The Effect of Topic Salience on Nonresponse, Self‐Reported Turnout and Party Choice: An Experimental Study from the Swiss Election Study (Selects)

Abstract

Political surveys are subject to nonresponse bias, leading to overestimated voter turnout and, at times, underestimated support for far‐right parties. This study examines whether altering the salience of a survey's topic at the recruitment stage affects survey participation and nonresponse bias, and with what trade‐offs. Using the panel survey of the Swiss Election Study (Selects 2023), respondents were randomly assigned to a high‐salience condition emphasizing federal elections or a low‐salience condition referring more broadly to societal and political issues. Low‐salience framing modestly altered the composition of respondents in the first wave, increasing the share of non‐voters and Swiss People's Party (SVP) voters, but also led to higher dropout in the second wave. Recruitment framing thus shapes who enters and remains in the survey: while it broadens initial participation, it reduces retention over time. Topic framing is therefore a relevant but constrained tool for addressing nonresponse bias in political surveys.