Surveys employing an Address-Based Sampling (ABS) design have the luxury of contacting and recruiting respondents via multiple modes of administration. Given the increasing popularity of ABS surveys, it is time to start research on potential mode differences in a mixed-mode ABS design. Empirical questions regarding what types of respondents are recruited by which modes are not as yet answered. This paper makes use of data from the 2010 Census Integrated Communications Program Evaluation (CICPE) to examine the differences in demographic characteristics and cooperativeness of respondents recruited by different modes. The CICPE was designed as a mixed-mode study using an ABS sample. Selected sample addresses with matched phone numbers were initially fielded as telephone cases; nonparticipants in the phone survey and sample without matched phone numbers were approached to complete an in-person interview. In later waves, panel respondents were offered self-administered paper questionnaire (SAQ) and web completion options in addition to the telephone and in-person options. This mixed-mode design provides a great opportunity to study mode differences. We will compare distributions of demographic characteristics and measures of cooperativeness by modes of interview (telephone, in-person, SAQ/web) to see if there were indeed differences by mode. In particular, CIPCE oversampled certain rare subpopulations (such as Native Hawaiian and Other Pacific Islander and American Indians and Alaska Natives). We will examine the extent to which these rare subpopulations were recruited by different modes.