Almost all libraries and information centres now provide some sort of access to networked information sources and services. There is probably no other title on quite this subject, so this book fills a distinct hole in the literature and will be gratefully received for that reason. Would that it could take us to the end of the road, but at the very least it provides some signposts to help us along the way.
A common theme throughout the book is that of convincing managers that evaluation of networked resources is a necessity. The problem lies in the rather amorphous common understanding of “evaluation” in the library context. Is it measurement, is it systems analysis, or is it something else? For that reason, I found the last chapter by Clifford Lynch valuable, for here he discusses “evaluating evaluation”. He makes the point that some intuitive judgement has its place, though as most grants now require an evaluation component then we have to produce something rather more tangible than our own opinions. Mostly, what we have to do is engage our customers (individuals and groups) in the design and development process, then have some sort of measurements in place to assess if we have achieved our customers’ expectations. Even so, that still may not satisfy some of the most quantitative‐minded members of funding panels.
For those who wish to evaluate electronic services by using an online customer survey (via the Web or e‐mail), a genuine problem is how to ensure they get a valid sample. Jonathan Lazar and Jennifer Preece have written a very useful guide to the conduct of electronic surveys “from idea to implementation” in which they cover the territory thoroughly. Naturally, as they are advocating electronic surveys, they have emphasised the advantages over the disadvantages, but experienced researchers should be able to choose what is appropriate. The most useful part of this chapter is the advice on sampling. Lazar and Preece provide examples of studies done with well‐defined populations (relatively easy) and on populations that are not well defined. For the latter they suggest three techniques: including demographic questions in the survey that facilitate an analysis of the respondents, examining the domain of the responses, and the enforcement of response diversity. None is perfect, but each is a part‐solution to the sampling problem.
Another very practical chapter comes from Jeffrey Rubin, on using log analysis techniques to evaluate networked services. Indeed, logs are one of the few sources of data that we have at the moment, so using them is obvious. How to make sense of the mass of data in the logs is a different matter, and I feel rather better able to make sense of them as a result of reading Rubin’s chapter. Each chapter has an extensive bibliography, and there is a full cumulated index. This is an essential purchase for all large libraries.
