Book excerpt: The development and widespread use of Web surveys have resulted in an outpouring of research on their design. In this volume, Tourangeau, Conrad, and Couper provide a comprehensive summary and synthesis of the literature on this increasingly popular method of data collection. The book includes new integration of the authors' work with other important research on Web surveys, including a meta-analysis of studies that compare reports on sensitive topics in Web surveys with reports collected in other modes of data collection. Adopting the total survey error framework, the book examines sampling and coverage issues, nonresponse, measurement, and the issues involved in combining modes. In addition, the concluding chapter provides a model for understanding the errors in estimates that combine data collected in more than one mode. Web surveys have several important characteristics that affect their ability to collect accurate survey data. Discussing these in detail, the authors address basic design decisions from input widgets to background colors. They additionally focus on the visual character of Web surveys, on their ability to automatically interact with respondents, and on the Web as a method of self-administration. The Science of Web Surveys is relevant for those with the practical goal of improving their own surveys and those with an interest in understanding an increasingly important method of data collection.