CCL Blog

What We Learned About User-Centered Technology for Census Outreach

Oct 1, 2018

Every 10 years, the federal government is required to conduct a census to count all the nation's residents. ​An accurate count of the U.S. population forms the basis for many important political, economic, and social decisions - ranging from congressional apportionment and redistricting to allocation of more than $675 billion dollars of federal funding each year for infrastructure, emergency services, and critical social programs.  

The Census Bureau uses a Master Address File (MAF) to send out information on completing the census and identify which residences require a follow-up. If an address is left off the MAF, it  is very unlikely an enumerator will be sent to follow up or that an estimated count for that household is included in the census. In low-income, minority, and immigrant neighborhoods, as many as ​2-3% ​of the places where people live may not be on this list.

Often, this is because they live in unconventional housing that does not show up in administrative records. In the past, the Census Bureau has done address canvassing to try to identify these units. Due to budget cuts, much of this will be done now through satellite images, making it difficult to catch clues like additional entrances and garage conversions. Therefore, people living in unconventional housing are at a higher risk of being left out of  Census 2020.

1. Technology solutions need to meet users where they are.

Solutions developed for Census need to be designed to be used by users who are CBO staff or volunteers in those HTC communities. Those users include low-income, immigrants, minorities, and non-English speakers.

These users need simple, lightweight technology solutions that leverage how they use technology every day. As an illustration, the most frequently asked question we received from community volunteers was a concern of how much of their cell phone data plan our tool would use. Downloading software was an issue for a similar reason. Many volunteers from HTC communities did not have space on their phones for mobile apps.

2. Technology tools for Census to be used by CBOs and their volunteers who serve minority or immigrant populations need to be available multiple languages.

With a significant number of CBO staff and volunteers also immigrants/non-English speakers, the translations need to be high-quality translations provided by humans, not Google translate.

3. Reporting data should be automated, not requiring manual formatting or uploading.

Data reporting will be a significant barrier for Census outreach. Many of the CBOs do not have the technical expertise or capacity to do reporting on their own. Even for CBOs that have technical capacity, it is time intensive and will detract attention from executing outreach efforts. Finally, manual reporting is slow, meaning that government planners will be making decisions with data that may not reflect the on-the-ground realities.

Automating reporting provides the following benefits:

  1. A wider range of CBOs can participate in outreach.
  2. More standardized data collection, which means more useful data for planners.
  3. Centralized and more efficient data processing.
  4. Lower coordination costs for intermediary organizations, such as a local governments and funders.
  5. Enable nonprofits to spend more time doing outreach and less time on reporting.

Technology solutions should provide data that adheres to security, format, and other requirements for easy, non-manual data sharing with reporting platforms such as the State of California’s SwORD.

4. Technology solutions need to be developed specifically for Census outreach.

The Census is a unique use case, requiring custom technology to be developed and simply repurposing other technology tools is not sufficient.

In particular, Census outreach is different from voter outreach.

Unlike political campaign technology based on existing public databases (e.g., voter file), there is no publicly available database for Census outreach. The voter database is limited legally in most states for use only for elections. The Census Bureau uses proprietary databases for its address canvassing and enumeration follow-ups that are not available to the public.

CBOs will have to use predictive modeling, live data streams of self-response rates, and live data reporting tools in order to ensure that they are coordinating and not duplicating efforts. CBOs will coordinate with each other to ensure they are maximizing coverage of HTC communities.