On the Measurability of Information Quality Ofer Arazy Department of Accounting and Management Information Systems, Alberta School of Business, University of Alberta, Edmonton, AB T6G 2E8 Canada. E-mail: ofer.arazy@ualberta.ca Rick Kopak School of Library, Archival and Information Studies, University of British Columbia,Vancouver, British Columbia, V6T 1Z1 Canada. E-mail: rkopak@interchange.ubc.ca The notion of information quality (IQ) has been inves- tigated extensively in recent years. Much of this research has been aimed at conceptualizing IQ and its underlying dimensions (e.g., accuracy, completeness) and at developing instruments for measuring these quality dimensions. However, less attention has been given to the measurability of IQ. The objective of this study is to explore the extent to which a set of IQ dimensions—accuracy, completeness, objectivity, and representation—lend themselves to reliable measure- ment. By reliable measurement, we refer to the degree to which independent assessors are able to agree when rating objects on these various dimensions. Our study reveals that multiple assessors tend to agree more on certain dimensions (e.g., accuracy) while finding it more difficult to agree on others (e.g., completeness). We argue that differences in measurability stem from properties inherent to the quality dimension (i.e., the availability of heuristics that make the assessment more tangible) as well as on assessors’ reliance on these cues. Implications for theory and practice are discussed. Introduction User assessment of the quality of Web-based information has received significant attention in the research-based lit- erature over the past decade. Two major reasons for this attention are (a) the phenomenal growth in the number of information sources available on the Web and (b) the highly accessible nature of this information by a diverse set of con- sumers. With the diminution of traditional gatekeeping on the “information production” side (e.g., editorial and peer- review processes), more and more of the available content is obtained from sources with mixed, and sometimes dubious, Received March 15, 2010; revised September 3, 2010; accepted September 16, 2010 © 2010 ASIS&T • Published online 5 November 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/asi.21447 provenance. A consequence of unreliable authority of sources and questionable quality of information is greater reliance on the ability of information consumers to make these quality judgments. Lankes (2008) described this as part of a larger trend toward “information self-sufficiency,” where more and more of our everyday decision making is based on receiving information that is “disintermediated.” The paradox resulting from this is that “end users are becoming more responsible for making information determinations, but because they have fewer physical cues to work with, they are becoming more dependent on the information provided to them by others.” (Lankes, 2008, p. 104). Information quality (IQ), as a concept, has been inves- tigated extensively in prior information science research, where much of the discussion has been devoted to the underlying dimensions (or attributes) of IQ, such as accu- racy, completeness, presentation, and objectivity (Hilligoss & Rieh, 2008; Lee, Strong, Kahn, & Wang, 2002; Liu, 2004; Rieh & Danielson, 2007; Wang & Strong, 1996). Largely, these investigations have focused on the salience of the various dimensions, studying whether one quality dimen- sion better represents users’ perceptions of IQ than does another dimension. These studies have shown that infor- mation consumers may perceive certain quality dimensions to be more important than are others, and for a variety of reasons, including domain expertise (Stanford, Tauber, Fogg, & Marable, 2002), gender (Flanagin & Metzger, 2003), or differences in information-seeking style (Rains & Karmikel, 2009). The objective of this study is to investigate the consistency with which different users assess dimensions of IQ, and to compare consistency levels across the various dimensions. We argue that IQ dimensions, by their nature, differ in the extent to which they lend themselves to reliable measure- ment, such that when multiple assessors (as users or readers of the information) analyze a set of information objects, the level of agreement reached will vary depending on the quality JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE AND TECHNOLOGY, 62(1):89–99, 2011