However, content sampling practices in this domain have often modeled digital search and content discovery in ways that resemble earlier search practices (such as the traditional library card catalog), with little consideration of the role of algorithmic systems in digital environments. As the scale and complexity of digitized data continues to expand beyond the bounds of traditional data processing techniques, researchers must rethink the methods and meanings we apply to collecting data and understanding media exposure in digital settings. Like YouTube and its parent company Google, many social media platforms (ex., Facebook) and search engines (ex., Bing) employ “smart” user-facing content algorithmic systems (or “algorithms” to use a shorthand in common use) for search and recommendation. While content discovery systems have access to a wealth of information about their users, users have comparatively little information about how these systems work.
Using our own experiences reviewing literature and sampling YouTube videos related to marijuana and tobacco co-use for a social media content analysis, we will (1) describe disconnects between media studies scholars’ concerns about algorithms and applications of social media content sampling in public health research; (2) highlight ways in which public health and Science and Technology Studies (STS) approaches to media can inform one another, and (3) advocate for further research focused on differential health impacts of algorithmic environments.