What Are PUMAs and Why Do They Matter?


TwitterFacebookCopy LinkPrintEmail

A term that few are familiar with is worth some attention in the coming weeks, as Connecticut’s population is divvied up into what are known in technical jargon as Public Use Microdata Areas, or PUMAs.

PUMAs are geographies that include multiple towns and allow the public to access data for an entire region that is not available in other ways. This data is called Public Use Microdata (PUMs).

The data allows people to analyze relationships between economic, demographic, or housing variables, or create new measures that aren’t included in the standard American Community Survey data tables, released periodically by the Census Bureau. PUMAs are the smallest geographic area available to PUMs data from the Census; next is statewide data. 

PUMAs allow data users to cross-tabulate across various topics while providing the option to disaggregate on a deeper level, to develop answers to an almost limitless range of questions that can have an impact on public policy decisions, business decisions, and much more.

 Nonprofits, business leaders, and municipal leaders, among others, can answer these kinds of questions about their region using PUMAs:  What is the number of children who are preschool age? What is the number of adults who are veterans in a region among different demographic groups?  What are the migration patterns in an area?  How many youth are not in school and not employed? 

Since the 2010 Census, the 26 PUMAs in Connecticut have been groupings of contiguous towns all contained within one of Connecticut’s eight counties.  Because planning in Connecticut is not done by counties, but by planning regions, the state is shifting to planning regions as the overarching boundaries to which PUMAs must adhere, rather than county lines. 

The delineation of new PUMAs is coordinated by State Data Centers (SDCs) in every state. In Connecticut, the organization designated as the SDC is the Connecticut Data Collaborative (CTData), a public-private partnership that advocates for the public availability of open and accessible data.

As it develops Connecticut’s new PUMA boundaries, which must be submitted to the U.S. Census Bureau by January 31, 2022, CTData is seeking public feedback, particularly from municipalities, businesses, nonprofits or organizations that have utilized PUMA-level data in the past, to make the new boundaries as useful as possible the future.  A population – much like a pie – can be sliced any number of ways. Where the lines are drawn will determine the data we’re able to obtain.

Creating new PUMA boundaries presents immediate challenges because of federal constraints that Connecticut must adhere to.  PUMAs need to be above 100,000 in population, and should be below 200,000, and be contained within the boundaries of Connecticut’s planning regions, where possible. They must encompass a single, geographically contiguous area, and may not cross state boundaries.

At the outset of this process, CTData sought to make combinations of towns more “logical” in terms of combining towns that have similar demographic composition, so that the data could be representative of each of the towns. However, given the planning region boundaries, there are some areas where this is possible, but many more where it is not.

To say that Connecticut is a patchwork quilt is an understatement.  It is a series of overlapping patchwork quilts, and no two quilts are alike. When we asked data users to recommend the most useful way to bring towns together within PUMA boundaries, suggestions included using boundaries set by a state agency or other regional groups.  

So, Health Districts, Department of Children and Families Regions, Community Action Agency regions, and the five state Congressional Districts were all considered. No two are alike.  Due to the numerous variations, it was impossible to match PUMA boundaries to these any of these regions, even if keeping them within planning regions wasn’t an issue.  Is one of these more relevant, for data purposes, than another?  That remains a pivotal question.   

In some regions of the state, because of the population requirements for PUMAs, the fit is easier than in others.  Some regions present additional questions in need of answers, and public input – particularly among data users – is extremely useful.  For example, when we’re looking at aggregated regional data:

  • In the Western planning region, should Darien be with Norwalk and Westport or with its northern neighbors New Canaan and Wilton? Should Danbury be with Ridgefield and Redding or with Bethel and Newtown?
  • Southwestern planning region can have two PUMAs, and could have all the shoreline towns in one PUMA and all the inland towns in another PUMA. Does this combination most appropriate, or is there a better way?
  • Northeastern planning region has fewer than 100,000 residents, so an additional town or two from Southeastern planning region needs to be included. Which one?
  • In the Capital region, does it make most sense for Bloomfield to be in the same PUMA as West Hartford and Simsbury, or with Windsor and Windsor Locks? Should more rural communities such as Marlborough accompany Glastonbury in a PUMA east of the city, or be included in an overall more rural, eastern part of the Capital Region COG?
  • There will be multiple Naugatuck Valley PUMAs – in which should Middlebury and Southbury be included, for example?  Which towns should they be accompanied by, for data purposes, that would be most useful?

As we continue to sift through the options and possibilities, public input is actively being sought as the deadline for decision nears.  A virtual Community Event is planned for January 5, to provide an opportunity for Connecticut residents to share their opinions, recommendations, and expertise.  Individuals interested in participating can sign up now.   Based on the current Census timetable, data utilizing the new configuration of PUMAs should be available online next summer.

In our Land of Steady Habits, change rarely comes easily.  In this instance, driven by the latest U.S. Census, change is imminent.  Our objective is to accomplish the best change possible, for data users and for residents of Connecticut.

Michelle Riordan-Nold is Executive Director and Sarah Eisele-Dyrli is Assistant Director of the Connecticut Data Collaborative.  Learn more about Public Use Microdata Areas (PUMAs) at www.ctdata.org/pumas.