization choices that would severely limit their exposure to information that might challenge their current points of view or prejudices. The key policy issue in this regard is whether or to what extent governments should enable users of their Web sites to insulate themselves from issues, ideas, or data that might differ from or contrast with their current views of the world. Governments that accede to the public's wishes in this regard are likely to be more popular than governments that do not. Over the long term, however, governance may suffer in that people's prejudices are confirmed or at least never challenged.

In our survey of public managers, respondents were asked four questions related to the issue of control over personalization of government-sponsored Internet applications. Respondents were asked to indicate on a seven-point scale the degree to which they agreed with statements regarding control of personalization by citizens or governments under certain circumstances. Given the constraints of space, only a limited number of control options could be explored.

The survey attempted first to identify if there was a substantial difference between public managers' support for citizens' control over personalization of non-policy information, such as notifications for bills, jury duty and meetings, and their control over policy-related information, such as information that might outline the rationale for the government taking a particular policy position. The majority of public managers participating in the survey supported citizens being able to control the personalization of the information that would be displayed (e.g. on their version of a government Web site or Internet messaging service). However, there was a slight fall-off in the level of support for citizen control of policy-related information when compared with non-policy information.

The responses to the control issues indicate that the public managers participating in the survey for the most part support a high level of citizen-controlled personalization. However, at least some of these same respondents also appeared to contradict themselves when they demonstrated support for governments being able to present citizens with opposing points of view on policy information. It would take a number of additional questions to be able to fully understand the real meaning of the contradiction between these two sets of responses. This contradiction suggests that public managers may not have yet been challenged to fully explore the policy options related to the control of personalization services or to come up with a philosophically consistent position in this regard.

The level of support for government customization of information was, however, much less substantial than for citizen personalization. In fact, the largest group of respondents indicated that they were essentially neutral on this issue. This finding suggests that a large portion of public managers probably have yet to form a strong opinion about the degree of threat posed by relinquishing control of personalization to citizens or the appropriate response to this threat. Though public managers generally desire to remain on the administrative side of the policy-administration divide, this finding (or neutral stance on government customization of policy information) suggests that public managers may be taking a wait-and-see attitude. That is, they may want to hold in reserve the government's right to customize policy information, but only exercise this right in cases where citizen personalization of policy information appears to be reinforcing existing prejudices rather than providing a more efficient search for truth.

The fourth survey question to deal with the issue of the control over personalization is one that is focused on the potential role of government as a protector of citizens bests interests from their own desires to compromise those interests for short term gain. Since the formation of the United States, governments have enacted laws to protect citizens in cases where individual citizens would choose not to be protected. In this regard, for example, we are not allowed to enter into contracts for voluntary servitude or for the sale of our own organs.

The related question for the future of the personalized public-sector Internet is whether citizens might be too willing to trade privacy for convenience. This question can further be divided into two parts -- an empirical part that asks whether trading privacy for convenience is likely to occur and a philosophical one that asks if any damage is actually done to the public interest when citizens are too free with their private information.

With respect to the first question, studies have shown that, although some citizens are privacy fundamentalists most are privacy pragmatists. That is, they will exchange a certain amount of private information for more or better service. Furthermore, a fair proportion of citizens are believed to be privacy libertines, persons behaving in ways that show no particular concerns about personal privacy at all.

With respect to the second or "so what if they do" question, we may need to be convinced that not protecting privacy will cause some harm to a cherished public value. If we accept that privacy is a public good as well as a private one, there is probably a case for government protecting citizens from being too free with their own private information.

We gathered public managers views on one aspect of this issue: Given the potential for citizens not to fully understand the extent and depth of information in existing government databases, governments should protect citizens privacy even in cases where citizen desire to relinquish their privacy rights in return for more convenience.

The pattern of responses to this question suggest that the public managers participating in the survey are generally sensitive to the potential for citizens to be overly eager to relinquish a certain degree of privacy in return for convenience. Again, despite respondents' stated support for "full citizen control over personalization," when we probed more deeply into the issue of control, this support was found to be somewhat soft around the edges.

What to Do
Our survey results indicate that the individuals who are most likely to influence the debate about the control of personalization of Internet services, i.e. public managers, are still somewhat divided among themselves and possibly within their own minds with respect to where control of personalization should lie. Given this division, there is certainly a need to begin a dialog on these issues at all levels of government.

One of the potential dangers is that the public sector will fail to come to any decisions regarding an appropriate allocation of control of personalization services among the stakeholders involved, and will thereby limit their options for the development of personalization services. Such a self-limitation would, in turn, essentially leave even more of the future of the Internet to private-sector providers of these services.