[House Hearing, 107 Congress] [From the U.S. Government Publishing Office] THE SUCCESS OF THE 2000 CENSUS ======================================================================= HEARING before the SUBCOMMITTEE ON THE CENSUS of the COMMITTEE ON GOVERNMENT REFORM HOUSE OF REPRESENTATIVES ONE HUNDRED SEVENTH CONGRESS FIRST SESSION __________ FEBRUARY 14, 2001 __________ Serial No. 107-7 __________ Printed for the use of the Committee on Government Reform Available via the World Wide Web: http://www.gpo.gov/congress/house http://www.house.gov/reform U.S. GOVERNMENT PRINTING OFFICE 75-289 WASHINGTON : 2001 ____________________________________________________________________________ For Sale by the Superintendent of Documents, U.S. Government Printing Office Internet: bookstore.gpr.gov Phone: toll free (866) 512-1800; (202) 512�091800 Fax: (202) 512�092250 Mail: Stop SSOP, Washington, DC 20402�090001 COMMITTEE ON GOVERNMENT REFORM DAN BURTON, Indiana, Chairman BENJAMIN A. GILMAN, New York HENRY A. WAXMAN, California CONSTANCE A. MORELLA, Maryland TOM LANTOS, California CHRISTOPHER SHAYS, Connecticut MAJOR R. OWENS, New York ILEANA ROS-LEHTINEN, Florida EDOLPHUS TOWNS, New York JOHN M. McHUGH, New York PAUL E. KANJORSKI, Pennsylvania STEPHEN HORN, California PATSY T. MINK, Hawaii JOHN L. MICA, Florida CAROLYN B. MALONEY, New York THOMAS M. DAVIS, Virginia ELEANOR HOLMES NORTON, Washington, MARK E. SOUDER, Indiana DC JOE SCARBOROUGH, Florida ELIJAH E. CUMMINGS, Maryland STEVEN C. LaTOURETTE, Ohio DENNIS J. KUCINICH, Ohio BOB BARR, Georgia ROD R. BLAGOJEVICH, Illinois DAN MILLER, Florida DANNY K. DAVIS, Illinois DOUG OSE, California JOHN F. TIERNEY, Massachusetts RON LEWIS, Kentucky JIM TURNER, Texas JO ANN DAVIS, Virginia THOMAS H. ALLEN, Maine TODD RUSSELL PLATTS, Pennsylvania JANICE D. SCHAKOWSKY, Illinois DAVE WELDON, Florida WM. LACY CLAY, Missouri CHRIS CANNON, Utah ------ ------ ADAM H. PUTNAM, Florida ------ ------ C.L. ``BUTCH'' OTTER, Idaho ------ EDWARD L. SCHROCK, Virginia BERNARD SANDERS, Vermont ------ ------ (Independent) Kevin Binger, Staff Director Daniel R. Moll, Deputy Staff Director James C. Wilson, Chief Counsel Robert A. Briggs, Chief Clerk Phil Schiliro, Minority Staff Director Subcommittee on the Census DAN MILLER, Florida, Chairman CHRIS CANNON, Utah WM. LACY CLAY, Missouri MARK E. SOUDER, Indiana CAROLYN B. MALONEY, New York BOB BARR, Georgia DANNY K. DAVIS, Illinois ------ ------ Ex Officio DAN BURTON, Indiana HENRY A. WAXMAN, California Jane Cobb, Staff Director Andrew Kavaliunas, Professional Staff Member Dan Wray, Clerk David McMillen, Minority Professional Staff Member C O N T E N T S ---------- Page Hearing held on February 14, 2001................................ 1 Statement of: Barron, William G., Acting Director, U.S. Bureau of the Census, accompanied by John Thompson, Associate Director, decennial census........................................... 26 Letters, statements, etc., submitted for the record by: Barron, William G., Acting Director, U.S. Bureau of the Census, prepared statement of.............................. 30 Cannon, Hon. Chris, a Representative in Congress from the State of Utah, prepared statement of....................... 24 Clay, Hon. Wm. Lacy, a Representative in Congress from the State of Missouri, prepared statement of................... 10 Maloney, Hon. Carolyn B., a Representative in Congress from the State of New York, prepared statement of............... 18 Miller, Hon. Dan, a Representative in Congress from the State of Florida, prepared statement of.......................... 5 THE SUCCESS OF THE 2000 CENSUS ---------- WEDNESDAY, FEBRUARY 14, 2001 House of Representatives, Subcommittee on the Census, Committee on Government Reform, Washington, DC. The subcommittee met, pursuant to notice, at 2:05 p.m., in room 2203, Rayburn House Office Building, Hon. Dan Miller (chairman of the subcommittee) presiding. Present: Representatives Miller, Cannon, Souder, Clay and Maloney. Staff present: Jane Cobb, staff director; Chip Walker, deputy staff director; Mike Miguel, senior data analyst; Andrew Kavaliunas, professional staff member; Dan Wray, clerk; Tim Small, staff assistant; Michelle Ash, minority counsel; David McMillen and Mark Stephenson, minority professional staff members; and Earley Green, minority assistant clerk. Mr. Miller. Good afternoon. A quorum being present, the subcommittee will come to order. We will have some opening statements before we hear from Mr. Barron, and then we'll have a chance to have some clarifications and some questions. Good afternoon. Welcome to the first hearing of the Subcommittee on the Census of the 107th Congress. I would like to welcome our new members to the committee, our new vice chair, Chris Cannon, who is not here yet; Mr. Barr from Georgia. I would also like to welcome Mr. Souder, a returning member. On the other side we now have a new ranking member, Mr. Clay of Missouri, and we welcome back Mr. Davis and Mrs. Maloney, and I wish Mrs. Maloney the best in her new role as a ranking member on the Financial Services Subcommittee. Mrs. Maloney. Thank you, Mr. Chairman. Mr. Miller. On December 28, 2000, Director Prewitt said the following, ``Never have we been so diverse, never have we been so many and never have we been so carefully measured.'' Ladies and gentlemen, today is a good news day. The Census Bureau has announced that the undercount of African Americans and Hispanics have been cut by more than half, and the undercount of American Indians reduced by more than two-thirds. Also, contrary to recent questionable studies, there has been notable improvement in the counting of infants and children. These inroads are remarkable. A great deal of gratitude goes out to the Nation for making this the most accurate and inclusive census in our Nation's history. Particularly the thousands of employees at the Census Bureau are to be congratulated for a job well done. The field staff and enumerators that went door to door did a tremendous job of counting our Nation. Also, the thousands of Census Bureau partners on the State and local level spearheaded an outreach effort that was extraordinarily successful in making people aware and getting people to respond to the census. Congress, too, should be proud of the 2000 census. Congress ensured that the funding levels needed to accurately count America were provided. Congress ensured that the Census Bureau developed an appropriate plan for the 2000 census and stuck to it. While at times some people thought our oversight to be burdensome, it is readily apparent that our focus on improving the census played an important role. Of course, we weren't alone in our oversight role. The Census Monitoring Board, the General Accounting Office and the Commerce Department's Office of Inspector General were also important to the success of the census. I am extraordinarily proud of the work of this subcommittee. The Members and staff have shown tireless dedication to their oversight responsibilities and, I believe, have made positive contributions to the largest of our Nation's civic ceremonies. While the news regarding the success of the census has been good, the political rhetoric surrounding the census threatens to taint the entire effort. In recent weeks the rhetoric of the postsampling community has reached an unfortunate yet familiar tone. I guess when you can no longer argue the facts, there's nothing left except to take cheap shots, race-baiting and name- calling. For months now relentless pressure has been placed on President Bush and Secretary Evans to make some sort of statement regarding the use of the controversial adjustment plan known as sampling. I have maintained that the administration is entitled to gather all the information they deem necessary to make an informed decision on this important issue. I still believe, and I agree with President Bush, that adjustment is not the answer. An actual head count is the best and most accurate way to conduct the census. Adjustment is a Pandora's box, filled with unintended consequences, legal uncertainty and inaccuracy. I would like for a moment to explore some of the unintended and unanticipated social costs of adjusting the census. Think about the people who took the time to fill out their census form and mail it in or those who responded to the enumerators who went door to door. These people did what former Director Prewitt stressed time and again was their civic duty. Under the sampling plan you can do your civic duty, live up to your civic and community obligation and still be counted as less than a whole person. I have always maintained, and this Congress has thoroughly demonstrated, we should do all that we can to count everyone. However, this effort should not come at the expense of those who dutifully answered the call of our Nation. And what about participation in this great civic ceremony in future censuses? It is a very slippery slope. After all, why stand up and be counted when you can sit down and be sampled? Of course, ultimately this issue will end up back in the courts. I firmly believe that sampling for redistricting is as illegal as sampling for apportionment and that the Supreme Court was clear in this regard. However, the Democrats continue to read that decision through rose-colored glasses. The legality of sampling is not one of degrees. A small adjustment is no more legal than a large adjustment. If the Census Bureau were to adjust the numbers, States should be cautious in their use, because their plans could be thrown out in court. Let us not forget, it is the Democrats' spin that tells us that the Supreme Court decision mandates sampling to be used for other purposes except apportionment, but a Congressional Research Service opinion, shortly after the High Court's decision, viewed the decision quite differently saying, ``A closer examination of other parts of the Court's opinion indicates that it did not interpret those other purposes as necessarily including, at least, intrastate redistricting.'' Joining us today will be Acting Director Bill Barron. Mr. Barron was a former career employee of the Bureau of Labor Statistics for 30 years. He was Deputy Commissioner from 1983 to 1998 before joining the Commerce Department as Deputy Under Secretary for the Economics and Statistics Administration in 1998, and he has served as Deputy Director of the Census Bureau since 1999. I don't envy at all the weight that has been hoisted upon his shoulders by a politically motivated rule put into place to remove control over the census from a possible Republican administration. This rule is illegal and poor public policy. It has put blinders on Congress and has hindered our ability to fulfill our constitutional obligation to oversee the conduct of the decennial census. For weeks now the internal committee within the Bureau has been meeting to make a recommendation to the Acting Director on whether or not to adjust numbers. Despite the constitutional, legal and political ramifications of this decision, our requests to observe the deliberations of these meetings have been continually denied. This has not been the transparent census the American people were promised. Furthermore, this decision will not be independently reviewed. Adjusted numbers could be released, and months from now we could find problems. That is what happened with the attempted adjustment in 1990. An adequate independent review of these numbers was conducted to determine the validity. This independent review found errors. Contrary to what others would have you believe, the National Academy of Sciences will be analyzing and evaluating the adjustment for months. While the panel seems to not have a problem with the concept of sampling, they have not concluded that the 2000 census should be adjusted. I worry that adjusted numbers will be released without adequate independent review. At this point I urge caution and restraint by all parties, especially those who have to begin their redistricting process. The old adage, buyers beware, will never be truer when it comes to States using adjusted numbers for redistricting purposes. Mr. Barron, I welcome you before the subcommittee. I know that you must feel like you've jumped from the frying pan into the fire. The decision you are to make regarding adjusting the count should not rest with career civil servants. I personally don't think this is appropriate given all the considerations outside of your expertise. Congress was right in putting this decision at the Cabinet level to begin with. Some would have us believe that this decision is simply one about statistics. Load the numbers in a computer and hit enter, and that's your answer, adjust or don't adjust. These people couldn't be further from the truth. The adjustment decision has far- reaching legal, political and social consequences and cannot be compared to the release of other government statistics such as the trade deficit or unemployment rate. This is not a decision to be made by a group of government civil servants, however well-intentioned, behind closed doors. Issues of this importance should be by those most accountable to the American people. This is why Congress originally and rightfully put this decision in the hands of the Commerce Secretary, a member of the President's Cabinet. Let me end by emphasizing how thrilled all of us should be with the inroads made with the undercount. Significant reductions occurred in the undercount rates virtually across the board for all minorities. The 2000 census is one we all can and should be proud of. Thank you. Glad to have you here. [The prepared statement of Hon. Dan Miller follows:] [GRAPHIC] [TIFF OMITTED] T5289.001 [GRAPHIC] [TIFF OMITTED] T5289.002 [GRAPHIC] [TIFF OMITTED] T5289.003 [GRAPHIC] [TIFF OMITTED] T5289.004 Mr. Clay. Thank you, Mr. Chairman. Mr. Chairman, if I may, let me first welcome our witness. May I use this? If I first can welcome our witness. Mr. Barron, I look forward to hearing your testimony today, and I have the distinct honor as serving as the ranking Democratic member of the Census Subcommittee. I also look forward, Mr. Chairman, to working with you to ensure the fairest and most accurate 2000 census possible. [The prepared statement of Hon. Wm. Lacy Clay follows:] [GRAPHIC] [TIFF OMITTED] T5289.005 [GRAPHIC] [TIFF OMITTED] T5289.006 [GRAPHIC] [TIFF OMITTED] T5289.007 [GRAPHIC] [TIFF OMITTED] T5289.008 [GRAPHIC] [TIFF OMITTED] T5289.009 Mr. Clay. Let me at this point defer to Mrs. Maloney in order for an opening statement, if that's OK with you. Thank you. Mrs. Maloney. Thank you so much, Lacy Clay, and welcome to everyone. As a first point, Mr. Chairman, the Republicans are today claiming victory based upon Census Bureau preliminary results. What may be lost upon some is that this claim of victory is based upon the results of modern statistical methods. It's based upon the finding of the very instrument of accuracy that the Republican Majority has continually opposed, the ACE, the accuracy coverage evaluation program. So it seems that the ACE is good enough for Republican press releases, but not good enough for counting people. Mr. Barron--and I thought Mr. Thompson was going to join you. Is he coming today? Mr. Barron. He is here. Mrs. Maloney. He is coming. Well, happy Valentine's Day to Mr. Miller and everyone here, and I am--I have got to begin by saying that I am sure that this is a very proud day for each of you, the professionals at the Census Bureau, in your long and devoted careers, serving your government. Your operational plan for the census 2000 is already recognized as a great civic milestone in our country's history. I have said from this dais for 2 years that we should let the professionals, the career experts, do their job, and you have done so magnificently. You redesigned the form, overhauled the national address list, created a national advertising program. You created a professional campaign, put in the field 520 local census offices and nearly 1 million part-time workers, designed a national partnership with more than 140,000 private, public and civic groups, and you did it on time and under budget. Congratulations. Your public service is commendable, and, in my view, you are patriots who served during very, very hazardous duty. I believe you deserve all of our gratitude and great applause. While I am encouraged and cheered by the new enthusiasm that I'm hearing from my friends on the Republican side of the aisle that they now have for the census professionals at the Bureau, I would like to offer a cautionary note and remind you of a very familiar refrain that we heard over and over in this hearing room. It was called, ``the rushed census.'' As a parent might advise a child on prom night, I want to give you some advice about the hugs and kisses you're receiving on this Valentine's Day from former critics. Sadly, it was not long ago they accused you of rushing the census. They used this subcommittee to investigate census professionals as individuals. While your enumerators were still at work, they broke a vow not to release preliminary data and held a press conference accusing you of criminal fraud. They sought to search all of your e-mails. They said, you are ``dangerous people.'' They held hostage two Federal budgets, blocked flood relief with antisampling language that the President had to remove with a Presidential veto, and shut down the government in order to pressure you to change your operational plan. From this panel they investigated former Director Dr. Ken Prewitt and questioned him about his political affiliations, about his contributions and about whom he associated with. If these valentine hugs today leave you open to the prospect of going on a date with them, I would really encourage you to be very cautious, and I mean that seriously. It may be the best census ever, but as I am sure you are about to tell us, we do not yet know that. Your numbers today are preliminary, and there is a much more critical story to be told in the details behind these numbers. They may change, they may improve, they may get worse, and now there are some in the administration who are rushing to prejudge these results. Once again preliminary data is being distorted, and once again we are on this side having the burden to call on the Chair to wait for all the facts before jumping the gun. We sincerely hope that it is the best census ever. The key to this question is not just how many were missed in the net calculation, how many were missed in total. Who was missed? Where do they reside? Were some groups missed at higher rates than others? What if a net of 3 million residents missed nationally, but that 1 million were in Florida, would not Florida insist on adjustments? The numbers released today tell us nothing about the inevitable geographic differences in census coverage. A successful effort to interfere with the modern scientific count to achieve a purely partisan advantage of one political party, as press accounts have suggested is under way, most recently in the Wall Street Journal, and I would like to put that article into the record, denies liberty and disenfranchises the unrepresented for 10 years. This is why we call this moment in our history the most important civil rights issue of the decade. I remind this committee of the recent election process in Florida. Those who felt denied access to the polls, or disenfranchised by having their ballot set aside, or stripped of their right to choose their political leadership, they still have recourse. Next year they will be able to go to the polls again in local, State and Federal elections and make their voices heard, and believe me, the whole world will be watching. But to those left out of the census, however, to those people that are disenfranchised in this census by a partisan intervention to ensure that they're not counted or recognized or represented, to them there is absolutely no recourse, not for 10 long years. Billions of dollars in Federal funding will be unfairly spent. Private investment will be redirected to those less deserving. Local planners and school boards will overlook again those uncounted unless we do everything we can to improve the census and ensure that it is as complete and accurate as possible. Let's address the so-called compromise that I keep reading about in the paper, and this compromise is the notion that more accurate adjusted data might be used to distribute Federal funds, but not to distribute Federal power and political power. Mr. Chairman, when it comes to political rights, there can be no compromise. Every American is entitled to his or her fair share of Federal dollars. This so-called compromise offers to give Americans that to which they are already entitled. If the numbers are accurate, they are accurate for every use. We are now on the verge across this Nation of redrawing every political jurisdiction in every State. Only those census numbers which give us the most complete accounting of everyone residing in our country should be used for this purpose. It is distressing to think that there are those who would seek to use this Federal Government, the very instrument of political empowerment and the last sentry for people of color, women and youth, to reverse those gains through manipulation of census numbers. We shall not have ended the poll tax, given suffrage to women, lowered the voting age to 18, ensured all qualified citizens the right to vote, arrested those who intimidated voters at the polls to now just turn away while millions are left uncounted, unrecognized and unempowered. We will ultimately learn if any political influence by this administration is used to interfere with the scientific process of a complete and accurate count. I am reminded, Mr. Chairman, of a very stirring moment that we shared together when we stood with former Director Prewitt in the Ronald Reagan Building and launched the advertising campaign for the census in November 1999. Over 1,000 people joined us. There were dozens of television cameras and hundreds of reporters witnessed when former Director Ken Prewitt announced that with all the modern scientific improvements in the census, that the key technology for the census was very simply a pen, a pencil. The struggle for full voting rights and political empowerment cannot and will not now be undone by the swipe of some political appointee's pen. Thank you. [The prepared statement of Hon. Carolyn B. Maloney follows:] [GRAPHIC] [TIFF OMITTED] T5289.010 [GRAPHIC] [TIFF OMITTED] T5289.011 [GRAPHIC] [TIFF OMITTED] T5289.012 [GRAPHIC] [TIFF OMITTED] T5289.013 [GRAPHIC] [TIFF OMITTED] T5289.014 Mr. Miller. Mr. Cannon for an opening statement. Mr. Cannon. Thank you, Mr. Chairman. Mr. Miller. We welcome you to join us on our subcommittee. We're glad to have you here today. Mr. Cannon. I do have a special interest, of course. The people of Utah feel like they were shorted a seat in this census count, and we need to take a look at that. I would ask unanimous consent to submit an opening statement for the record so we can move on. Mr. Miller. Without objection. Mr. Cannon. Thank you very much, Mr. Chairman. [The prepared statement of Hon. Chris Cannon follows:] [GRAPHIC] [TIFF OMITTED] T5289.015 [GRAPHIC] [TIFF OMITTED] T5289.016 Mr. Miller. Mr. Clay, you want to add anything else? Mr. Clay. Sure. If it's OK, I'd like to submit an opening statement, too. Mr. Miller. Without objection, the opening statements will be included. Before we get started, Mr. Barron, if you would rise and raise your right hand--and Mr. Thompson, do you want to go ahead and get sworn in in case you're called upon? [Witnesses sworn.] Mr. Miller. Thank you. Let the record reflect that Mr. Barron and Mr. Thompson answered in the affirmative, and on behalf of the subcommittee, we welcome you today, and it is indeed a pleasure to have you. I think it's going to be a good news hearing, and I look forward to hearing your testimony. Mr. Barron. STATEMENT OF WILLIAM G. BARRON, ACTING DIRECTOR, U.S. BUREAU OF THE CENSUS, ACCOMPANIED BY JOHN THOMPSON, ASSOCIATE DIRECTOR, DECENNIAL CENSUS Mr. Barron. Thank you, Mr. Chairman. Mr. Chairman and members of the subcommittee, it's indeed a pleasure to testify before you today on the status of census 2000 operations. I have testified previously before you, Chairman Miller, when you were on the Labor-HHS Appropriations Subcommittee and I was the Deputy Commissioner at the Bureau of Labor Statistics, but this is the first time that I've presented formal testimony to you since coming to the Census Bureau nearly 2 years ago. It's indeed a pleasure. As you noted, I'm accompanied by John Thompson, the Associate Director for the decennial census. I'd like to begin my testimony by discussing the success of census 2000, as you requested in your letter of invitation. Over the last 2 years, former Director Prewitt on many occasions very eloquently has reported to you on various operational successes of the census. These successes included the fact that we have completed every planned operation on schedule. We achieved higher than expected mail response rates, and we met our hiring goals, implemented a highly efficient and accurate data processing system and so on. These operational successes culminated in the release on December 28, 2000, 3 days ahead of schedule, ahead of the legal deadline, of State population totals that are to be used for the purpose of apportioning seats in the House of Representatives. At the same time, the Census Bureau announced the resident population of the 50 States and the District of Columbia. The release of the apportionment counts fulfilled the requirement under Article I, Section 2 of the Constitution, but the apportionment numbers and all the successes that we achieved over the last 2 years were always subject to this caveat. While we knew we had conducted a good census operationally, we would not know whether we succeeded in improving the counts until the count of the population results from the Accuracy and Coverage Evaluation survey were compiled. All of the ACE operations have now been completed, and we've produced the first results from dual system estimation; that is, comparing the ACE results to the census. Today we are removing the caveat and announcing that preliminary estimates from the ACE indicate that the census was not only an operational success, but was also successful in improving coverage of the population and in reducing undercounts for some population groups. I would call your attention to the two tables attached to my testimony. Table 1 is preliminary estimated coverage of census 2000 based on the Accuracy and Coverage Evaluation survey. Table 2 is estimated coverage of the 1990 census based on the Postenumeration Survey [PES]. While we are confident of having made improvements in coverage, I again want to emphasize that these are preliminary estimates of the extent of the improvement. We still have quite a bit of review and analysis, and additional refinements will undoubtedly lead to more precise calculations. I also want to add another note of caution. The race and origin groups for 1990 are not directly comparable to the preliminary estimation groups for census 2000 because of the different racial and origin reporting requirements of the two censuses. However, in the interest of openness and transparency, we believe it's warranted to share these data because they do tell an important story about the success of census 2000 in improving accuracy. We can make several observations looking at these tables. Overall coverage has been improved from 1990 to 2000. Significant reduction occurred in the undercount rates for non- Hispanic Blacks and Hispanics. For American Indians on reservations, the undercount in census 2000 will be well below the 12.2 percent figure that was reported for 1990. Also, while there remains a difference in coverage rates for owners and renters, the undercount of renters appears to have been significantly reduced from 1990 to 2000. And coverage for children under 18 years old has improved notably. Finally, for American Indians off reservations, Asians and Native Hawaiians and other Pacific Islanders, there are no comparable data for 1990. While the preliminary estimates indicate there are still undercounts for some groups, I believe we have taken significant steps toward improving census accuracy. The full extent of these improvements will be documented with greater specificity in the weeks ahead, but it's a characteristic of the Census Bureau that even when we have good news to tell, we will be measured and deliberate until we complete our evaluations. I want to emphasize that the committee of Census Bureau professionals charged with recommending whether or not to use the results of the Accuracy and Coverage Evaluation survey to adjust the census is still receiving and reviewing detailed tabulations and reports designed to assess the quality of both the census and the ACE. This committee, the Executive Steering Committee for Accuracy and Coverage Evaluation Policy, is analyzing census data and the ACE results to determine whether the use of the ACE to adjust the census figures would improve results at the level of redistricting. This committee is scheduled to make its recommendations by February 28th, with a final decision expected by March 5th. If I may, Mr. Chairman, I'd like to reflect briefly on some of those elements of the good census that contributed to significant improvements in coverage. First, the census used a multifaceted marketing campaign to aggressively encourage householders to complete and mail back their census forms and to include themselves in the census. This included partnerships, paid advertising, working with schools, improved questionnaire design and providing multiple ways to respond. These cumulative efforts were successful in marketing the census. Approximately two-thirds of households answered the census by mail, exceeding our expectations on mail response. Second, because of our resourceful recruiting plan, research on pay rates and recruiting, and the attractive wages that we could afford to offer because of the full census funding that the Congress provided, we were able to hire and train enough highly skilled temporary staff through the course of the census to complete all operations on time. Third, because of the timely completion of nonresponse followup, we had the time and the resources to conduct eight other operations designed to improve coverage, plus conduct additional reenumeration in selected areas. We called these operations ``quality counts.'' If we had stopped at the end of nonresponse followup, we would have provided an incomplete estimate of the population. The ``quality counts'' operation helped us improve coverage and the census estimates. Fourth, for census 2000, the Census Bureau has used digital imaging and optical character recognition technology for the first time to recognize handwritten answers in addition to blackened circles. This was a vast improvement over previous computer systems and allowed us to process the data faster and introduced quality assurance steps to be sure we had captured the data accurately. Our improved data capture systems, with the ability to capture names, also meant that we could offer multiple options for responding to the census with confidence that we could find and remove duplicate questionnaires. Mr. Chairman, census staff at headquarters, in the regional and field offices and in the processing centers, as well as our partnering contractors, can rightly take pride in these achievements, but many others share the credit for a successful census 2000, and they deserve our thanks: the American public, who helped to exceed expectations on mail response and opened their doors to census enumerators; the temporary census workers, who were dedicated, enthusiastic and resourceful, and who braved tough situations to get the job done; our many partner organizations throughout the country, who worked so hard giving of their time and energy to participate in this great national event; and the Congress, for oversight and your support in providing sufficient resources to offer attractive pay rates for temporary census workers and to conduct the other operations so important to our success. Mr. Chairman, in closing I want to bring to your attention a letter of January 17, 2001, from Mr. Chris Mihm of the General Accounting Office to then Secretary of Commerce Norman Mineta announcing that census 2000 has been removed from the GAO's list of high-risk Federal Government programs. That census 2000 was ever on this list is a reminder to all of us of the great challenges the Census Bureau faced and overcame in conducting a successful census. In the letter Mr. Mihm underscores the need and importance of planning for the 2010 census. Doing so will require completing the 2000 evaluations, and that will shed further light on what worked well or what didn't work well in this census, eliminating the long form from the decennial census by collecting data in the American Community Survey, improving the accuracy of our geographic data base and our master address file, and reengineering the census process through early planning. Mr. Chairman, I encourage you and members of the subcommittee to support 2010 planning, which needs to start right away. This concludes my testimony, Mr. Chairman. I'll now be happy to try and answer any questions that you and your colleagues may have. Mr. Miller. Thank you, Mr. Barron. [The prepared statement of Mr. Barron follows:] [GRAPHIC] [TIFF OMITTED] T5289.017 [GRAPHIC] [TIFF OMITTED] T5289.018 [GRAPHIC] [TIFF OMITTED] T5289.019 [GRAPHIC] [TIFF OMITTED] T5289.020 Mr. Miller. Before we start, I apologize for not having enough seats in the room for everyone, but I see that there are three seats here in the front row that you're welcome to come up to, and there's at least one seat I see in the second row there. There may be some other seats, so if you want to--it's not like classroom or something where you get stuck on the front row or something. And if there are any other empty seats, let someone standing know. Are there any other empty seats anyone can identify? Thank you. Mr. Barron, this is a good news day, and it's been a tough decade, and I agree with you that we need to start planning for the 2010 census. And one of the many objectives we'll have in this subcommittee over the next 2 years is preparing for the American Community Survey and how it addresses the problem of the long form. We look forward to having some hearings on that. Counting overseas Americans is certainly something we're going to look at for the 2010 census, as well. So I look forward to future hearings on these issues. By all accounts so far, the actual head count is reported to be a tremendous success. Let me give you several quotes. An article from last Thursday's San Antonio Express news says, ``Former census Director Kenneth Prewitt said before stepping down from his job last month that Census 2000 is the most accurate census in the country's history.'' On December 28, former Director Prewitt said, ``Never have we been so diverse, never have we been so many and never have we been so carefully measured.'' On December 28th, former Commerce Secretary Mineta said, ``The participation by the people of this country in Census 2000 not only reversed a three-decade decline in response rates, but also played a key role in helping produce a quality census.'' On September 27, 2000, former Director Prewitt said, ``This is truly a marvelous achievement for America and a rousing compliment to the American people for their unprecedented participation in making the census a success.'' In your estimation, was the 2000 census a quality census, as Secretary Mineta suggested? Mr. Barron. Well, Mr. Chairman, I think with the evidence we have today, we are heading closer and closer to being able to say that in all respects this was a very good census. We had operational evidence prior to now, but with the preliminary results today, we're beginning to see that we addressed one of the most serious issues about the census over the past decade, and that's the differential undercount. Now, clearly we've not eliminated it, but if one thought that the differential was a serious matter, and I think it was, if we indeed have reduced it, as these numbers are seeming to indicate, then that's a serious success. So that would be my position, sir. Mr. Miller. What new efforts did the Census Bureau try in 2000 that may have made this a better census than prior ones? Specifically, what steps did the Bureau take to make inroads with the traditionally undercounted populations? Mr. Barron. Well, I think the effort began very early on, Mr. Chairman, with an effort to improve the address list, which I think reduces the undercount generally. I think the evaluations from 1990 indicated that a sizable proportion of the undercount stems from problems with the address list, and very early on the Bureau, working with the Postal Service, sought to address that issue. Later on, especially after the funding was provided, the Bureau was able to engage in some rather innovative efforts, such as the paid advertising campaign--the first paid advertising campaign, in the statistical system that I'm aware of. We had partnership activities throughout the country that I think also were very creative. Those, I think, were the major new, new items, Mr. Chairman, that helped make this census a success. Mr. Miller. Compare the 1990 census and 2000 census with respect to hiring a temporary work force. You mentioned that topic briefly in your opening statement, but why was this aspect of the 2000 census considered more successful than in 1990? In 1990, as in 2000, we had a very full employment economy, and there was great concerns that we were going to have an adequate work force, a quality work force, a good work force, and that was one of the successes. And what would you like to comment on that? Mr. Barron. Thank you for helping me, Mr. Chairman, because that's another area of creative difference between this census and 1990. This time we had a pay policy that was actually based on salary surveys produced by the Bureau of Labor Statistics. I had nothing to do with that, but it was a policy that had been arrived at before I left the BLS, and that enabled the Census Bureau to go out and offer attractive wages, attract and retain people, and avoid, as I understand, one of the more difficult problems encountered in 1990 where, with turnover, the Census Bureau simply didn't have the staff to get the job done. So, you're right, the pay policy had to also be funded, so I thank the Congress for providing the funds. This was another major change between 1990 and 2000 that contributed very significantly to the Census Bureau's success. Mr. Miller. Something else that was successful in the 2000 census was also the use of the computers, reading the data on forms. It was a huge success. Mr. Barron. Right. That's contracted out this time. Mr. Miller. There were concerns early on about it because it was such a huge volume, it had never been attempted before. Mr. Barron. Huge volume, and I think whenever you do something once every 10 years, and when the Census Bureau didn't have the time, Mr. Chairman, and there was no funding to begin early in the decade--and that's why I made the point about the ACS and 2010 planning--I think that's what made the concerns so great about computer systems. Those things typically take time to test, and I think by the time the funding did arrive, we were up against some pretty tight deadlines, but through a remarkable effort through Census Bureau staff and contractors, this worked extremely well. Mr. Miller. You have something else? Mr. Barron. Well, I just wanted to observe, that I'm very concerned about the people that work on this in 2010. I'd like them to work in an environment that's perhaps a little less tense than the environment that those who have worked on the 2000 census will work on, and I simply can't help but think that, you know, in April 2000, the employment-to-population ratio in the United States reached its all-time historical peak of almost 65 percent, 64.8, and what that meant was that this census was conducted during a time when I think people in this country were feeling pretty good about their job prospects. So I think we did benefit from a good economy at the time the census was launched, and I think we need to be humble enough to acknowledge that, and that's just one more reason why 2010 planning needs to get started. Mr. Miller. Hopefully in 2010 there will be better facilities out of Suitland. Mr. Barron. I hope so, too. Mr. Miller. I hope those plans are moving ahead. Mr. Clay. Mr. Clay. Thank you, Mr. Chairman. In 1940, the Census Bureau first discovered that African Americans were missed by the census at rates far higher than the population as a whole. They discovered this when more black men registered for the draft in World War II than were counted by the census. Since 1940, every decennial census has shown the same result. Mr. Barron, you've released figures today that indicate a net national undercount of about 1 to 1.4 percent. That would translate to about 3 to 4 million people net missed, slightly better than the 1990's; 4.4 net missed. Can you elaborate a little bit on that? Mr. Barron. Yes, Mr. Clay, I can. I think when all is said and done, we're going to see that the undercount for 2000 is reduced. Any undercount is unsatisfactory to us at the Census Bureau. Anybody missed is unsatisfactory, but I think the fact that we made progress is something that we need to note. I think we're going to see numerical improvement in the undercount for Blacks and African Americans. I think we're going to see improvement in the undercount for Hispanics, and I think we'll see improvement in the undercount for renters and children. Now, those are important improvements, Mr. Clay, but they don't mean that the institution for a second will say, ``Oh, gee, our work is done, we don't care about this anymore. That's not going to happen. The Census Bureau is going to stay committed to working on the undercount problem as long as there's an undercount in existence. Mr. Clay. Mr. Barron, on a more personal basis, I represent St. Louis, MO. In 1990, you had an undercount of 8,490 approximately. What is your estimation of the undercount of St. Louis, MO, in the 2000 census? Mr. Barron. Mr. Clay, we're not there yet. The numbers that I presented today are just net national totals, and indeed the point you're making is a very excellent one. Part of what we need to do is to continue our analysis to examine this below the national level, and we realize there could be differences that occur there that are very significant. Mr. Clay. Do you have any more detailed information you can share with us about geographic differences in the undercount rates? For example, do large metropolitan areas have higher undercount rates than the suburbs or rural areas of the Nation? Mr. Barron. Not today, but very shortly--on the computer sites we've created to assist all the oversight folks, as well as the National Academy of Sciences, look at our work--there will be more detailed geographic information provided. We don't have that today, though, sir. Mr. Clay. Can you give us any information on the overcounts of any population groups? Mr. Barron. Well, from the tables that are attached to my testimony, we can see that relative to 1990, while there is a net overcount of older folks, it seems to be a little bit better, but again, that's going to be another number that we're going to need to examine more carefully by region in the days ahead as we continue our analysis. Mr. Clay. Final question, Mr. Barron. A lot of energy has been spent talking about accuracy at the block level, but that really misses the point. When you use corrected census counts to create congressional districts of 650,000 or other large aggregations of census tracts and blocks, the corrected counts are generally more accurate than the census counts. Can you explain to us how the errors in census data get larger as you put blocks together, while the kind of error in the ACE gets smaller as you put blocks together? Mr. Barron. As you accumulate data from a sample, and you have more sample observations, you're going to reduce what statisticians call ``variance,'' and so that's going to improve estimates from a system that's generated by something like the ACE. In other cases, such as the census, if you're having a systematic bias in what you're doing in your observations, that's not going to be improved when you start adding geography together. So that, I think, would be the heart of the difference. The fact that data get better as you add it up from a block to a higher level of geography, I mean, that's true of, I think, every statistic that the Federal Government produces, Mr. Clay, so that's not a problem that's peculiar to the ACE. That's the nature of statistics. Mr. Clay. But you do have confidence in the ACE? Mr. Barron. We have confidence in it in concept. We have to look at the particular set of data that we have for this year, for this census and this ACE, and see how that plays out in terms of being able to improve accuracy for redistricting, and that's the challenge that's ahead of us over the next 2 weeks on the committee that's mentioned in my testimony and the chairman's. Mr. Clay. Thank you, Mr. Barron. Thank you, Mr. Chairman. Mr. Miller. Mr. Cannon. Mr. Cannon. Thank you, Mr. Chairman. Mr. Barron, thank you for being with us today. As I mentioned in my opening statement, my real concern--I have several concerns, but one of the core concerns is the difference in the counting of the two groups of foreign citizens. You have, for instance, missionaries and others who were not counted, but Federal employees and dependents who are temporarily serving abroad who were. The failure to treat these two groups of American citizens equally when the Bureau tabulated the apportionment population for the 50 States resulted in Utah being deprived of an additional Member of Congress, to which I think the State is entitled. Are you familiar with the issue of Utah and this count and the missionary issue? Mr. Barron. I'm familiar with how the Bureau arrived at the decisions for coverage in 1990 and 2000, Mr. Cannon, and I understand the circumstances you've described to me. Mr. Cannon. Thank you. Would you explain the legal authority upon which the Bureau relied in setting two apportionment groups? Particularly upon what basis did you decide that the second apportionment group, consisting of those Americans who were living outside of the United States temporarily, would only include Federal employees, military and their dependents? Mr. Barron. Mr. Cannon, I'm not sure that I can give you a legal explanation for the decision. Let me tell you what the substantive basis for it was. Basically, the Census Bureau decided to continue for the 2000 census a practice that was adopted for the 1990 census. Just prior to the conduct of the 1990 census there was a lot of interest emanating from the Congress that was presented to the Census Bureau indicating that it would be a good idea to add military, and given how so many Federal employees also work in the military, that was then expanded to include Federal employees as well. That was announced and done for 1990. After the 1990 census, there did not seem to be any particular controversy or concern about that decision, and so for 2000, as I understand it, it was announced that practice would be continued, and that's how we got to where we are. I'm not aware of a legal process that arrived at that decision, sir. There may be, and I can check on that for you and get back to you, but I'm not familiar with it. Mr. Cannon. We will come back to this on the particular legal issue, but were not people like missionaries, Mormon missionaries, counted in prior censuses? Mr. Barron. My understanding, sir, is that if you go back to the 1920's or 1930's, I think, that there were questions asked of households about members of that household who might be temporarily in residence overseas. I'm not aware that there was a specific question per se dealing with missionaries, but I think they would have been included by the type of question that I just cited where if you went to someone's home and asked perhaps the parents about someone who was temporarily overseas, they would have been included on the census form, and perhaps it would have been noted that they were a missionary. But I don't think there was ever a specific question aimed at missionaries going back over time. Mr. Cannon. Of course, this is not specific, but if you had information about people overseas, you would have had to have made a decision either to ask a question or have the information whether or not to count those people within the State. Have those issues been dealt with over time? Mr. Barron. Well, they haven't been dealt with extensively, Mr. Cannon. As a result of language attached to our appropriations bill, the Bureau is committed to reexamining this issue. This came up prior to the situation with respect to Utah, and we think it clearly is something that warrants review. There are some real challenges. As I think you've heard, there are challenges doing the census here. Move this to an overseas situation, and I'm certain there are some very serious challenges in doing that as well. So we're going to be looking at this and reporting back to the Congress because we understand the unhappiness, and we'd like to find a way to deal with this if we possibly can. I don't know that we can, but we're going to try. Mr. Cannon. Will we have another round, Mr. Chairman? Mr. Miller. Yes. Mr. Cannon. Just finishing up then very briefly, you mentioned the statutory language requiring the Bureau's report to Congress on counting overseas Americans. Obviously this is a matter of urgent concern because we have to choose Congressmen, and there's a lawsuit ongoing. Is there any way that report can be speeded up, do you think? Mr. Barron. I'll look into it, Mr. Cannon. I seriously doubt that it can be done in time to be helpful in the legal process. Mr. Cannon. Thank you. I yield back, Mr. Chairman. Mr. Miller. In June 1999, we had a hearing on this issue of counting of overseas Americans, and I know we'll have another one this year, and this is one of the issues that Mrs. Maloney and I totally agree on. The Bureau felt at that time that it was too late to incorporate all of the overseas Americans, but language was put in the appropriations bill in the Commerce- Justice Committee, which I serve on, to have them come up with a plan. And so, you know, hopefully we'll have a plan later this year, a proposal, and as Mrs. Maloney said, we should at least go through and do a test of that plan sometime in the next few years so we are prepared for the year 2010. Mrs. Maloney. Mrs. Maloney. And just to add on that particular point on which we do agree, I filed legislation yesterday that would compel the Census Bureau to come forward with a plan and to conduct a trial run so that we are ready for the next census. Everyone talks about it, and then between the censuses nothing happens. So we want to make sure we have a trial run, and we have been supportive of each other's legislative efforts in that area. I would like to begin by noting the census document that is on display, which quotes former Director Kenneth Prewitt, and I would like to respectfully request that Dr. Prewitt be called back to speak for himself so that his words are not misinterpreted. It is a very important position that he holds. I would like to place into the record, when there was conflicting statements about his intentions, I called him on February 8th, and I'd like to read what he said to me: For the record, I have said that we believe that the 2000 census was an operational success, but at no time have I stated or characterized the accuracy of the 2000 census. He further stated, only the accuracy coverage evaluation will tell us exactly how accurate the 2000 census was. In fact, I have counseled all interested parties not to characterize the accuracy of the census 2000 numbers until the Census Bureau has reported on the results of the ACE program, end quote. And I'd like to place this into the record and request that he be called in to speak for himself. Would that be appropriate since you are using his words on a flier up there? Mr. Miller. I'll see if we can have time for another hearing on this issue, but go ahead. Mrs. Maloney. Mr. Barron, there have been press reports that the new administration has assured House Republican leaders that, ``sampling is dead,'' and I'd like permission to place into the record the various press reports that have stated that, Mr. Chairman. And I'd like to ask Mr. Barron, who do you think is best qualified to make the decision on the accuracy of the corrected versus uncorrected census numbers, the scientists and professionals at the Census Bureau, the nonpartisan expert professionals at the Census Bureau, or politicians? Who do you think is best equipped to make that decision? Mr. Barron. That's a tough question for a 33-year veteran of the statistical system but in another sense it's not. On statistical issues I think the Census Bureau has the expertise to decide matters of accuracy. On legal and other issues, the Census Bureau doesn't, and that would be my answer, Mrs. Maloney. Mrs. Maloney. Has there been any pressure on you to interfere with your professional work to make a decision one way or the other? Mr. Barron. No, there hasn't. Mrs. Maloney. There has been no pressure at all? Mr. Barron. Absolutely not. Mrs. Maloney. On the professionals, and you think that speaks for the whole Department, you think the professionals are being left alone to do professional work without political interference? Mr. Barron. Yes, ma'am. In fact, the Secretary has encouraged me to go do what I need to do. I think the statements made by Director Prewitt had been made before he left, and combined with the fact that we now have some information showing statistically we've done a good job--not a final word, but some preliminary sense--I think that's impressed the Secretary. So he's encouraged me to do what I need to do. Mrs. Maloney. Mr. Barron, that's encouraging. That's encouraging. Mr. Barron, just based on the numbers that you released today, it appears that the undercount rate for African Americans is twice that of whites, and for Hispanics it is three times that rate, and I personally find that quite troubling. Do you have any comment? What does that say about the ability of traditional census-taking methods to eliminate the differential undercount which your numbers show there is? Mr. Barron. I share your concern that it's troubling, and I think we need to continue to look at and investigate ways to solve that problem. It may be that the only way that can ever finally resolve this is using some kind of survey adjustment, but we don't know that yet. By the same token, Mrs. Maloney, if having an undercount for African Americans two or three times that for Whites is a serious problem, I think we need to note that, and note the fact that it's less than seven times, which is where we were in 1990. So I think we need to note the improvement without slacking off on our commitment to solve the problem, and I think that's where we are. Mrs. Maloney. And could you go through with us what the next steps are in solving the problem? Mr. Barron. Well, there's two things. We're going to continue the process for 2000 of examining whether the use of the accuracy and coverage evaluation survey improves data. The first tasks are the estimates for congressional districts. At some later point we would look at other levels of geography and hopefully address the issue of whether these data would be used for survey controls. Longer term, my reference to beginning planning for 2010, the adoption of the American Community Survey, and using new technology to improve the address list--I think all of those things need to be looked at longer term to see if we can make further inroads into a problem that we agree with you needs to be addressed. Mrs. Maloney. What factors will weigh most heavily on the decision of whether or not to adjust for the undercount? Mr. Barron. I think in its most general sense, Mrs. Maloney, the committee is going to be looking at whether at the level of geography known as a congressional district, the data we've assembled enable us to bring any value added to the estimates we get from an unadjusted census. Now, we have been reviewing and are continuing to review loads of information. I note one of my colleagues described it as hog heaven in terms of the amount of data we have. He's got a different definition of heaven than I do, but I'm in there with him and other colleagues examining data on the quality of the census. We are looking at the quality of the data on the accuracy and coverage evaluation survey. There are techniques and methods that statisticians, very noted statisticians, have developed to evaluate both and how they integrate and relate to one another. So that's what we're doing, and it's painstaking work that's now happening virtually every day, and we have 2 weeks to finish. Mrs. Maloney. But in your preliminary---- Mr. Miller. We'll have another round. Mrs. Maloney. This is a followup just real briefly, and then I know my time is up. In your preliminary numbers, which is a range, it's roughly 3 million net undercount? Mr. Barron. Yes, at the national level. Mrs. Maloney. Approximately. And for this particular census, the ACE is really the only instrument available to correct this problem; isn't that correct? Mr. Barron. That's correct. Mrs. Maloney. OK. Thank you. Mr. Miller. Mr. Souder, thank you for returning to the committee, and you're recognized. Mr. Souder. Thank you. We've been hearing various versions of these arguments for at least 3 years when Speaker Hastert chaired--when the census was under Brouder, and then with Chairman Miller coming and taking this, and I don't think the basic positions are likely to change. I wanted to first congratulate the Census Bureau on their hard work and becoming more accurate in reaching many of the people we haven't been able to reach and to thank all the public organizations who helped with that. I had plenty of frustrations in my area. The Chicago office came down, attempted to address some of those, because I didn't feel they were hiring the minorities they needed to hire to reach the hard-to-reach population, and we tried to address that, and we had meetings with the African American Ministerial Alliance and Hispanic leaders in the community to try to address that question. Indiana still loses a seat, but we attempted to try to address the question, and I'm sure we still have an undercount, and I'm very concerned about that. But as we've seen over and over on sampling, and anybody with a business background understands the kind of basic principles as statistical deviations that in the aggregate may be accurate, but what we're talking here now about is how to disaggregate the data to do congressional districts, State legislative districts, city council districts, potentially even township trustee in Indiana districts based on the population. And the fundamental question is in the hard-to-count population, people who, for example, may be illegal immigrants but still need to be counted, people who may be homeless, people who may be moving from the law, there are--in addition to people who may just have moved and got lost in the system, but there are a lot of the people who don't want to be counted, and they are very hard to count. We see in some school districts in my area or some schools where they will have a transient population in that school of 75 percent of the kids will move across school jurisdictions in a given year. Now, my question is, in the adjustment, would anything guarantee that if we got the--a more accurate total count that those people would actually be in the location where they would be put? In other words, if so and so was missed because they were like in Fort Wayne, we have 125--at one point we had 125 houses they claimed were crack houses, but that really means any given night four to six would be in a count, so if you estimate that any abandoned house that was used as a crack house has four to six people in it, you're going to be off by 120 some houses. So where do you put the additional people, and is there any guarantee in the adjusted count that the people you would add back in actually are a name and a place that would have any accuracy at the township trustee level, or any even reasonable accuracy at a city council level, and even to some degree a congressional level, because certainly that--in the last time we found that people that supposedly had been undercounted in New York were actually in Milwaukee and Indianapolis when they actually studied postcensus. Mr. Barron. Mr. Souder, I don't believe that we can speak to accuracy at these very detailed levels of geography that you mentioned. Indeed, the issue that's before the committee is whether utilizing dual system estimation can improve data for congressional districts, which I think are like 650,000 people. And, by the way, I think at that level, in concept, they could, but at these very detailed levels I'm not knowledgeable, and I don't believe---- Mr. Souder. And, in fact, since the block deviation--census block deviation is greater in the estimating than it is in an actual accurate hard count, could it not be just as likely that the deviation would be greater or it would be more off under adjustments than it would be in unadjusted? Mr. Barron. I think as you begin to add the data, respectfully, I don't think I'd make that conclusion, but for 2000 I have not seen the results yet. Mr. Souder. In other words, as you move it up to a sample size--in other words, you're arguing that potentially, depending--and assuming no gerrymandering in a congressional district, which, of course, most districts are gerrymandered, and therefore it isn't even a statistical block that you can do a deviation off of, but what you're saying is that 650,000, it might be reliable, and what I'm saying is this census data is far more than just that. It breaks down into every kind of unit. I think personally I agree with Chairman Miller that constitutionally we have to have a hard count in the Federal level. What you're really not talking about are congressional districts. I'm pretty sure we'll win the court decision on that. The Constitution says hard count, but at the State and local level your count matters, too, and by putting in adjusted figures that could indeed be off even more than the unadjusted figures, you're potentially skewing the whole political system because some people believe it may give them an advantage. What you're saying is that the bigger the data, possibly the adjusted would have an impact, but that as a statistician, the smaller that is, the less impact that will be. Mr. Barron. Yes. I think at low levels of geography, the census data are pretty noisy to begin with, and I think we would like to demonstrate there's a block level improvement. Mr. Souder. Given the fact that we're moving ahead, what reason is there not to release the unadjusted data now? Mr. Barron. Well, we're not going to know exactly the degree of confidence that we have in the block level data unadjusted until we finish this process. Now, it is true that's virtually complete, but it's not totally complete until we finish the ACE process. So that's the best answer I can give you on that, Mr. Souder. Mr. Souder. Thank you. Mr. Miller. We'll go on to a second round. I would like to refer to a quote by Mr. Prewitt, Director Prewitt, that was made at a public forum with the Brookings Institute, and it's published actually on the Web page of the Brookings Institute, and I think you have a copy in front of you. Let me read the quote from Director Prewitt: ``Just one final word, envision a two-by-two table on which you have a sample survey, the accuracy and coverage evaluation, good and bad. So you've got four possibilities. You've got the possibility of a good census and a good ACE. Under those circumstances would we adjust? Probably not, because why pay the social cost and the confusion and concerns and so forth? If you start with a good census, you don't need to adjust, so you wouldn't.'' I'm looking at another exhibit, the chart that Mr. Prewitt said to visualize. You see that there's only one instance out of four where adjustment might be recommended, and that's if you have a bad census. But you have testified that we have a good census, and former Director Prewitt and Secretary Mineta have indicated that it was not only a good census, but a quality census, a marvelous achievement for America, the most accurate census in the country's history. Mr. Barron, why does a good census make adjustment harder to justify? Mr. Barron. What we're looking at in our committee, Mr. Chairman, is to see whether or not we can make a good census even better, and I don't know how that fits into Dr. Prewitt's description here, but I'm sure he knew that could come to pass as well. So I don't think there'd be any difference between Dr. Prewitt and me on that matter. We think we at least need to examine whether, given that we have a very excellent Accuracy and Coverage Evaluation survey, whether that can add to the accuracy with which we can produce estimates for congressional districts, which are a pretty sizable hunk of geography. Now, that's what we're engaged in right now. We don't have any preconceived notion about that other than, in concept, it ought to work, but in practice we need to see. So that's where we are, and it's really an effort to see if we can make a good census even better. Mr. Miller. But if you have a bad census and good ACE, it's easy to argue the adjustment. You could argue the adjustment, not counting the legal and social costs of it. I mean, it's more than just a statistical question, as I have said. Mr. Barron. Right. Mr. Miller. But with a good census it really is harder to justify adjustment. Mr. Barron. A good census makes our job tougher in terms of discerning, I guess some statisticians might call it, ``signal from noise.'' We have done a very good job. So now as we approach--I don't want to say perfection because I'm not sure anybody will ever get to ``perfection,'' and I don't want the 2010 census people to come find me wherever I am and lynch me-- but I think if you have done a very good job applying the ACE, distinguishing improvement from noise becomes tougher, but that's the task ahead of us. Mr. Miller. Good job of counting 99 percent of the people, and so the question is do we add this other factor or problem. Mr. Barron. Right. Mr. Miller. Let me refer to this quote of Director Prewitt that said that you would not adjust if you had a good census. ``You've got the possibility of a good census and good ACE, under those circumstances would we adjust? Probably not, because why pay the social costs and the confusion concerns and so forth.'' Let's think about the social costs, the confusion and concerns for just a minute. We know that a traditional head count has always been legal, but an adjusted census may not be. We know that a traditional head count has always been constitutional, but an adjusted one may not be. We know that the traditional 2000 census is probably the lowest error in history, but that adjustment would have its own error. Let me stop there and ask you a question about sampling error. Would you explain sampling error and the problems it can introduce into the adjustment process? Mr. Barron. Well, in general, sampling error would be the difference between the value of a population if you could count or enumerate every member of that population, and the noise you get from having selected a sample. I think one of the issues in play here in anticipating what might have happened from 1990 is that our statisticians have designed and we have executed an Accuracy and Coverage Evaluation survey, which is an excellent survey. I mean, it's really been an impressive survey effort with double the sample size of the 1990 PES. So in this particular case we've, I think, confronted head on the issue of sampling error as applied to the issue in front of us for adjustment in 2000. We're not done looking at it yet, but I think variance or sampling errors are probably not going to be an issue for us as we look at this adjustment issue. Mr. Miller. In the census that we've just conducted, there is no sampling error. Mr. Barron. There are other kinds of errors. Mr. Miller. But they're not sampling errors. Mr. Barron. Right. Mr. Miller. But there's no sampling error. I mean, we have the nonsampling error. Mr. Barron. Right. Mr. Miller. When you adjust, you introduce sampling error; is that correct? Mr. Barron. Yes, there would be an error. And if we can't discern improvement from no improvement, then that would be a reason for us not to adjust. On the other hand, if we think we can improve the estimates, then from a statistical perspective, not these other perspectives, but from a statistical perspective, we would recommend adjusting. Mr. Miller. By adjusting you're introducing the whole error issue of sampling error to your data. Mr. Barron. Right. Mr. Miller. Mr. Souder was talking about the block level data all the way up to the State population and the Nation's population. I mean, you're introducing a whole new set of problems, and that's part of the tough choices that you're facing, but---- Mr. Barron. It is. Mr. Miller. Not counting the legal and social cost problems, but you're introducing a totally new set of problems that have never been in the census as sampling error. Isn't that a serious concern, introducing sampling error into the whole set of numbers? Mr. Barron. We will be able to frame that and describe that for people, Mr. Chairman. We're still in process, so I haven't reached a conclusion on this, but my sense is that sampling error is probably the least of our concerns because we will be able to frame that and describe that to people as we do for the Current Population Survey or other sample surveys that the government does. So I'm not--you know, I don't believe that that's a problem for us here. Mr. Miller. OK. Thank you. Mr. Clay. Mr. Clay. Mr. Barron, as you testified, the Bureau released the apportionment total for the States on December 28th. Do you have any comments you'd like to share with us, or did you see any surprises in the data as it came back? Mr. Barron. I suppose the only surprise, Mr. Clay, was that the total population estimate was bigger perhaps than people had anticipated, bigger than some of the estimates from other demographers, including our own. So I think that was perhaps a bit of a surprise, but I think we were very pleased with the results, and we had worked with folks from the States to look at those numbers before they were published, and I think we were very pleased. Mr. Clay. Mr. Barron, have you discussed with the Secretary the rule currently in place which governs the decisionmaking process on whether or not to release corrected data? Mr. Barron. Only in a very, very general way. I have spent much more time talking to the Secretary about the building in Suitland that we inhabit that needs improvement, and I have talked with him about the data that we released today not from the perspective of the rule at all, but from the perspective of I think we've done a good job, you should know about this. I will also acknowledge that in briefing him on this, which was at my request, by the way, my sense was the data was going to leak, and I guess it did. Mr. Clay. And so you are operating under the assumption that those rules will remain in place until your process is completed? Mr. Barron. Mr. Clay, I have no information to the contrary. Mr. Clay. Thank you. Thank you. Mr. Miller. Mr. Cannon. Mr. Cannon. Thank you, Mr. Chairman. You mentioned earlier that you were part of discussions about the legality--or not legality, but the practicality of counting people who are temporarily overseas. Who else was part of those discussions? Mr. Barron. I'm sorry, Mr. Cannon, if I misspoke. I was describing to you my understanding of discussions that took place. I was not involved in those personally. I was recounting history to you as best I understand it. Mr. Cannon. But you heard that, I take it, from other Bureau personnel? Mr. Barron. Yes. Actually on the issue of adding the military and Federal civilians in 1990, there is some correspondence on that between the then Secretary of Defense, who I believe is now the Vice President, and others encouraging the Census Bureau to do so. Mr. Cannon. You don't know if the issues of the violation of the apportionment clause or the free exercise clause or the equal protection clause of the Constitution were every considered in that process then? Mr. Barron. No, sir. I'm sorry, I can't speak to that. Mr. Cannon. And on the same line you--were you or Director Prewitt ever advised of a potential violation of either the Constitution or those clauses or the restoration of the Religious Freedom Act as it related to temporary missionaries in particular? Mr. Barron. No, sir, I'm not aware of any briefings or conversations on that. Mr. Cannon. Apparently someone has said somewhere that counting Americans temporarily overseas would be infeasible, unreliable, and prohibitively expensive. Do you know if there's any factual basis in the Department for that conclusion? Mr. Barron. Well, there was a hearing on the issues surrounding counting Americans overseas last--I think it was last--I'm sorry if my memory's failing me on exactly when that was. Mr. Cannon. I'm more concerned not with the hearings, but with the internal discussions in the Department. Mr. Barron. It is extremely difficult. So the flavor of those remarks is something that I think I have heard, because it's a daunting task to figure out how we might count Americans overseas for the most part. Mr. Cannon. The daunting task would include the difficulty of counting military and their dependents and other Federal workers? Mr. Barron. At least in that case there's a source we know about, and I think in some other cases, probably the Mormon Church, they probably have very good records. In lots of other cases, there are no records. Mr. Cannon. Did you ever communicate with the Mormon Church or any other churches about their records of Americans overseas temporarily? Mr. Barron. No, sir. Mr. Cannon. Let me jump to another matter, which I'm sure will please the chairman and others. I am disturbed that there's possible politicalization of the Bureau or of staff there with apparently their own agenda opposed to that of the administration and the congressional majority. Last week, for instance, a Census Bureau staff member in the public affairs office sent to various State Governor liaisons a highly political press release by Congresswoman Maloney. It actually was an attack on the Bush administration. I think there's an attached copy of that, which is, you know, fine for--you know, Congresswoman Maloney and I disagree on issues, and that's appropriate here, but probably not for the Bureau. Is it--is there--do you have policy dealing with what kinds of things should go out from the Bureau? Mr. Barron. There is a policy, Mr. Cannon. First, let me say that I apologized for that, and I apologize again. That was inappropriate. The policy is that we use that mechanism to distribute Census Bureau press releases and products, and since that matter was brought to my attention, we emphasized that policy. I think most people understand that. We had a lapse in judgment here, which I regret. Mr. Cannon. And what was the response with that employee? Was he reprimanded? Was the matter investigated? Mr. Barron. The matter has been investigated. I must say to you I've had such an exciting week, I don't know much beyond that, but I am confident that it's not going to happen again. Mr. Cannon. Do you know if anything similar has occurred with the employee who sent out that press release in the past, or is this just a one time matter? Mr. Barron. I don't know, sir. My understanding is this was a one-time matter. Mr. Cannon. Great. Thank you. I appreciate your being here, your candid answers and very clear answers, and I yield back, Mr. Chairman. Mr. Miller. Mrs. Maloney. Mrs. Maloney. Thank you. Thank you, Mr. Chairman. Mr. Barron, I would like to go through a series of questions and see if you have a number to give us, and if you don't know, then for the record just merely state that you don't know. First of all, do you today know the total number of erroneous enumerations? Mr. Barron. No, ma'am, we don't. Mrs. Maloney. You do not know? Mr. Barron. We do not know. Mrs. Maloney. Do you know the total number of duplicates? Mr. Barron. We have some estimates of duplicates, and I think if you give me a second, I can give you that number. Now, duplicates does cover a number of things, but because in this census we went out of our way to be inclusive in terms of constructing our address list, we anticipated that we could have a problem with duplicate addresses, and our estimate of this is that we had about 2.4 million of those. And having done some, I think, very creative and innovative work, we looked at that number, and we have deleted about 1.4 million of them as being duplicates and have maintained or retained about 1 million records. There's probably error in this somewhere, Mrs. Maloney, but we're pretty confident that we've done a good job with dealing with this issue. Mrs. Maloney. Is the ACE program the only way to correct these duplicates? Mr. Barron. The ACE is a very excellent way of dealing with this problem, and it also in its estimation structure deals with the variance that's attributed to duplicates as we go through the census process. Mrs. Maloney. Do you know the net undercount for New York, California or Florida? Mr. Barron. No, ma'am, we do not know that at this point. Mrs. Maloney. Do you know the total number of African Americans missed? Mr. Barron. We have a range estimate, but as our press release today indicates, we don't have a precise number for that now, but we do have a range. I can give you the numbers associated with the range if you wish. Mrs. Maloney. I have that, but the point is you just have a range, but it could be defined more with the ACE program, correct? Mr. Barron. No. As a result of the ACE, this is going to be the range. So we're pretty confident that the final number is going to be within that range, and we will have a point estimate as we did---- Mrs. Maloney. There's no way that you're going to define this any more; you're just going to have a range? Mr. Barron. No. We'll end up with what we had in 1990, which is a point estimate. There always is a range surrounding a point estimate. Since we're still in process, we thought we would do what was done in 1990, by the way. In 1990, when this information was first produced, a range was published, and we've replicated that process this time. Mrs. Maloney. Do you have the total number of children missed? Mr. Barron. We don't have a number now, no. Mrs. Maloney. Do you know whether the improvements in counts of minorities took place in cities, suburbs or rural areas, or what the differentials are in those counts and those geographic areas? Mr. Barron. I don't, but the information that we're going to put up for your staff and others to look at as part of our oversight of our process in conducting the final stages of the census will be available very shortly. Mrs. Maloney. So you don't have that information now, but it will be available shortly? Mr. Barron. Yes, yes. Actually on all these things it will be available shortly, but that one is coming up very shortly, probably within a day or so. Mrs. Maloney. About how many different metrics or measures of accuracy will the ESCAP committee consider before making a decision on whether or not to correct the raw count? Mr. Barron. I don't have a count for you, Mrs. Maloney. There's a pretty strong array of measures that statisticians have arrived at to examine both the accuracy of the census, the accuracy of ACE, and when you try to adjust the census count with the ACE, whether you're in effect adding value or not. So there is a set of measures, but I have not counted. Mrs. Maloney. So at this point there's still a considerable amount that the Census Bureau does not know, so you need to complete your work basically? Mr. Barron. We need to complete our work. Mrs. Maloney. And if you were asked today to give your recommendation on whether to adjust, would you have enough data to make that decision? Mr. Barron. Today? Mrs. Maloney. Uh-huh. Mr. Barron. No, ma'am. I think we have enough data coming, but what we've not done is completed our analysis and review of that data, and that's the real challenge we have before us over the next 2 weeks. Mrs. Maloney. And you will have that information. So you're still very much in a work in progress which you need to really get more information, to get more accurate information? Mr. Barron. Very much so. Mrs. Maloney. I just want to note that in this subcommittee press release, Dr. Prewitt is quoted not once, twice, not three times, but four times, and I'd just like to reiterate my request that he be called to this committee to speak for himself. Thank you. Mr. Miller. We all use quotes from different sources. This was one that was pulled off the Web. It's a public forum. It's kind of a funny thing to question, you don't get a quote to use. But one of the issues that was brought up was the count of children in 1990 was a problem, and I believe we've had a great success on children. Would you comment about that first? Mr. Barron. Well, while we don't have the final numbers, and our press release does deal with a range, it looks like we've made very significant improvement in what's been recognized as a serious problem, which is the undercount of children. Last time the undercount was over 3 percent, and this time we estimate that when our efforts are completed, the undercount of children is going to be somewhere within the range of 1.2 to 1.9 percent. So that's a big improvement over the past. It is still there. It's still a problem, and again, just because we've improved something doesn't mean I'm saying that we're going to rest on our laurels. But it's an improvement, and we ought to, I think, recognize it. Mr. Miller. That is a huge improvement, and we're very pleased about that. You're operating under a regulation that specifically delegates the adjustment decision from the Secretary of Commerce to the Census Bureau Director, correct? Mr. Barron. Yes. Mr. Miller. For the record, is a position of Census Bureau Director a political position? Mr. Barron. Yes, it is. Mr. Miller. Are you a political appointee or career civil servant, and how long have you served in the government? Mr. Barron. I'm a career civil servant, and I began in the government June 17th, 1968. Mr. Miller. For the record, are you serving as the Acting Census Bureau Director, or have you been confirmed as Census Director? Mr. Barron. I'm an Acting Census Bureau Director. Mr. Miller. Now, since there is no official Census Bureau Director, how does the law work? Does the adjustment decision now fall to you, the Acting Director, a civil servant? Mr. Barron. My understanding is that it does, Mr. Chairman. My further understanding is that there was a law passed within the past several years that dealt mostly with recess appointments, but it somehow reinforces the idea that I would be an Acting Director. I'm not sure I understand that. I'm just doing the best I can, but I think the role I have been playing in census 2000, at least in terms of managing us through to completion, means I can do that, but in terms of Directors and Acting Directors I don't define those things. Mr. Miller. The census numbers you released in December were unadjusted State-level population totals for apportioning the 435 seats in the House of Representatives. The numbers to be released next month are the population totals at various geographic levels of redistricting, for redistricting. The million-dollar question before you is whether these numbers should be adjusted. What are the various geographic levels census data is distributed for, and how many people are in those different geographic levels, starting with blocks and working our way up? Mr. Barron. I'm not sure I'm going to be able to answer that question completely. We do start with blocks, tracts, and there are various geopolitical units going up. Mr. Miller. What's the average size of a block and a tract, for example? Mr. Barron. I believe a block is--you know, I don't know. Mr. Miller. Mr. Thompson maybe could even chime in, the average block--how many tracts. Mr. Thompson. About 9 million blocks. Mr. Miller. About 9 million blocks. Mr. Thompson. They vary in size, but we use an average sometimes of 30 housing units per block. There's 60,000 census tracts. And we use an average of around 1,000 housing units. Mr. Miller. About how many you say within a tract, about 1,000? Mr. Thompson. 1,000 to 2,000 housing units. Mr. Miller. Whereas in a block there's varying numbers. Mr. Thompson. Well, blocks vary considerably in size. For planning purposes sometimes use an average of 30 housing units per block. In urban areas it's bigger than 30. In rural areas it's less than 30. Mr. Miller. For the adjustment decision about to be made, what's the smallest number of geography that the Bureau considers important to know whether adjustment improves accuracy? What are you focused on, what level of accuracy? Mr. Barron. The level that we're focused on for this process that we're engaged in is the level of the congressional district. That's all that we're focused on at this time. At some subsequent point, we'd like to address the issue of what is the best data set to use for survey controls and things like that, but that would be later. Right now it's the congressional district decision that is the object of our attention. Mr. Miller. But your focus is on block level accuracy, tract level accuracy; what's the next level above tracts in the hierarchy? What's the accumulation of tracts? Mr. Thompson. At that point there's really not a higher level--you have counties, you have places, you have voting districts. Mr. Miller. So is your focus on tracts or blocks? Mr. Barron. The focus is on congressional districts. That's the focus. Mr. Miller. And congressional districts are made of blocks? Mr. Barron. Some of blocks. Mr. Miller. And the Constitution, as the Supreme Court ruled, is one man, one vote, and every congressional district has to be essentially the exact same size as another congressional district in that State, and in order to get that exact number--I mean, our districts cannot vary within our States by more than a couple of people. You have to work with block data, right? I mean, I guess you don't deal with redistricting, but the point is block data, one block at a time is what we work with. We don't have a defined county we work with. We work with a block here and block there, and we add the blocks together, and the concern we have with block level data is the accuracy is not there. Back in 1990, since this is--the Cape report said that for population areas of 100,000 or less, you cannot tell if adjustment improves accuracy; is that right? At populations of 100,000 or less, you cannot tell whether that improves accuracy? Mr. Barron. I believe that's what that report said, yes, and you know, the issue before us now, Mr. Chairman, is whether when you get to something that's as large as a congressional district, can you tell. I mean, that's the decision that we're facing, and we think you can, but that remains to be proven. Mr. Miller. Is block level data better adjusted or unadjusted? I mean, that's what we work with. Mr. Barron. I would say---- Mr. Miller. Maybe going back to 1990. I know you don't have the answer for 2000. Mr. Barron. We're looking at the issue of whether congressional districts are better when you add the blocks together. This other issue is something I'd have to go consult with my experts on and get back to you. Mr. Miller. One of the problems Mr. Souder brought up, and it's something that's not our focus constitutionally, we're looking at congressional districts, but I have got cities in my district that are very small. The total population of Anna Maria Island--there's three cities on one island in my area. There's just a couple thousand people in each city, and they have to draw city council districts, and there's no question, but, you know, you're working with a total city population of 2,500 people, and you divide it up into five city council districts, what would they use, adjusted or unadjusted data? Mr. Barron. I don't think I have a recommendation for them when you work with something that small. Mr. Miller. When you only have 2,500 people in the total population, you have real problems if you try to use adjusted data, I would think. It's well below 100,000. Mr. Barron. I think at that level everything these people have access to is going to have a certain amount of noise, and what we're doing is trying to provide better estimates for a higher level of aggregation. Mr. Miller. Are you more focused on block or tract level? I know you keep saying you're focused on congressional districts, but you have got to look at those two levels of data. Which is the one you're trying to decide or making the decision process on? Mr. Barron. The decision process that we're working on basically takes blocks and adds them together, and then we determine whether that estimate, having aggregated the data, whether that estimate is better. Mr. Miller. Mr. Clay. Mr. Clay. Mr. Chairman, just one issue to raise; was wondering would we have an opportunity to hear from Dr. Prewitt at some time in the near future? Mr. Miller. That's definitely a possibility. He's always a very good witness. Mr. Clay. He sounds like he's the expert. Being a freshman Member of Congress, I'd like the opportunity to hear from him if possible. Mr. Miller. Director Prewitt deserves great congratulations for making the census so successful. We didn't always agree on every issue, but one thing I advocated was make sure we had all of the resources that the Bureau could use to get the most accurate census possible. Since I serve on that appropriations subcommittee, we agreed there. And so Director Prewitt did an outstanding job leading the Census Bureau during a very difficult time, and so I commend him on that, and if it's possible, we'll love to have him back. Mr. Clay. I would hope, just in conclusion, that--I would hope we would leave no American uncounted because it's so important as far as Federal resources, as far as highway redistricting that we have an actual count and an accurate one. And I think that's the key to a good census is that it is accurate, and I'd like to ensure somehow or have a level of comfort that we do have the most accurate count possible. Mr. Miller. Well, since 1790, when Jefferson did the first census, I think we can all probably say that this is the best census ever because it's hard even to comprehend what the census was like when the U.S. marshals did it--I think it was the U.S. marshals in 1790 or in the 1890's or something. And so we have a lot to be pleased with right now, but it's not perfect. Mr. Clay. No, it's not perfect. Mr. Miller. But we sure have made tremendous strides, and you all are to be congratulated. Mrs. Maloney. Mrs. Maloney. Yes. Mr. Thompson, getting back to the line of questioning of the chairman on the block level versus larger geographic areas, is there a legislative district in America that is made up of one block, Mr. Thompson? Mr. Thompson. I don't know if there's a legislative district made up of one block or not. I don't know the answer to that. Mrs. Maloney. Well, I can tell you that legislative districts, even the smallest ones, school districts, city council districts, are made up of literally thousands of blocks. That's what makes up a representative area. It's not one block, it is thousands of blocks, and have you not testified that the higher levels there are of geographic area, that ACE becomes more accurate? Mr. Thompson. Yes. Mrs. Maloney. Well, I think that's an important point. Now, I think that one of the themes that we keep hearing here today is the people talk that, you know, that we have a measurement that is more accurate, but it is this ACE measurement that is also the measurement used to correct for the undercount. So if it's accurate for the measurement, then it should be accurate for the use of correcting for the undercount. And I just have heard talks today, you know, at one point we think that maybe there's going to be an undercount of roughly 3 million, and at one point there was an undercount of 4 million, and we know that 8 million was missed in the last census in 1990, and 4.4 were counted twice. But I think the main point is how many blacks and Hispanics does it take to be missed in a census for the Republican Party to agree that they should be corrected for the undercount, because in my--my position is that every person counts, and whether it's 8 million or 4 million or 3 million or 2 million, all of these people should be counted, and if you're not counted, the injustice is more extreme, I would argue, than what happened in Florida. And the Florida fiasco will have a chance next year to go and vote again, and believe me, I truly believe that the corrections will be made, and every vote hopefully will be counted. But if people are missed in this census, it will be 10 years before the correction can be made. If we don't correct for the undercount, there will be missed Federal dollars, there will be missed Federal representation, not to mention data is less accurate. So I just want to make a point that we need to count every American, and if there is an undercount, it should be corrected in general fairness and for what's at stake in representation and distribution of Federal funds. Thank you, Mr. Chairman. Mr. Miller. Mrs. Maloney, being from the big city of New York City, your jurisdiction areas within there are much larger than my area in southwest Florida, which is a beautiful area. But we have communities incorporating cities that have 1,500 people in it, and so you start dividing up those areas--and there aren't a lot of blocks in those areas, and so there's no question I think that I would advise small communities to be very cautious if they were looking at adjusted data. But when you go through the redistricting process, it's one block at a time. It's not like it's a defined area like the State of Florida, the State of New York or the city limits of New York City. It is adding a block here, taking a block there, so we get the exact same numbers. And so you work with block data. And what you will argue is you add them all together, and the errors average out, but the bottom line is you're working with one block at a time, and that block by itself is not accurate when you start adjusting. So there's a legitimate debate about the accuracy of that data at the block level for redistricting purposes, and that's really what the constitutional purpose of the census is. Mr. Barron, I've finished. Does anyone else have anymore questions? Mrs. Maloney. I have one closing question, and I'd like to ask Mr. Barron it if I may. Mr. Director, John Kennedy and many others have said that defeat is an orphan, but success has 1,000 fathers. Real quickly, I'd like to ask you where some of the enhancements and improvements for this census 2000 plan came from. I want to know who were the key players. Was it the Census Bureau professionals, the National Academy of Sciences, Congress, the Commerce Department, the White House? I'd like you to designate where the idea and where these improvements came from, the redesigning of the form to make it more user-friendly? Mr. Barron. That was a Census Bureau idea. Mrs. Maloney. The optical scanning technology to read handwriting? Mr. Barron. That was a Census Bureau proposal. Mrs. Maloney, if I may, since John was involved in this more than I was, I was not there at the time, I want to ask him to respond to some of these if I may. Mrs. Maloney. OK. Certainly. Mr. Barron. Am I right so far? I'm two for two. The partnership program. I think the partnership program was a Census Bureau proposal, probably Commerce Department, but also the Census Bureau has strong advisory committees involving its partners. So I think that goes back in history. So maybe that's something that originates with lots of people. Mrs. Maloney. But basically the Census Bureau. Mr. Barron. Well, perhaps both. Mrs. Maloney. The advertising campaign. Mr. Barron. I don't know. Mr. Thompson. That was generated by the Census Bureau, but we got a lot of advice from our advisory committees and the Congress about paid advertising and how important it was. So I think that's something that came up in partnership, so I wouldn't want to claim it was the Census Bureau's idea only. Mrs. Maloney. But the Census Bureau in collaboration with advisers---- Mr. Thompson. We certainly endorsed it and embraced it and went forward with it. Mrs. Maloney. Higher pay scales for enumerators. Mr. Thompson. That was our basic research that went into that. Mrs. Maloney. The three-part mailing of an advance letter, the form and the postcard? Mr. Barron. Census Bureau. Mrs. Maloney. So basically the major enhancements came from the Census Bureau. Congratulations on a job done well so far, and I wish you well in your remaining days to complete the work before you and to come forward with more accurate numbers and your recommendations. Mr. Barron. Thank you. Mr. Miller. I concur with Mrs. Maloney to congratulate the people at the Census Bureau for coming up with these ideas and implementing a very successful census. I am sorry that some of the recognition did not arrive for the individuals, but hopefully we will have some way we can make sure individuals are given the recognition they deserve. Mr. Barron, you are in a terribly unfortunate position today. An extremely important decision, one that has far- reaching ramifications, is now weighing heavily on the shoulders of a dedicated, highly capable, career civil servant. I'm sure Congress did not mean for this to happen. The regulation seems to have intended--the regulation seems not to have intended it either, but both Congress and the regulation put the adjustment decision with political appointees, and you are not a political appointee. You are a civil servant with 30 years of outstanding service at the Bureau of Labor Statistics before going to the Commerce Department 2 years ago. Thank you for the job you've done. Thank you, Mr. Thompson. On behalf of the subcommittee, I would like to thank you for appearing before us today. I ask unanimous consent that all Members' written opening statements be included in the record. Without objection, so ordered. In case there are additional questions that Members may have for our witnesses, I ask unanimous consent for the record to remain open for 2 weeks for Members to submit questions for the record, and that the witness submit written answers as soon as possible. Without objection, so ordered. Thank you again. Meeting adjourned. [Whereupon, at 3:50 p.m., the subcommittee was adjourned.] [Additional information submitted for the hearing record follows:] [GRAPHIC] [TIFF OMITTED] T5289.021 [GRAPHIC] [TIFF OMITTED] T5289.022 [GRAPHIC] [TIFF OMITTED] T5289.023