Open Sharing of Behavioral Research Datasets: Breaking Down the Boundaries of the Research Team

Abstract

Behavior is the lynchpin underlying many of the most vexing problems in public health. Behavior can contribute to the progression or prevention of disease, define a disorder or mark recovery, and provide mechanisms for therapeutic intervention. Clinicians and health researchers have many robust and reliable tools at hand—from blood assays to brain or whole body images—for measuring people’s physical health. But, the tools for measuring healthy and at-risk behaviors are severely limited. Here, we argue that a readily available, inexpensive, and profoundly powerful tool already exists for capturing the complexity and richness of many health-related human behaviors– video recording. Video captures subtle, real-time dimensions of behavior that standardized observational or self-report instruments ignore or grossly simplify. Video also captures the nuances and vital details of experimental procedures more completely and accurately than do written protocols or the “methods sections” of published articles. Although video contains identifiable information such as faces and voices, video data can be securely but openly shared with other researchers using the policy framework and technology developed by the Databrary (databrary.org) digital data library. None of the rich information contained in video must be lost. Trained human coders can systematically extract the essential information using fast, efficient, low-cost annotation tools, and video can be subsequently mined and reused by other researchers to ask additional questions. By making commonplace secure video sharing about health-related behaviors, we can break down barriers between diverse and geographically dispersed research teams, increase scientific transparency, and hasten the pace of disease prevention and health promotion.