As computer access becomes increasingly prevalent, web surveys have become increasingly popular. The primary advantage to using the Internet over more traditional modes of data collection is that web surveys eliminate the need for an interviewer or data-enterer and, consequently, reduce costs and potential for error. However, not everyone has internet access and levels of computer literacy vary, so important subsets of the population may be excluded if an internet survey is the exclusive mode of data collection. The Health Care Survey of Department of Defense Beneficiaries, traditionally a mail survey, recently gave respondents the option of completing the survey over the Internet. This paper describes the characteristics of beneficiaries who chose to respond online and compares them to those of beneficiaries who responded by mail to determine if web respondents differ from mail respondents and whether particular subsets of the population prefer to respond online. Differences in key estimates are also examined as an indicator of whether response bias may be introduced in an estimate due to the mode of administration effect.