Science and survey management
Section 1. Introduction
Surveys are in trouble these days, faced with the twin dilemmas of rising costs and falling response rates (e.g., Tourangeau, 2017; Williams and Brick, 2018). Both trends have been apparent in the United States since the 1970s (Atrostic, Bates, Burt and Silberstein, 2001; Steeh, Kirgis, Cannon and Dewitt, 2001), but seem to have accelerated in the last ten years or so. The same trends hold throughout the developed world (de Leeuw and de Heer, 2002). It seems fair to say that survey researchers do not really know what hit them (although see Brick and Williams (2013), for a thoughtful exploration of the possible causes behind these trends). But it is clear that fewer and fewer people want to do surveys these days; the downward trend in response rates mainly reflects increasing resistance to surveys among members of the general public.
Partly in response to this global industry-wide crisis, researchers have taken a closer look at the impact of falling response rates on the accuracy of survey estimates and have also proposed various measures to counter declining response rates. For example, more and more surveys have begun to offer incentives, make use of advance letters, and increase the number of contact attempts they make.
But another trend has been the use of a range of methods to improve the management of surveys to reduce the potential for error, data collection costs, or both. In Section 2, we review these efforts, generally known as responsive and adaptive designs. In Section 3, we look at another method for reducing cost and increasing efficiency in face-to-face surveys. This method ‒ optimal routing ‒ involves survey managers giving field interviewers detailed instructions about which cases to try to interview and what route to follow in their next venture into the field. In Section 4, we look at another development with the potential to improve the performance of interviewers with the computer audio-recording of interviews, or CARI (Hicks, Edwards, Tourangeau, McBride, Harris-Kojetin and Moss, 2010). CARI allows central office staff the opportunity to hear how the interviewers are administering the questions in the field and make midcourse corrections in their performance. Research has shown that field interviewers depart from script more often than telephone interviewers do (Schaeffer, Dykema and Maynard, 2010; West and Blom, 2017), presumably because telephone interviewers can be monitored and given feedback in real time. In this fourth section, we describe two experiments in which central office staff provided rapid feedback to field interviewers ‒ feedback provided within two or three days of the interview. What these techniques have in common is replacing the judgment of interviewers and field staff with the evidence- based prescriptions of survey managers ‒ that is, they are attempts to replace management art with management science. Finally, Section 5 presents some conclusions.
- Date modified: