Survey Representativeness Among Multiple Modes of Administration Using Random Assignment

The purpose of this study was to assess the extent to which demographic and occupational characteristics of individuals influenced response rates and completion rates to a survey administered by three different modalities. We used an experimental design in which employee workgroups were randomly assigned to one of three response modes: paper-and-pencil, internet, and interactive voice response (IVR). Overall response rates by mode were 40.5%, 33.6% and 37.3%, respectively. African American and Asian racial groups had lower response rates when assigned to survey modes other than paper-and-pencil, as did staff with direct customer contact and hourly occupations. These findings suggest that some groups within organizations may be in danger of being under-represented using survey modes of administration other than paper-andpencil.