Slides (2,55 MB)   |   This talk on YouTube   |   MTE playlist

A Short Introduction to Computational Social Science and Digital Behavioral Data

This talk will provide an introductory overview of the emergence of computational social science (CSS) as a new research area combining multidisciplinary methods and new types of data that promise to be valuable complements to surveys. The term digital behavioral data summarizes a broad variety of data captured by web-based platforms (most prominently platforms for online communication, but also e.g. shopping portals or dating sites) and other digital technologies like smartphones, fitness devices, or RFID sensors. Digital behavioral data result from traces that humans are leaving when using these platforms, i.e. the data is typically not a direct product of a scientifically predesigned research setup.

The talk showcases examples of digital behavioral data and how they have been used in past CSS research to learn about or predict behavior, characteristics, or opinions of platform users. It provides the basis for a series of talks from the area of CSS that will take a closer look at individual strategies for collecting digital behavioral data, different methods data analysis and different use cases for social science research.

Speaker

Dr. Katrin Weller leads the team Social Analytics and Services and is deputy head of the Computational Social Science department at GESIS. She holds a PhD in information science and is interested in understanding scientific communication structures in online environments. Her work also questions data archiving and sharing surrounding social media data. She was Digital Studies Fellow at the US Library of Congress.

Slides (4,16 MB)   |   This talk on YouTube   |   MTE playlist

Digital Traces of Human Behavior from Online Platforms – Research Designs and Error Sources

Peoples’ activities and opinions recorded as digital traces online, especially on social media and other web-based platforms, offer increasingly informative pictures of the public. They promise to allow inferences about populations beyond the users of the platforms on which the traces are recorded, representing real potential for the Social Sciences and a complement to survey-based research. But the use of digital traces brings its own complexities and new error sources to the research enterprise. Recently, researchers have begun to discuss the errors that can occur when digital traces are used to learn about humans and social phenomena.

This talk discusses various strategies for critical reflection on the limitations, implications, and consequences of using digital traces for measuring social constructs. Inspired by the Total Survey Error (TSE) Framework developed for survey methodology, we introduce a conceptual framework to diagnose, understand, and document errors that may occur in studies based on such digital traces. While there are clear parallels to the well-known error sources in the TSE framework, the new “Total Error Framework for Digital Traces of Human Behavior on Online Platforms” (TED-On) identifies several types of error that are specific to the use of digital traces. By providing a standard vocabulary to describe these errors, the proposed framework and this talk advances communication and research concerning the use of digital traces in scientific social research.

Speakers

Dr. Fabian Flöck leads the Data Science team at the Computational Social Science Department at GESIS. He is interested in the validity and transparency of automated measurement in social science contexts, but also researches interactive data analysis services, collaborative content creation and digital communication processes. He studied communication sciences and sociology, and subsequently acquired a PhD in computer science.

Indira Sen is a doctoral candidate at GESIS, working at the intersection of Computational Social Science and Natural Language Processing. She has a bachelor’s and master's degree in Computer Science, and currently works on measuring social constructs like political attitudes and hate speech from social media data and understanding the limitations inherent to this task.

Slides (3,39 MB)   l   This talk on YouTube   |   MTE playlist

Combining Survey Data and Digital Behavioral Data

The use of digital behavioral data (DBD) is one of the key features of computational social science. These data have several advantages compared to other types of data, such as survey data. For example, DBD are generally less costly to collect and less prone to being influenced by social desirability than survey data, and they allow for capturing behavior immediately when it occurs instead of ex-post, thus, reducing problems related to recall. At the same time, however, DBD also have specific limitations. These include a lack of information about the individuals (e.g., regarding their demographics, personality traits, or attitudes) or the fact that they provide only limited information about offline activities. Yet the measurement of individual attributes and attitudes as well as self-reported behaviors across domains is what surveys are well-suited for. Combining survey data and DBD therefore leverages the unique strengths of these two data types, while also addressing some of their respective limitations.

This talk discusses the benefits of combining survey data with DBD and the challenges associated with this approach. We will present different ways of linking surveys and DBD and address key challenges regarding linking procedures, informed consent, and data privacy.

Speakers

Dr. Johannes Breuer is a senior researcher in the team Data Linking & Data Security at GESIS where his work focuses on data linking and the use of digital trace data. He holds a Ph.D. in psychology and his research interests include the use and effects of digital media, computational methods, data management, and open science.

Dr. Sebastian Stier is a senior researcher in the Computational Social Science department at GESIS. He holds a PhD in political science and conducts research in the fields of political communication, comparative politics, and computational social science.

Slides (3,31 MB)   |   This talk on YouTube   |   MTE playlist

Research Ethics and Data Protection in Social Media Research

Information from social media and other web platforms are collected and used as a new type of research data across academic disciplines. Examples are the study of online communication or learning about users’ behaviors and opinions. While this promises numerous research opportunities, it not only leads to novel methodological challenges, but also raises various questions about research ethics and data protection. By now some guidance exists for researchers who study social media communication, e.g., material offered by professional research associations. But there is still a lot of work in progress and standards or best practices may not exist for every case. With this presentation we give a starting point to reflect upon critical questions when working with this new type of research data and point out useful resources as references.

The presentation will shed light on how researchers, social media platforms and users are entangled in complex relations that contribute to the challenges of research ethics. A specific focus will be placed on data protection as a core concept. Exemplary research scenarios will be used to illustrate critical questions along different phases in a prototypical research process.

Speakers

Dr. Katrin Weller leads the team Social Analytics and Services and is deputy head of the Computational Social Science department at GESIS. She holds a PhD in information science and is interested in understanding scientific communication structures in online environments. Her work also questions data archiving and sharing surrounding social media data. She was Digital Studies Fellow at the US Library of Congress.

Oliver Watteler is a senior researcher at GESIS – Leibniz Institute for the Social Sciences, where he works in the department Data Services for the Social Sciences (DSS). He holds a master's degree in history and political science and is responsible for data acquisition. Together with colleagues he advises and trains on topics of research data management. His focus is on legal conditions of RDM, especially data protection, on which he also publishes. He is a member of the Leibniz Association's Data Protection Working Group and vice member of the GESIS Ethics Committee.

 Slides (6,41 MB)   |   This talk on YouTube   |   MTE playlist

Introduction to Online Data Acquisition

Scientists from various disciplines increasingly want to include web data in their work. The acquisition of such online data is an essential step in most computational social science research projects. There are multiple ways to collect data from online platforms, such as using web application programming interfaces (web APIs), web scraping, and data donations. Some methods and online platforms are better suited to explore different research questions, and the data that can be obtained vary according to policies and the structure of the platform forcing researchers to balance the alternatives. Therefore, it is important to understand the ecosystem to decide the best way to proceed, also considering the implicit policies that each API imposes in the process. Online data acquisition offers multiple opportunities to address societal questions that would otherwise be impossible to answer, and researchers should be able to easily evaluate the different alternatives taking into consideration their scope and limitations.

This talk presents methods of collecting online data, illustrates how access varies across different platforms such as Twitter and Crowdtangle, and shows the opportunities that the methods and platforms offer. First, the talk will introduce general web concepts including terminology related to data collection. Then, it will showcase some of the possibilities to access data from online platforms, including direct API (e.g. Twitter), third-party sources (e.g. CrowdTangle) and annotation services (e.g. Amazon Rekognition), as well as examples of how this data has been used to answer different research questions.

Speaker

Dr. Roberto Ulloa is a postdoctoral researcher at the Computational Social Science department of GESIS. He researches the role of institutions in shaping societies, and online platforms as forms of digital institutions.

 Slides (4,37 MB)   |   This talk on YouTube   |   MTE playlist

Auditing Algorithms: How Platform Technologies Shape our Digital Environment

In digital environments algorithms play a major and often a decisive role. Algorithm auditing is an important research method for identifying problematic behavior in online platforms (such as social networks, search engines, job matching sales platforms). As these platforms gain influence in society, questions regarding their fairness have been raised. For example, how do algorithms reinforce discrimination against different societal groups? Do recommendation systems lead to polarization and help spread extremists views? Are certain sources favoured by search engines? Multiple auditing approaches have been applied to investigate these questions. Although most of the methods can be applied to relevant social science cases, the use of automated browsing (virtual agents) stands out as a promising venue for a permanent monitoring of algorithmic systems that can be leveraged to keep the companies behind them accountable.

This talk compares different methods for algorithm auditing, and showcases different results from past investigations. It will focus on a family of methods using virtual agents to systematically scrape online platforms at a large scale, allowing for experimental designs that can control factors that affect algorithms such as different levels of personalization.

Speaker

Dr. Roberto Ulloa is a postdoctoral researcher at the Computational Social Science department of GESIS. He researches the role of institutions in shaping societies, and online platforms as forms of digital institutions.

 Slides (2,92 MB)   |  This talk on YouTube   |   MTE playlist

The German Federal Election: Social Media Data for Scientific (Re-)use

Social media have developed into important arenas for political debate and strategic communication by political actors. For social scientists, observing political communication over party competition and social media behavior of elites and regular citizens alike allows for drawing inferences about new political phenomena and longstanding research questions. In particular, a systematic monitoring of political campaigns and elections provides a high number of observations with a fine-grained granularity at a limited cost, when compared to population surveys.

However, research on social media comes with its own set of challenges, from collecting to storing, filtering, coding and interpreting the resulting large data sets. In this session, we elaborate on five best practices: the identification and validation of the target population, the most efficient and legal collection of data using APIs and open source tools, the design of databases to store these types of data, efficient ways to put humans in the loop for coding and validation, and finally, the management and maintenance of these processes.

Speakers

Dr. Marius Sältzer is a Postdoctoral researcher at GESIS CSS in the Project “Negative Campaigning in German Elections in Social Media”. He received his PhD from University of Mannheim on how political elites use social media to signal their positions and priorities to the electorate and other political actors. His research revolves around the dimensions of political conflict in political communication; methodologically Marius focuses on Text-as-Data.

Dr. Sebastian Stier is a senior researcher in the Computational Social Science department at GESIS. He holds a PhD in political science and conducts research in the fields of political communication, comparative politics, and computational social science.

 Slides (4,37 MB)   |  This talk on YouTube   |   MTE playlist

Introduction to Text Mining

Language is the essence of our social life, and its use is at the center of many phenomena, studied by computational social scientists and political communication researchers alike. Examples include opinion dynamics, political polarization, or the biased representation and discrimination of social groups. The statistical analysis of large text corpora from traditional news and social media platforms greatly extends our methodological toolkit to study these phenomena. 

In this talk, Arnim Bleier gives an overview of the use of text as data and corpus statistics in the social sciences. The session will provide an introduction to areas of natural language processing such as feature extraction and supervised as well as unsupervised modeling of text. 

Speaker

Dr. Arnim Bleier is a postdoctoral researcher in the Department Computational Social Science at GESIS. His research interests are in the field of Computational Social Science, with an emphasis on natural language processing and reproducible research. In collaboration with social scientists, he develops models for the content, structure, and dynamics of social phenomena.

Slides (3,06 MB) (3.31 MB)   |  This talk on YouTube   |   MTE playlist

Social Network Analysis with Digital Behavioral Data

Our uses of digital technologies like social media platforms, email, or cell phones leave massive amounts of behavioral traces that are most interesting for social research. Such Digital Behavioral Data (DBD) consists of genuinely relational records. It requires a shift of perspective from persons to micro events as units of observation and brings established techniques like Social Network Analysis to center stage. In this session, I will propose a definition of DBD, characterize it as records of transactions, and discuss its properties in contrast to survey data. Next, I will draw upon empirical examples to describe how Social Network Analysis with DBD can address social dynamics that have traditionally been studied on either the micro or macro level. Social Network Analysis with DBD allows studying identities and social formations as processes. The key for doing so is that transactions are traces of individual behavior that are highly resolved and allow the reconstruction of emerging patterns. I will close with thoughts about methodological challenges and the significance of relational theories for the consolidation of Computational Social Science.

Speaker

Dr. Haiko Lietz is a sociologist with an engineering background. His dissertation at the University of Duisburg Essen is about network theory and analysis of scientific practice (2016). He is a post-doc researcher at the Computational Social Science department at GESIS. His research interests are in analytical and relational sociology and complex systems approaches.

Slides (5,41 MB)  |  This talk on YouTube   |   MTE playlist

Altmetrics – Analyzing Academic Communications from Social Media Data

Citation indexes have been used during the last years as an instrument to measure the scientific success of researchers. Although, they have been criticized for not showing immediate impact and for not being suited to assess the impact of non-traditional contents, like survey and research data or computer software and tools. Therefore, alternative metrics (altmetrics) to measure scientific impact and success are being explored. These new measures are often based on online news media, reference managers, and social media networks, including social bookmarking systems.

The presentation will shed light on existing aggregators of altmetrics from social media data and how these indicators can be used. Additionally, we will showcase some of the possibilities of indicators collected from Wikipedia and YouTube, as well as the challenges that occur with the data collection and interpretation.

Speakers

Olga Zagovora is a doctoral candidate at the Computational Social Science department at GESIS. Prior to joining GESIS, she studied computer science, web and data science. Her research focuses on the evaluation of alternative metrics for measuring scholarly communication and scientific impact. She has experience working with big data for social science research.

Dr. Katrin Weller leads the team Social Analytics and Services and is deputy head of the Computational Social Science department at GESIS. She holds a PhD in information science and is interested in understanding scientific communication structures in online environments. Her work also questions data archiving and sharing surrounding social media data. She was Digital Studies Fellow at the US Library of Congress.

Slides (1,80 MB)  |  This talk on YouTube   |   MTE playlist

Political Behavior and Influence Dynamics in Online Networks

Millions of people comment daily on current events on a variety of platforms ranging from diverse social media to the news sites. Reading others’ comments can shape one’s own opinions about the story, its author, and the media outlet and help spread opinions and claims which counter the mainstream narrative. Online discussions and user networks from these discussions create a large amount of datasets, which can be used to investigate human behavior online. This talk presents typical steps taken when performing research on such vast datasets, as Tweets and comments from news websites, and introduces some of the Network Science methods employed in the search for answers to research questions regarding the political behavior and influence in online networks.

Speaker

N. Gizem Bacaksizlar Turbic is a postdoctoral researcher at the Computational Social Science department at GESIS. She received her Ph.D. from the College of Computing and Informatics, University of North Carolina at Charlotte (2019). Her research areas include complex adaptive systems and social-political networks. Gizem studies the effects of social influence on political behavior and group dynamics. Her methodological expertise is in network science, agent-based modeling, and system dynamics.

 Slides (711 kB)   |   This talk on YouTube   |   MTE playlist

SocioHub – A Collaboration Platform for the Social Sciences

SocioHub is a web portal for sociology that (1) allows users to search for literature, research data, researchers, and research groups, and (2) improves communication between researchers and within research groups through a collaborative network. Users can create profiles that include their scholarly accomplishments. Research groups (and projects) can use various tools for internal collaboration and also use their group page as their website.

This talk will demonstrate the possibilities for researchers and research groups through a live tour. Researchers who want to use SocioHub for themselves or for their research groups and projects are invited to participate interactively with feedback and questions. Furthermore, our team will be available for further consultation appointments for participants.

SocioHub is a joint project of the University and City Library of Cologne (USB) and GESIS - Leibniz Institute for the Social Sciences and has been funded by the German Research Foundation (DFG) since July 2016.

Speaker

David Brodesser is a product manager at the „Knowledge Technologies for the Social Sciences“ department at GESIS. His focus is on agile and user centered development processes for software services.

 Slides (577 kB)   |   This talk on YouTube   |   MTE playlist

Pollux – Literature and Research Tools for Political Scientists 

With more than 8 million references of articles, books, book chapters and datasets, Pollux is the central academic search engine for political science. Individual titles are either immediately available electronically or listed in the portal with a reference to the owning library. In addition it provides information on open access publishing and research data management.

In this talk, we will introduce the services and research tools we offer in Pollux. We then provide an overview of our collaborative partners and our feedback loop with the political science research community. We end our talk with a discussion on the challenges of discoverability of text corpora, which is one aspect for the future development of Pollux.

Pollux is a joint project of the State and University Library Bremen (SuUB) and GESIS – Leibniz Institute for the Social Sciences and has been funded by the German Research Foundation (DFG) since July 2016.

Speakers

Regina Pfeifenberger is Public Relations and Open Access Officer at the Specialised Information Service for political sciences at the State and University Library Bremen (SuUB).

Wolfgang Otto is a research associate at the Knowledge Technologies for the Social Sciences department at GESIS and software developer in Pollux.