Navigating Statistical Uncertainty

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

Census data have been an integral part of planning practice from the original 1909 Plan of Chicago (IL) to present-day planning efforts (Burnham & Bennett, 1993 ). Today, the U.S. Census Bureau’s (USCB) American Community Survey (ACS) serves as a key data set in contemporary planning practice. ACS data, however, are estimates based on a sample rather than complete counts and are drawn from a smaller sample than the decen-nial long-form data. Therefore, ACS data possess a greater amount of statistical uncertainty or unreliability, referred to as margin of error (MOE). This statistical uncertainty is magnified for smaller geographies (e.g., census tracts) or subpopulations (e.g., poverty rate for children) and for cross-tabulations (e.g., race/ethnicity by income).

We know little about how planners understand and communicate the statisti-cal uncertainty of ACS data, although we believe that doing so is part of a plan-ner’s ethical responsibility under the AICP Code of Ethics. We address this gap in the literature by interviewing seven planners in depth and surveying 200 planners to explore their familiarity with statistical uncertainty in ACS data, to understand how and to what extent they use ACS data in research and practice, and to iden-tify their approaches for conveying statistical uncertainty to policymakers and the public. Our results suggest that some planners have a limited understanding of the nuances of statistical uncertainty. Many planners discount the importance of the MOE in using ACS data for policy-specifi c decision-making tasks and rarely, if ever, communicate statistical uncertainty to clients, policymakers, and the public. Just 27% of the planners we surveyed indicated they would warn the end user about unreliable ACS data.

There is an apparent disconnect between the ethical responsibilities set forth by the AICP Code of Ethics and actual on-the-ground practice. The planning academy, through Planning Accreditation Board curricular requirements, and the profession, through professional development training, should underscore the importance of understanding and conveying to users the MOE in ACS data. We also suggest the developers of popular web re-sources that use demographic data (e.g., Social Explorer, Policy Map, and Tableau) report the associated statistical uncertainty. We also recommend that planners follow fi ve guidelines when using ACS data: Report the MOE of ACS estimates, indicate when they are not reporting the MOE, provide context for the (un)reliability of ACS data, con-sider alternatives for reducing statistical uncertainty, and conduct a test of statistical significance when comparing ACS estimates over time.

Sentence Structure

1. ACS data possess a greater amount of statistical uncertainty or unreliability, referred to as margin of error (MOE).
2. There is an apparent disconnect between the ethical responsibilities set forth by the AICP Code of Ethics and actual on-the-ground practice.
3. A gap is something that remains to be done or learned in an area of research; it's a gap in the knowledge of the scientists in the field of research of your study. Every research project must, in some way, address a gap–that is, attempt to fill in some piece of information missing in the scientific literature.

We address this gap in the literature by interviewing seven planners in depth and surveying 200 planners to explore their familiarity with statistical uncertainty in ACS data, to understand how and to what extent they use ACS data in research and practice, and to identify their approaches for conveying statistical uncertainty to policymakers and the public.

Phrases & Words

4. Our results suggest that some planners have a limited understanding of the nuances of statistical uncertainty. Many planners discount the importance of he MOE in using ACS data for policy specific decision-making tasks and rarely, if ever, communicate statistical uncertainty to clients, policymakers, and the public.

• discount <=> underscore/underline =estimate
  the importance of

5. Somewhere (=approximately) 大约
   ▸ somewhere between 50 and 100 people
   50到100人之间
   ▸ somewhere around 10 o'clock
   10点钟左右

Conclusion

American Community Survey (ACS) Data or other data relating to demographics especially that of small geographies or subpopulations is uncertain and unreliable statistically for urban planning .