Degraded Data in Nonprofit Homebrew Databases

Paper Title

Degraded Data in Nonprofit Homebrew Databases

Publication Info

• Topic area: Information management in nonprofit organizations using imperfect database ecosystems.
• Keywords: Homebrew databases, nonprofit organizations, data quality, degraded data, volunteer management, data imagination, sociological imagination, information systems, resource constraints, data practices.

Background and Problem

• Problem / challenge: Nonprofit organizations often rely on messy, imperfectly interoperable ecosystems of information systems, referred to as "homebrew databases," which result in degraded data quality. Little is known about how these systems affect the data managed within them.
• Significance: Understanding degraded data in nonprofit contexts is crucial for improving data practices, organizational decision-making, and long-term usability of data, while respecting resource constraints and organizational priorities.
• Motivation and related work: Previous research has characterized the infrastructural challenges of homebrew databases but has focused more on systems-level analysis than on the nature and impact of the data itself. This paper addresses the gap by exploring how data quality is affected and how organizations navigate competing priorities.

Solution

• Proposed approach: Investigating the nature of data managed in nonprofit homebrew databases through semi-structured interviews with volunteer administrators.
• Novelty:
  1. Identification of five genres of degraded data: incomplete and selectively omitted data, out-of-date and out-of-sync data, “bulk” data with low fidelity, “anecdotal” numbers and “modest” data, and “garbage” data.
  2. Analysis of three alternate priorities that informants prioritize over data management: managing time and money, managing technology, and managing people and relationships.
  3. Introduction of the concept of "data imagination," inspired by Mills’ sociological imagination, to help organizations balance the value and cost of data.
• Procedure and key techniques:
  • Conducted semi-structured interviews with 19 informants from 14 nonprofit organizations.
  • Used inductive thematic analysis to identify patterns in data degradation and management priorities.
  • Explored implications for fostering a "data imagination" through design and education.

Results

• Concrete findings:
  • Five genres of degraded data were identified, including incomplete data due to evolving schemas, out-of-date data from fragmented systems, bulk data that sacrifices fidelity, anecdotal numbers created under chaotic conditions, and fabricated "garbage" data entered to bypass system errors.
  • Informants prioritized managing time, money, technology, and relationships over improving data quality.
• Advantage over baselines: Provides a nuanced understanding of the trade-offs between data quality and organizational priorities, highlighting the legitimacy of these trade-offs rather than dismissing them as inefficiencies.
• Experiments / evaluation:
  • Interviews lasted an average of 68 minutes and covered themes such as data work, collaboration, technical infrastructure, and challenges.
  • Informants represented diverse nonprofit sectors and roles, ensuring breadth in perspectives.
• Limitations and future work:
  • Findings are specific to nonprofit organizations in the United States and may not generalize to other contexts.
  • Future work could explore how fostering a "data imagination" impacts decision-making and data practices in resource-constrained environments.

Summary

This paper investigates the quality of data managed in nonprofit homebrew databases, identifying five genres of degraded data and three alternate priorities that informants prioritize over data management. It introduces the concept of "data imagination," inspired by Mills' sociological imagination, to help organizations better understand the broader social and temporal contexts of their data practices. The findings highlight the strategic trade-offs nonprofit organizations make to balance resource constraints, technology challenges, and human relationships, offering insights for designing systems that support informed decision-making about data quality and management priorities.