Companion Guide: Interactive Tool to Reduce Racial Bias in Big Data Studies
This report is a companion resource for the beta version of AcademyHealth’s publicly available Interactive Tool to Reduce Racial Bias in Big Data Studies.
This report is a companion resource for the beta version of AcademyHealth’s publicly available Interactive Tool to Reduce Racial Bias in Big Data Studies. The Tool itself prompts people to think critically about ways they might intentionally adjust their methods for collecting, accessing, and using data to avoid unintentionally perpetuating race- or ethnicity-based bias – especially in their research using big data sources. Questions presented in the Tool also prompt learning and critical thinking that urges researchers to researchers to understand key considerations for ethical use of race or ethnicity data, as well as language for communicating transparency about potential biases in their work. Used to accompany the Tool or as a standalone resource, the Companion Guide provides helpful resources and references. It also details actionable steps researchers can take to proactively, intentionally align their approaches with best practices for mitigating bias in big data studies.
Corresponding to sections of the Tool, the Companion Guide focuses on six steps in the research workflow where bias can potentially be introduced or perpetuated:
- designing the study (e.g., aims, audiences) and defining its parameters;
- selecting high-quality big data sources and ensuring adequate documentation;
- accounting for key considerations related to race and ethnicity attributes;
- identifying appropriate proxy variables;
- addressing data completeness limitations; and
- assessing the representativeness of big data sources.
Guidance provided in this resource was informed by a literature scan of peer-reviewed and other supporting publications. This scan was intentionally designed to review and reference work from experts and authors reflecting diversity across various dimensions. For additional information about this Companion Guide or the Tool, contact HSRInnovation@academyhealth.org.