|
|
|
-- Blackmon, M.H., Mandalia, D.R., Polson, P.G., & Kitajima, M. (2007)Blackmon, M.H., Mandalia, D.R., Polson, P.G., & Kitajima, M. (2007). Automating Usability Evaluation: Cognitive Walkthrough for the Web Puts LSA toWork on Real-World HCI Design Problems. In Handbook of Latent Semantic Analysis (University of Colorado Institute of Cognitive Science Series) (Thomas K. Landauer, Danielle S. McNamara, Simon Dennis, Walter Kintsch, Eds.), Chapter 18, pp. 345-375. Lawrence Erlbaum Associates (January 30, 2007).
Automating Usability Evaluation: Cognitive Walkthrough for the Web Puts LSA toWork on Real-World HCI Design ProblemsWhen people navigate a relevant Web site for information needed to solve a problem, they encounter two subproblems. The first is navigating within the Web site to find a relevant web page(s) or a relevant document downloadable from the Web site. The second subproblem is comprehending the retrieved information. For the past 5 years, we have addressed the first subproblem by developing a usability evaluation method (UEM) called the cognitive walkthrough for the Web (CWW). This chapter focuses primarily on how CWW employs LSA to identify and repair usability problems that impair navigation of large, complex Web sites. By now we have collected a large amount of evidence demonstrating that CWW reliably and validly predicts the usability problems that impede navigation of a Web site to retrieve information (Blackmon et al., 2002, 2003, 2005). The special genius of LSA is its versatility to switch among a variety of different semantic spaces. Each LSA semantic space used by CWW is constructed from a scientifically sampled corpus of documents, emulating Zeno, Ivens, Millard, and Duvvuri (1995). Scientifically sampled corpora ensure that the semantic space faithfully represents the background knowledge and general reading ability of a particular population, such as 3rd-, 6th-, 9th-, or 12th-grade general reading knowledge of American English. This versatility of LSA enables us to apply CWW to evaluate the usability of a given Web site for a diverse array of user populations. For example,CWW might simulate navigating a one-size-fits-all online encyclopedia to answer a range of information needs, predicting that users with college-level general reading knowledge would be successful, but that users with 3rd- or 6th-grade reading knowledge would experience frustration and probably fail to find the information they needed in the same online encyclopedia. Recently, one of us (Mandalia, 2004) extended the research program in a new direction by addressing the second subproblem users face: whether users can comprehend the content that they find. Mandalia created a new LSA-based tool that integrates with the workflow of content writers. A section near the end of this chapter describes how this tool supports content writers in improving the comprehensibility of content material made available on the Web site, and targeting one or more particular user groups whose level of background knowledge differs markedly from the writers' own level of background knowledge.
|