Due to cost and coverage challenges, many research organizations are now conducting surveys via the web as part of mixed-mode designs. The increasing use of the web in combination with modes such as telephone and face-to-face raises several important research questions. Are there differences in respondents who opt to complete the survey in each mode? Are there differences in response quality across modes? A recent AP- NORC Center survey of how 12 neighborhoods in New York and New Jersey are recovering from Superstorm Sandy provides an opportunity to explore such questions about mixed-mode surveys. All households in the address-based sample were first sent an advance letter directing them to a URL to complete the survey. After a two-week delay, cases with matched landline telephone numbers were dialed, and a sample of cases without matched telephone numbers were contacted by field interviewers. Reminder mailings were sent to a randomly selected set of non-respondents with $2 or $5 pre- paid incentives. Of the 1,009 completed interviews, 316 were via phone interviews, 393 were via face-to-face interviews, and 300 were via web self-administered questionnaires. Although respondents were not randomly assigned to a mode, the surveys were identical across mode and all of the responses were collected between June 28 and September 9, 2014. The analysis shows that there are clear demographic differences between respondents who opted to complete the survey via the web than those who completed the survey with phone or field interviewers. Measures of data quality such as breakoffs, item nonresponse and straight -lining are relatively similar across the three modes. However, the survey results show that the web can reduce social desirability bias among respondents. The findings from this study provide researchers with insights into the potential benefits and challenges of including web surveys in mixed-mode designs.