Personal academic webpage Shiya Wu
Bayesian Adaptive Survey Design
The increasing effort and costs required to achieve survey response have led to a stronger focus on survey data collection and the rise of adaptive survey design. Adaptive survey design, by means of monitoring before or during data collection, allow for adaption of survey design features to strata identified based on auxiliary information, in contrast to non-adaptive or uniform designs. In other words, each population stratum will receive a different treatment such that survey designs are tailored to optimize response rates. In the survey context, adaptive survey designs provide a flexible mathematical framework to obtain a tradeoff between survey quality and costs, given a specified quality as the objective and under cost and quality as constraints. The optimal strategies can be found through specialized computer programs. To support this endeavour, the dissertation research will focus on prior elicitation and design optimization in a Bayesian context, which give an opportunity to include data collection from expert knowledge and historic survey data, to reveal uncertainty at some extent, and even to formulate complicated cost and quality indicators.
On 16 June 2023 Shiya Wu defended her thesis Improving Predictions of Response Propensities for Effective Adaptive Survey Design (ASD) at the University of Utrecht.
Prof. dr. J.G. Schouten (Statistics Netherlands and M&S, UU), Dr. M. Moerbeek (M&S, UU)
26 October 2017 – June 2023