GESIS Training Courses

Scientific Coordination

Verena Kunz

Administrative Coordination

Noemi Hartung
Tel: +49 621 1246-211

Social Media-Based Field Experiments

Mannheim, B6, 4-5
General Topics
Data Analysis
Course Level
Students: 200€
Academics: 300€
Commercial: 600€
Additional links
Lecturer(s): Florian Foos, Asli Unan

About the lecturer - Florian Foos

About the lecturer - Asli Unan

Course description

The proliferation of social media use has provided new avenues for social science research, both in relation to the quantitative description of social media data and when it comes to identifying the effects of social media interactions or experiences on political and social outcomes. Moreover, due to the COVID-19 pandemic much of social and political life moved online, providing new opportunities for social media-based field experimentation.
This workshop takes up this development and introduces how social media can be used as a platform to conduct digital field experiments. Social media-based field experiments hold a lot of promise but equally pose numerous methodological and ethical challenges. These challenges are also common in other contexts, but often aggravated as social media is all about interaction; hence how can the non-interference assumption plausibly hold? Previous studies have shown that the effects of social media interventions are small, but if they exist (and can be identified), they are potentially important because interventions are scalable. Additionally, social media-based field experiments face the challenge that many platforms only allow researchers to target users via geographical or demographic clusters, with important implications for experimental design. Moreover, targeting individuals on social media and measuring outcomes has important ethical and data protection implications. How can we deal with these common challenges in our experimental designs? This workshop gives best-practice advice and sets out how different research teams have dealt with these questions. The workshop includes hands-on applications, such as the implementation of experimental treatments on social media and experimental data analysis.
Organisational structure of the course
  • Lectures and hands-on exercises and assignments on participants' own laptops.
  • Work on designing an experiment based on your own research idea
  • Lecturer and TA will be available for individual consultations on participants' projects, and to support work on exercises and assignments.

  • Target group

    Participants will find the course useful if:
  • They are graduate students, early career researchers, or faculty in Political Science, Sociology, Economics, Management, Psychology, or related fields who want to learn how to design, conduct and analyze experiments on social media.

  • Learning objectives

    By the end of the course participants will:
  • Understand the range of topics that lend themselves to social media-based experimentation
  • Understand common assumptions and trade-offs in experimental methodology and how they apply in social media environments
  • Be aware of ethical and data-protection issues surrounding social media-based experimentation
  • Know the strengths and weaknesses of various social media platforms for experimentation
  • Be able to implement experimental treatments on various social media platforms
  • Know how to measure outcomes on social media and offline
  • Be able to conduct random assignment and analyze common datatypes collected via social-media experiments
  • Develop experimental design based on their own research ideas

  • Prerequisites

  • Familiarity with R and R Studio (preferably with tidyverse grammar) (e.g. importing data frames, working with objects, manipulating variables)
  • Graduate level causal inference and statistics class (familiarity with the Neyman-Rubin potential outcomes framework as well as robust working knowledge of linear regression and hypothesis testing).
    Software and hardware requirements
    R and R Studio (the latest version)


    Recommended readings