Educational researchers and policy makers have come to rely on data from sample surveys. However, survey research on educational issues poses some special challenges. In many respects, the survey methodology issues in educational research are the same as those throughout the social and behavioral sciences. These issues concern obtaining the best quality data for the lowest cost, whether quality pertains to representation or measurement. A major factor in survey quality and cost is the mode of data collection. In this article, the author discusses recent work on web surveys which is an important, emerging mode of data collection. He focuses on measurement issues in web surveys, in particular how the interactive character of the web can be exploited to promote better and more uniform understanding of survey questions, and can promote completion of questionnaires once respondents begin them. He explores three types of interactivity of web surveys that can be implemented with available technology requiring relatively simple programming. These include: (1) the impact of different forms of survey completion progress indicators; (2) the effectiveness of various forms of respondent-initiated definitions; and (3) the notion of a respondent's inactivity that prompts an action by the survey itself. (Contains 2 figures.)