Building an ethical data strategy from the ground up

Taylor Schooley is the Research & Policy Manager and LA County’s new Division of Youth Diversion and Development (YDD). Lauren Greenawalt is a New America Public Interest Technology (PIT) Fellow. Together, they are working to build YDD’s data strategy and understanding of data ethics.


Even a first time arrest can keep a young person from opportunities to thrive and increases their chances of future involvement in the justice system. In Los Angeles County, about 80% of all youth arrests are eligible for diversion away from the juvenile justice system. Community and County leadership created Los Angeles County’s Division of Youth Diversion and Development (YDD) to fund and coordinate a network of community-based alternatives to arrest.We’re about to launch the first year of this program, which will allow law enforcement officers to refer kids to YDD-funded providers rather than arresting them. The new program means developing new data and data sources. And that means we’ve started to think a lot about data ethics––we want to ensure our data practices align with our values and mission.

To paraphrase the historian Melvin Kranzberg, data itself “is neither good nor bad; nor is it neutral,” so our decisions about how we use data are critical. Data has certainly been used for good purposes in the burgeoning field of youth diversion; we used data from other initiatives to shape our model and we’ve seen other diversion programs improve their policies after reviewing program data. We also know data practices can harm youth, as arrest records are commonly used to reduce future opportunities for leniency in the justice system and to exclude youth from employment and education opportunities. We’re focused on data ethics before we launch the program or collect data for two central reasons.  

First, risks to our values and goals present themselves from the moment data is captured. When we create new sources of data, we also create the opportunity for it to be misused, intentionally or unintentionally. Think, for example, of the risks present whenever a company asks for your social security number. That data could hurt you if it is stored in a place where someone could misuse it. Likewise, there are risks to collecting sensitive data about youth who participate in diversion. Since we aim to advance equity and youth development, it’s incumbent on us to to minimize these potential risks.

Second, considering data ethics from the beginning means we’ll be better equipped to evaluate future data decisions we may not be able to anticipate today. Government actors who designed data collection years ago might not have predicted that such data would be used in algorithms that determine benefit eligibility or predict cases of fraud. Standing in 2019, we don’t know how people may want to use YDD data in 2029—or even 2020. While we can’t predict every potential application of data, we can lay the groundwork now to ensure future data decisions continue to align with our mission and goals.

So what does it look like to consider data ethics before launching a new program?

We started by talking to our stakeholders about their views on data. We convened a working group and conducted additional interviews with law enforcement officers, community based providers, youth, and community advocates to understand their priorities and concerns regarding data collection and sharing. People were excited about the possibility that data could help measure progress towards YDD goals and highlight areas where new policies or programs might be needed. Some stakeholders also thought that data collection and sharing could improve service delivery for participating youth. At the same time, our stakeholders worried that diversion data could be used to increase or unduly influence decisions about a young person’s care in other systems. Some expressed concern that a singular focus on data could lead us to focus on what could be measured—rather than on what matters.

These conversations spurred us towards more research. We reviewed literature and reports from other initiatives to get a wider sense of the additional risk and opportunities posed by data. And, we turned to additional experts to understand how we could reduce those risks or expand those possibilities. This groundwork led us to draft baseline commitments for the ethical use of data for the YDD program. We’ll be able to deliver on some of these commitments immediately, like documenting our data collection, sharing and analysis plans. Others commitments will take more planning to execute, like designing a process to evaluate and mitigate risks when considering data-related requests. Taken together, our commitments present a unified vision for addressing risks and opportunities, and will help the community hold YDD accountable for aligning all data decisions with our mission and values—whether those decisions are made in the next month or in 10 years.

Resources for developing your own data ethics

If you’re looking to consider data ethics in developing a program or policy, you don’t have to do it alone. We leaned heavily on the work of other initiatives during this process. The resources below were particularly helpful::

  • Responsible Data is an online community for those aiming to reduce unintended consequences of working with data. Their email list is a great way to stay up-to-date on the international conversations around responsible data.
  • DataEthics, a Danish think tank, produced a handbook, which defines key data-ethics terms, outlines principles, and provides a questionnaire to help organizations determine if their practices align with this guidance.
  • Upturn’s Data Ethics report highlights key issues related to data ethics and potential risk mitigation strategies.While the report is targeted towards philanthropists, it’s also instructive for government leaders,
  • New Zealand’s Privacy Commissioner and the Chief Data Steward developed these principles for the safe and effective use of data and analytics which informed YDD’s ethical data commitments and could similarly shape other jurisdictions’ principles.
  • Youth Justice Coalition’s explanation of its RealSearch program is an important resource on the pitfalls and danger of many types of research—including data analysis. This article informed our own commitments to how we’ll engage the community in data analysis and is useful for all those who plan to make data-driven decisions.