As the web becomes more commonly used for survey data collections, many researchers are facing the decision of whether or not to move their new or ongoing research programs to this new mode. Response rates and data collection costs drive many mode decisions, and in this regard, use of the web is no different. Many web surveys have been found to obtain inadequate response rates, and the cost of building the infrastructure to conduct interactive web surveys can be prohibitive. As e-mail samples mature, and the research industry learns how to best capture the attention of the sample members (through appropriate contacts and incentives), the time has come where web surveys can enjoy the same, or better, response rates that have been received with mail or phone data collections. Likewise, as more efficient processes and systems emerge for the sole purpose of online data collections, the costs of conducting a web survey will become more manageable. This paper will provide results from a data collection effort conducted at the University of Michigan in Spring 2001. The data collection, a survey of student drug and alcohol use, was conducted as a mode experiment. Two randomly selected samples were assigned to complete a web survey, or a mail (paper) survey. Each sample consisted of 3500 undergraduate students. The methods used (number of contacts, questionnaire content, etc.) for each mode were comparable to maintain the integrity of the mode comparison. The results of this study and mode comparison demonstrate how successful a web based data collection could be. The response rate for the web mode concluded at over 20% higher than that of the mail mode. The overall costs for each mode showed that when using a web data collection facility with established infrastructure, the web data collection is more efficient.