How low can we go? Declining survey response rates to new zealand electoral roll mail surveys over three decades

Abstract

ABSTRACT Quality survey research relies on the willingness of the public to participate, but survey response rates are decreasing over time. This research note explores response rate data from several national probability sample mail surveys drawn from a consistent sampling frame (the New Zealand electoral roll) over the past 30 years. The aim of this work is to document the scale of the response rate decline problem in mail survey studies, predict future response rates, and track nonresponse bias through comparing self-reported voter turnout in surveys to official turnout. We use data from the New Zealand Election Study, the International Social Survey Programme, the New Zealand edition of the World Values Survey, and the New Zealand Attitudes and Values Study. Our findings demonstrate that there has been a consistent downward trend in response rates and an increase in nonresponse bias for self-reported voter turnout. Response rates in the early 1990s were as high as 72%, before dropping to the 60% range around the year 2000, with response rates effectively halving from their early 1990s rates by 2015. We comment on the implications for future surveys, voter enrolment, local government turnout, and political polling. Ultimately, these results leave us with the question of how long traditional mail survey research will be able to continue.