A growing challenge in survey research is completing surveys with younger adults. Increased cell phone use in younger cohorts has made typical RDD telephone surveys increasingly ineffective. Researchers turn to other types of samples to help with coverage and response rate issues, often using selfadministered surveys (web or mail). Each of these has their own pros and cons. For example, web surveys tend to be cheaper to conduct, have shorter field periods, provide greater flexibility with questionnaire design, and may be associated with lower levels of item-missing data and higher reports of threatening behaviors. Their limitations include coverage and response rates. In the current study we evaluate the effects of a sequential, web-mail mixed-mode design on a variety of
outcomes. A total of 7,200 adults between the ages of 21 and 35 were randomly sampled from driver‘s license records in eight counties in Wisconsin and asked to complete a questionnaire about attitudes, beliefs, and behaviors about drinking alcohol. Our experiment explores how different modes and preincentives influence data quality. For the mixed-mode experiment, sample members were randomly assigned to web-mail or mail-web treatments. Respondents initially received either a postal letter containing a URL link to the survey (web-mail) or a mailed survey (mail-web) along with a pre-incentive followed by two additional contacts. For those that did not respond, the mode was then switched for two more contacts. Within each mode, respondents were randomly to 1 or 2 dollar pre-incentive condition. Our analysis examines the effects of mode and incentives on response rates, costs, survey reports about drinking behaviors, item-missing data, and nonresponse bias in a sample of young adults assigned to the four experimental groups described above (e.g., web-mail with $2, web-mail with $1, mail-web with $2, and mail-web with $1). Preliminary analyses indicate web-mail condition fared better than anticipated.