1. Plans for the Research and Testing Phase of the 2020 Census Presentation to: Council of Professional Associations on Federal Statistics December 3, 2010 Frank Vitrano U.S. Census Bureau

2. Agenda 2010 Census Update 2010 Census Update December Data Releases December Data Releases 2020 Planning – Organization 2020 Planning – Organization 2020 Planning – Research and Testing 2020 Planning – Research and Testing Operational Design Alternatives – Establishing Where to Count – Enumeration – IT and Operational Infrastructure Policy and Communication Challenges Policy and Communication Challenges Conclusion Conclusion 2

3. Completed data capture of all questionnaires to be included in the census counts on September 7. Completed data capture of all questionnaires to be included in the census counts on September 7. Completed closeout of the Local Census Offices on November 12. Completed closeout of the Local Census Offices on November 12. Completed the processing and review of all of the captured data, including count imputation and final review of the counts. Completed the processing and review of all of the captured data, including count imputation and final review of the counts. 3 2010 Census Update

4. Demographic Analysis Population Estimates (December 6) - These will be national totals broken down by age, gender, and black/non black status. These are not 2010 census data. These do attempt to estimate the population of 2010, but it uses birth registration data, death registration data, estimates of immigration and emigration, in and out migration. We've used this method throughout the decade and update these regularly. American Community Survey five year period estimates (December 14) – This release will reflect the years 2005 through 2009. These will be very rich data describing characteristics related to social issues, education, marital status, and so on; economic issues, income, employment, occupation, and housing and demographic characteristics. They'll be released for the first time for areas with less than 20,000 people. 670,000 geographical areas will get their own estimates. These are not 2010 census data. These are from the American Community Survey and will continue to be released on an annual basis. Population Totals for the Country (by December 31)- These data are from the 2010 Decennial Census. The release will also include the state population totals and the updated number of members of the House of Representatives by state. 4 December Data Releases

5. The Census Bureau has proposed two key organizational changes in order to better focus on conducting a more efficient and effective 2020 Decennial Census. 1. New 2020 Directorate – Will be in charge of planning the 2020 Census. 2. Research and Methodology Directorate – Will focus on the evaluation of alternative methodologies for conducting surveys and censuses. 5 2020 Planning - Organization

6. New 2020 Directorate It will start out as a small directorate, and then over the years grow into that organization needed to lead the Census Bureau in completing the 2020. It will start out as a small directorate, and then over the years grow into that organization needed to lead the Census Bureau in completing the 2020. At its inception it will contain the American Community Survey Office and a small research and planning division. At its inception it will contain the American Community Survey Office and a small research and planning division. Much of its early work will be developing an integrated plan for research and testing, in addition to defining effective integration with the new research directorate, the CFO's office, and the IT enterprise activities. Much of its early work will be developing an integrated plan for research and testing, in addition to defining effective integration with the new research directorate, the CFO's office, and the IT enterprise activities. A central focus of the 2020 planning will be efforts to reduce the cost of the census. We also want to gain efficiencies in the planning process by using the ACS as a principal test vehicle for the 2020. A central focus of the 2020 planning will be efforts to reduce the cost of the census. We also want to gain efficiencies in the planning process by using the ACS as a principal test vehicle for the 2020. 6 2020 Planning - Organization

7. Research and Methodology Directorate A sustainable, vibrant research directorate requires that its researchers ally with program staff and collaborate with experts outside the Bureau to work on key problems: A sustainable, vibrant research directorate requires that its researchers ally with program staff and collaborate with experts outside the Bureau to work on key problems:  To contribute to the overall institution, the researchers need to find generalizable knowledge through their work, allowing a solution discovered in one area to be applicable in others.  To contribute to the programs, the researchers need to create real innovation within practical work processes, solving economic and social measurement problems.  To be sustainable as an enterprise resource, the Research Directorate must be a valuable and valued partner to program areas. To complement this internal effort, the National Science Foundation (NSF) and the Census Bureau have announced a research solicitation, which invites teams of researchers from U.S. universities to submit proposals for conducting long-term interdisciplinary research and educational activities on methodological questions of interest to the research community. To complement this internal effort, the National Science Foundation (NSF) and the Census Bureau have announced a research solicitation, which invites teams of researchers from U.S. universities to submit proposals for conducting long-term interdisciplinary research and educational activities on methodological questions of interest to the research community. 7 2020 Planning - Organization

8. Research and Testing 8 2020 Planning

9. 9

10. The Challenge and the Target 10 Notes: Adjusted for projected FY 2010 savings of $1.6 billion; projected cost per housing unit for 2020 assumes no change in design and real cost growth from 1990-2000 and 2000-2010 averaged (58.3%); includes the costs for the 2002-2012 MAF/TIGER Enhancement Program and 2001-2013 American Community Survey. (*) One of the goals of the 2020 Census is to conduct the census at a lower cost than the 2010 Census (per housing unit on an inflation-adjusted basis), while maintaining quality.

11. 11 Four Cost-Drivers Increased population diversity and decreased willingness to cooperate with self-response and non-response follow-up Limited 2000-2009 updating of the Master Address File and TIGER maps led to a design incorporating “last-minute” (2009) updates Failure to link acquisitions, schedule, and budget Demand for absolute accuracy

12. Increased population diversity and decreased willingness to cooperate with self-response and non- response follow-up Expanding, automating, and tailoring self-response methods, including an Internet option, and using administrative records for follow-up + Potential Solutions To The Four Cost-Drivers 12

13. Limited 2000-2009 updating of the Master Address File and TIGER maps led to a design incorporating “last- minute” (2009) updates Continuously updating the address frame and targeting address canvassing and map updates using multiple sources and geographic partnerships, as described in FY 2011 Geographic Support Systems budget initiative 13 Potential Solutions To The Four Cost-Drivers

14. 14 The Geographic Support System (GSS) initiative is an integrated program of improved address coverage, spatial feature updates, and enhanced quality assessment and measurement of address and spatial data and systems. The objective of the GSS initiative is to enable continuous update of the, Census Bureau’s Master Address File/Topologically Integrated Geographic Encoding and Referencing (MAF/TIGER). This will ensure that more accurate and timely address and geospatial data is available for Census Bureau programs such as the American Community Survey, other current surveys, and other small areas estimates programs. The GSS initiative will also support the goal of a targeted (vs.nationwide) address canvassing for the 2020 Decennial Census. To maintain a current and complete inventory of streets, roads, accurate geographic boundaries, and addresses in the MAF/TIGER database the Census Bureau obtains address and geographic information from many external sources, including the U.S. Postal Service, state, local, and tribal governments. The GSS initiative supports ongoing geographic partnership efforts with federal, state, local, and tribal governments as well as commercial entities so that the Census Bureau can acquire the most up to date address and spatial information available. GSS Initiative

15. Failure to link acquisitions, schedule, and budget A rolling approach to planning, budget, and schedule 15 Potential Solutions To The Four Cost-Drivers

16. Failure to link acquisitions, schedule, and budget Strong program management strategies to reduce program risk 16 Potential Solutions To The Four Cost-Drivers

17. Failure to link acquisitions, schedule, and budget Full-Life Cycle Work Breakdown Structure 17 Potential Solutions To The Four Cost-Drivers

18. Demand for absolute accuracy Building consensus among stakeholders regarding tradeoff between accuracy and cost 18 Potential Solutions To The Four Cost-Drivers

19. Operational Design Alternatives Drive the Research Establishing Where to Count Enumeration IT and Operational Infrastructure 19

20. Establishing Where to Count Three Options Supplemented by a full address canvassing in 2019 Supplemented by a targeted address canvassing in 2019  All options: continuous updating of address and spatial information in the MAF-TIGER database along with expanding the list of updating sources No (special) address canvassing for the 2020 Census 20

21. Enumeration Six Options 21

22. Enumeration 22

23. Enumeration 23

24. Enumeration 24

25. Enumeration 25

26. Enumeration 26

27. Enumeration 27

28. IT and Operational Infrastructure Three Options 28

29.  Modification of current approach Local Census Offices494  Modest HQ IT Improvements IT and Operational Infrastructure 29

30. IT and Operational Infrastructure Just a small number of depots (Logistics Support Centers)  Centralized approach to managing through automating field work  Shared IT platforms 30

31.  Centralized management where feasible due to geographic challenges  Shared IT platforms IT and Operational Infrastructure Fewer Local Census Offices 31

32. Six Key Operational Design Alternatives 32

33. Enumeration Research Questions  What is the best mix of modes and strategies by demographics and geography?  What contact frames can be linked to the physical address?  What technology will be feasible for self-enumeration in 2020?  How can household follow-up be improved or eliminated?  What language services and technologies are needed? Self-Response Research 33

34. Enumeration Research Questions  How can we develop and maintain an independent administrative records database? Non-Response Research with Administrative Records  How can administrative records be used to minimize census field work?  What is the quality and the coverage of responses supplied by administrative records? 34

35. IT and Operational Infrastructure Research Questions  What technologies will be available to support IT and field infrastructure?  How can we effectively automate and streamline field operations to take advantage of design changes?  How can we modernize universe control, HQ management, data processing?  What are our workload management system options and how can we best acquire and implement them?  How can we use available technology to improve QC? 35

36. Establishing Where to Count Research Questions  How can we build a usable statistical model for Master Address File (MAF) errors, error components, and their magnitude?  How can the model be used to measure the quality of the Master Address File?  How can we improve Local Update of Census Addresses (LUCA) procedures for compatibility with continual updating, improved quality measurement, and expanded partnerships with local governments, and a targeted address canvassing operation? 36

37. Policy and Communication Challenges  Cost versus Quality Tradeoff  Meeting Stakeholder Expectations  Public Concern about Privacy and Confidentiality  Public Confidence in the Census 37

38. Conclusions   The purpose of the research and testing phase is to determine how much change is possible; critical design decisions are required by 2015 to avoid past problems  To what extent can the 2020 Census target address canvassing in 2019?  How will we collect and integrate respondent information (which response modes will we use and to what extent will we use administrative records)?  What will be the extent of field infrastructure (in other words, how far should we go toward centralization)?  Investment is required early in the cycle to reduce cost and risk as the decade progresses (late design changes add risk and cost)  The more substantial a change to the decennial design that is contemplated, the earlier research must be conducted. 38