The Effects of Mixed-Mode Data Collection on Nonresponse in the Icelandic National Election Study 2021

Abstract

Nonresponse rates have increased over time in Icelandic surveys. The Icelandic National Election Study has relied on telephone responses for more than four decades, but due to a deteriorating survey climate, a mixed-mode data collection protocol was adopted in 2021. On the one hand, respondents without listed telephone numbers were offered participation through a web survey. On the other hand, individuals who refused participation over the telephone were offered the choice of completing a shorter version of the survey online. This article outlines the reasons for transitioning to a mixed-mode data collection protocol, as well as analysing fieldwork outcomes. The paper focuses on response rates, representation, and measures of key target variables (vote choice). The results indicate that transitioning to a mixed-mode protocol did not reverse the trend of decreasing response rates over time. However, the new protocol did avert an even worse outcome in terms of the number of responses gathered. Web responses resulted in a more representative response set and slightly reduced bias in estimates of a key target variable (vote choice). These results indicate the potential for reducing nonresponse errors through mixedmode data collection in Icelandic social surveys, but also highlight the need for further actions to ensure the continued quality of data in the Icelandic National Election Study.