ABSTRACT
This chapter explores the question whether it is beneficial to adjust by design, when adjustment during estimation can also be performed, employing the same variables. It limits the focus to nonresponse, since it is not standard practice to adjust for systematic measurement bias. The chapter presents some empirical evidence for the efficacy of adaptive designs to decrease bias and discusses theoretical reasons why they may be able to do so. It also discusses theoretical conditions for bias reduction after nonresponse adjustment. The chapter focuses on inference and discusses nonresponse adjustment in the wake of adaptive designs. The associations between auxiliary variables and survey variables can be estimated with the highest precision, once all response data become available. Random strategy allocation and paradata that are subject to variation may lead to larger standard errors for estimators that need to be accounted for in the adjustment and when making inferences.
