Navigating Statistical Uncertainty: How Urban and Regional Planners Understand and Work With American Community Survey (ACS) Data for Guiding Policy

Authors:
Publisher website
Google Scholar
Abstract
Problem, research strategy, and findings: The American Community Survey (ACS) is a crucial source of socio­demographic data for planners. Since ACS data are estimates rather than actual counts, they contain a degree of statistical uncertainty—referred to as margin of error (MOE)—that planners must navigate when using these data. The statistical uncertainty is magnified when one is working with data for small areas or subgroups of the population or cross-tabulating demographic characteristics. We interviewed (n = 7) and surveyed (n = 200) planners and find that many do not understand the statistical uncertainty in ACS data, find it difficult to communicate statistical uncertainty to stakeholders, and avoid reporting MOEs altogether. These practices may conflict with planners’ ethical obligations under the AICP Code of Ethics to disclose information in a clear and direct way. Takeaway for practice: We argue that the planning academy should change its curriculum requirements and that the profession should improve professional development training to ensure planners understand data uncertainty and convey it to users. We suggest planners follow 5 guidelines when using ACS data: Report MOEs, indicate when they are not reporting MOEs, provide context for the level of statistical reliability, consider alternatives for reducing statistical uncertainty, and always conduct statistical tests when comparing ACS estimates.
Funding Information
  • the National Science Foundation (1132008)

This publication has 22 references indexed in Scilit: