Keynote Speech by Dr. Roger Tourangeau
Nonresponse Bias: Three Paradoxes
A number of papers have shown (or claim to show) that nonresponse bias is not a function of the nonresponse rate for a survey, even though the mathematics of nonresponse is clear—the rate of nonresponse should be related to the level of nonresponse bias. Thus, the first paradox is why nonresponse rates predict nonresponse bias so poorly. I will begin my talk by examining several theories about why the relation between nonresponse rate and nonresponse biases is apparently so weak.
Because of this weak relation, several alternatives to the nonresponse rate have been proposed as measures of survey quality. The meta-analysis by Groves and Peytcheva, however, indicates that most of the variance in nonresponse bias is within-survey rather than across surveys. This suggests that no single number can predict the average level of nonresponse bias for a survey. Thus, the second paradox is that no survey-level measure is likely to be useful for assessing the overall quality of a data collection effort or managing the field work.
Finally, many researchers have proposed that adaptive designs can be used to improve data quality. However, given the first two paradoxes, it is not clear how well researchers can redirect efforts to improve the quality of a survey. With this in mind, I will review both empirical efforts and simulation studies to evaluate responsive and adaptive designs, revealing that the gains are quite small or nonexistent.
Special Session
Addressing Nonresponse Bias during Fieldwork
Dr. Tourangeau’s talk motivates the topic of our special session on the successes and challenges of addressing nonresponse bias during fieldwork. Authors are encouraged to share their analyses and theories supporting or refuting any of the three paradoxes. Papers contributing to this session could: quantify the relationship between response rates or alternate measures of survey quality and nonresponse biases, present innovative field strategies to address nonresponse error and the results of those efforts, or highlight responsive and adaptive design techniques. Analyses may result from empirical work or simulations but we also encourage theoretical papers intended to stimulate discussion and ideas for future work.
