- This event has passed.
CUPC Research Presentation – Seth Spielman
March 2, 2017 @ 12:30 pm - 1:30 pm
Dr. Seth Spielman, Department of Geography, CU Boulder
“Bad Neighborhoods or Bad Data?
Why tract-level data from the US Census Bureau is broken and how to fix it”
The American Community Survey (ACS) is the primary source of demographic and economic data about American neighborhoods. The ACS is used to allocate over $400B in federal spending. It might be single most important dataset in America. However, the ACS has major problems, for many policy relevant variables the data is so imprecise its almost unusable. For example, in over 70% of census tracts estimates of childhood poverty have a Margin of Error greater than the estimate. How should one use data that says 30% of children in this place are in poverty +/- 35%?
Our research indicates that the most common solution to the ACS’ problems is the "Ostrich Method" – simply ignoring the margins of errors when using the ACS. This talk both diagnoses the problem with ACS data and presents two novel ways of addressing the ACS’ problems. Both “solutions” involve rethinking how we use surveys to describe places.
The “solutions” are computationally intensive and technical but we’ve made some open source software and worked with carto.com to distribute these methods in a usable, accessible form. Attendees will hopefully gain a richer understanding of the ACS and new perspectives on the relationship between the conceptualization and measurement of social space