The increasing popularity of online surveys in the social sciences led to an ongoing discussion about mode effects in survey research. The following article tests if commonly discussed mode-effects (e.g. sample differences, data quality; item-non response, social desirability and open-ended question) can indeed be reproduced in a non-experimental mixed-mode study. Using data from two non-full-probabilityrandom samples, collected via an online and face-to-face survey concerning itself with opinions on migration and refugees, most assumptions found in experimental literature can indeed be replicated via research data. Thus, the mode effects need to be accounted for if the usage of mixed-mode designs is necessary, especially if online surveys are involved.

Mixed probability and non-probability

Face-to-face interview: PAPI; Self-administered questionnaire: Web-based