In order to gain competitive advantage through cutting edge technology, the Web survey is considered as a strategic tool in the data collection process. It offers great potential for collecting survey data in the most cost effective way, and the speed of collecting data makes it a viable option for surveys with a short field period. However, the research community has raised a number of interesting questions related to the successful implementation of web-based interview, including whether the web mode allows effective contact and follow-up procedures, and the impact of sensitive questions in break-off patterns, among other concerns. This paper will evaluate break-off patterns of respondents in the Survey of Doctorate Recipients (SDR) project from various perspectives. SDR project is sponsored by the National Science Foundation and the National Institute of Health, and the study collects data from science and engineering doctorate recipients who earned their degrees from institutions within the United States. Analysis will include an evaluation of break-off duration of complete cases in order to develop effective follow-up procedures. We will also review incomplete cases to determine if there is any specific break-off pattern in any particular section of the questionnaire. Selected variables will be used in the analysis to measure behavior across demographic profiles. Finally, the paper will introduce the concept of minimizing potential break-off through the intelligent use of userinterface and questionnaire programming techniques.