Abstract:
|
Collecting information from sampled units over the Internet or by mail is much more cost-efficient than conducting interviews. These methods make self-enumeration an attractive data collection method for surveys and censuses. However, self-enumeration data collection can produce low response rates compared to interviews. To increase response rates, nonrespondents are subject to follow-up treatments, which influence the resulting probability of response. Because response occurrence is intrinsically conditional, we primarily record response occurrence in discrete intervals, and we characterize the probability of response by a discrete time hazard. This approach facilitates examining when a response is most likely to occur and how the probability of responding varies. Because response rates are presumed to be low, a widely used approach is to consider a second-phase of data collection, where only sub-sampled nonrespondents are followed-up. However, in practice, data collection from self-enumeration and from follow-ups are done in parallel, which makes sub-sampling from nonrespondents difficult to apply. In this case, excluding late self-enumeration responses ? not obtained from the follow-up subsample after follow-up has been started ? is recommended in the literature to avoid a nonresponse bias. Finally, we study the estimator of the finite population total that use all observed responses. Simulation results on the performance of the proposed estimators are also presented.
|