What do people do online?
Using data donation to understand digital behavior.

a workshop at the SPP Junior Researcher Meeting

Frieder Rodewald

University of Mannheim & Institute for Employment Research

Sebastian Prechsl

Institute for Employment Research & LMU Munich

October 22, 2025

3️⃣ What is data donation?

The researcher’s perspective.

🤔 What are methodological decisions researchers have to take in data donation studies?

🔑 Key decisions

  • Which theoretical questions do I want to answer?
  • How do I operationalize key variables via my data donation tool?
  • How do I integrate the tool in surveys & recruit participants?

Survey Request & Download Data Extract Data Inspect Data Consent

Decisions we took

  • Goal: Feasability to implement data donation in a labor market panel, develop best-use practices, and understand non-response bias
  • Issues: Data availability, low response rates and bias (e.g., (Keusch et al. 2024))
    • Behavioral intentions as "willingness to donate" high (79-52% of survey respondents)
    • Actual behavior as "participation in data donation" low (37-12% of survey respondents)
    • Well known intention-behavior gap; where 🤑 seems to help (Kmetty et al. 2025)
  • Sample: A non-probability panel (online access panel)

Survey

  • Survey concerning…
    • online platform usage (YouTube, Instagram, and LinkedIn),
    • labor market characteristics,
    • and common indicators for non-participation.

Data request & download

Data extraction

🛠️ Strategy to make the extraction work

  1. Take a look at the DDP; download it, best for multiple time periods and for different languages
  2. Understand the structure of the JSON or CSV.
  3. Get an example file running.
  4. Write the code for the extraction script.
  5. Test your script, first locally and then in the wild.
  6. Adapt your script.

📢 Task: Try it yourself.

Take a look at your downloaded data. What do you see; anything caught your eye?

Feel free to work in groups of 2-3 people for 5 minutes.

Different degrees in standardization for DDP content (Hase et al. 2024)…

  • Documentation
    • DDP structure?
    • Measurements?
  • Completeness & scope
    • Missing data?
    • Limited time frames?
    • Language sensitive?

🚨 Key issues (Hase et al. 2024)

  • Missing documentation by platforms (e.g., file structure)
  • Sudden changes in DDPs
  • Differences across languages & devices
  • Insufficient in-tool classification (e.g., LLM integration)

Example: Extract list of subscriptions

subscriptions.csv (before processing)

...
    "subscriptions": {
        "extraction_function": ef.extract_subscriptions,
        "possible_filenames": ["Abos.csv", "subscriptions.csv"],
        "title": {
            "en": "Which channels are you subscribed to?",
            "de": "Welche Kanäle haben Sie abonniert?",
            "nl": "Op welke kanalen ben je geabonneerd?",
        },
    },
...
def extract_youtube_content_from_zip_folder(zip_file_path, possible_filenames):
    """Extract content from YouTube data export zip file using filenames"""

    try:
        with zipfile.ZipFile(zip_file_path, "r") as zip_ref:
            # Get the list of file names in the zip file
            filenames = zip_ref.namelist()
            # Look for matching files
            for possible_filename in possible_filenames:
                for filename in filenames:
                    if possible_filename in filename:
                        try:
                            # Process based on file extension
                            if filename.endswith(".json"):
                                with zip_ref.open(filename) as json_file:
                                    json_content = json.loads(json_file.read())
                                    return json_content
                            elif file_name.endswith(".csv"):
                                with zip_ref.open(file_name) as csv_file:
                                    csv_content = pd.read_csv(csv_file)
def extract_subscriptions(subscriptions_csv):
    """Extract YouTube channel subscriptions"""

    # Define column name
    if "Kanaltitel" in subscriptions_csv.columns:
        channel_column = "Kanaltitel"
    else:
        channel_column = "Channel Title"

    # Define description
    channel_name = "Subscribed Channel"

    # Create DataFrame with just the channel names
    subscriptions_df = pd.DataFrame({channel_name: subscriptions_csv[channel_column]})
    return subscriptions_df

subscriptions.csv (after processing)

Data inspection

Data donation

  • A data donation platform helps to guide them through the process
  • The process should be made as easy as possible for participants

📚 Workshop Takeaways

  • There are many ways in which researchers can learn about people’s online behavior through digital trace data
  • People can request, download and finally donate their (anonymized) data form most online platforms
  • Data donation can be burdensome for participants
  • Data quailty heavily depends on the platform and what kind of data you extract from participants

🙃 Thank you for participating; happy to talk with you about data donation (and anything else) throughout the next days.

EXTRA: Can I extend the data?

  • Manual annotation by participants during data donation
  • APIs/scraping to extend collected data
  • Text-as-data methods for classification

EXTRA: Errors in representation

For example …

  • Coverage error: Who is (not) represented in the sampling frame? (e.g., social media users vs. YouTube users)
  • Sampling error: Who is (not) represented in the sample? (e.g., non-probability samples)
  • Non-response error: Who does (not) want to participate in the data donation?
  • Compliance error: Who is (not) able to participate in the data donation?

🤔 What do you think: Which participant characteristics may correlate with non-response or non-compliance?

EXTRA: What’s next for data donation studies?

Advancing the method

  • 📸 Multimodal & cross-platform data (Wedel, Ohme, and Araujo 2024)
  • In-tool, local classification (e.g., local ML/LLMs?)
  • Workflow/UX-perspective

Data as a political tool

  • Platforms do (willingly?) not provide data according to the GDPR/DSA (Hase et al. 2024)
  • The EU has started to sanction platforms like X/TikTok
  • DSA may become the subject of larger geo-political debates with the USA (Seiling, Ohme, and De Vreese 2025)

🚀 Can we improve & apply the method?

  • Can the method actually be applied for empirical research? (few examples, like (Thorson et al. 2021; Wojcieszak et al. 2024))
  • Requires interdisciplinary perspectives (e.g., addressing bias, integration in probability-based panels)

🤔 Questions?

References

Hase, Valerie, Jef Ausloos, Laura Boeschoten, Nico Pfiffner, Heleen Janssen, Theo Araujo, Thijs Carrière, et al. 2024. “Fulfilling Data Access Obligations: How Could (and Should) Platforms Facilitate Data Donation Studies?” Internet Policy Review 13 (3). https://doi.org/10.14763/2024.3.1793.
Keusch, Florian, Paulina K. Pankowska, Alexandru Cernat, and Ruben L. Bach. 2024. “Do You Have Two Minutes to Talk about Your Data? Willingness to Participate and Nonparticipation Bias in Facebook Data Donation.” Field Methods 36 (4): 279–93. https://doi.org/10.1177/1525822X231225907.
Kmetty, Zoltán, Ádám Stefkovics, Júlia Számely, Dongning Deng, Anikó Kellner, Edit Pauló, Elisa Omodei, and Júlia Koltai. 2025. “Determinants of Willingness to Donate Data from Social Media Platforms.” Information, Communication & Society 28 (7): 1324–49. https://doi.org/10.1080/1369118X.2024.2340995.
Seiling, Lukas, Jakob Ohme, and Claes De Vreese. 2025. “Wird Europa Den DSA in Verhandlungen Mit Trump Opfern?” Tagesspiegel, March.
Thorson, Kjerstin, Kelley Cotter, Mel Medeiros, and Chankyung Pak. 2021. “Algorithmic Inference, Political Interest, and Exposure to News and Politics on Facebook.” Information, Communication & Society 24 (2): 183–200. https://doi.org/10.1080/1369118X.2019.1642934.
Wedel, Lion, Jakob Ohme, and Theo Araujo. 2024. “Augmenting Data Download Packages – Integrating Data Donations, Video Metadata, and the Multimodal Nature of Audio-visual Content.” Methods, Data, Analyses (Online First), October, 32 Pages. https://doi.org/10.12758/MDA.2024.08.
Wojcieszak, Magdalena, Ericka Menchen-Trevino, Bernhard Clemm Von Hohenberg, Sjifra De Leeuw, João Gonçalves, Sam Davidson, and Alexandre Gonçalves. 2024. “Non-News Websites Expose People to More Political Content Than News Websites: Evidence from Browsing Data in Three Countries.” Political Communication 41 (1): 129–51. https://doi.org/10.1080/10584609.2023.2238641.