class: center, middle, inverse, title-slide # Data Ethics Considerations for Special Populations ### Sondra Stegenga ### CascadiaR 2018 --- #  --- #  --- # 87 Million  --- # 17.6 Million  --- #  --- # Predictive Analytics in Child Welfare? (Hurley, 2018)  --- # Data Protection Policies Not Keeping Pace .pull-left[ + Varied and often contradictory federal, state, and agency specific policies + IRBs updating but still not comprehensive ] .pull-right[  ] --- # Systematic Scoping Review on Big Data Use 🛠 Methods & Tools .pull-left[ + Mixed Methods + Qualitative Content Analysis + Prisma Scoping Review Guidelines ( Prisma-ScR coming soon) ] .pull-right[  ] --- #  --- # What Can We Do as Researchers and Data Scientists? .pull-left[ + ***Collaborate*** to gain domain knowledge in area of analysis to ensure accuracy of findings and decrease harm to potentially sensitive groups that may be under-represented + ***Utilize mixed methods*** with large scale data to improve accuracy of findings and ensure underrepresented groups have voice + Engage in ***Training*** in data ethics ] .pull-right[  ] --- # What Can We Do as Researchers and Data Scientists? + ***Promote policies*** for data ethics that include considerations and biases related to vulnerable and underrepresented populations + Advocate for a ***balance of data access and data privacy*** to ensure protection while also avoiding exposing more subject groups and vulnerable populations + Continue to ***encourage diversity*** in the field of data science and research --- # What Can We Do as Researchers and Data Scientists? **Take Time to Reflect!** + Who is missing? + What are the potential biases? + What are the potential impacts of assumptions and findings? + Who should I be collaborating with to ensure the domain knowledge to answer the prior questions? + How could this be misinterpreted? How do I help to communicate and advocate for appropriate use? --- # What’s Next?? .pull-left[ + Develop ***ethical standards*** for reporting of big data use in research (e.g. consistent reporting of data linkage rates) + Need for ***ethical guidelines*** in the field of data science – contribute your expertise! + Include data ethics in training for data scientists ***with particular attention to special populations*** ] .pull-right[  ] --- #  --- # Special Thanks to @datalorax for Slidex! [https](https://github.com/datalorax/slidex) [://github.com/datalorax/slidex](https://github.com/datalorax/slidex)