Interactive Information Retrieval in Digital Environments

Web log analysis, like my own area of webometrics, is a relatively new Web-based, primarily quantitative set of approaches of growing importance to information scientists. With the shift in service provision to theWeb, libraries and archives need to engage with quantitative techniques to assess the impact and effectiveness of their online resources and information. Moreover, information scientists need to researchWeb use as an important aspect of information behavior, andWeb log analysis should be a part of this. Hence, the release of this large volumewithmultiple log file analysis methods and case studies is particularly welcome to support new teaching in the subject. While the use of Web server log file analysis programs is covered in the professional literature— often described as Web analytics—this volume gives a researcher’s perspective in terms of both methods and case studies. The heart of the book covers Web server log file analysis from individual Web sites, but transaction logs based on user actions across multiple Web sites (recorded by tracking software on PCs or by Internet service providers), and the analysis of search engine query logs also are included. The book is explicitly research-oriented, in contrast to commercially orientedWeb analytics manuals. Nevertheless, several chapters are written or coauthored by commercial experts, which gives a welcome additional perspective and emphasizes the wider relevance of the book, despite its research orientation. This reference work is split into five main sections. The first section (Perspective, Issues, and Directions) introduces the key ideas and generic issues such as ethics. Chapter I sets a research context for “transaction log analysis,” drawing on the concept of behaviorism. It particularly emphasizes the advantages of unobtrusiveness, for example, in avoiding the Hawthorne Effect. In contrast, chap. II gives an overview of Web log analysis from a historical perspective, drawing upon pre-Web transaction log analysis. This chapter gives insights into methods and methodological issues. Chapter IV discusses the value of surveys as a complement to Web log analysis. This is a useful addition, although its study does not really engage in a detailed analysis of the kind of extra information that surveys can deliver. Chapter IV makes an intelligent case for limitations for the use ofWeb log data in research. It mainly addresses the commercial Web analytics literature, so the chapter is actually not primarily tackling the academicWeb log analysis that is the main theme of the book. In addition, I am not sure that “the unspoken belief within web-traffic measurement is that those keystrokes and mouse clicks represent the sum total of what there is to know about a web site visitor’s experience” (p. 70) is fair and think that the mostly unspoken assumption is probably that keystrokes and mouse clicks provide sufficient evidence on their own to make intelligent decisions for Web site design optimization purposes and perhaps also for Web marketing. Nevertheless, this chapter is valuable context for all academic Web log file analysis researchers, and should serve as a reminder for commercial practitioners and researchers that Web log files do not tell the full story of the user experience. This is well captured by the chapter’s final paragraph. Concluding section I, chap. V surveys ethical and