The Effects of Survey Mode and Sampling in Belgian Election Studies

Ce contenu n'est pas disponible dans la langue sélectionnée.


National probability election surveys are more and more abandoned. Decreasing response rates and the escalating costs of face-to-face and telephone interviews have strengthened election scholars’ reliance on non-probability internet samples to conduct election surveys online. In a number of countries, experiments with alternative
ways of recruiting respondents and different interview modes have been well documented. For other countries, however, substantially less is known about the consequences of relying on non-probability internet panels. In this paper, we investigate the effects of survey mode and sampling method in the Belgian context. This is a particularly
important and relevant case study because election researchers in Belgium can draw a sample of voters directly from the National Register. In line with previous studies, we find important differences in the marginal distributions of variables measured in the two surveys. When considering vote choice models and the inferences that scholars would draw, in contrast, we find minor differences.



Election study; Belgium; survey mode effects; representativeness; non-probability


Supplementary materials can be accessed here.

Ce contenu a été mis à jour le 3 octobre 2018 à 11 h 19 min.