This work is licensed under a Creative Commons Attribution 4.0 International License.
(Also available in a Spanish translation by James Galea.)
Projects that strive to improve the availability of government information and increase public participation in government must deal with the challenges listed here. This list delves underneath the short-term legal, organizational, and cultural barriers that every agency deals with—and which are certainly not to be trivialized—to look at more fundamental dilemmas preventing the implementation of desirable goals.
For instance, the difficulty of persuading government workers to reveal early discussions of policies is often ascribed to a culture of insular decision making in which the most common term for “public revelation” is “leak.” But there are also legitimate sources of resistance to revealing early policy decisions. Remedies must face the deeper challenge listed here as stakeholders and must respect the problems of timing and groupthink.Projects involving participation must avoid frustrating the public by framing questions unambiguously, clarifying the subject matter boundaries, and describing how the agency will be responsive and produce results from comments. Even when requests for comments try to establish reasonable expectations, many members of the public come without knowing the constraints (financial, political, informational, and other) that agencies face.
Public feedback must be solicited at the points where it can do the most good, a delicate decision. At very early stages of a project, the questions being asked are too broad to solicit useful feedback from most members of the public, and the range of answers will be too diffuse to extract meaningful trends. At late stages of a project, crucial decisions have already been locked in. The Administrative Procedure Act assures the public some time to comment on any decision—but at what point do assumptions and definitions become an actual decision?
Strict adherence to formal procedures impinges on creativity and generally slows down change. It also, however, protects underrepresented populations and keeps policy-makers from impetuously driving off the tracks. Although procedures must change, they should do so in ways that respect the checks and balances they were instituted to protect; the goal should be to strengthen due process and make implementation easier, not to weaken it.
Research on the wisdom of crowds shows that the crowd is not so wise when a few people’s early decisions strongly influence those who come later. Along these lines, many agencies are reluctant to reveal early discussions because then may unduly influence, or be thought to influence, the thoughts of public advisors. Similarly, passing arguments from the agencies to the legislative and judicial branches can be inimical to the separation of powers between government branches, because the legislative and judicial branches are responsible for correcting the mistakes of the administrative branch and therefore must be somewhat independent in their thinking. On the flip side of the coin, some agencies lapse in their internal information sharing, which is important to maintain institutional memory.
A universal taxonomy is the rainbow toward which knowledge management experts are eternally peddling. Ideally, every agency would have a single way to determine a person’s income for entitlement programs, a single moniker for a function performed by different agencies in different jurisdictions, and so forth. Given that life is more complex, how much effort should be invested in creating taxonomies like XBRL? How much can different models be tied together?
It’s hard to know how much information is enough. In military and disaster situations, real-time access to imperfect data is preferable to more precise data that arrives late. But even in an office environment, the quest for clean and complete data might not be worth the effort. Furthermore, collecting data from the public must contend with inconsistent and inaccurate reports.
See for instance the discussion of airline on-time data in Open Government Initiative Discussion Phase: Transparency Principles.
Enormous systems are hard to digitize and program formally. This dilemma has dominated the software engineering field from the beginning, but it is particular relevant to the sizes of systems used in government. The usual solution is a “gentle slope” of gradual conversion to the new system. But this can introduce irksome difficulties when the public or staff have to move back and forth between old and new systems. Furthermore, the old system must be available in case the new one fails.
See for instance Summary of a Workshop on Information Technology Research For Federal Statistics, 2000, published by the National Academies Press, pp. 34-40.
Proponents of data sharing and public engagement argue that a prudent investment of staff time and other resources can pay for itself via the resulting efficiencies and innovations. But the costs in technology and staff time are immediately evident, whereas the savings are nearly impossible to measure, particularly because successful engagement changes organizational structures and decision-making systems to the point where no simple mapping exists between resources spent and resources saved.
Personal data aggregated by the government is often anonymized before being offered to the public. Experiments have shown, though, that data mining can identify individuals through the intersection of supposedly anonymous, aggregated data. Furthermore, some organizations provide information under strict confidentiality, some data is sensitive when related to crime and security, some people will simply withhold data when they feel their privacy isn’t guaranteed, and many government agencies have existing rules concerning privileged information. How can privacy be preserved as data is aggregated? Can copyrighted data, provided by a company under a confidentiality agreement for narrow use, be added to databases for public data mining?
See for instance Building the Virtual State (2001) by Jane Fountain: the discussion of US Business Advisor, pp. 141-155, and the discussion of Maneuver Control System, Version 2.0, pp. 167-192.
Online forums encourage the unabashed sharing of early thoughts and ill-formed judgments for the purposes of soliciting feedback. Good forums design in mechanisms to prevent these ungaurded statements from being misquoted and misused. However, the statements can severely hurt the speaker’s and employer’s goals if taken out of context, which can easily happen if they are quoted outside the forum.
Collaboration and participation require two to tango (and a lot more, in many organizations). If an agency puts in place a data source that no one uses, of sets up communication channels that are ignore, the effort is wasted. Each party to changes is likely to wait for another to move first. Furthermore, it’s hard to know which parts of each initiative should be funded or staffed by each party, and it’s likely that none of the parties will commit enough funds or personnel.
Although some governments desire public contributions of information and ideas, many contributors want to keep rights to their contributions. Should everything contributed enter the public domain, and will some contributors withhold useful information under such rules? Can contributors be recognized individually for their contributions? The complexity of these question is compounded by the choice some contributors might make to serve up information of value on separate, private sites for the purposes of rapid updating, commercial services, or other considerations. Government staff are also evaluated on the basis of contributions that are easier to measure in the form of stand-alone reports than collaborative undertakings.
Public forums, online or offline, attempt to combat one of the perennial failings of democracy: inadequate involvement by unorganized or resource-poor constituencies. But these forums call require additional safeguards to address two other common problems: domination by well-organized and persistent special interest groups, and the complementary tyranny of the majority against groups that would be disproportionately hurt by adoption of a policy.
Generally accepted research shows that people are most effective working in local groups bound by multiple ties at different levels. This is even true in legislative bodies, where representatives who come to know each other personally cross ideological boundaries to find common causes (probably the only real antidote to partisanship). However, online participation tends to flatten relationships and bring together people who lack the neighborhood ties. Can online participation work in the long run? Can online forums be designed to replicate the bonds that people form naturally when they are colocated physically? Ideally, the communication resources of online media could be used actually to strengthen those bonds.
Due process requires record-keeping and notice of contacts between government workers and outsiders that are related to decision-making proceedings. How broadly should such regulations be applied in an age of weblogs and comment boxes? Tools can automate and streamline the notification process, but each communication still entails overhead. And we can’t compare the influence of a citizen meeting a staff person face-to-face with a citizen who logs in to Regulations.gov and posts a comment that may be overlooked.
Laws, regulations, jurisdictions, and interagency ties have become intractably complicated. The reason for the importance of lobbyists (a widely and unfairly derided group) comes from their being the only people to understand this complexity. If the public is to be effective participants in decision-making, how can they untangle the complexity?
Long-standing laws and traditions have granted authority to elected officials, and the functionaries they appoint, to represented the public. Although the Federal Advisory Committee Act regulates the formation of advisory committees and logistics such as expensing and open records, no tradition grants any authority to public forums that gather to advise or pressure the elected government. Elected officials and functionaries may well disagree with and overrule the consensus reached in a public forum, which is just as much at risk of falling prey to the passions of the moment and capture by special interests. When people feel their organizing and lobbying don’t produce the results they want, was it because they were not effective enough at organizing? Was their proposal just plain bad? Or were the government officials out of touch? The forums’ recourses now are the familiar ones: public pressure and the hope of changing governments through elections.
Raw data yields no insights; its interpretation must be informed by the goals and parameters under which the data was collected and processed. This is why hospitals and doctors warn against publishing data about their performance; they fear that patients can’t judge it make accurately without taking into account such things as the severity of medical conditions and the expected outcomes. Data correlations can also produce invalid results because of small, unnoticed inconsistencies. How can the release of data include guidelines and procedures for valid interpretation, without reducing public inventiveness? Can open, public commentary on the data be self-correcting?
People like to have rights and entitlements without responsibilities. Constituencies sometimes demand action and then undermine it by refusing to play their part in the solution. How can increased public participation in decision-making be tied to commitments to respect the decisions and help implement them?
My thanks to people who have suggested changes to these descriptions, especially Robert Cannon (Senior Counsel for Internet Law in the Federal Communication Commission’s Office of Strategic Planning and Policy Analysis), and Mark Drapeau (government consultant and co-founder of the Government 2.0 Club).
Andy Oram (home page)
Other articles in chronological order
Index to other articles
Originally posted March 1 2009; last updated June 18, 2009
This work is licensed under a Creative Commons Attribution 4.0 International License.