The institutionalized population and social surveys

Schanze, Jan-Lucas

PDF
2021-08_Doctoral-thesis_JL-Schanze_final.pdf - Published
Download (2MB)

URL:	https://madoc.bib.uni-mannheim.de/60175
URN:	urn:nbn:de:bsz:180-madoc-601753
Document Type:	Doctoral dissertation
Year of publication:	2021
Place of publication:	Mannheim
University:	Universität Mannheim
Evaluator:	Wolf, Christof
Date of oral examination:	16 August 2021
Publication language:	English
Institution:	Außerfakultäre Einrichtungen > Leibniz-Institut für Sozialwissenschaften (GESIS)
Subject:	300 Social sciences, sociology, anthropology
Keywords (English):	institutionalized population , elderly population , hard-to-survey , hard-to-interview , social surveys , bias , data quality , monte carlo simulation , propensity score matching , health , voting behavior
Abstract:	People living in institutions, such as prisons, refugee accommodations, or retirement homes are excluded from most social surveys. The exclusion is justified given their small share within the population and their label as hard-to-survey. This cumulative dissertation examines whether the exclusion might bias estimates derived from social surveys in Europe. Both the quantitative size and the statistical distinctiveness of institutionalized residents regarding their socio-demographic characteristics, health status, and political behavior are analyzed. The dissertation concludes that the elderly population living in institutions is especially critical due to their large relative size in many European countries. For this part of the population, bias in health estimates and turnout can be expected, however, bias in underlying socio-demographic characteristics is less pronounced. Moreover, one chapter of the dissertation investigates whether institutionalized residents living in retirement and nursing homes are really harder-to-interview than residents living in private households. The results confirm most indicators of response behavior and data quality fare worse in institutions compared to private households, but when controlling for socio-demographic and health-related variables, differences between the two types of housing diminish. As an overarching conclusion, the dissertation emphasizes that institutionalized persons are not the only hard-to-survey population in terms of coverage and interviewing. Therefore, potential adaptations made by social surveys might also benefit the group of hard-to-survey respondents living in private households.