Mobile- and landline-onlys in dual-frame-approaches: Effects on sample quality

Abstract

Introduction: Mass media, politics, and commercial enterprises are increasingly using social survey data to substantiate their reasoning and interventions. Although the number of telephone surveys is decreasing more than one third of all studies of market and social research in Germany is conducted on the telephone (ADM 2011). This raises questions about the quality of the gathered data. One of the central variables to look at is the quality of the realised sample. Reflecting the number of sample units interviewed as a percentage of the number of eligible sample units, response rates are considered reliable indicators of data quality. However, irrespective of the mode declining response rates question this assumption (Schnell 1997, de Heer 1999, Däubler 2002, de Leeuw and de Heer 2002, Curtin et al. 2005, Stoop 2005). In addition to the number of ineligible cases (including people not part of the target population or unable to participate) the quality of surveys of the general population suffers from nonrespondents. A growing number of contacted persons refuses participation in market and scientific surveys. Numerous studies have examined different aspects of this phenomenon in detail and offered solutions to the problem (Blasius and Reuband 1995, Hüfken 2000, de Leeuw and Hox 2004, Meier et al. 2005, de Leeuw et al. 2007, Schnauber and Daschmann 2008).