The relationship between response rates, sampling method and selection bias Evidence from the European Social Survey
Several studies have shown that, contrary to most researchers' expectations, high response rates are not correlated with low bias in survey data. In this paper we show that that relationship, between response rates and bias, is moderated by the type of sampling method used. When interviewers are involved in selected the sample of households for the survey, high response rates can in fact be a sign of high bias. We suggest that this relationship is due to interviewers' incentives to select households with high response propensities.
Eckman, S., & Koch, A. (2018). The relationship between response rates, sampling method and selection bias: Evidence from the European Social Survey. Public Opinion Quarterly.