DEZIM SUMMER SCHOOL 2024

Multilevel Analysis - 13. & 14.08.

From a social science perspective, individual attitudes and behavior are significantly influenced by contextual conditions. These contextual conditions can take the form of political institutions, regional opportunities, characteristics of neighborhoods, families or social groups in general. Hierarchical regression models provide an analytical procedure for looking at individuals clustered in groups (such as states, neighborhoods, families) and thus for simultaneously considering individual and contextual determinants of individual attitudes and behaviors.

In a first step, the two-day course provides an overview of context effects in the social sciences and related research designs and analytical procedures of multilevel analysis. In a second step, hierarchical regression models based on linear regression will be derived step by step and explained using examples. These include random intecept models, random coefficient models and cross-classified models. In a third step, the methods will be practically applied in hands-on sessions to exemplary questions and sample data using Stata. In particular, comparative country analyses with the European Social Survey will be used, a typical application of multilevel analyses in recent years. Depending on the participants' interests, the focus can be on specific methods and applications or address concrete topics from ongoing research projects.

• Required software: Stata

Martin Kroh is Professor of Methods of Empirical Social Research with a focus on quantitative methods at the Faculty of Sociology at Bielefeld University. He is also a member of the board of the Institute for Interdisciplinary Research on Conflict and Violence (IKG) and is involved in projects within the FoDiRa Research Network as well as in projects on institutional racism at the Research Institute for Social Cohesion. His research interests are the relationship between social and political inequalities as well as longitudinal and experimental research data.

Web-Scraping with Python - 15. & 16.08.

In this 2-day workshop, participants will gain foundational knowledge and practical skills in web scraping using Python. The course is designed to introduce essential concepts and techniques required for collecting data from the web. Throughout the workshop, participants will delve into understanding the structure of web pages, grasp the principles of web scraping and crawling, and acquire efficient methods for data extraction.

• Required software: Python

Dongrui Jiang has served as the lecturer for the course "Introduction to Python Programming" at the TU Berlin Summer and Winter University since 2020. With over 5 years of experience as a researcher leveraging Python for data processing, she brings a wealth of practical knowledge and expertise to the field. Passionate about teaching and sharing insights, she conducts training sessions covering basic Python grammar, as well as advanced topics in web scraping and data analysis. Participants of her previous Python courses have offered positive feedback on the practical skills they acquired, noting the tangible impact these skills have had on their professional endeavors.

Introduction to qualitative methods - 19.08.

The course is divided into four thematic blocks. The first block provides a brief introduction to epistemological principles in contrast to quantitative research approaches. Building on this, the variety of qualitative designs, data collection methods and analytical procedures will be outlined. Central features of qualitative research and quality assurance criteria are discussed. The second section takes an in-depth look at qualitative data collection, in particular guided interviews. In the third section, role plays are used to try out how to conduct interviews and then reflect on them together. This is followed by a brief introduction to qualitative content analysis. In the fourth block, the first steps of the analysis are tested using an exemplary text excerpt. Finally, the presentation of qualitative results will be discussed and illustrated using selected examples.

• Required software: none

Dr. Elena Tsarouha is a sociologist and has extensive knowledge of empirical social research, especially in the field of qualitative methods. She works at Esslingen University of Applied Sciences in the accompanying research for the introduction of the new nursing training programs. Previously, she worked in the field of occupational medicine, social medicine and healthcare research at the University Hospital of Tübingen in various research networks on the topics of stress-preventive leadership and mental health in the workplace. She completed her doctorate at the European University of Flensburg on the subject of examinations at universities and worked in qualitative and quantitative methodology.

Introduction to Statistics - 20.08.

This course covers four topics: empirical social research, descriptive statistics, inferential statistics and their interrelationships. These will be explained using examples from various fields such as sociology, politics and economics. Various well-known data sets such as Allbus, GSS, PISA will be used during the workshop to perform various analyses.

• Required software: R and R-Studio

William Tarazona has been working as a research assistant at the European University of Flensburg in the Department of Central Methodology since 2005, where he teaches various courses in statistics. His courses range from Introduction to Statistics to Statistics with Excel and SPSS as well as Simulations. In 2011, he received a university-wide teaching award for the best course at the University of Flensburg.

Introduction to questionnaire & item construction - 21.08.

The one-day workshop "Questionnaire and Item Construction" offers interested participants the opportunity to learn the basics of questionnaire development, evaluation and testing. The aim is to enable participants to assess the quality and suitability of existing questionnaires and items for specific purposes and to develop and test their own questionnaires. The course focuses on interactive exercise units in which course participants apply their knowledge under guidance. The exercises are framed by theoretical principles (e.g. total survey error framework, measurement error, cognitive response process, measurement of latent constructs) and best-practice guidelines as well as empirical findings from questionnaire research (e.g. on the arrangement of questions in the questionnaire or the number of response options). A special focus is on the consideration of the survey mode in questionnaire construction (e.g. visual design effects in online questionnaires, interviewer effects in telephone or face-to-face interviews) as well as the basic techniques of question evaluation and the use of cognitive interviews for questionnaire optimization. Participants are invited to bring their own research questions that they would like to translate into questionnaire modules. However, this is not a prerequisite for participation.

• No software required.

Dr. Carina Cornesse studied sociology in Mainz and Frankfurt am Main and completed her doctorate at the University of Mannheim. Her research work focuses in particular on the topics of data collection and the use of panel data. She is currently working as a scientific project manager at the German Institute for Economic Research (DIW Berlin). She also works as a research assistant at the University of Bremen and as a lecturer in data collection methodology at the Free University of Berlin. She is a board member of the European Survey Research Association, Associate Editor of the Journal of Survey Statistics and Methodology and a cooperation partner of the Center for Panel Survey Sciences.

Applied Panel Data Analysis - 23.08.

Panel data analysis offers important advantages over basic inferential analyses using cross-sectional data. In particular, it allows the identification of causal effects under comparatively weak assumptions and the analysis of individual developmental trajectories. This course provides a general introduction and application of various models and estimators commonly used for panel data analysis. With a focus on causal inference, we will discuss the assumptions and application of various estimators including pooled ordinary least squares (POLS), random effect (RE), fixed effect (FE), first difference (FD), and difference in differences (DID). Model specification for the consistency of the estimators will be emphasized. We will also cover the modeling techniques of growth curves and event study designs using impact functions, which allow the investigation into life course trajectories and effect dynamics. The central goal is to enable participants to critically evaluate different panel research designs and to independently apply these methods to their research. To strike a balance between statistical theory and practical application, we will have several practical exercises using the statistical package Stata and hands-on advice for researchers.

Chen-Hao Hsu has started pursuing his PhD at the University of Bamberg in October 2019, where he has also served as a lecturer for several courses on Applied Panel Data Analysis and Multilevel Analysis in the Chair of Methods of Empirical Social Research. He has contributed significantly to research on family demography and labor market sociology, working as a research associate for the ERC-funded SECCOPA project from 2021 to 2023, as well as a research associate in the Professorship of Demography. He has not only published extensively in reputable journals but also presented his findings at numerous international conferences.

• Required software: Stata