Workshop Synthesis: Comparing and Combining Survey Modes

Abstract

This paper summarizes the discussions held during an in-depth six-hour workshop on the challenges of combining data from different survey modes with the anticipated aim of identifying current research needs. The main theme of the workshop was mixing survey modes as a way to meet the challenge of low response rates. However, the use of multi-mode surveys introduces new sources of bias: not all households have access to certain survey media (coverage bias); the response rate using one or another of the survey modes is correlated with social demographics (non-response bias); the sampling frame is dependent on the mode (sampling bias) or the instrument itself may affect the responses (measurement bias). The aim of this report is present the workshop's discussion on the identification of research needs with related to combining data from different survey modes.