Humboldt-Universität zu Berlin - Department of Social Sciences - Berlin Graduate School of Social Science

Cohort

BGSS Generation 2024

Title

Novel Statistical Approaches to Predicting Response Patterns in Mixed-Mode Longitudinal Household Surveys

Supervisor

Prof. Dr. habil. Sabine Zinn (Sozialwissenschaftliche Methoden mit dem Schwerpunkt Survey-Methodik, Humboldt-Universität zu Berlin)

Abstract

As long as there are surveys, there has been unit nonresponse. If the nonresponse is selective, biased estimates can be the result. My research is aimed at quantifying and preventing unit nonresponse in household panel studies given an ever-changing survey landscape and changing circumstances. It is particularly central to the conflict of cost versus data quality; ever so present. The proposed work covers the modeling of nonresponse with novel methods, as well as a predictive survey mode allocation to households, in order to prevent attrition. The project aims to model interactions between specific subsamples and field work, like survey mode effects, after data collection while also optimizing data collection and cost by providing predictions for survey mode preferences within households before the data is being collected.

Institution

German Institute for Economic Research (DIW)

(personal website: https://www.felixsuettmann.com/)