For further information please consult the course catalogue of Universität Potsdam. https://puls.uni-potsdam.de/qisserver/rds?state=verpublish&status=init&vmfile=no&publishid=83629&moduleCall=webInfo&publishConfFile=webInfo&publishSubDir=veranstaltung close

"Regression Analysis is too little, too late" - this is the conclusion of a book by Richard Berk, in which he criticises the uncritical use of increasingly complicated regression models (not only) in the social sciences. In it, Berk calls for more reflection on decisions that precede the actual data analysis. The aim of the seminar is to sharpen the view on all decisions of a social science investigation: Which substantive statements are sought (descriptive vs. causal inference), which units of investigation must be selected - and how. How can measurement errors be avoided, and what consequences result from them. How "certain" are the results actually achieved, etc. The seminar is a reading course, in which first of all important texts on the design of the examination should be read together. In the second part of the seminar the reading is then selected research papers which are to be critically questioned against the background of what has been learned. Translated with www.DeepL.com/Translator (free version) close