Design Research in Digital Government: A Query Prosthesis for Voters Scott P. Robertson Information and Computer Sciences Dept. University of Hawai'i at Mānoa +1 808 956-2023 scott.robertson@hawaii.edu ABSTRACT The internet is an increasingly important source of information about political candidates and issues. Voters and candidates alike are taking advantage of a wide variety of internet tools. Many internet users continue to be unskilled at the use of search tools and find a restricted subset of available information. A simple assistive technology was developed to help voters formulate queries. This “query prosthesis” changed the search and browsing patterns of participants in a mock voting study and provides guidelines for more comprehensive interfaces to help voters. An argument is made for greater application of design science in digital government research. Categories and Subject Descriptors H.5.2 [Information Interfaces and Presentation]: User Interfaces – interaction styles, prototyping, user-centered design. K.4.1 [Computers and Society]: Public policy issues. General Terms Design, Experimentation, Human Factors. Keywords Digital government, electronic voting, internet and politics. 1. INTRODUCTION Americans continue to increase their use of the internet to obtain information about politics and make decisions about how to vote. Regular surveys by the PEW Internet and American Life Project have shown continuous increases in use of the internet for a range of political purposes over several election cycles [16,23,24]. In 2006, Americans reported using a variety of online sources to obtain information about political campaigns including news portals, TV network websites, local and international news websites, newspaper websites, state and local government websites, blogs, issue-oriented websites, news satire websites, and listservs [23]. This year, “campaign internet users” [23] will undoubtedly also be relying heavily on social networking websites such as MySpace and Facebook and participating in virtual worlds such as Second Life. Candidates are now campaigning vigorously in these venues and news organizations have begun to watch them carefully. While there are many sources that people report using on the internet, it is not clear how people find those sources or how broadly individuals sample from multiple sources. News portals are designed to guide users to information, but it is carefully selected and filtered information. Search tools are undoubtedly often used, and have no bias by design, however, the usefulness of search tools is constrained by the ability of users to formulate good queries, select relevant results, and sample broadly from the result set. Studies of search behavior in other domains suggest that people are not particularly good at any of these tasks [15,32,34,35]. Robertson and his colleagues [27,28,29,30,31] have been studying browsing behavior of voters in mock voting situations with emphasis on their use of search tools, query formulation, and browsing patterns. This paper discusses how a simple design enhancement to the interface of a search tool profoundly changed the information available to voters and the searching and browsing behavior of voters. This paper amplifies data reported in [30], adds new data, and discusses the experiment in the context of the paradigm of design research. 1.1 Design Research Design research has not been common in the area of digital government although it is widely practiced in computer science and human-computer interaction (HCI) [3,4,9,22,33]. Design research [33] focuses on the interface between “inner” and “outer” environments, in this case between voter cognition and behavior and the digital information available about candidates in an election. Manipulation of the artifacts that mediate interaction between these two environments (i.e. design) reveals something about the nature of the environments (science) and yields design insights as well (engineering). Design research, in one way or the other, underlies a considerable amount of activity in the area of human-computer interaction (HCI) [3]. Much of the research on human behavior in the area of digital government has involved survey work or experiments on how people use information tools such as television, newspapers, friend networks, and the internet. Design research in this area would involve creation or modification of an information tool and subsequent analysis of the impact of the new artifact on behavior and cognition. In a series of studies [27,28,29,30,31], Robertson and his colleagues have argued for the importance of “voter-centered design” and described studies of voters on the internet (see also [10]). “Design research” in this area involves the incremental and iterative development of a portal that helps voters to find The Proceedings of the 9th Annual International Digital Government Research Conference 73