Completed projects

Instagram, political preferences, and federated learning

Niek de Schipper, Laura Boeschoten, and Daniel Oberski, Utrecht University

What people see on social media is hard to reconstruct from self-reports alone. This study paired a questionnaire on political preferences, media use, life satisfaction, and trust in government with donated Instagram data, capturing which creators participants had actually interacted with in the four months before the request. Its primary aim was methodological: to test a federated learning approach for data donation, in which analysis runs across donations without pooling the raw data centrally. Of 1,103 panel members willing to take part, 481 completed a donation.

Data collection: May 2025
Recruitment: Participants were recruited through the Ipsos panel in the Netherlands, with reminders and helpdesk support provided by Ipsos. The response rate was 30%.
Instructions: Participants received in-flow instructions via Port for requesting and downloading their Instagram data.
Software: Port
Processing: GitHub
Data: The metadata of the survey and the metadata of the donated Instagram data are archived on the ODISSEI Portal (CC0 1.0). Contact the authors for access; see the readme.txt for the codebook.

Behind the screens: Exploring Netflix via Data Donations

Dennis Nguyen and Karin van Es, Utrecht University

Netflix is often hailed as a disruptor of the traditional media industry, primarily due to its extensive data collection and analysis capabilities, as well as the effectiveness of its recommendation system. This study used data donations to investigate Netflix viewing behaviours, moving beyond the platform's own selective data sharing to gain an independent understanding of streaming habits.

Data collection: January -- February 2024
Recruitment: Participants were recruited through Ipsos I&O.
Instructions: Instructions to request data and download data.
Ethics: Reviewed by the Ethical Review Board of the Faculty of Humanities of Utrecht University. See the invitation letter, consent form, and privacy policy.
Software: Port
Processing: GitHub
Data: Contact the PI.
Funded by: Funded by PDI-SSH D3I.

References

Lotz, A. D. (2023). “Netflix Data Dump: Cautions Regarding Gifts of Data.” LinkedIn
Van Es, K. (2023). “Netflix, (Re)Claiming Television: Myths and Horseless Carriages.” Keynote at ECREA TV Studies Section Conference, Potsdam, Germany, 26 October 2023.

When news resonates: How societal issues permeate people’s lifeworld

Tim Groot Kormelink (PI), Kasper Welbers, Helen Arts, Summer Skillen & Niek de Schipper

While digitalization and platformization have made it challenging to capture the full range of people’s engagement with news and information, paradoxically, they have also enabled unprecedented access into people’s practices and experiences. Harnessing these developments, this study combines data donations with WhatsApp-diaries and in-depth interviews to answer the following question: How and under which circumstances do societal issues ‘break through’ the continuous stream of information and resonate with citizens?

Data collection: Spring 2025
Recruitment: Participants were recruited via flyers and social media posts.
Ethics: Approved by the Ethics Review Board, Vrije Universiteit Amsterdam (ref: ETO21.52).
Software: Port
Processing: Github
Data: Contact the PI.
Funded by: Funded by NWO (VI.Veni.221C.068).

The Google Family Home -- Exploring the Use of Google Assistant in Families via Data Donations

Rebecca Wald, Jessica Piotrowski, Theo Araujo, Johanna M.F. van Oosten, Amsterdam School of Communication Research, University of Amsterdam

Smart speakers have become increasingly popular in family households. This study investigated how Dutch families with young children use Google Assistant at home, comparing self-reported use with actual registered interactions obtained through data donations.

Data collection: May -- June 2024
Recruitment: Participants were recruited through Motivaction.
Ethics: Approved by the university’s Ethics Review Board (ref: FMG-8809).
Software: Port
Processing: Available on GitHub upon completion.
Data: Anonymized dataset will be available on OSF.
Funded by: Funded by the Communication in the Digital Society Initiative.

References

Wald, R., Piotrowski, J. T., Araujo, T., & Van Oosten, J. M. F. (2023). Virtual assistants in the family home. Computers in Human Behavior, 139. DOI
Wald, R., Piotrowski, J. T., Van Oosten, J. M. F., & Araujo, T. (2024). Who are the (non-)adopters of smart speakers? Tijdschrift Voor Communicatiewetenschap, 51. DOI
Wald, R., Van Oosten, J. M. F., Piotrowski, J. T., & Araujo, T. (2024). Smart Speaker Data Donations in Families. In IDC ‘24. DOI

Harnessing digital data to study 21st-century adolescence

Amy Orben (PI), Laura Boeschoten, Daniel Oberski, Sebastian Kurten, Amanda Ferguson, Valerie Yap, Amelia Leyland-Craggs

Large cohort studies have failed to gather comprehensive digital data about participants' online lives. This project tested the feasibility of deploying the D3I data donation infrastructure in the MRC UK Adolescent Health Study, collecting TikTok and Instagram data from secondary school students.

Data collection: March -- May 2024
Recruitment: Six secondary schools in the UK were visited in person, identified through Gov.uk and existing partnerships.
Instructions: TikTok guide, Instagram guide.
Ethics: Approved by Cambridge Psychology Research Ethics Committee. Information sheet, consent form.
Software: Port
Processing: TikTok, Instagram on GitHub.
Data: Contact the PI.
Funded by: Funded by SUAG/091 and SUAH/031.

An investigation of Facebook groups dedicated to rare diseases

Annemiek Linn and Kaiyang Qin, University of Amsterdam

This study mapped the collective wisdom of a Facebook group dedicated to FOXP1 syndrome, a rare disease. Using data donations of posts and comments, combined with survey data, the study explores how caregivers leverage online platforms for knowledge sharing and community support.

Data collection: January -- February 2024
Recruitment: Participants were recruited from the Friends and Family of FOXP1 Facebook group.
Instructions: Procedure
Ethics: Approved by the Ethics Review Board, University of Amsterdam (ref: FMG-4422). Informed consent, privacy policy.
Software: Port
Processing: GitHub
Data: Contact the PI.
Funded by: Funded by NWO.

Project SOCIAL Social Online Connectedness In Adolescents’ Lives

Loes Pouwels & Michelle Pingel

This project aims to better understand how the effects of social media use on social connectedness differ across adolescents, social contexts, and time scales. In which context does what apply, when, and to whom? This project combines data donations from Snapchat, WhatsApp, Instagram, and TikTok with experience sampling data, survey data, and peer nomination data.

Data collection: January 2025-June 2026
Recruitment: 4 Cohorts of participants were recruited via ongoing collaborations with schools
Instructions: Contact the PI
Ethics: The Ethics Committee Social Sciences (ECSS) has approved the study (23N.010144).
Software: Port
Processing: Contact the PI
Data: Contact the PI
Funded by: This research is funded by a Veni grant (VI.Veni.221G.105) awarded to J. Loes Pouwels by the Dutch Research Council (NWO).

Social Media Use of Individuals in the Context of Organizational Posts

Sarah Marschlich

This study investigates how individuals engage with organizational posts on social media and whether their engagement differs depending on the platform. Using data donations from Facebook, Twitter, and Instagram combined with semi-structured interviews, it explores how motivations and perceived platform affordances shape user engagement with organizational content.

Data collection: January -- August 2024
Recruitment: Participants were recruited from a lab recruitment system from UvA.
Instructions: OSF
Ethics: Approved by the Ethics Review Board, University of Amsterdam (ref: FMG-2664_2023). Materials.
Software: Port
Processing: GitHub
Data: Contact the PI.
Funded by: Funded by the Digital Communication Methods Lab (digicomlab), University of Amsterdam.

Assessing WhatsApp Networks with Donated Data

Rense Corten

Mobile instant messenger services play an increasingly important role in social life, yet we know surprisingly little about these networks. This study used data donations from LISS panel respondents to collect WhatsApp contact and group metadata, producing one of the first datasets on WhatsApp usage from a broad, high-quality sample.

Data collection: February -- April 2023
Recruitment: Participants were recruited through the LISS Panel.
Instructions: Documentation
Ethics: Approved by the Ethical Review Board of the Faculty of Social and Behavioral Sciences of Utrecht University and of the LISS panel.
Software: Port
Processing: GitHub
Data: Contact the LISS panel.
Funded by: Funded by the 2022 ODISSEI-LISS panel grant and NWO VIDI (195.152).

Mapping the digital food environment using YouTube data

Kaiyang Qin, Saar Mollen, Wilma Waterlander, and Eline Smit

Social media represents an emerging dimension of the digital food environment. This study used YouTube data donations to map how users interact with food-related content, examining whether people are exposed to more unhealthy than healthy food content and the role of both advertiser targeting and user self-selection.

Data collection: April -- November 2023
Recruitment: Participants were recruited from Panelinzicht and a lab recruitment system from UvA.
Instructions: Instructions
Ethics: Approved by the Ethics Review Board, University of Amsterdam (ref: FMG-472). Informed consent, privacy policy.
Software: Port
Processing: GitHub
Data: Contact the PI.
Funded by: Funded by the Healthy Future initiative of the University of Amsterdam.