Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 1 of 63 IN THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF MARYLAND LA UNIÓN DEL PUEBLO ENTERO, et al., Plaintiffs, v. No. 8:19-cv-02710-PX-PAH-ELH DONALD J. TRUMP, in his official capacity as President of the United States, et al., Defendants. DECLARATION OF ALBERT E. FONTENOT, JR. Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 2 of 63 I, Albert E. Fontenot, Jr., make the following Declaration pursuant to 28 U.S.C. § 1746, and state that under penalty of perjury the following is true and correct to the best of my knowledge and belief: EXECUTIVE SUMMARY I. 1. I am the Associate Director for Decennial Census Programs at the U.S. Cen- sus Bureau, and I submit this declaration to: • Explain the magnitude, complexity, and planning involved in the 2020 decennial census, including the tightly integrated nature of census operations and processing; • Detail the changes made to the original design in light of the COVID-19 pandemic; and • Discuss the impacts of extending field operations past their current end date of September 30, 2020. II. QUALIFICATIONS 2. I am the Associate Director for Decennial Census Programs, in which ca- pacity I serve as adviser to the Director and Deputy Director of the Census Bureau on decennial programs. In this role, I provide counsel as to the scope, quality, management and methodology of the decennial census programs; provide executive and professional leadership to the divisions and central offices of the Decennial Census Programs Directorate; and participate with other executives in the formulation and implementation of broad policies that govern the diverse programs of the Census Bureau. I have served in this capacity since November 12, 2017. 1 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 3 of 63 3. I began my career with the Census Bureau after retiring from a successful 40-year career as a senior executive in the private sector with midsize manufacturing companies where I was responsible for providing visionary leadership, developing innovative corporate growth and development strategies. I served as Vice President of Marketing, Vice President of Research and Development, and, for the last 14 years, as President and Chief Executive Officer. 4. In addition to a successful corporate career I served as Adjunct Professor in the MBA program in the Keller Graduate School of Management from 2005–2013 where I taught Leadership and Organizational Development, Marketing Management, Corporate Finance, Statistics, and Marketing. I earned a BA in management and MBA in management and finance from DePaul University and Doctor of Ministry in pastoral ministry from Bethel Theological Seminary 5. I served as a as a commissioned officer in U. S. Army and was decorated in combat in Vietnam. After leaving active service, I remained in the US Army reserve attaining the rank of Major. 6. After retirement from private sector corporate management, I began my ca- reer with the Census Bureau in 2009 as a Field Operations Supervisor in Southern California for the 2010 Census. I quickly rose through the ranks and managed the Nonresponse follow-up operations for the 2010 Census as Area Manager responsible for census activities in Los Angeles County, the State of Hawaii, San Bernardino County and 2 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 4 of 63 Riverside County California. After 2010, I served in positions of increasing responsibility as Survey Supervisor, Senior Supervisory Survey Statistician, Assistant Regional Director for the Los Angeles Region, and Regional Director for the Chicago Region. I moved from the field to the Census Bureau headquarters to assume the position as Chief of the Field Division and subsequently Assistant Director of Field Operations, Assistant Director for Decennial Census Operations, then Associate Director for the Decennial Census. 7. From 2012–16, I represented the Field Directorate on the team that devel- oped and wrote the Operations plan for the 2020 Decennial Census. 8. I have in-depth firsthand knowledge about the planning, management, and execution of Census Bureau field operations and effective mission-oriented leadership. I serve as the Chairman of the Census Crisis Management Team; I served as a member of the 2020 Census Design Executive Guidance Group; I am a member of the Census Data Quality Executive Guidance Group; and I chair the 2020 Census Operations Planning Group. Additionally, I represent the Decennial Census Program in our engagement with two of the three committees that advise the Census Bureau: the Census Scientific Advisory Committee and the National Advisory Committee. A COMPLEX DESIGN AND BUDGET FOR THE 2020 CENSUS III. 9. The Census Bureau goes to extraordinary lengths to count everyone living in the country once, only once, and in the right place, including those in hard-to-count 3 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 5 of 63 populations. This is the core mandate of the Census Bureau, and has been the most significant factor informing every decision made in designing, planning, testing, and executing the decennial Census. 10. The Census Bureau’s mandate in conducting the decennial census is to count everyone living in the United States, including the 50 states, the District of Columbia, and the territories of Puerto Rico, American Samoa, Commonwealth of the Northern Mariana Islands, Guam, and U.S. Virgin Islands. To that end, we expend significant funds, efforts, and resources in capturing an accurate enumeration of the population, including those who are hard to count. In particular, the 2020 Census operational design considers population groups that have historically been hard to count, as well as population groups that may emerge as hard to count. 11. The planning, research, design, development, and execution of a decennial census is a massive undertaking. The 2020 decennial census consists of 35 operations utilizing 52 separate systems. We monitor and manage the status and progress of the 2020 Census—the operations and systems in large part using a master schedule, which has over 27,000 separate lines of census activities. Thousands of staff at Census Bureau headquarters and across the country support the development and execution of the 2020 census operational design, systems, and procedures. In addition, the 2020 Census requires the hiring and management of hundreds of thousands of field staff across the country to manage operations and collect data in support of the decennial census. 4 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 6 of 63 12. The 2020 Census operational design is tailored to enumerate all persons, including hard-to-count populations. Almost every major operation in the 2020 Census contains components designed to reach hard-to-count populations. This includes: census outreach, census content and forms design, finding addresses for enumeration, field infrastructure, multiple modes for self-response, Non-Response Follow-Up (NRFU) operations that enumerate households that did not self-respond to the census, and other operations designed specifically for the enumeration of population groups that have been historically hard to count. The best explanation of the many integrated operations designed to reach these populations is set forth in Appendix B to Version 4.0 of the 2020 Census Operation Plan, available at https://www.census.gov/programs-surveys/decennial-census/2020-census/planning-management/planning-docs/operational-plan.html. Examples include: • Verifying address lists using address data provided by community organizations, satellite technology, and in-person address listers checking addresses in communities nationwide; • In-person enumeration using paper questionnaires in areas such as Remote Alaska; • Hand-delivering 2020 Census materials to areas impacted by natural disasters, such as those impacted by Hurricane Michael in Florida; • Conducting a special operation to count persons in “Group Quarters.” Group Quarters include places such as college or university student housing, nursing homes, and corrections facilities; • Working with local partners to identify locations, like shelters and soup kitchens, to best count people experiencing homelessness; and 5 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 7 of 63 • Creating culturally relevant advertisements targeting hard-to-count communities. 13. The Operational Plan cited above is an overall plan that integrates numer- ous sub-operations. Details on most of these sub-operations can be found on our website at https://www.census.gov/programs-surveys/decennial-census/2020-census.html. A partial list of the major operations for which we have posted detailed operations plans includes: a. Count Review b. NRFU c. Integrated Communications Plan d. Intended Administrative Data Use e. Formal Privacy Methods f. Post Enumeration Survey (PES) g. Integrated Partnership and Communications h. Count Question Resolution i. Forms Printing and Distribution j. Response Processing k. Evaluations and Experiments l. Counting Federally Affiliated Americans Overseas m. Field Infrastructure and Logistics n. Data Products and Dissemination 6 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 8 of 63 o. Geographic Delineations p. Local Update of Census Addresses Operation q. Update Enumerate Operations r. Archiving s. Internet Self Response t. Non-ID Processing u. Update Leave v. Address Canvassing 14. The Census Bureau obtained approval under the Paperwork Reduction Act from the Office of Management and Budget for the data collections involved in the 2020 Census. The Operational Plan is a project management document and, as in prior censuses, we did not obtain clearance for it. We presented information about our plans as we developed them in quarterly public Project Management Reviews, and we obtained input on our plans from both our Census Scientific Advisory Committee and National Advisory Committee. We consulted with other agencies throughout the decade about data security, postal delivery, acquisition of records, and the like, though we did not ask other agencies to review or approve our project management plans. 15. We allocate vast resources to ensure as complete and accurate a count as possible. Research and testing, in addition to the Census Bureau’s collective knowledge and experiences, has resulted in an effective approach to reach all population groups. 7 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 9 of 63 16. The complexity and inter-related nature of census operations is echoed in the budget for the 2020 Census. The overall budget estimate for the 2020 Census—covering fiscal years 2012 to 2023—is $15.6 billion. This represents enough funding to successfully complete the 2020 Census in virtually all possible scenarios, including the current challenging circumstances. In fact, the Government Accountability Office (GAO) recently reviewed this budget estimate1 and determined that, as of January 2020, the estimate substantially or fully met GAO’s standards and best practices for a reliable cost estimate in terms of credibility, accuracy, completeness, and documentation quality. It is rare for civilian agencies to be so designated, and we are proud that the Census Bureau has achieved this status. 17. As of this writing, the Census Bureau has been appropriated in aggregate just under $14 billion to use for the 2020 Census, covering fiscal years 2012 through 2020. This is $4.4 billion greater in appropriated dollars than the $9.6 billion actually expended from fiscal years 2002 to 2010 for the 2010 Census. 18. Combined, prior to the COVID-19 pandemic operational adjustments, there remain just over $2 billion in contingency funds that have been appropriated, but which we have not needed to use. With only minimal exceptions, Congress appropriated these This is known as the 2020 Census Life Cycle Cost Estimate (LCCE) Version 2.0. An executive summary of that estimate is publicly available at https://www2.census.gov/programs-surveys/decennial/2020/program-management/planning-docs/lifecycle-cost-estimate_v2.pdf. 1 8 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 10 of 63 funds to allow us to flexibly and quickly respond to any and all risks to the 2020 Census that might be realized and have an impact on the operations. 19. That is exactly what the Census Bureau has done in these challenging times. We have always planned to exhaust any resources necessary to fulfill the Census Bureau’s mission in counting everyone living in the United States once, only once, and in the right place. In all scenarios, the focus of our resources includes the hard-to-count. We have designed and implemented the 2020 Census to enumerate the most willing and able to respond in our most efficient and cost effective manner, thereby freeing the majority of our resources to reach hard-to-count communities using a bevy of techniques specifically tailored to reach them. CENSUS STEP 1: LOCATING EVERY HOUSEHOLD IN THE UNITED STATES IV. 20. The first operational step in conducting the 2020 Census was to create a Master Address File (MAF) that represents the universe of addresses and locations to be counted in the 2020 Census. This operation constitutes a significant part of the 2020 Census, and our plans to enumerate every resident once, only once, and in the right place. 21. A national repository of geographic data—including addresses, address point locations, streets, boundaries, and imagery—is stored within the Census Bureau’s Master Address File/Topologically Integrated Geographic Encoding and Referencing (MAF/TIGER) System. The MAF/TIGER System provides the foundation for the Census Bureau’s data collection, tabulation, and dissemination activities. It is used to generate 9 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 11 of 63 the universe of addresses that will be included in a decennial census. Those addresses are then invited to respond, typically through an invitation in the mail. The MAF/TIGER System is used to control responses as they are returned to the Census Bureau and to generate a list of nonresponding addresses that will be visited in person. Finally, the MAF/TIGER System is used to ensure that each person is tabulated to the correct geographic location as the final 2020 Census population and housing counts are prepared. 22. For all of these reasons, the Census Bureau implemented a continuous pro- cess for address list development in preparation for the 2020 Census. There are two primary components to address list development—in-office development and in-field development. In-office development involves the regular, on-going acquisition and processing of address information from authoritative sources, such as the U.S. Postal Service (responsible for delivering mail to addresses on a daily basis), and tribal, state, and local governments (responsible for assignment of addresses to housing units). In-field address list development involves individuals traversing a specified geographic area and validating or updating the address list based on their observations and, if possible, interaction with residents of the housing units visited. 23. Between 2013 and 2019, the Census Bureau accepted nearly 107 million ad- dress records from government partners. Over 99.5 percent of those records matched to addresses already contained in the MAF, many of which were obtained from the U.S. 10 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 12 of 63 Postal Services’ Delivery Sequence File (DSF). The remaining 0.5 percent of address records from partner governments represented new addresses and were used to update the MAF. In addition, partners submitted over 75 million address points that were either new or enhanced existing address point locations in TIGER. Over 257,000 miles of roads were added to TIGER using data submitted by partners. 24. For the third decade, as mandated by the Census Address List Improve- ment Act of 1994, the Census Bureau implemented the Local Update of Census Addresses (LUCA) Program to provide tribal, state, and local governments an opportunity to review and update the Census Bureau’s address list for their respective jurisdictions. In 2018, participants from over 8,300 entities provided 22 million addresses, of which 17.8 million (81 percent) matched to addresses already in the MAF. The Census Bureau added 3.4 million new addresses to the MAF, nationwide, as a result of LUCA. 25. Between September 2015 and June 2017, the Census Bureau conducted a 100 percent in-office review of every census block in the nation (11,155,486 blocks), using two different vintages of imagery (one from 2009, which was contemporary with the timing of address list development and Address Canvassing for the 2010 Census, and one concurrent with the day on which in-office review occurred) and housing unit counts from the MAF. The 2009-vintage imagery was acquired from a variety of sources, including the National Agricultural Imagery Program as well as publicly available imagery from state and local governments. Current imagery was acquired through the National 11 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 13 of 63 Geospatial Intelligence Agency’s Enhanced View Program, through which federal agencies can access imagery of sufficiently high quality and resolution to detect individual housing units and other structures, driveways, roads, and other features on the landscape. 26. During the in-office review, clerical staff had access to publicly available street-level images through Google Street View and Bing StreetSide, which provided the ability to see the fronts of structures, as if standing on the sidewalk. The technicians categorized blocks as passive, active, or on-hold. Passive blocks represented stability, meaning the technician verified the currency and accuracy of housing data in the office. Active blocks represented evidence of change and/or coverage issues in the MAF. On-hold blocks represented a lack of clear imagery. In these latter two instances, In-Field Address Canvassing was required. At the end of the initial review in June 2017, 71 percent of blocks were classified as passive, suggesting a need for in-field review of only 29 percent of blocks. 27. However, since the 2020 Census was still several years away when In-Office Address Canvassing completed its initial review of the nation, the Census Bureau continued the in-office review to ensure the MAF was keeping up with changes on the ground. The Census Bureau used information from the U.S. Postal Services’ DSF and partner governments to identify areas experiencing recent change and triggered these areas for re-review. Between July 2017 and March 2019, the additional review resulted in 12 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 14 of 63 the categorization of nearly 87.9 percent of the 11.1 million census blocks as passive, indicating a need for in-field review of only 12.1 percent of census blocks. 28. In-Field Address Canvassing occurred between August 2019 and October 2019. Of the 50,038,437 addresses in the universe, fieldwork validated 44,129,419 addresses (88.2 percent). The remainder were removed from the universe as deletes, duplicates, or non-residential addresses. There were 2,685,190 new addresses identified during fieldwork, of which 1,553,275 matched addresses already in the MAF as a result of contemporaneous in-office update processes. In other words, even the hardest to count areas that required fieldwork to verify the addresses, resulted in only a small percentage of additions to the existing MAF. 29. The design for address list development in the decade leading up to the 2020 Census was the most comprehensive in history. Extensive partnerships with tribal, federal, state, and local governments provided multiple opportunities to validate and update the MAF using the most authoritative sources available. This process of continual assessment and update using partner-provided data created a strong foundation on which to implement the use of satellite imagery to validate existing addresses or detect change during In-Office Address Canvassing. This suite of in-office methods allowed the Census Bureau to focus In-Field Address Canvassing resources in the hardest to validate census blocks. 13 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 15 of 63 The MAF/TIGER System created the foundation for the 2020 Census. The 30. Census Bureau believes that the Census Bureau’s MAF/TIGER System is the most complete and accurate in history. V. CENSUS STEP 2: ENCOURAGING SELF-RESPONSE THROUGHOUT THE 2020 CENSUS In order to encourage everyone in the United States to self-respond, the 31. Census Bureau designed, tested, and implemented and Integrated Communications Program, the IPC. The two major components of this program are the ICC, the Integrated Communications Contract, and the IPP, the Integrated Partnership Program. A. 32. Advertising and Media The ICC is the major contract that supports all components of the commu- nications campaign for the 2020 Census. For the 2020 Census, the push to educate people and motivate response to the 2020 Census represented the largest advertising campaign in U.S. government history. 33. The budget for the 2020 Integrated Communications Contract is currently funded at a higher level than in the 2010 Census, adjusted for both inflation and population growth. The cost of the 2010 Census Integrated Communications Contract, in 2020 constant dollars, would be $456 million, while the Census Bureau currently plans to spend approximately $695 million on the 2020 Census Integrated Communications Contract. The $695 million spent on the communications program will mean an 18% increase in per-person spending over the 2010 amount. 14 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 16 of 63 34. To run the ICC in connection with the Census Bureau, a contract was awarded to VMLY&R, a major legacy-advertising firm with over 80 years of experience. Known as Team Y&R, or TYR, by the Census Bureau, the contracting team includes 13 subcontractors. TYR includes firms with expertise in reaching and working with the major audiences that will receive advertising through the media outlets directed toward their population groups, including the Black/African American, Hispanic/Latino, Asian, American Indian and Alaska Native, and Native Hawaiian and Other Pacific Islander populations. By relying on firms with these individual skill sets, the Census Bureau was able to better tailor the media and messaging toward individual groups and gauge the response before going live with the advertising. It also allowed for more creative risktaking, and less of a one-size-fits-all approach. 35. Every part of the 2020 Census communications program was grounded in research. Based on the commitment to being a data driven campaign, beginning in 2018, we extensively researched how people perceived the census and what would motivate them to complete it. Models were developed to predict areas and audiences of low response across the country. These models were then translated into “low response scores” that help the Census Bureau anticipate respondent behavior so that messaging, media, and other communications activities could be deployed to maximize impact. 36. As a result of that research, we mounted a media campaign with stories in news media across the country in print, social, and digital media. The campaign was 15 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 17 of 63 tested in over 120 focus groups across the country, and driven by efforts to reach historically undercounted audiences. More than 1,000 advertisements, in English and 43 other languages, were developed to communicate the importance of responding to the 2020 Census. This compares to roughly 400 separate creative pieces created in 2010. A sample of these creative pieces can be seen on the Census Bureau’s YouTube channel website. 37. On March 29, 2019, the Census Bureau launched 2020census.gov—a key in- formation hub about the census, how to complete it, and how it will affect communities across the country. Three days later, on April 1, 2019, we held a press conference to unveil the campaign platform: "Shape Your Future. START HERE." On January 14, 2020, we unveiled highlights of the public education and outreach campaign. That same day, we began airing ads to reach 99 percent of the nation's 140 million households, including historically undercounted audiences and those that are considered hard to reach. 38. The massive multimedia campaign sought to engage stakeholders and part- ners, support recruitment efforts and the Statistics in Schools program, and communicate the importance of the census through paid advertising, public relations, social media content, and the new web site. This was the first census where we made a significant investment in digital advertising, and spending time and resources targeting online sites including Facebook, Instagram, paid search engines, display ads, and programmatic advertising. 16 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 18 of 63 The push to have a greater digital presence allowed the Census Bureau to 39. reach a mobile audience, tailor messages, micro-target, and shift campaign ads and messages as needed. Online media, particularly search engines and social networking sites, made up a significant portion of digital connections. Nearly every person living in the United States was reached an average of 40 times throughout the campaign, from television, radio, newspaper and online ads, as well as outdoor locations such as billboards and bus stops. The Census Bureau adapted its outreach strategies in response to delayed 40. census operations due to COVID-19, increasing advertising and outreach to specific areas of the country with lower response rates. We quickly adjusted our messaging, pivoting from our original campaign to encourage people to respond online from the safety of their own homes. The use of micro-targeting allowed the Census Bureau to tailor its messaging, including directing appropriate messages to hard-to-reach communities and those who distrust government, both of which have been traditionally undercounted. This targeting continues through NRFU as we encourage the public to cooperate with enumerators. This targeting has allowed us to make each dollar spent on the advertising campaign more effective than in any previous census. B. 41. Partnerships with Community Organizations The second major element of the Integrated Communications Program is partnerships. There are two prongs to the Partnership Program, the National Partnership 17 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 19 of 63 Program that works from Census Bureau headquarters mobilizing national organizations, and the Community Partnership and Engagement Program, that works through the regions at the local level to reach organizations that directly touch their communities. The National Partnership Program and Community Partnership and Engagement Program are more integrated than ever before, and numbers involved for both programs significantly exceed the totals reached in prior censuses. 42. Census partners include national organizations like the National Urban League, the Mexican American Legal Defense Fund, the National Association of Latino Elected Officials (NALEO), the National Association for the Advancement of Colored People (NAACP), and the U.S. Chambers of Commerce. Major corporations also become census partners. At the local level, partners can be churches, synagogues and mosques, legal aid clinics, grocery stores, universities, colleges, and schools. 43. Partners are the trusted voices in their communities; they have a profound impact on those who listen when they say the census is important and safe. We depend on our partners to seal the deal with communities that may be fearful or distrustful of the government. Even with all the Census Bureau’s innovation and improvements to the self-response system, we have learned—and confirmed through research—that when communities and leaders recognize the importance of participating in the census, this message is better conveyed to households within those communities. The best, most trusted information comes from a person of trust. 18 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 20 of 63 VI. CENSUS STEP 3: SELF-RESPONSE 44. The design of the 2020 Census depends on self-response from the American public. In an effort to ensure the most efficient process to enumerate households, the Census Bureau assigns every block in the United States to one specific type of enumeration area (TEA). The TEA reflects the methodology used to enumerate the households within the block. There are two TEAs where self-response is the primary enumeration methodology: TEA 1 (Self-Response) and TEA 6 (Update Leave). 45. TEA 1 uses a stratified self-response contact strategy to inform and invite the public to respond to the census, and to remind nonresponding housing units to respond. Invitations, reminders, and questionnaires will be delivered on a flow basis unless a household responds. These mailings are divided into two panels, Internet First and Internet Choice. Internet First emphasizes online response as the primary self-response option. Mailings to the Internet First panel begin with an invitation letter that alerts the housing unit to the beginning of the 2020 Census and provides the Census ID, 2 the URL for the online questionnaire, and information for responding by phone. 46. Internet Choice is targeted to areas of the nation that we believe are least likely to respond online. Historical response rates from other Census Bureau surveys, A Census ID is a unique identifier assigned to each address in a decennial census; the Census ID is used to track whether an address has self-responded or to track the address through nonresponse data collection and, ultimately through response processing and data tabulation. 2 19 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 21 of 63 internet access and penetration, and demographics are used to determine those areas least likely to respond online. Mailings to the Internet Choice panel begin with an invitation letter that alerts the housing unit to the beginning of the 2020 Census and provides the Census ID and the URL for the online questionnaire, information for responding by phone, and also a paper questionnaire. Housing units in Internet Choice areas have the choice to respond on paper beginning with the initial contact. All nonresponding housing units, regardless of panel, receive a paper questionnaire after the initial mailing and two separate reminder mailings. 47. Update Leave (TEA 6) is conducted in areas where the majority of the hous- ing units do not have mail delivery to the physical location of the housing unit, or the mail delivery information for the housing unit cannot be verified. The purpose of Update Leave is to update the address list and feature data, and to leave a 2020 Census Internet Choice package at every housing unit. The major difference from TEA 1 is that a Census Bureau employee, rather than a postal carrier, delivers the 2020 Census invitation to respond, along with a paper questionnaire. Housing units also have the option to respond online or by phone. 48. Self-response began in March 2020 and will continue until the end of data collection. The total self-response period for the 2020 Census will be longer than the 2010 self-response period. 20 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 22 of 63 VII. CENSUS STEP 4: NONRESPONSE FOLLOWUP (NRFU) 49. NRFU is the field operation designed to complete enumeration of nonre- sponding housing unit addresses. The primary purpose of NRFU is to conduct in-person contact attempts at each and every housing unit that did not self-respond to the decennial census questionnaire. 50. After giving everyone an opportunity to self-respond to the census, census field staff (known as enumerators), attempt to contact nonresponding addresses to determine whether each address is vacant, occupied, or does not exist, and when occupied, to collect census response data. Multiple contact attempts to nonresponding addresses may be needed to determine the housing unit status and to collect decennial census response data. 51. The 2020 Census NRFU operation is similar to the 2010 Census NRFU op- eration, but improved. In both the 2010 Census and the 2020 Census, cases in the NRFU workload are subject to six contact attempts. In both the 2010 and 2020 NRFU, the first contact attempt is primarily an in-person attempt. In the 2010 Census, these six contact attempts could be conducted as three in-person attempts and three attempts by telephone. By comparison, each contact attempt in the 2020 Census NRFU will be either a telephone or an in-person contact attempt (however the vast majority of attempts will be in-person). 21 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 23 of 63 52. In both the 2010 Census and 2020 Census NRFU, if upon the first contact attempt an enumerator determines an address is occupied and the enumerator is able to obtain a response for the housing unit, then the housing unit has been counted, and no follow-up is needed. 53. If upon the first contact attempt, the enumerator is not able to obtain a re- sponse, the enumerator is trained to assess whether the location is vacant or unoccupied. Enumerators will use clues such as empty buildings with no visible furnishings, or vacant lots, to identify an address as vacant or non-existent. 54. In both the 2010 and 2020 Census, a single determination of a vacant or nonexistent status was not sufficient to remove that address from the NRFU workload; a second confirmation is needed. If a knowledgeable person can confirm the enumerator’s assessment, the address will be considered vacant or non-existent and no additional contact attempts are needed. A knowledgeable person is someone who knows about the address as it existed on census day or about the persons living at an address on census day. A knowledgeable person could be someone such as a neighbor, a realtor, a rental agent, or a building manager. This knowledgeable person is known as a proxy respondent. 55. If a knowledgeable person cannot be found to confirm the status of vacant or non-existent, use of administrative records may provide confirmation of the enumerator’s assessment. The Census Bureau does not rely on a single administrative records 22 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 24 of 63 source to determine an address is vacant or non-existent. Rather, multiple sources are necessary to provide the confidence and corroboration before administrative records are considered for use. When used in combination with an enumerator’s assessment of vacant or non-existent, corroborated administrative records provide the second confirmation that a nonresponding address is vacant or non-existent. 56. If upon the first in-person contact attempt, the enumerator believes the ad- dress is occupied but no knowledgeable person is available to complete the enumeration, and the Census Bureau has high quality administrative records from trusted sources, the Census Bureau will use the administrative records to complete the enumeration. We consider administrative records to be of high quality if they are corroborated with multiple sources. Examples of high-quality administrative records include Internal Revenue Service Individual Tax Returns, Internal Revenue Service Information Returns, Center for Medicare and Medicaid Statistics Enrollment Database, Social Security Number Identification File, and 2010 Census data. 57. Regardless of whether administrative records are used as a confirmation of vacancy or non-existent status or for the purposes of enumerating an occupied housing unit, the Census Bureau will, as a final backstop, send a final mailing encouraging occupants, should there be any, to self-respond to the 2020 Census. 23 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 25 of 63 58. The vast majority of nonresponding addresses in the NRFU workload will require the full battery of in-person contact attempts to determine the status of the nonresponding address (vacant, occupied, does not exist) and to collect 2020 Census response data. The full battery of in-person contact attempts also includes the ability to collect information about persons living in a nonresponding housing unit from a proxy respondent on the third unsuccessful attempt to find residents at that address. 59. The Census Bureau designed the 2020 Census NRFU operation to leverage automation and technological advances to control and track the NRFU workload and improve the efficiency of enumerators and the process of collecting census responses. The 2020 Census design for NRFU replaces paper-based NRFU operations used in past Censuses, providing a faster, more accurate, more efficient and more secure means of data collection. 60. The Census Bureau has developed a robust and modern Field Operational Control System to handle many tasks and makes many decisions historically made by individuals. The Field Operational Control System creates daily enumerator workloads using an optimizer that takes into account the location of cases in the workload, the number of attempts a case has received, the time of day to contact an address to maximize response, travel time and mileage from case to case, the location of enumerators, and the hours each enumerator is available to work for a given day. Cases are sorted in the optimal order to ensure enumerators travel to their cases and conduct interview attempts 24 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 26 of 63 in the most efficient manner possible. Workloads are generated each night and transmitted to enumerators to work the next day. 61. Enumerators use government furnished iPhones to receive daily assign- ments and to collect census data. Use of iPhones and the capabilities afforded by automation allow for near real-time case status updates, transmission of response data, and increased enumerator efficiency. Enumerators receive daily workloads of nonresponding addresses, as generated by the optimizer. Enumerators work the addresses in the order prescribed by the optimizer, to determine the Census Day status of the housing unit and, when occupied, to enumerate the housing unit. The data collection application on the iPhone guides the enumerators through their activities for completing interviews. It provides enumerators with scripting for the introduction and the specific census questions and provides extensive help screens for answering questions a respondent may ask during the interview. At the end of each day, enumerators uses case management capabilities on their iPhone to enter work availability for the upcoming five days, as well as to enter/verify payroll information, including hours worked, mileage, and other expenses incurred during their shift. 62. To summarize, the operational design for NRFU evolved over the course of the decade. Use of administrative records, field management structures, systems, procedures, data collection tools and techniques were proven in tests occurring in 2013, 2014, 2015, 2016, and 2018. 25 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 27 of 63 VIII. 63. CENSUS STEP 5: QUALITY CONTROL DURING DATA COLLECTION The Census Bureau is committed to a quality NRFU operation and has in place several programs to monitor and promote a quality data collection, such as the NRFU Reinterview Program, the Decennial Field Quality Monitoring Operation, and the Coverage Improvement Operation. 64. The NRFU Reinterview Program involves contacting a small number of households to conduct another interview—to help us ensure that enumerators are conducting their jobs correctly and are not falsifying responses. We have streamlined this operation as part of the Replan, using information collected from the mobile devices used by enumerators. The data from these mobile devices tell us where the enumerators were physically located while they were conducting the interviews, how long they spent on each question in the interview, time of day of the interview, and other detail data about the interview process. Having this information—which is new for the 2020 Census— provides management with information on how the census takers are doing their jobs, and allows us to select reinterview cases in a targeted fashion. 65. A second quality check program, new for the 2020 Census, is the Decennial Field Quality Monitoring operation. This operation monitors overall adherence to field procedures in order to identify unusual patterns. We used this near real-time data analysis successfully during the Address Canvassing operation in 2019, and it is currently 26 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 28 of 63 active in the NRFU operation. The goal of the program is to identify and investigate potential quality issues. In this program we examine data from individual field representatives and larger scale data, scanning for the possibility of both individual and systemic data quality problems. The program monitors outlier metrics, and produces reports that we analyze on a daily basis. Management staff use these reports to investigate suspicious activities and follow up as needed. 66. Another quality check operation, the Coverage Improvement Operation, seeks to resolve erroneous enumerations (people who were counted in the wrong place or counted more than once) and omissions (people who were missed) from all housing unit data. Coverage Improvement will attempt to resolve potential coverages issues identified in responses from the Internet Self-Response, Census Questionnaire Assistance, and NRFU operations, as well as from the paper questionnaires. 67. The Census Bureau believes that these quality programs (Reinterview, De- cennial Field Quality Monitoring, and Coverage Improvement), taken together, provide a robust quality check for our data collection operations. We believe that our quality program remains an effective deterrent to poor performance, and an appropriate method to identify enumerators who fail to follow procedures. None of these programs, to date, reveals a pattern of substandard data collection. 68. The Census Bureau has also formed a Data Quality Executive Guidance Group that brings together the Census Bureau’s experts in the fields of census operations, 27 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 29 of 63 statistical methodology, acquisition and utilization of administrative records, and in the social, economic and housing subject areas. The group’s mission is to provide direction and approvals about quality assessments of changes to the operational plans and of the 2020 Census data during and post data collection. We plan to release Demographic Analysis estimates of the population in December, prior to the release of the apportionment counts, as previously planned. 69. Finally, as noted by the Director in his August 3, 2020 statement, under the Replan the Census Bureau is working to meet a similar level of household responses as in prior censuses. In short, the Census Bureau has robust programs in place to monitor data quality and has no indication that its NRFU operation is collecting “substandard” data. CENSUS STEP 6: POST-DATA COLLECTION PROCESSING IX. 70. The next major step in the census, after the completion of data collection operations, is post processing. Post processing refers to the Census Bureau’s procedures to summarize the individual and household data that we collect into usable, high quality tabulated data products. Our post processing procedures and systems are meticulously designed, tested and proven to achieve standardized, thoroughly vetted, high quality data products that we can stand behind. 71. The 2020 Census leveraged significant advances in computing technology that have occurred since the 2010 Census. Internet data collection, use of smart-phones 28 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 30 of 63 for field data collection, digital input of phone data collection, and state-of-the-art paper data capture have enabled the Census Bureau to consolidate and prepare the raw census data for processing more rapidly than ever before. Additionally, our computer applications include built-in quality controls that guide respondents through the data collection process and help to ensure higher data accuracy at the point of data input than ever before. 72. The computer processing systems at Census Headquarters have also been optimized in partnership with industry leaders using the latest hardware, database, and processing technology available. Taking advantage of this processing power and speed, we have been able to accelerate our processing time to fit within the re-planned schedule. 73. Nonetheless, post data collection processing is a particularly complex oper- ation, and the steps of the operation must generally be performed consecutively. It is not possible, e.g., to establish the final collection geography for the nation prior to processing housing units and group quarters that are added or corrected during NRFU. Similarly, it is not possible to unduplicate responses prior to processing all non-ID responses. In this sense, the post data collection activities are like building a house – one cannot apply dry wall before erecting the walls, any more than one could lay floor tile before the floor is constructed. There is an order of steps that must be maintained. 74. As part of developing the Replan Schedule, we looked at the possibility of starting the post data collection processing activities on a flow basis and reaffirmed that 29 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 31 of 63 there is little opportunity to begin until data collection operations close everywhere. The only processing step we could adjust in the schedule was initial processing of addresses, which we advanced by 26 days. It is not possible, however, to begin final census response processing in one region of the country while another region is still collecting data. 75. The information below provides additional detail about the post data col- lection activities under the Replan Schedule. A. Incorporate address updates from the field data collection operations into MAF/TIGER Original Dates: February 10 – August 10, 2020 Replan Dates: February 6– September 24, 2020 During the data collection operations, the census field staff can update address and physical location information and add addresses. These updates are incorporated into our address and geo-spatial MAF/TIGER databases. Once updated, each address must be associated to the correct state, county, tract, block group and block. Since it is critical to associate each address to the correct geography, we verify that the address and geospatial updates are incorporated correctly. B. Produce the Final Collection Geography MAF/TIGER Benchmark Original Dates: August 14 – September 1, 2020 Replan Dates: September 5 – 25, 2020 30 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 32 of 63 In preparation for the producing the final collection geography data files needed for producing the apportionment counts and redistricting data products, we create a benchmark of MAF/TIGER, which is a snapshot of the databases. C. Produce the Final Collection Address Data Products from MAF/TIGER Original Dates: September 2 – 14, 2020 Replan Dates: September 26 – Oct 14, 2020 Once the benchmark has been created, the final collection geographic data files are produced and verified. D. Produce and review the Decennial Response File 1 (DRF1) Original Dates: September 15 – October 14, 2020 Replan Dates: October 14– November 8, 2020 The verified final collection geography data are integrated with the response data. Integration of these data is also verified to ensure accuracy. The next set of activities involves the standardization of the collected information. • First we determine the final classification of each address as either a housing units or a group quarters facility. Addresses can change from a housing unit to group quarters and vice versa. Initial status is set at the start of the data collection operations as either a housing unit or group quarters. During the enumeration operations, we collect information that informs us on the classification. For a small number of addresses the classification may change, for example a housing unit may 31 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 33 of 63 have been turned into a small group home. Based on the information collected we determine the status of every address as either a housing unit of group quarters. • Next, we identify each unique person on the housing unit returns. • As part of NRFU operation, we conduct a reinterview of a sample of cases to ensure quality. We incorporate the results of the reinterview. • As part of the Internet self-response option and telephone operation, respondents can provide their data without their Census Identification Number (ID). These cases are assigned an ID which associates them to the final collection geography. • Some group quarters will provide the information electronically. These files can contain duplicate records, so we need to remove the duplicates. • We also determine the population count for all group quarters. • We collect data in many ways, for example on-line, over the phone, on a paper questionnaire, electronic administrative files, and in person using an electronic questionnaire. As a result, we need to standardize the responses across the modes of collection. • Finally, for the operations that collect data on a paper questionnaire, some housing units have more people than can fit on one paper questionnaire. The census field staff will use multiple paper questionnaires to enumerate the house. These continuation forms are electronically linked to form one electronic form. E. Produce and review the Decennial Response File 2 (DRF2) Original Dates: October 14 – November 4, 2020 Replan Dates: November 9 – 30, 2020 Once the previous step has been verified, we incorporate the results from the Self-Response Quality Assurance operation. As part of the group 32 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 34 of 63 quarters operations, we enumerate domestic violence shelters. Their locations and data are high sensitive and are handled with special procedures both in the field and in processing. Their data are incorporated at this point in the process. Finally, for a small number of addresses we receive multiple returns, for example where one person in a house completes the form on-line, and other completes the paper questionnaire. For these cases, we select a form that will be used as the enumeration of record. F. Produce and review the Census Unedited File (CUF) Original Dates: November 4 – 30, 2020 Replan Dates: December 1 – 14, 2020 Once the previous step has been verified, we incorporate administrative records data as the response data for housing units were we do not have an enumeration and have high quality administrative records data. Next we determine the status for every housing unit as occupied, vacant or non-existent. Non-existent units are removed from future processing. For every occupied housing unit, the population count is determined. For each person with write-in responses to the race and Hispanic origin questions, we merge in the information from automated and clerical coding operations. The coding operations assign a numerical value to the writein responses. At this point in the post-data collection activities, for every 33 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 35 of 63 housing unit and group quarter their location (state, county, tract, block group and block) is assigned, their status (occupied, vacant or non-existent) is determined, and in occupied addresses the number of persons is known. In addition, at the person level the demographic information (relationship, age, date of birth, sex, race and Hispanic origin along with write-in code values) and at the housing unit level housing information (tenure) is determined. For the majority of these items, the respondent provided the information. However, for a small number of people and addresses the information may be missing or inconsistent with other provided information, for example the Person 1’s spouse is five years old. The result of these processes is a file that contains records for every housing unit and group quarters along with person records for the people associated with the addresses. Note that some of the demographic information and response to the tenure question may be missing. G. Produce, review and release the Apportionment Counts Original Dates: December 1 – 28, 2020 Replan Dates: Dec 15- 31, 2020 Once the CUF has been verified, the process goes down two paths. The first path is to determine the apportionment counts. Since every housing unit and group quarters has a population count and linked to a state, we 34 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 36 of 63 can tabulation the state level population counts. In addition, we merge in the count of the Federally Affiliated Overseas population and the results of the Enumeration of Transitory Locations for each state. To ensure accuracy in the apportionment numbers, the state counts including the overseas population and apportionment numbers are verified by multiple independent ways. The results of the independent verifications are compared and reconciled, if necessary. CENSUS STEP 0: RESEARCH AND TESTING OF THE 2020 CENSUS DESIGN X. 76. The operational design of the 2020 Census, discussed above, has been sub- jected to repeated and rigorous testing. Given the immense effort required to conduct the census, the importance of the results, and the decade of work by thousands of people that goes into planning and conducting the decennial census, the Census Bureau expends a significant amount of effort to evaluate its planning and design to ensure that its operations will be effective in coming as close as possible to a complete count of everyone living in the United States. Design and testing of the 2020 Census was an iterative process: after each test, we revised our plans and assumptions as necessary. 77. Below are eight significant tests conducted prior to the 2020 Census. Seven of the tests listed below directly contributed to the support of the NRFU operational design or the infrastructure needed to support it. The eighth test pertained to In-Field Address Canvassing. 35 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 37 of 63 A. 2013 Census Test. The 2013 Census Test explored methods for using administrative records and third-party data to reduce the NRFU workload. Key objectives of the 2013 Census Test included: i. Evaluate the use of administrative records and third-party data to identify vacant housing units and remove them from the NRFU workload; ii. Evaluate the use of administrative records and third-party data to enumerate nonresponding occupied housing units to reduce the NRFU workload; iii. Test an adaptive design approach for cases not enumerated with administrative records and third-party data; and iv. Test methods for reducing the number of enumeration contact attempts as compared with the 2010 Census. B. 2014 Census Test. The 2014 Census Test built upon the results from the 2013 Census Test specific to administrative records and third-party data usage to reduce the NRFU workload. Key objectives of the 2014 Census Test included: i. Testing various self-response modes, including the Internet, telephone, and paper, and response without a preassigned census identifier; 36 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 38 of 63 ii. Testing the use of mobile devices for NRFU enumeration in the field; iii. Continuing to evaluate the use of administrative records and third-party data to remove cases (vacant and nonresponding occupied housing units) from the NRFU workload; iv. Testing the effectiveness of applying adaptive design methodologies in managing the way field enumerators are assigned their work; and v. Examining reactions to the alternate contacts, response options, administrative record use, and privacy or confidentiality concerns (including how the Census Bureau might address these concerns through micro- or macro-messaging) through focus groups. C. 2014 Human-in-the-Loop Simulation Experiment (SIMEX). Key findings included: i. Determination that the field management structure could be streamlined and the supervisor-to-enumerator ratios increased; ii. Messaging and alerts within the operational control system provided real-time and consistent communication; and 37 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 39 of 63 iii. Smartphones were usable by all people—even those with little technology experience were able to adjust and adapt. D. 2015 Optimizing Self-Response Test. The objectives of this test included: i. Determining use of digital and target advertising, promotion, and outreach to engage and motivate respondents; ii. Offering an opportunity to respond without a Census ID (NonID Processing) and determine operational feasibility and potential workloads around real-time Non-ID Processing; and iii. Determining self-response and Internet response rates. E. 2015 Census Test. The 2015 Census Test explored reengineering of the roles, responsibilities, and infrastructure for conducting field data collection. IT also tested the feasibility of fully utilizing the advantages of planned automation and available real-time data to transform the efficiency and effectiveness of data collection operations. The test continued to explore the use of administrative records and third-party data to reduce the NRFU workload. Key objectives included: i. Continue testing of fully utilized field operations management system that leverages planned automation and available realtime data, as well as data households have already provided to 38 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 40 of 63 the government, to transform the efficiency and effectiveness of data collection operations; ii. Begin examining how regional offices can remotely manage local office operations in an automated environment, the extent to which enumerator and manager interactions can occur without daily face-to-face meetings, and revised field staffing ratios; iii. Reduce NRFU workload and increase productivity with the use of administrative records and third-party data, field reengineering, and adaptive design; and iv. Explore reactions to the NRFU contact methods, administrative records and third-party data use, and privacy or confidentiality concerns. F. 2016 Census Test. The 2016 Census Test tested different supervisorto-enumerator staffing ratios and incremental improvements and updates to the field data collection software that guided an enumerator through interviews. The 2016 Census Test also allowed the continued evaluation of the use of administrative records to reduce the NRFU workload. Key NRFU objectives included: i. Refining the reengineered field operations; ii. Refining the field management staffing structure; 39 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 41 of 63 iii. Testing enhancements to the Operational Control System and field data collection application; and iv. Testing scalability of Internet and Non-ID Processing during self-response using enterprise solutions. Objectives related to self-response included: i. Testing provision of language support to Limited English Proficient populations through partnerships and bilingual questionnaires; ii. Testing the ability to reach demographically diverse populations; iii. Testing deployment of non-English data collection instruments and contact strategies; and iv. Refining Real-Time Non-ID processing methods, including respondent validation. G. 2018 End-to-End Census Test. The 2018 End-to-End Census Test focused on the system and operational integration needed to support the NRFU operation. Nearly all 2020 system solutions supporting the NRFU operation were deployed. The test also allowed continued evaluation of the NRFU contact strategy. The objectives of this test included: 40 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 42 of 63 i. Testing and validating 2020 Census operations, procedures, systems, and field infrastructure together to ensure proper integration and conformance with functional and nonfunctional requirements. H. Address Canvassing Test (conducted in the fall of 2016). The Address Canvassing Test examined the effectiveness of the In-Office Address Canvassing through the results of the In-Field Address Canvassing. The objectives of the test included: i. Implementing all In-Office Address Canvassing processes; ii. Evaluating the effectiveness of online training for field staff; iii. Measuring the effectiveness of In-Office Address Canvassing through In-Field Address Canvassing; and iv. Integrating multiple information technology applications to create one seamless operational data collection, control, and management system. CURRENT STATUS OF 2020 CENSUS OPERATIONS XI. 78. As of September 10, 2020, over 97 million households, 65.7 percent of all households in the Nation, have self-responded to the 2020 Census. Combining the house- 41 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 43 of 63 holds that self-responded with those that field staff have enumerated under NRFU reveals that as of September 10, 2020 the Census Bureau has enumerated 90.1 percent of the nation’s housing units. 79. The Census Bureau is now roughly 4 ½ weeks into the 7 ½ week schedule for conducting the NRFU operation. Under the Replan Schedule, NRFU is scheduled to last 7 ½ weeks, not 6 weeks as some of Plaintiffs’ declarations state. As of September 10, 2020, we have completed roughly 75.1% of the NRFU workload. We were helped in achieving this result by the fact that we got a “head start” on data collection by beginning NRFU at select offices in July at a “soft launch.” When we began NRFU in all areas on August 9 we had already enumerated over 7.4 million households. Additionally, over 90% of the households in 49 states, Washington D.C., and the Commonwealth of Puerto Rico have been enumerated. 80. While the number of enumerators hired and deployed has not been at the level anticipated, current progress indicates that we will nonetheless be able to complete NRFU before September 30. We currently have over 231,000 enumerators actively deployed, and we are conducting continuous replacement training sessions to increase that number. 42 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 44 of 63 81. The productivity rate for our enumerators thus far is above the planned rate. Our plans assumed a productivity rate of 1.55 cases/hour, and 19 hours/week average hours worked, but as of September 10, 2020 we have experienced a productivity rate of approximately 2.19 cases/hour, and 20.0 hours/week averaged work hours. 82. As the Director stated on August 3, 2020, the Census Bureau intends to meet a similar level of household responses as collected in prior censuses, including outreach to hard-to-count communities. We are, however, facing significant risks to complete all states by this date, due to factors beyond the Census Bureau’s control, such as wildfires in the western part of our country, major storms, resurgence of COVID-19 restrictions and other similar disruptions. REPLANNING THE CENSUS – MULTIPLE TIMES XII. 83. The Census Bureau’s planning for the 2020 Census was, in my professional opinion, excellent. Our plan was comprehensive and thoroughly tested. In March 2020, however, it became clear that COVID-19 was a serious health issue, and we were forced to change our plans around the time we began our self-response operation. 84. On March 18, 2020 the Census Bureau initially announced a two-week sus- pension of field operations to protect the health and safety of our employees and the American public because of the COVID-19 Pandemic. Self-response continued during this period through Internet, telephone and paper questionnaires. On March 28, 2020 the Census Bureau announced an additional two week suspension, until April 15, 2020. 43 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 45 of 63 85. At that time, the career professional staff at the Census Bureau undertook the project of replanning each of the field operations based on our best predictions of when we could safely begin sending staff into the field to interact with the public. On April 13, 2020 staff finalized the plan to adjust field operations, and I presented the plan to the Secretary of Commerce and Department of Commerce management. The plan involved delaying our key high personal contact operations by 90 days. Update Leave, which had started on March 15 and been stopped because of COVID-19 on March 17, would resume pursuant to a new schedule beginning on June 13 and concluding on July 9. In-person Group Quarters operations which had been scheduled from April 2 – June 5 would be rescheduled from July 1–September 3, and our largest field operation, NRFU, which was scheduled from May 13–July 31, would be moved to August 11- October 31. We rescheduled self-response to conclude with the end of Field Operations so instead of ending on July 31 as indicated in the original plan, it was extended to October 31. This schedule required Congress to provide legislative relief from the statutory deadlines of December 31, 2020, for the submission of the Apportionment counts to the President, and March 31, 2021, for the delivery of redistricting data to the states. A request statutory relief from Congress was made for 120 days to enable us to complete the field operations and post enumeration processing. 86. On April 13, 2020, the Secretary of Commerce and the Director jointly an- nounced the new Census Schedule and stated that they would seek statutory relief from 44 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 46 of 63 Congress of 120 additional calendar days. This new schedule set a completion date for field data collection and self-response of October 31, 2020. For clarity, I will refer to this as “the COVID Schedule.” The COVID Schedule assumed Congressional action and called for the delivery of apportionment counts to the President by April 30, 2021 (120 days after the statutory deadline) and redistricting data files to the states no later than July 31, 2021. 87. Once it became apparent that Congress was not likely to grant the requested statutory relief, in late July the career professional staff of the Census Bureau began to replan the Census operations to enable Census to deliver the apportionment counts by the statutory deadline of December 31, 2020. On July 29, the Deputy Director informed us that the Secretary had directed us, in light of the absence of an extension to the statutory deadline, to present a plan at our next weekly meeting on Monday, August 3, 2020 to accelerate the remaining operations in order to meet the statutory apportionment deadline. I gathered all the senior career Census Bureau managers responsible for the 2020 Census at 8:00 a.m. on Thursday, July 30 and instructed them to gather their staff of professional demographers, survey analysts, statisticians, and programmers and begin to formalize a plan to meet the statutory deadline. At that time I consulted with the Associate Director of Communications and we directed that the COVID Schedule be removed from our website while we replanned. 45 We divided into various teams to Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 47 of 63 brainstorm how we might assemble the elements of this plan, and held a series of meetings from Thursday to Sunday. We developed a proposed replan that I presented to the Secretary on Monday August 3. 88. In developing the proposed replan we considered a variety of options and evaluated risk for each suggested time-saving measure. We evaluated the risks and quality implications of each suggested time-saving measure and selected those that we believed presented the best combination of changes to allow us to meet the statutory deadline without compromising quality to an undue degree. The challenge was to shorten the field data collection operation by 30 days, and to conclude the post processing operation in only 3 months, as opposed to 5 months in prior schedules. We began with a review of the status of all field outreach operations, and assessed the impacts of possible revisions on the Census Bureau’s ability to complete those operations within the compressed timeline. The six million housing units in the Update Leave Operation (which provides Census invitations to housing units that do not receive regular US mail) had been completed in early July, and we had received over two million self-responses and the remaining housing units would be moved into the NRFU operation to be visited by enumerators for personal interviewing. The Group Quarters enumeration operation which had begun on July 1st was on track to be completed on schedule by September 3, 2020 and would not be negatively affected by compressing the balance of the Field Schedule. The enumeration of persons staying in transitory locations (Campgrounds, RV 46 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 48 of 63 parks, marinas and hotels without a home elsewhere) was scheduled to be conducted from September 3–September 28. That operation could be conducted as planned within the replan schedule timeline. 89. The COVID-19 pandemic had precluded the Census Bureau from sending staff to conduct our Service Based Enumeration (SBE) operation. SBE is conducted at emergency and transitional shelters, soup kitchens and regularly scheduled food vans and targeted non-sheltered outdoor locations (TNSOL), and is designed to insure that people experiencing homelessness are counted); it was originally scheduled to be conducted March 30–April 2. We had conducted an extensive consultation in May and early June with a panel of 67 national service providers, federal and state agencies to determine the best time frame to conduct this operation to best replicate the weather, migratory behaviors and other factors affecting this population. The overwhelming consensus of the stakeholders, and the input from Census experts, was that the best time to conduct this operation would be mid-late September. Based on that stakeholder consultation we selected September 22-24 to conduct the SBE and TNSOL operations with appointments made with service providers in early September. A review of this operation indicated that we could conducted it in the replan as currently scheduled without disruption. 90. We also reviewed NRFU, our largest and most critical operation. The Cen- sus Bureau had conducted soft launches of all our major operations (during a soft launch 47 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 49 of 63 a small portion of the operation starts early to insure that all the planned and tested systems work as designed under real field conditions with real respondents and actual newly hired temporary employees). The NRFU Soft Launch was planned with six offices that could be safely started based on COVID risk profiles (developed using CDC, HHS, State and Local health guidance), availability of staff, and provisioning of Personal Protective Equipment. The original plan was to begin the operation in one office from each of our six regions starting on July 16th (Cycle 1a) and to follow on July 23rd (Cycle 1b one week later) with six additional offices picked from coastal areas that would be prone to Hurricane risk. As the plan developed we were unable to take offices from all of the areas in the original plan because of high COVID risk and state and local stay at home orders, however we were able to select 6 offices for each cycle and these offices commenced NRFU field operations without incident on the planned dates. On July 14, as the pandemic controls began to be lifted, and our concerns grew over lack of action on a waiver of the December 31, 2020 apportionment statutory deadline, we decided to expand NRFU operations to all offices that could meet the safety, health, and staffing requirements – to start those offices in advance of the initial planned start date of August 11, 2020. We deployed NRFU operations in 35 additional offices on July 30, 2020 and 39 additional offices on August 6, 2020. We then made the decision to pull forward all remaining offices from August 11 to August 9. All ACOs had begun NRFU operations by 48 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 50 of 63 August 9 and we had enumerated over 7.4 million housing units before the Replan Schedule’s official start date of August 11. 91. Concurrent with the early start of NRFU operations, we observed higher levels of overall staff productivity resulting from the efficiency of the Optimizer (a software program that both schedules work for our enumerators and then routes them in the most effective routing). The increased productivity that we observed during the soft launch period was a factor in our ability to design the replanned field operations to end by September 30, 2020. The bonus plan to increase hours also contributed to our ability to create a replan to meet this deadline. We presented the Replan Schedule to the Secretary on August 3, he accepted it, and the Director announced it that same afternoon. For clarity, I will refer to this schedule as “the Replan Schedule.” 92. The Replan Schedule intends to improve the speed of the NRFU operations without sacrificing completeness. Under the Replan Schedule, the Census Bureau has responded to the shortened calendar period for NRFU operations by taking steps to increase the ability of its employees in the field to work as efficiently as possible. This involves increased hours of work per enumerator, spread across the total workforce, to get the same work hours as would have been done under the original time frame. We incentivize this behavior by providing monetary bonuses to enumerators in who maxim- 49 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 51 of 63 ize hours worked, and retention bonuses to those who continue on staff for multiple successive weeks. Successful completion of NRFU is dependent on hours worked, not days worked. We have aimed to improve the effectiveness of our count by continuing to 93. maintain an optimal number of active field enumerators by conducting additional training sessions, and keeping phone and tablet computer devices for enumeration in use for the maximum time possible, thereby decreasing the inefficiency created by training new enumerators. 94. The Census Bureau was able to adopt the Replan Schedule because the de- sign of the 2020 Census allows a more efficient and accurate data collection operation in a shorter timeframe than was possible in the 2010 Census. Improvements that make this possible include use of our route and case optimization software, use of handheld devices, and streamlined processing. Additionally, it is worth noting that largely because of the schedule delays, the self-response period for the 2020 Census will be longer than the self-response period for the 2010 Census. 95. The Replan Schedule also necessitated some changes to the content and tim- ing of our post processing operation. These changes include: • We shortened address processing from 33 to 20 days. This required eliminating 13 days of processing activities that will be deferred until the creation of the redistricting data products. 50 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 52 of 63 • We cancelled the internal independent review of the final list of addresses that will be used to tabulate 2020 Census data (what we call “the MAF Extract”). • We eliminated redundant quality control steps, and the multiple file deliveries that supported those steps, in order to enable a state-by-state flow of deliveries for processing. (Previous procedures delivered data to the next step only when the entire country had been reviewed by multiple teams). • We optimized employee assignments to ensure maximum staff resource usage during this shortened production period – i.e., implemented a sevenday/week production schedule. • We compressed the time allotted for subject matter expert review and software error remediation, cutting 21 days from the schedule. 96. These changes increase the risk the Census Bureau will not identify errors during post processing in time to fix them. 97. Nevertheless, the Census Bureau is confident that it can achieve a complete and accurate census and report apportionment counts by the statutory deadline following the Replan Schedule. The 2020 Census operational design is tailored to enumerate all persons, including hard-to-count populations. 98. The Census Bureau has kept the Office of Management and Budget in- formed about schedule developments for both the COVID Schedule and the Replan Schedule, and has filed nonsubstantive changes that have been published in the Federal Register. OMB was not required to approve the changes to the operational plan, nor did it. As with the 2018 Operational Plan, we did we not ask other agencies to review or approve either the COVID Schedule or Replan Schedule. 51 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 53 of 63 XIII. 99. IMPACTS OF GRANTING A PRELIMINARY INJUNCTION On September 5, 2020 the District Court for the Northern District of Califor- nia entered a Temporary Restraining Order (TRO) enjoining the Census Bureau from “implementing the August 3, 2020 Replan or allowing to be implemented any actions as a result of the shortened timelines in the August 3, 2020 Replan, including but not limited to winding down or altering any Census field operations, until the Court conducts its September 17, 2020 hearing on Plaintiffs’ PI motion.” In compliance with the TRO, the Census Bureau issued the guidance attached as Exhibit A to its field staff. It also took action at the Headquarters level to ensure that no action in the period of September 5–17 would be contrary to the TRO. 100. If a Court grants a Preliminary Injunction, the Census Bureau will need to replan the remaining census operations again. 101. The timing of any Court order changing the schedule is particularly im- portant, as prior to the Temporary Restraining Order, the Census Bureau had already begun to take steps to conclude field operations. As I will explain further, the fact that we are concluding field operations in ACOs that have completed their workload is a normal part of the NRFU operation, and is not specific to the Replan Schedule. 102. The Census Bureau manages its nonresponse follow up operation (NRFU) out of “Census Field Supervisor areas” or “CFS areas” within each of the nation’s 248 52 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 54 of 63 ACOs. CFS areas are supervisory work assignment areas consisting of 4,000-5,500 housing units. As of September 10, 2020, roughly 25% (3,782) of CFS areas nationwide are eligible for what we call “the closeout phase,” and 3,461 are actually in the closeout phase, and roughly 70 have actually reached conclusion. The closeout phase refers to the process of focusing our best enumerators to resolve the remaining cases in that area. CFS areas are eligible for closeout procedures when they cross the 90% completion mark. Under the Replan, all CFS areas would have become eligible for closeout procedures on September 11. This does not mean that all CFS areas would have been moved to closeout procedures on that date, only that regional directors could have made this decision. Under the TRO, we have directed that no CFS area can be moved into closeout procedures until it reaches 90% completion. The Census Bureau is continuing to work across the nation to obtain responses from all housing units, and has not begun closeout procedures for any CFS area with under 90% completion. 103. It is a normal and planned part of the NRFU operation for an ACO to move into the closeout phase and complete operations. We used closeout procedures in NRFU in the 2010 Census and always planned to do the same for the 2020 Census. If we have not wound down in some areas, it is because we are still counting. Some ACOs have greater initial workload, and some started earlier than others –therefore, moving to completion varies by ACO and is a reflection of workload and local conditions and results in 53 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 55 of 63 the allocation of enumerator resources from areas that are complete to areas that require more work. 104. As of September 10, 2020 we are finished with 75.1% of the NRFU field work and over 90.1% of the housing units in the nation have been enumerated and those numbers increase daily. Additionally, 30 states have 90% or more of their housing unit enumeration completed, and in 13 additional states, including Washington D.C. and the Commonwealth of Puerto Rico, we have completed over 85% of the housing units in those states. As we complete areas, staff are offered an opportunity to assist by enumerating in other areas that are not yet complete. Some staff elect that option, others choose not to go outside of their home area, and as their area is completed, they are released. As we complete more field work, the number of staff who are still active and the need for staff declines, and our ability to ramp up is severely hampered. 105. Reduced numbers of field staff would be a barrier to reverting to the COVID Schedule were the Court to rule later in September. Prior to issuance of the TRO, based on progress to date, we had already begun terminating some of our temporary field staff in areas that had completed their work, as is standard in prior censuses. It is difficult to bring back field staff once we have terminated their employment. Were the Court to enjoin us tomorrow we would be able to keep more staff on board than were the Court to enjoin us on September 29, at which point we will have completed enumeration of most of the nations’ housing units and terminated many more employees. 54 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 56 of 63 106. Were the Court to enjoin us, we would evaluate all of the changes we made for the Replan Schedule and determine which to reverse or modify. We would go through each and every aspect of remaining operations and determine how best to use the remaining time to maximize the accuracy and completeness of the census results. 107. Finally, we wish to be crystal clear that if the Court were to extend the data collection period past September 30, 2020, the Census Bureau would be unable to meet its statutory deadlines to produce apportionment counts prior to December 31, 2020 and redistricting data prior to April 1, 2021. The post processing deadlines for the Replan Schedule are tight, and extending the data collection deadline would, of necessity, cause the Census Bureau to fail to be able to process the response data in time to meet its statutory obligations. We have already compressed the post processing schedule from 5 months to only 3 months. We previously planned and tested our post processing systems assuming that we would follow a traditional, sequential processing sequence, and the 3month schedule necessary for the Replan Schedule has already increased risk. We simply cannot shorten post processing beyond the already shortened 3-month period. 108. As I have tried to make clear in this Declaration, the decennial census is a massive, complex, and interrelated endeavor. Particularly troubling is the prospect of continual, conflicting, and evolving court orders from this this and other courts, including appellate courts. While Census Bureau staff have demonstrated considerable resilience and flexibility during this difficult year, some certainty as to the amount of time 55 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 57 of 63 available to conclude data collection and post processing will increase the likelihood of a successful outcome. XIV. Commitment to Transparency and High Quality Enumeration 109. In my role as Associate Director, I remain committed to transparency about 2020 Census operations. The Census Bureau has been posting detailed information on its website about both self-response and NRFU completion progress: https://2020census.gov/en/response-rates/self-response.html https://2020census.gov/en/response-rates/nrfu-completion.html https://2020census.gov/en/response-rates/nrfu.html The 2020 Census is the first to post NRFU workload completion rates, which is now available at the ACO level and may be seen at https://2020census.gov/en/responserates/nrfu-completion.html. By “completion rate” or “completed” we are referring to the percent of housing units in the ACO that we have resolved as occupied with an interview (either resident or proxy), or confirmed as a vacant unit or deleted (as not a housing unit). 110. I have briefed staff for House and Senate leadership every Friday since April (except for August 7), and I have provided a transcribed briefing to Congress. We produce a massive amount of documents and other information to the Office of the Inspector General and the General Accounting Office every week, and these organizations interview Census Bureau staff on almost a daily basis. 56 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 58 of 63 111. In my role as the Associate Director, I remain committed to conducting a high-quality field data collection operation as explained above, and the ultimate goal of a complete and accurate census. I have read the foregoing and it is all true and correct. DATED and SIGNED: ____________________________________ Albert E. Fontenot, Jr. Associate Director for Decennial Census Programs United States Bureau of the Census 57 Case Document 117-1 Filed 09/11/20 Page 59 of 63 Exhibit A 1001 Case Document 117-1 Filed 09/11/20 Page 60 of 63 Fw: Important Information related to Census Bureau's Compliance with today's Federal Court Order James Christy FED) ?@census.gov> Sat 9/5/2020 11:59 PM 1 attachments (205 KB) TRO ORDER 9-5-20.pdf; FYI James Christy U.S. Census Bureau Ho? oeu? census.gov Connect with us on Social Media Shape Your Future Start Here 2020census.gov From: James Christy FED ?@census.gov> Sent: Saturday, September 5, 2020 11:57 PM To: James Christy Subject: Important Information related to Census Bureau's Compliance with today's Federal Court Order Guidance for Census Bureau Field Employees A federal district court for the Northern District of California issued a temporary restraining order at 9:29 PM EDT on 9/5/2020 in the case of National Urban League v. Ross, No. 20-05799. The Order provides that the Census Bureau and the Commerce Department "are enjoined from implementing the August 3, 2020 Replan or allowing to be implemented any actions as a result of the shortened timelines in the August 3, 2020 Replan, including but not limited to winding down or altering any Census ?eld operations, until the Court conducts its September 17, 2020 hearing on Plaintiffs? PI motion.? The Census Bureau and the Commerce Department are obligated to comply with the Court?s Order and are taking immediate steps to do so. The Bureau and the Department are also in the process of preparing additional guidance and will distribute that guidance shortly. Enumeration will continue. The Order is attached for reference. Direct any questions to your regional management. James Christy U.S. Census Bureau Ho? census.gov Connect with us on Social Media Shape Your Future Start Here 2020census.gov DISTRIBUTION: All RDs, DRDs, ARCMs, Area Managers ACOMs 9/6/2020. 12:02 AM 9/7/2020 ail- Jam CE LAF Case otcumen11??U1S./ Fi ed Page 61 of 63 Fw: Guidance for Field Managers James Christy FED) _@census.gov> Mon 9/7/2020 12:02 AM To: FLD Regional Directors?@censusgow Cc: ?@census.gov> 2 attachments (130 KB) Guidance for Field Managers related to Action Required following the 9?5 Court Order.pdf; Guidance for Field Managers related to Action Required following the 9-5 Court Order.docx; The attached documents the actions the Census Bureau will take or has taken to implement the September 5 Temporary Restraining Order. It includes adjustments you will need to make in your operations. Please note that it continues on to a second page. At this point, please do not share this document with your staff. We will discuss this on a conference call tomorrow. Look for an invite shortly. James Christy U.S. Census Bureau ?Cell census.gg Connect with us on Social Media Shape Your Future Start Here 2020census.gov From: Albert Fontenot FED).lcensus.gov> Sent: Sunday, September 6, 2020 11:52 PM To: James Christy Subject: Fwd: Guidance for Field Managers Jamey Let me know when you send to the RDs. Al Albert E. Fontenot Jr. Associate Director, Decennial Census Programs United States Department of Commerce Bureau of the Census Office Office Cell 1/1 Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 62 of 63 Guidance for Field Managers related to Action Required following the 9/5 Court Order The following are actions the Census Bureau will take or has taken to implement the September 5 Temporary Restraining Order enjoining the Bureau and the Commerce Department from: “implementing the August 3, 2020 Replan or allowing to be implemented any actions as a result of the shortened timelines in the August 3, 2020 Replan, including but not limited to winding down or altering any Census field operations, until the Court conducts its September 17, 2020 hearing on Plaintiffs’ PI motion.” Field Managers will take the following actions to come into compliance with the September 5 Order and to preserve continuity of field operations pending further litigation developments: • • • Refrain from releasing data collection staff (enumerators and CFSs) in Area Census Offices where operational progress indicates the area is in Phase 2 or the Closeout phase of the NRFU operation. The Optimizer will continue to assign available work to staff who enter work availability. Continue to release staff for reasons related to performance, quality concerns or conduct, as was appropriate before the Replan. The fixed date of September 11, 2020, as the date on which all CFS areas become eligible for the Closeout phase has been removed. We will implement Phase 2 only after we have enumerated 85% of housing units in a CFS area and we will implement the Closeout phase only when we reach 90%, which allows us to collect more quality data than the pre-Replan thresholds. Do not implement Phase 2 or the Closeout phase until these benchmarks have been met. The workload assigned for follow-up will reflect the following reversions to pre-Replan status: o We are restoring field verification of self-reported vacant housing units. o We will resume making six contact attempts to confirm vacant housing units, instead of the one contact attempt set forth in the Replan. o We will resume making six contact attempts on vacant/delete cases with conflicting information, instead of the one contact attempt set forth in the Replan. o We will resume making six contact attempts on addresses designated for reinterview and SRQA contacts, instead of the one contact attempt set forth in the Replan. o We will reintroduce random sample reinterview cases, as had been used to supplement analytical sample reinterview cases before the Replan. Continue to take the following actions, consistent with pre-Replan procedures: • • • • • Continue to conduct and schedule replacement training in areas that are showing low completion rates. Continue to have staff travel from areas that are at higher levels of completion to areas that are underperforming because of insufficient staffing numbers. This includes within and across regional boundaries. Continue to use the outbound telephone enumeration option for areas that are difficult to reach for in-person interviews, and as a supplement to in-person interviewing activities. Continue using “pop count only” during the final enumeration attempts (which occur in the last part of the Closeout Phase), as we have done in every prior decennial census. Continue to close CFS areas when enumeration of all housing units in that area is complete. Case 8:19-cv-02710-PX-PAH-ELH Document 117-1 Filed 09/11/20 Page 63 of 63 Additionally, continue to utilize enumerator and CFS awards programs as designed to maintain the maximum feasible level of staffing and staff footprint throughout the field.