Youth Mental Health Study

Social media platforms (Meta, X, TikTok, Google) allow people to download their personal data, creating opportunities for research through voluntary data donation. However, this data is often difficult to collect and analyze due to format inconsistencies and platform delays. The Youth Mental Health Study (YMHS) addressed these challenges to investigate digital footprint changes related to mental health service use.

The Youth Mental Health Study (YMHS) extended the Passive Data Kit framework to add infrastructure for gathering social media data using custom importers crafted for each service that transformed personal data downloads (such as Google Takeout) into individual data points representing actions on a variety of networks. This made proprietary data formats usable for researchers. It also introduced reminder infrastructure to the PDK platform in the form of e-mail and SMS reminders to participants to upload their information after it had been compiled for them by each service, which can take days or weeks in some cases.

Since social networking data is usually very sensitive, YHMS adopted the same approach used on mobile devices in the LifeSense project: extracting features - as opposed to raw data - as uploaded data files were processed and encrypting the original files using robust asymmetric cryptography. The features extracted included various text properties, sentiment analysis and content scoring, and (later) toxicity estimates (using the open-source Detoxify classifier).

These features were provided to researchers using the standard Passive Data Kit server. In addition to the features extracted from the data, this project introduced the Codebook component to Passive Data Kit, which allows researchers to create maintainable data codebooks derived from data collected in the field.