Meet the Experts

Registration (via Zoom)   |  

Slides (3,47 MB) |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

A first step towards identifying and closing data gaps is to take stock of data that already exists. We researched all data archives of CESSDA ERIC, the Consortium of European Social Science Data Archives, and found 66 LGBTIQ+ datasets in 9 of the 35 member and partner archives.

In this talk, we will present these data, their temporal and spatial coverage, and which LGBTIQ+ identities were sampled. Furthermore, we take a closer look at the datasets’ keywords and topics. We recorded whether a given keyword is related to LGBITQ+ identities. To describe the topics covered in the studies and to identify topic gaps, all keywords were mapped to the CESSDA Topic Classification. A sentiment analysis with SentiStrength classifies whether a keyword has negative or positive/neutral sentiment. Our paper is available: Recker, Jonas & Perry, Anja (2024). Data on the Margins – Data from LGBTIQ+ Populations in European Social Science Data Archives. Data Science Journal, 23: 39, pp. 1–21. https://doi.org/10.5334/dsj-2024-039

The data are published in this dataset: Perry, Anja, & Recker, Jonas (2024). Data on the margins – Data from LGBTIQ+ Populations in European Social Science Data Archives. GESIS, Cologne. Data File Version 1.0.0, https://doi.org/10.7802/2650.

Presenters:

Dr. Jonas Recker

Dr. Anja Perry

Registration (via Zoom)   |  

Slides |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

Social scientists increasingly use digital behavioral (DBD) data. Collecting and processing such data is typically associated with a set of practical/methodological, ethical, and legal challenges. Resources for large-scale or continuous DBD collections of DBD are often not available to individual researchers or small(er) projects. As a research infrastructure institute, GESIS can engage in such efforts. These are implemented in the new service Web Data for the Social Sciences which has three key components:

1) Data Collection: Continuous or time-limited topical data collections from different platforms;

2) Data Offers: Created based on the collections and made accessible to the community;

3) Community Engagement.

The first two new collections are a continuous crawl of Telegram channels and a comprehensive collection web data from various sources data from the German candidates for the 2024 European Parliament election. In this talk, we will introduce the new service and present the two first data collections.

Presenters:

Dr. Johannes Breuer

Dr. Dimitar Dimitrov

Registration (via Zoom)   |  

Slides |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

In social science research, the traditional “trait-paradigm” has focused on stable variables, between-person differences, and long-term dynamics. However, new research designs that are possible and scalable through mobile data collections consider that measurement does not take place in a “situative vacuum” and that measurement scores are influenced by not only stable, person-specific variables but also within-person, short-term processes as well as combinations of between- and within-person variability that should not be overlooked

Making use of mobile data collection thus implies following an intensive-longitudinal approach. It involves collecting detailed, frequent individual-level data, often daily or multiple times per day; intensive-longitudinal designs usually involve collecting rich, multi-faceted data, including self-reports, behavioral observations, or environmental assessments using built-in sensors. The primary goal of these methods is to understand intra-individual variability, situational aspects, state-like variables, and fast-lived dynamics.

Within this talk, we will outline the promises and pitfalls of different data obtained through mobile intensive-longitudinal methods. We will first introduce the idea of within- and between-person effects and show how mobile methods and intensive-longitudinal designs facilitate new designs and tackle new research questions in social science research. We will also show how these designs can be conducted using GESIS services such as the GESIS AppKit.

Presenters:

Dr. Lukas Otto

Dr. Mareike Wieland

Registration (via Zoom)   |  

Slides |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

To better understand the determinants and effects of digital technologies, approaches for linking surveys and digital behavioral data have been developed recently. In particular web tracking data, detailed web browsing histories of participants, can enrich established survey-based research designs with precise measures of digital behavior. As part of the GESIS Panel Infrastructure, the newly established GESIS Panel.dbd enables researchers to collect linked survey and digital behavioral data (e.g., via the new service GESIS Web Tracking). Study submissions will be fielded free of charge after a successful peer review. A unique feature of the sample (n ~ 7000) is its hybrid sampling design, which features nonprobability and probability samples, allowing for substantial and methodological research.

The MtE session will showcase research that linked surveys and web tracking, present the data collection infrastructure and introduce the pathways how the academic community can participate in the data collections and use the collected data.

Presenters:

Dr. Frank Mangold

Dr. Joachim Piepenburg

Registration (via Zoom)   |  

Slides  |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

Recent advances in the application of Large Language Models (LLMs) to address Social and Political Science research questions have highlighted their potential not just in the “classic” scenarios in which NLP techniques were used in the pre-LLM era (e.g., classification) but as a source of new data through generation or, even more ambitiously, simulation of population samples.

In our Meet the Expert talk, we will start by giving an overview of research directions, technical and ethical challenges, and application scenarios for the use of LLMs as a source of new data for the Social Sciences. We will continue with some questions and insights concerning the validity and robustness of such approaches. Finally, we will discuss ongoing research in the CSS department which targets precisely the intersection between LLMs and survey data.

Presenters:

Jun-Prof Dr. Gabriella Lapesa

Leon Froehling

Registration (via Zoom)   |  

Slides  |   Presentation on YouTube   |   MTE Playlist

The Lecture will be held in English.

In recent decades, the number of large-scale surveys that have included measures of the Big Five personality traits in their standard questionnaires has grown sharply both in Germany and internationally. Consequently, a vast, heterogeneous, high-quality data base is now readily available for secondary analyses.

In our talk, we will provide an overview of 25 public large-scale surveys assessing the Big Five personality dimensions, with the aim to increase researchers’ awareness of the availability and analytic potential of these data. All data presented are surveys of the adult population, conducted in Germany, based on probabilistic samples with a minimum sample size of 1,500 respondents, and assessing all Big Five personality dimensions with a validated Big Five instrument. We describe the study designs, the measures used to assess the Big Five, and the research potential of these data sets.

Presenters:

Prof. Dr. Beatrice Rammstedt

Dr. Lena Roemer