Stefan Zins Stefan.Zins@iab.de
Version: MZES Social Science Data Lab, 2025-04-30
Even if it is not apparent to many daily users of statistics, analyzing survey data can be one of the most challenging tasks in inferential statistics. Survey methodologists and statisticians have long acknowledged this pointing to the so-called Total Survey Error (TSE). Although there is an awareness of the complex error structure, all too often statistical inference is done with methods that implicitly assume rather simple errors that are straightforward to estimate. This is particularly true concerning the usage of survey weights. Often considered a nuisance, but necessary for unbiased estimates for so-called descriptive statistics, their effect on standard errors is seldomly considered using the appropriate methodology. The lecture will give basic insights into the construction and logic of survey weights and the appropriate methods are when using them for statistical inference.
📝 Slides
Stefan Zins studied economics at the University of Trier and received his doctorate in statistics in 2015. Since April 2019 he has been working as a specialist in the Statistical Methods Department at the IAB. From 2008 to 2013 he worked as a research assistant at the Chair of Economic and Social Statistics at the University of Trier, and from 2013 to 2016 as a research assistant in the Survey Design and Methodology Department at the Leibniz Institute for the Social Sciences (GESIS), where he was head of the Survey Statistics team from 2016 to 2019. His research focus is survey statistics.