On July 19th, the CONNECT Research Program in Complex Systems and Health Disparities hosted Bias in Big Data: Advancing the Conversation on SGM Health, a free, half-day workshop on big data, bias, and health justice. Selling out in less than a week, the workshop was packed with 50 in-person attendees, and over 900 viewers joining through two livestreams. Workshop participants spanned multiple disciplines and 20 universities, representing community members, community-based organizations, students, public health agencies, activists, academics, and federal research bodies including the NIH and CDC.
The event kicked off with CONNECT’s director, Dr. Michelle Birkett, engaging the audience in thinking through how bias is introduced during each step of the research process. Dr. Birkett emphasized that bias begins with who has the power to ask questions and what issues are considered worthy of study. She concluded by noting, “None of these problems with research are new. But big data can amplify them – and hide them.”
Keynote speaker and founder of Data for Black Lives Yeshimabeit Milner addressed the historical context of algorithm development and big data. Milner explained that, “Some narratives were created by data, and can only be disrupted by data.” Big data has been misused to develop the narrative of a ‘risk factor,’ which has been leveraged against people of color and the LGBTQIA community. Milner eloquently pushed back on this narrative, stating, “Race is not a risk factor. Racism is. LGBTQ identity is not a risk factor. Homophobia and transphobia are.”
Dr. Gregory Phillips II, director of ISGMH’s Evaluation, Data Integration, and Technical Assistance (EDIT) program, presented on gaps, challenges, and opportunities in SGM data, and EDIT associate director, Dr. Lauren Beach, addressed the ethical considerations of conducting research utilizing big data on SGM populations. Dr. Phillips explained that currently the right questions about sex, sexual orientation, and gender idenitity are not being asked, and therefore we miss out on vital information relevant to the persistent health disparities experienced by SGM communities. Dr. Beach highlighted the risk of big data without ethics, urging researchers to follow the ethical recommendations coming from community members and stakeholders.
The workshop concluded with a discussion of how best to promote quality training in data science, and highlighted emerging voices who are doing high-quality work in the field, such as CONNECT’s 2018 data science & SGM health equity paper competition winners, including Austin Eklund. Eklund, a PhD candidate at SUNY: University at Albany, spoke on the importance of identity and discussed his winning paper, which leveraged cultural and historical context to investigate individual and structural predictors of HIV testing among Latinx MSM, using substance use as a moderating factor.
The workshop was CONNECT’s first step in fostering a space where multidisciplinary and community-informed conversation could occur on how biased data impacts the real lives of marginalized communities – and what can be done to address this.
To join this conversation, follow @CONNECT_ISGMH and visit the #BiasInBigData Website, where you can learn more about this work, other scholars and organizations in this area, and view the livestream of the workshop.