Primer 8: Tracking, Monitoring, and Clarifying Decision and Data Provenance
Capturing data-handling and decision-making processes to ensure coordination
Posted on 22nd of March 2021 by Andrew Young, Andrew Zahuranec, Stefaan Verhulst, Kateryna Gazaryan
Data and decision provenance are key to reducing data risks while re-using data in the most impactful way. By identifying all the decisions that impact the collection, processing, analyzing, sharing, and re-use of data and the parties that impact these decision points, organizations can understand why systems exist as they do and react appropriately when systems do not produce intended results. Decision and data provenance bring to light gaps in data analysis strategies and can lead to improved awareness of the factors that influence decision-making.
Implement Processes to Identify Blindspots in Current Decision-making Processes: Implementing decision provenance does not require complex technical interventions—a common concern for organizations, especially smaller or less-resourced ones. Instead, process-oriented approaches, such as systems and constituency mapping, data responsibility journey mapping, decision trees, and other record keeping methods can provide frameworks to bring greater transparency to the decision-making process.
Determine Who Is Responsible vs. Who Is Accountable: Distinguishing between those working with the data and those who determine the strategies for its use is important. It allows organizations to implement mechanisms that can address harms where they occur and ensure accountability for their actions throughout the process of collecting, transferring, storing, processing, analyzing, and using data.
Communicate Decision Provenance to Relevant Internal and External Stakeholders: Decision and data provenance processes can be useful in identifying the influences on data, but these systems are useless if the people who can use them are unaware of them. Organizations need to ensure that all those stakeholders involved in data processes are tracking their decisions and that all those stakeholders involved in assessing applications of data are aware that a paper trail of decision-making exists.
Decision Provenance: Harnessing Data Flow for Accountable Systems: A paper by Jatinder Singh, Jennifer Cobbe, and Chris Norval, which introduces the concept of decision provenance and explores its potential through a tech-legal perspective.
Decision Provenance Mapping: A methodology, developed in the context of the Responsible Data for Children Initiative and The GovLab, supporting users in capturing decision-making processes and assessing .
Towards a Sociology of Institutional Transparency: Openness, Deception and the Problem of Public Trust: An article by Sara Moore which examines open government initiatives and emphasizes the importance of trust in achieving government openness.
Data Lineage: A Wikipedia article describing the term, which includes “data origin, what happens to it, and where it moves over time.”