Dealing with various flavors of missing data in ex-post survey harmonization and beyond

PDF
Dissertation_Haensch_Madoc_Version.pdf - Published
Download (2MB)

URL:	https://madoc.bib.uni-mannheim.de/59940
URN:	urn:nbn:de:bsz:180-madoc-599404
Document Type:	Doctoral dissertation
Year of publication:	2021
Place of publication:	Mannheim
University:	Universität Mannheim
Evaluator:	Keusch, Florian
Date of oral examination:	15 June 2021
Publication language:	English
Institution:	School of Social Sciences > Statistik u. Sozialwiss. Methodenlehre (Juniorprofessur) (Keusch 2016-2021) School of Social Sciences > Statistik u. Sozialwissenschaftliche Methodenlehre (Kreuter 2014-2020)
License:	Creative Commons Attribution 4.0 International (CC BY 4.0)
Subject:	300 Social sciences, sociology, anthropology
Keywords (English):	multiple imputation , missing data , survey harmonization , survey data , sensitivity analysis
Abstract:	This dissertation overcomes several missing data problems in ex-post survey harmonization. Although ex-post survey harmonization projects have become more common in the social sciences in recent years, methodological research on this topic is still relatively rare. The dissertation's main body consists of four chapters, each of which presents a study exploring solutions for missing data problems not only in ex-post survey harmonization but also in related research fields. The study presented in Chapter 2 deals with multiple imputation approaches for discrete-time survival analysis. Chapter 3 presents a study examining multiple imputation approaches to handling systematically missing partner variables. The study presented in Chapter 4 is on the topic of sensitivity analysis in the case of unobserved confounders. And finally, the study in Chapter 5 is on the topic of weighting pooled survey data. The last chapter concludes the dissertation with a discussion of the results, scientific contributions, and limitations of the aforementioned studies, plus a broader outlook on the issue of study heterogeneity.