                                                              1

 1    UNITED STATES DISTRICT COURT
      EASTERN DISTRICT OF NEW YORK
 2    ------------------------------x
      
 3    THE CITY OF NEW YORK,
      THE STATE OF NEW YORK,
 4    THE PEOPLE OF THE STATE OF
           CALIFORNIA EX REL. JOHN K.
                      _______        
 5         VAN DE KAMP, ATTORNEY GENERAL,
      THE CITY OF LOS ANGELES,
 6    THE CITY OF CHICAGO
      THE CITY OF HOUSTON,
 7    DADE COUNTY, FLORIDA,
      THE U.S. CONFERENCE OF MAYORS,
 8    THE NATIONAL LEAGUE OF CITIES,
      THE LEAGUE OF UNITED LATIN
 9         AMERICAN CITIZENS,
      THE NATIONAL ASSOCIATION FOR THE
10         ADVANCEMENT OF COLORED PEOPLE,
      MARCELLA MAXWELL,
11    DONALD H. ELLIOTT,
      JOHN MACK,
12    OLGA MORALES,
      TIMOTHY W. WRIGHT III,
13    RAYMON G. ROMERO,
      ANTONIO GONZALEZ,
14    ATHALIE RANGE,
      JERRY ALAN WOOD, and
15    CAROLYN SUE LOPEZ,
      
16                   Plaintiffs, and          89 Cic. 3474 JMcL
      
17    THE STATE OF TEXAS,
      THE CITY OF PHOENIX, ARIZONA,
18    THE STATE OF NEW JERSEY
      THE STATE OF FLORIDA,
19    THE CITY OF CLEVELAND, OHIO,
      THE CITY OF DENVER, COLORADO,
20    THE CITY OF INGLEWOOD, CALIFORNIA,
      THE CITY OF NEW ORLEANS, LOUISIANA,
21    THE CITY OF OAKLAND, CALIFORNIA,
      THE CITY OF PASADENA, CALIFORNIA,
22    THE CITY OF PHILADELPHIA, PENNSYLVANIA,
      THE CITY OF SAN ANTONIO, TEXAS
23    THE CITY OF SAN FRANCISCO, CALIFORNIA,
      BROWARD COUNTY, FLORIDA,
24    THE STATE OF ARIZONA,
      THE CITY OF BALTIMORE, MARYLAND,
25    THE CITY OF BOSTON, MASSACHUSETTS,
                                                              2

 1    UNITED STATES DISTRICT COURT
      EASTERN DISTRICT OF NEW YORK
 2    ------------------------------x
      THE CITY OF LONG BEACH, CALIFORNIA,
 3    THE CITY OF SAN JOSE, CALIFORNIA,
      LOS ANGELES COUNTY, CALIFORNIA,
 4    SAN BERNARDINO COUNTY, CALIFORNIA,
      THE DISTRICT OF COLUMBIA,
 5    THE NAVAJO NATION,
      THE STATE OF NEW MEXICO,
 6    THE CITY OF TUCSON, ARIZONA,
      and THE COUNTY OF HUDSON, NEW JERSEY,
 7    
                     Plaintiff-Intervenors,
 8    
                 v.                           
 9    
      UNITED STATES DEPARTMENT OF COMMERCE,
10    C. WILLIAM VERITY, as Secretary of the
           United States Department of Commerce,
11    ROBERT ORTNER, as Under Secretary for
           Economic Affairs of the United States
12         Department of Commerce,
      BUREAU OF THE CENSUS,
13    JOHN G. KEANE, as Director of the
           Bureau of the Census,
14    RONALD W. REAGAN, as President of the
           United States, and
15    DONNALD K. ANDERSON, as Clerk of the
           United States House of Representatives.
16    
                     Defendants.
17    
      THE STATE OF WISCONSIN,
18    THE STATE OF OKLAHOMA,
      
19                   Defendant-Intervenors.
      
20    ------------------------------x
      
21    
      
22    
      
23    
      
24    
      
25    
                                                              3

 1                                            May 11, 1992
                                              10:00 a.m.
 2    
      Before:
 3    
                 HON. JOSEPH M. McLAUGHLIN
 4    
                                              District Judge
 5    
                     APPEARANCES
 6    
      DAVID B. GOLDIN
 7    SANFORD M. COHEN
      ROBERT S. RIFKIND
 8    PETER L. ZIMROTH
      JAVIER P. GUAJARDO
 9    GEORGE L. WAAS
      MICHAEL S. BOKAR
10    LOUIS M. SOLOMON
      YEORYIOS APALLAS
11    ADA TREIGER
      JESSICA HEINZ
12    STANLEY A. BOWKER
           Attorneys for Plaintiffs
13    
      ANDREW J. MALONEY
14         United States Attorney for the
           Eastern District of New York
15    THOMAS W. MILLET
      MICHAEL SITCOV
16    JASON R. BARON
      DAVID M. GLASS
17    JUDRY L. SUBAR
           Assistant United States Attorneys
18    

19               -o0o-

20               (Case called)

21               THE COURT:  Good morning, ladies and gentlemen.  

22    This day has been a long day coming.  I would like to say 

23    that I have been looking forward to it, but candor forbids 

24    that.  

25               Let me just make a preliminary statement and then 
                                                              4

 1    we'll get down to brass tacks.  The Constitution, of course, 

 2    requires that a decennial census be taken and by statute 

 3    Congress has delegated to the Secretary of Commerce the task 

 4    of taking the census "in such form and content as he may 

 5    determine, including the use of sampling procedures and 

 6    special surveys."  

 7               The Bureau of the Census is the agency within the 

 8    department that discharges this responsibility.  

 9               In October of 1987, the Secretary of Commerce 

10    announced that there would be no statistical adjustment in 

11    the 1990 census.  Plaintiffs commenced with this action in 

12    November of 1988 seeking an injunction and challenging the 

13    methodology by which the census was to be conducted.  

14               Since that suit has begun, various states and 

15    other municipal bodies have intervened.  If my mathematics 

16    are correct, and regrettably mathematics is my weakest suit 

17    which makes me eminently qualified to understand all of this 

18    testimony, if my mathematics are correct, we have seven 

19    states as parties, seventeen cities, four counties, the 

20    District of Columbia and the Navajo Nation.  We also have 

21    several amicus briefs.  

22               The defendants moved to dismiss the application 

23    for the injunction and the court denied the motion to 

24    dismiss holding that the plaintiffs had standing to assert a 

25    challenge to the census and that the court would review the 
                                                              5

 1    Secretary's determination under the arbitrary and capricious 

 2    standard of review.  

 3               A hearing was then scheduled on the injunction, 

 4    and was set to go forward in the summer of '89 when at the 

 5    11th hour the parties entered into a stipulation that was 

 6    subsequently approved by the court.  

 7               That stipulation and order, which is pivotal to 

 8    this suit, provided that the 1987 decision of the former, 

 9    the now former Secretary of Commerce that he would not 

10    adjust would be vacated, and that the new Secretary, 

11    Secretary Mosbacher, would consider de novo and "with an 

12    open mind" whether adjustment was warranted.  

13               The stipulation further provided that the 

14    Secretary was to decide whether to adjust and make that 

15    decision by July 15, last summer, that his decision would be 

16    consistent with certain procedures including guidelines 

17    articulating what the defendants believed to be the relevant 

18    technical and policy considerations.  

19               The stipulation also mandated the establishment 

20    of a special advisory panel consisting of 8 members who were 

21    to advise the Secretary on whether to adjust.  

22               After the Commerce Department adopted the 

23    guidelines required by the stipulation, the plaintiffs 

24    challenged that as inadequate and they sought a declaratory 

25    judgment that a statistical adjustment would not violate the 
                                                              6

 1    Constitution or federal statutes.  

 2               The defendants countered that the case presented 

 3    a nonjusticiable political question.  The court ruling on 

 4    those motions held that a challenge to the accuracy of the 

 5    census is not a political question, and that statistical 

 6    adjustment would not violate either the Constitution or the 

 7    laws of the United States.  

 8               The court further held that while the guidelines 

 9    promulgated were somewhat vague, they did not violate the 

10    stipulation and order.  

11               The census then went forward and the Secretary 

12    announced on July 15 that he had reached a decision not to 

13    adjust.  The suit, of course, was revitalizeed, discovery 

14    proceeded apace.  All of that discovery is now substantially 

15    completed.  

16               The plaintiffs requested that the court hold a 

17    trial.  The defendants objected to conducting a trial, 

18    arguing that because the case arose under the Administrative 

19    Procedure Act the scope of the court's review should be 

20    limited to the administrative record.  

21               On February 18th of this year the court ordered a 

22    trial.  The court held that, recognizing the elementary 

23    review of an agency decision is generally limited to the 

24    administrative record, and where a court finds the record is 

25    inadequate the preferred course is to remand to the agency 
                                                              7

 1    and not conduct a trial, nevertheless the court concluded 

 2    that the evidentiary hearing is necessary in this case for 

 3    several reasons.  

 4               First, there have been several allegations, some 

 5    of them serious, that the administrative record is a 

 6    self-serving post-op compilation of documents, assembled for 

 7    the purpose of strengthening the defendant's litigation 

 8    position.  The court therefore must consider the integrity 

 9    of the administrative record to insure that these documents, 

10    and only these documents, were considered by the Secretary.  

11               The second purpose of the hearing is to enhance 

12    the court's understanding of the record, replete as it is 

13    with references to technical jargon and arcane procedures.  

14    To this end, the testimony of expert witnesses hopefully 

15    will prove helpful.  Their testimony should also assist the 

16    court in determining whether the Secretary considered all 

17    the relevant factors in making his decision.  

18               Finally, the court is fully aware of the 

19    important constitutional questions this case presents.  

20    Those considerations led the court to conclude that a trial 

21    was necessary.  

22               I want to assure the parties that they will 

23    receive an attentive and open-minded hearing on all of the 

24    issues in the case.  We are ready to begin, the parties have 

25    agreed that each side would like to make an opening 
                                                              8

 1    statement.  

 2               Are the plaintiffs ready?  Mr. Rifkind?

 3               MR. RIFKIND:  Good morning, your Honor.  I think 

 4    I have been authorized to speak at least for most of the 

 5    large army that sits on the groom's side of this chapel.  

 6               I appreciate your opening statement, sir, and I 

 7    wish to adopt it as a chapter of mine.  Your arithmetic is 

 8    significantly more accurate than the census and you will 

 9    have no difficulty with the arithmetic presented in case.

10               THE COURT:  Point won.

11               MR. RIFKIND:  I might just begin by identifying 

12    some of the lawyers who have been principally active and 

13    will be principally active in the course of these 

14    proceedings on the plaintiff's side.  In addition to myself 

15    there is Mr. Peter Zimroth of Arnold & Porter.

16               THE COURT:  I noticed his picture in full pose on 

17    the front page of today's Law Journal aspiring to his 

18    Napoleonic image.  

19               MR. ZIMROTH:  That was not my doing, your Honor.  

20    That was the photographer who insisted.

21               MR. RIFKIND:  In the same photograph, your Honor, 

22    doubtless saw Mr. Louis Solomon.  Regrettably not present in 

23    that photograph was Mr. Yeoryios Apallas of the Attorney 

24    General's Office in California, Mr. David Goldin of the New 

25    York Corporation Counsel's Office and Mr. Sanford Cohen of 
                                                              9

 1    the New York Attorney General's Office and also Ms. Jessica 

 2    Heinz of the Los Angeles City Council's Office and Ms. Ada 

 3    Treiger of the Los Angeles County Attorney's Office and they 

 4    are all here today and can be photographed during the break.  

 5               I appreciate the tour of the history of this case 

 6    that the court has set forth and I'd now like to put in a 

 7    little focus the array of evidence as we see it coming in on 

 8    the ground that some sense of the order might help elucidate 

 9    what is, I appreciate, complex.  I'm going to refrain, if 

10    your Honor will forgive me, from trying to define or 

11    articulate some of the more technical processes that go on.  

12    That is precisely what the witnesses when they get on the 

13    stand will do for you, and I will refer to some of the key 

14    terms in passing and ask you just to suspend disbelief until 

15    you hear from the witnesses.  

16               Now, your Honor, on Friday June 28, 1991, the 

17    Director of the Census Bureau, Dr. Barbara Bryant, submitted 

18    to Secretary Mosbacher her conclusions and her 

19    recommendation on the question of whether to adjust the 1990 

20    census.  That 30 page report from the Director, which we 

21    have marked Plaintiffs' Exhibit 55 here, is a key document 

22    in this case.  In that report the Director recommended that 

23    the 1990 post-enumeration survey be used to statistically 

24    adjust the 1990 census and the report sets forth the grounds 

25    for that recommendation in some detail.  
                                                              10

 1               It begins with a key fact or pair of facts.  On 

 2    the one hand, undeniably, the enumeration in 1990 found the 

 3    resident population of the United States to be approximately 

 4    248,000,000.  

 5               On the other hand, two separate independent 

 6    analyses, one based on demographic analysis and one based on 

 7    the post-enumeration survey, each found the resident 

 8    population to be approximately 253,000,000.  

 9               Thus both the Bureau's PES as we call it and its 

10    demographic analysis indicate an undercount of approximately 

11    5,000,000.  

12               The similarity of those two independent findings, 

13    the Director reported, is "strong evidence that these 

14    residents of the United States actually exist.  Not 

15    adjusting would be denying that these 5,000,000 people 

16    exist.  Such denial would be a greater inaccuracy than any 

17    inaccuracies that adjustment may introduce."  

18               Secondly, Dr. Bryant's report pointed out that 

19    both demographic analysis and PES disclose significantly 

20    higher undercount rates for blacks than for non-blacks.  

21    Moreover, the black-non-black differentials in undercount 

22    rate, as measured by both, were consistent with the 

23    differentially higher undercount rate for blacks measured by 

24    the Census Bureau's demographic analysis in each of five 

25    previous decennial censuses going back to 1940.  
                                                              11

 1               And I've put up on the lefthand side over here a 

 2    chart showing swiftly the pattern of differential rates of 

 3    undercount from 1940 to 1990, that is there is a three and 

 4    had a half to four and a half percent spread in each year 

 5    between the black and the non-black rate of undercounts.  

 6    Those are percentage points in the lefthand column.  

 7               Now, Dr. Bryant's report recognizes that 

 8    demographic analysis, because of the birth and death data on 

 9    which it is based, does not disclose estimates for 

10    minorities other than blacks.  Non-black Hispanics, Asians 

11    and Pacific Islanders and American Indians are all swept 

12    together in those figures under the general rubric "other," 

13    which is largely Caucasians in the United States.  

14               The PES on the other hand, reveals that the 

15    undercount rate for Asians and Pacific Islanders was 3.1 

16    percent, for American Indians 5 percent and for Hispanics of 

17    any race 5.2 percent and those are reflected in the black 

18    bars over here.  If I might just point out, Hispanic, black, 

19    Asian Pacific Islanders, national average, non-black.  

20               The Director also pointed out that demographic 

21    analysis can only provide nationwide data and cannot 

22    allocate the missing five million people to specific dates 

23    and substate areas.  That is what the PES was designed to 

24    do.  

25               Dr. Bryant concluded and advised the Secretary 
                                                              12

 1    that as a result of the decade of work since a 

 2    post-enumeration was first tried in 1980, the 1990 PES 

 3    proves to be, she said, a high quality survey of 157,000 

 4    households.  The Census Bureau, she said, subjected the PES 

 5    and resulting estimates to test after test to find fatal 

 6    flaws in procedures.  The Census Bureau did not find fatal 

 7    flaws.  Evaluation shows, she said, that the PES is of 

 8    sufficiently high quality to use as an adjustment tool in 

 9    the professional judgment of the Census Bureau Undercount 

10    Research Committee, this survey and the selected PES model 

11    for adjustment improved the count over the census.  

12               Now, there's a balancing involved in this 

13    judgment.  And it's one that the Census Bureau took very 

14    seriously.  

15               On the one hand, as the report says, the Census 

16    Bureau mounted the most extensive effort every to enumerate 

17    blacks and other minorities in the course of the regular 

18    enumeration.  Despite this effort, it says, the undercount 

19    differential was not reduced below its historical level and 

20    indeed, there is indication that the problem might be worse 

21    in the year 2000.  

22               On the other hand, and I quote, "The quality of 

23    the 1990 PES is excellent.  Thus, for the first time in 

24    history, a tool exists with which to correct the census 

25    enumeration and to make it more accurate."  
                                                              13

 1               Dr. Bryant reported that a majority of the 

 2    Undercount Steering Committee comprised of nine senior, 

 3    career statistical and demographic experts in the Bureau 

 4    believe statistical adjustment leads to an improvement in 

 5    the census count as enumerated.  

 6               Both the Director and the Undercount Steering 

 7    Committee's report which accompanied the Director's report 

 8    acknowledged that the recommended corrections were 

 9    imperfect.  The question before them was whether the 

10    enumeration was better or worse when adjusted by the PES.  

11               On the one hand, they acknowledge that the PES, 

12    like all sample surveys, inherently introduces some measure 

13    of random variability.  On the other hand, the unadjusted 

14    census is biased as against, as they put it, the same well 

15    identified subpopulations as previous censuses and one of 

16    the major questions that is going to emerge in the 

17    testimony, your Honor, is how one weighs those two 

18    conflicting sources of error in the two methods, whether one 

19    attaches equal weight to bias as to random variability and 

20    we will submit at the end of the case that in our society, 

21    random variability is a great deal to be preferred to bias.  

22               The Bureau conducted the most elaborate set of 

23    tests and analyses to determine whether the inaccuracy of 

24    the uncorrected counts was more or less than the inaccuracy 

25    of the corrected counts.  All of those studies showed 
                                                              14

 1    emphatically that the unadjusted enumeration at the state 

 2    level and at all significant substate levels was lower.  

 3    Not, perhaps at the very lowest level, at the block level as 

 4    they call it, but as Dr. Bryant acknowledged, nothing short 

 5    of a perfect census could say that the adjusted counts were 

 6    more accurate in each of the 4,000,000 blocks into which the 

 7    nation is divided.  And we will be showing your Honor that 

 8    the block level issue is really an irrelevance and a red 

 9    herring.  

10               Both the Director and the Undercount Steering 

11    Committee acknowledged that further research in the months 

12    and years ahead might well lead to improvements in the 

13    adjusted data.  But both asserted that such improvements 

14    will still show that the adjusted data are more accurate 

15    than the census.  

16               Dr. Bryant reported that she had sat through 

17    months of deliberation of the committee as an ex officio 

18    member as it evaluated the PES as the model of adjustment 

19    that had been prespecified.  "I have listened to research 

20    teams present results of 19 studies to evaluate the quality 

21    of the PES and 11 studies of the demographic analysis."  On 

22    the basis of all that she stated her recommendation.  

23               "As Director of the Bureau of the Census, I 

24    Barbara Everitt Bryant recommend to Secretary of Commerce 

25    Robert A. Mosbacher that results of the 1990 
                                                              15

 1    post-enumeration survey be used statistically to adjust the 

 2    1990 census."  And shortly thereafter the Bureau produced 

 3    the vast array, I think it's 60 rolls of data tapes, which 

 4    encompass the corrected data for every area of the country, 

 5    usable for redistricting the apportionment and so on.  

 6               Now, those tapes which are curiously not part of 

 7    the so-called administrative record, are something that the 

 8    government has regarded as top secrets next to the secrets 

 9    of the hydrogen bomb, and I have been lucky enough to 

10    convince them into letting us have them but only under the 

11    wraps of a strict confidentiality order under which I may 

12    not disclose them or their contents to anybody.  

13               We have nevertheless in an attempt to honor that 

14    stipulation marked them as an exhibit, sort of in absentia, 

15    and I thought consistent with that I might just, to make it 

16    realistic to your Honor, disclose exactly what it is I'm 

17    talking about.  That is what the Bureau produced and it 

18    contains corrected data for every block in the country.  

19               I should say that I'm not the only one who had a 

20    difficult time in getting ahold of this result of the 

21    Bureau's efforts.  The Commerce Department declined to honor 

22    a congressional subpoena to produce them and has taken the 

23    position that it is dangerous information which will confuse 

24    the public, and that the states might be tempted to use the 

25    data in actually redistricting their lines which would be a 
                                                              16

 1    very dangerous thing to do, although elsewhere the 

 2    government has taken the position that it is indifferent to 

 3    what the states do, they're free to reapportion themselves, 

 4    redistrict themselves in any way they like.  So there's a 

 5    strange inconsistency in saying that they may not have this 

 6    data if they want to use it, even though a great deal of 

 7    money was spent producing it, but they're free to do 

 8    whatever they want except they can't have it.

 9               THE COURT:  Is there a suit pending in the Ninth 

10    Circuit on those tapes?

11               MR. RIFKIND:  There is, indeed.  There are two 

12    suits at least in the Ninth Circuit, I think, last count, 

13    trying to get that.

14               THE COURT:  What is the status of that?

15               MR. RIFKIND:  My understanding is that the 

16    district court has held that the state was entitled to have 

17    it and that's been argued on appeal and I think we're 

18    waiting to hear from the Ninth Circuit.

19               THE COURT:  If they don't get enjoined.

20               MR. RIFKIND:  Right.

21               THE COURT:  By the Supremes.

22               MR. RIFKIND:  Right.  A number of other states 

23    have desparately tried to get this stuff by one route or 

24    another, FOIA and so on, but I gather Congress is very 

25    apprehensive that the states might actually decide to use 
                                                              17

 1    this data.  Although as I say they have elsewhere said it's 

 2    a matter of indifference to Commerce whether they use it or 

 3    not.  

 4               Now, the reports of the Director and the 

 5    Undercount Steering Committee did not spring out of the 

 6    blue.  As we will show, they are the result of intensive 

 7    work by the Bureau on the undercount problem over a long 

 8    period of time.  I'm not going to try to retrace that long 

 9    history today but I think it might be worth noting just to 

10    put it in perspective that an early milestone is the classic 

11    article published by Professor Ansley Coale of Princeton in 

12    1955.  Applying the methods of demographic analysis which he 

13    had a large hand in developing, Dr. Coale there found 

14    significant flaws, undercounts in the census and of these, 

15    he wrote, the most striking errors, and the ones which seem 

16    most likely to make a difference are the rates of 

17    underenumeration among nonwhites.  If we interpret these as 

18    representing nonwhite persons omitted, we are led to such 

19    views as that, for instance, the sections of the country 

20    with a high percent of nonwhites contain a larger fraction 

21    of the total population than the census shows.  

22               Following up on Professor Coale's work, the 

23    Bureau in a thoroughly professional really quite dedicated 

24    manner steadily refined and improved the quality of its 

25    demographic analysis and resolutely published, unlike most 
                                                              18

 1    bureaucracies, for public and professional scrutiny the 

 2    evidence of the shortcomings of their own work in regular 

 3    reports of undercoverage of the census.  

 4               As a result, we have the measures of 

 5    undercoverage that the chart on the left there discloses, 

 6    going back to 1940.  

 7               Thus, the differential rate between blacks and 

 8    non-blacks is no novelty, it is one of the most thoroughly 

 9    documented phenomena and concern about that problem has 

10    resulted in intensive effort by the Bureau as I say over a 

11    long period of time.  

12               Starting in 1980 they focused on the use of 

13    sampling methods in solving that problem.  Now sampling is 

14    not new to the census.  As you will hear from the witnesses, 

15    it has been used to solve a variety of problems for a long 

16    time.  Indeed in 1970, more than a million and a half people 

17    were added to the census by appropriate sampling methods.  

18               But in 1980 they made a stab of solving the 

19    general problem of differential undercount by that means.  

20    At the end of that effort, and it was a very intensive and 

21    public effort, much debated, much discussed, the Director of 

22    the Bureau decided that the tool was at that stage 

23    inadequate and I want to emphasize that Secretary of 

24    Commerce Klutznick then made it quite clear that the 

25    decision was a decision for the Director to make.  
                                                              19

 1               But even while that battle was under way, the 

 2    Bureau turned to the question of whether statistical 

 3    methodology could be designed to deal with the undercount 

 4    problem in the future.  And you will hear from Dr. Kirk M. 

 5    Wolter who was chief of the Bureau's Statistical Research 

 6    Division from 1983 to 1988 and is now a vice-president of 

 7    A.C. Neilsen.  He also served as a member of the Special 

 8    Advisory Panel created pursuant to your Honor's order in 

 9    this case.  You'll also area from Dr. Barbara Bailar who was 

10    the Bureau's Associate Director for Statistical Standards 

11    and Methodology and is now the Director of the American 

12    Statistical Association.  

13               Both Dr. Bailar and Dr. Wolter, it should be 

14    noted, testified against New York's claim for an adjustment 

15    in 1980.  They will describe for you the careful, methodical 

16    step-by-step process in which the problem of undercount was 

17    assessed, the post-enumeration survey was designed and the 

18    methods for assessing its reliability were established.  

19               Not only was there extensive research within the 

20    Bureau, the Bureau in its normal fashion invited the most 

21    extensive outside professional assistance.  And you will 

22    hear from Dr. John Rolph of the Rand Corporation who was a 

23    member of the Panel on Decennial Census Methodology of the 

24    National Academy of Science who will report to you on the 

25    collaborative work with that body as will Dr. Stephen E. 
                                                              20

 1    Feinberg who is a member of the Committee of National 

 2    Statistics of the National Research Council.  All of those 

 3    groups supported vigorously the Bureau's efforts in solving 

 4    this problem.  

 5               In light of all that effort, as your Honor is 

 6    aware, in 1986 John G. Keane, the Director of the Bureau 

 7    announced that they intended to decide by the following year 

 8    whether or not to adjust.  In the course of the following 

 9    year they concluded that it was technically feasible to do 

10    so and Dr. Keane, the Director was about to announce that 

11    when his decision was abruptly overturned by the Commerce 

12    Department and the press release was cancelled and a month's 

13    period of awkward silence ensued until Dr. Ortner, the 

14    Undersecretary of Commerce, announced in October of 1987 

15    that there would be no adjustment because, among other 

16    things, the undercount in 1990 would sink to de minimis 

17    proportions, and that, of course, precipitated this suit and 

18    your Honor has summarized the proceedings in that in a way 

19    that I'm happy to adopt.  

20               I might add only one thought at this point.  The 

21    stipulation and order that was entered in this case in July 

22    of 1989, provided among other things that the population 

23    count produced prior to the time that the decision was 

24    actually made on adjustment must all bear a caveat, warning, 

25    subject to revision, whatever the words were.  And not only 
                                                              21

 1    did all the Bureau's releases substantially carry that 

 2    warning, but even the letter communicating the apportionment 

 3    of Congress to each of the several states contained the 

 4    warning.

 5               THE COURT:  What do you say is the significance 

 6    of that?

 7               MR. RIFKIND:  Well, I emphasize that point 

 8    because I understand an intervenor from Oklahoma suggests 

 9    that they have some vested right in the status quo 

10    regardless of how erroneous it may be.  And I just want to 

11    emphasize the point that when they received their 

12    certificate, it came with a legend from the Clerk of the 

13    House of Representatives who was a party to this case, 

14    because of pending litigation with respect to the current 

15    census, "I am also notifying you that population counts 

16    which form the basis for the number of representatives set 

17    forth herein are subject to possible correction for 

18    undercount or overcount.  The Department of Justice is 

19    considering whether to direct and will publish" and so on.  

20               Everyone was on notice.  

21               And that brings me back to June 28, 1991, when as 

22    a result of our stipulation and order and all the efforts of 

23    the Bureau, the Director submitted her report.  And our next 

24    step is a very short one, because approximately five days 

25    later, on or about July 3, Deputy Undersecretary of Commerce 
                                                              22

 1    Mark Plant was instructed by Undersecretary Darby to 

 2    beginning writing the Secretary's decision rejecting 

 3    adjustment.  

 4               The speed with which that decision was reached 

 5    boggles the mind.  On the one side you have all the months 

 6    of intensive effort that I've described and on the other 

 7    side apparently we have the lightning fast acuity of Robert 

 8    Mosbacher.  

 9               So incredibly fast is that, that even the 

10    defendants blink at it and they will say that the decision 

11    was not really made on July 3.  They will say that Mark 

12    Plant was only told to write a decision rejecting the Census 

13    Bureau's recommendation, the decision was still being 

14    considered, but a writing had to be prepared as a cautionary 

15    matter.  Why a writing only rejecting adjustment, you might 

16    say?  Ha, they say, because the stipulation and order 

17    requires a written opinion if there's a decision rejecting 

18    adjustment, it doesn't require an opinion, a written 

19    decision document if it was in favor of adjustment.  That's 

20    the explanation.  

21               We will show that that is an incredible story, 

22    especially in view of the Commerce Department's reported 

23    concern that any decision be clearly articulable and 

24    articulated and explained to the American public and the 

25    professional community.  Indeed the Secretary's own 
                                                              23

 1    guideline, number 8, states that the ability to articulate 

 2    clearly the basis and implications of the decision whether 

 3    or not to adjust was to be a factor in the decision.  The 

 4    general rationale for the decision must be clearly stated.  

 5    So far as appears, your Honor, no one at Commerce was 

 6    assigned to work on the task of articulating clearly or 

 7    otherwise a decision in favor of adjustment.  Apparently no 

 8    one thought it was worth the effort.  

 9               And so it seems only reasonable, I think, to 

10    suppose that the decision against adjustment has been 

11    reached by the time that Mark Plant was told to write such a 

12    decision, if not indeed long, long before.  

13               We will show that the senior commercial officials 

14    had long before expressed their hostility to adjustment.  

15               But even on their account of the story, at least 

16    by July 9, General Counsel Willkie testifies, he knew what 

17    the decision was.  

18               Now, June 29 to July 9 is pretty fast, even at 

19    that.  What happened in the interval?  The Secretary met on 

20    July 1 with Director Bryant.  He met on July 2 with General 

21    Counsel Willkie, his Chief of Staff, his Counselor and Mark 

22    Plant.  And he met on July 8th with Dr. Bryant, four members 

23    of the Undercount Steering Committee, Undersecretary Darby, 

24    Mark Plant, four lawyers including General Counsel Willkie 

25    and Assistant Attorney General Gerson.  That meeting appears 
                                                              24

 1    from the very manicured minutes that have been produced to 

 2    have produced a lively formal sort of debate from which the 

 3    Secretary departed before it had reached its conclusion.  

 4               And that's that.  By the following day even on 

 5    Willkie's account the matter had been decided.  

 6               Who or how anyone, let alone the Secretary, had 

 7    given consideration to what is represented to be an 

 8    administrative record so-called consisting of 1522 documents 

 9    and 18,490 pages, is designed to remain an unresolved 

10    mystery, since defendants have blocked us from deposing and 

11    have refused to call anyone who could really know.  

12               Several dozen of those documents I should point 

13    out are all dated after July 9, including importantly the 

14    final loss function analysis that added significant 

15    confirmation of the Bureau's recommendation.  

16               Now, we have been unable to examine Secretary 

17    Mosbacher or Undersecretary Darby and Mark Plant doesn't 

18    know when the decision was made until he saw the Secretary 

19    sign it on July 15th.  And so it is virtually impossible to 

20    know what the Secretary did or did not take account of.  

21               It seems evident, however, that Mark Plant 

22    doesn't know either.  And therefore, his certificate, 

23    executed last month, representing that that 18,000 page 

24    so-called administrative record constitutes an 

25    administrative record of the decision, is I submit 
                                                              25

 1    worthless.  

 2               What we do know about the administrative record 

 3    is that it was put together after the decision in accordance 

 4    with instructions from defendant's counsel and that as trial 

 5    counsel have told the court, its contents are central to the 

 6    theories of defendant's counsel regarding this case and 

 7    their advice concerning the contours of that record 

 8    necessarily was given with an eye to litigation.  

 9               Now, in the meanwhile, the writing of this 

10    opinion proceeded in the hands of a very large crew.  I 

11    won't identify all the participants but they included one 

12    representative of Commerce, of the Census Bureau, Mr. Peter 

13    Bounpane who they are calling and it included eight lawyers, 

14    four from the Department of Justice, including happily Mr. 

15    Sitcov and Mr. Millet.  

16               And on July 15th the Secretary issued his 

17    decision and this case is about that decision, and he says 

18    in his opening words "Reaching a decision on the adjustment 

19    question has been among the most difficult decisions I have 

20    ever made."  

21               And well he might say so.  Our evidence will show 

22    that it is utterly groundless, that all its convoluted 

23    obscurity is a sham, dressed up to look like scientific 

24    analysis but utterly falacious, root and branch.  Indeed, 

25    your Honor, if it were a brief submitted by counsel in this 
                                                              26

 1    court, counsel would be subject to serious rebuke for 

 2    serious and patent misuse of the record.  

 3               We will focus both on what the decision omits and 

 4    what it includes.  What it omits is critical and what it 

 5    includes is erroneous.  

 6               Our first witness will be Dr. Eugene P. Ericksen, 

 7    Professor of Statistics and Sociology at Temple University 

 8    and co-chairman of the Special Advisory Panel.  Dr. Ericksen 

 9    will show that the Secretary's decision fails to take 

10    account of what is known about the systematic bias that is 

11    built into the very structure of the enumeration.  That 

12    structural bias is of a fundamental importance both in 

13    comparing the relative accuracy of the corrected and 

14    uncorrected counts, and in confirming the reliability of the 

15    PES.  

16               Indeed, the Secretary ignores the fact, confirmed 

17    by the government accounting office, that the 5,000,000 net 

18    undercount to which we have referred is a net figure, 

19    composed, resultant from the combination of some 25 to 

20    30,000,000 erroneous enumerations and omissions.  In short, 

21    truly massive error in the enumeration.  

22               The Secretary's decision asserts at the very 

23    outset that if we change the counts by a computerized 

24    statistical process, we abandon a 200 year tradition of how 

25    we actually count people.  Dr. Ericksen and others will 
                                                              27

 1    demonstrate that that statement reflects a profound 

 2    misconception of the census.  Their testimony will make it 

 3    clear that the Secretary's reported concern for public 

 4    confidence in the enumeration amounts to a hopeless attempt 

 5    to persuade the public that the emperor is fully clothed.  

 6               Dr. Wolter and Dr. Bailar, to whom I previously 

 7    referred, will show that the Secretary's decision ignores 

 8    the consensus that formed around the fundamental design of 

 9    the PES.

10               THE COURT:  Are they the second and third 

11    witness?

12               MR. RIFKIND:  They are the second and third.  I 

13    say this with a little baited breath because you know with 

14    academicians and so on, I can summon them but their order 

15    sometimes is at their discretion, especially since I'm not 

16    paying them.  But I believe -- this is the order we intend 

17    to put them on.

18               THE COURT:  I seem doomed to be in the throes of 

19    dealing with academicians all my life.

20               MR. RIFKIND:  Not bad company all things 

21    considered.  

22               They will show that the Secretary's decision 

23    ignores the problems that were solved in the design process.  

24    And ignores the fact that the PES in operation performed as 

25    specified.  So that the Secretary, in effect, changed the 
                                                              28

 1    rules after the game.  

 2               We then expect to hear from Dr. Leo Estrada an 

 3    associate professor of urban planning at UCLA and a member 

 4    of the Special Advisory Panel.  He will show that the 

 5    Secretary's decision ignores the evidence that the Bureau's 

 6    demographic analysis is competent to give, the evidence I 

 7    referred to earlier, while relying on selected bits of 

 8    demographic data for purposes for which the Secretary was 

 9    advised demographic analysis is unsuited and inappropriate.  

10               Dr. John Rolph, who I also referred to earlier, 

11    will testify that the faults found by the Secretary in one 

12    of the statistical processes the Bureau planned and used to 

13    enhance the reliability of the adjustments, the so-called 

14    smoothing process, is unwarranted.  He will show that the 

15    Bureau's operations in that regard were properly planned, 

16    properly executed, in a perfectly appropriate and 

17    professional manner.  

18               We will then hear from Dr. Franklin Fisher, a 

19    Professor of Economics at the Massachusetts Institute of 

20    Technology, who will testify about the Bureau's loss 

21    function analysis, that is the systematic process by which 

22    errors in the enumeration are weighed against errors in the 

23    adjusted counts.  He will show that the Bureau's loss 

24    function analysis provides powerful evidence of the superior 

25    accuracy of the adjusted counts and were ignored by the 
                                                              29

 1    Secretary.  He will also show that the Secretary utilized a 

 2    different, utterly inappropriate and senseless loss function 

 3    to support his decision and perhaps most strikingly 

 4    miscalculated that loss function which he elected to use.  

 5               We will also show that the Secretary's decision 

 6    rests on concerns about the distribution among the states of 

 7    those persons who are not counted either by the census or 

 8    PES.  

 9               The notion that one could ignore 5,000,000 people 

10    the PES adds because there may be others who still remain 

11    uncounted, it seems to me, in lay terms, is perfectly 

12    absurd.  But there is a great deal of straining at gnats and 

13    swallowing camels in this case.  

14               Dr. Fisher will testify that the Secretary's 

15    ostensible concern on this front would require perfectly 

16    incredible assumptions to be of any significance.  We will 

17    call Dr. Bruce Cain, Professor of Political Science at 

18    Berkley and Associate Director of Berkley's Institute of 

19    Governmental Studies.  He will speak to a variety of 

20    statements in the decision concerning the political impact 

21    of adjustment.  He will focus particularly on the 

22    disproportionate harm that the undercounting of ethnic and 

23    racial minorities causes in the political process and the 

24    serious, lopsided, partisan effects that flow from the 

25    Secretary's decision.  
                                                              30

 1               Dr. John Tukey of Princeton, a member of the 

 2    Special Advisory Panel and one of the nation's most renowned 

 3    statisticians, will testify about his recommendation to the 

 4    Secretary and the irrelevance of the Secretary's purported 

 5    concern about accuracy at lowest block level.  

 6               And finally Dr. Stephen Feinberg to whom I have 

 7    previously referred will analyze the decision document in 

 8    detail, showing that it is unsupported by the evidence 

 9    relevant to the issue of adjustment, is not a reasoned 

10    analysis of the evidence and that the Secretary either 

11    misconstrued or misstated the evidence or focused on 

12    irrelevance for purely speculative concerns.  

13               Your Honor, the arbitrariness and capriciousness 

14    with which the Secretary acted is enough to overturn his 

15    decision and reestablish the decision of the expert agency.  

16    Although I think we do not need to demonstrate any motive on 

17    the part of the Secretary and although we have been barred 

18    from inquiring of the Secretary, we intend to show your 

19    Honor that there is ample evidence of the real rationale at 

20    work in the decision, although not apparent in the decision 

21    document.  

22               When he announced his decision, the Secretary 

23    stated flatly, "I did not contact the White House, the White 

24    House had no input in this, and I neither talked to the 

25    president nor anyone else at the White House during this 
                                                              31

 1    decision-making period, nor did they seek to have input in 

 2    it."  

 3               We will show that the guilty flee where none 

 4    pursueth.  We know that statement to be false.  We know that 

 5    in March 1991, four high ranking officials at the Department 

 6    met with the White House Deputy Chief of Staff, Andrew Card.  

 7    We are told that the purpose was to discuss the correction 

 8    decision, surely a benign purpose, because, as General 

 9    Counsel Willkie testified, they discussed the preliminary 

10    PES figures.  

11               But there were no preliminary PES figures in 

12    March of 1991.  The explanation is pretextural, and there is 

13    no reference to these discussions at all in the 

14    administrative record.  

15               At that meeting, Card told the Commerce 

16    Department people to take the high road.  Was such a 

17    recommendation necessary to senior officials of the 

18    Department of Commerce?  What does Mr. Card know about 

19    Commerce that we don't know.  Or is this a "but that would 

20    be wrong" sort of statement, a cover story.  We will show 

21    that it was.  

22               The very next morning, after being cautioned to 

23    take the high road, John Sununu called General Counsel 

24    Willkie, the only time that year he called him, and told 

25    Willkie that he was very skeptical of adjustment, that he 
                                                              32

 1    thought that there was an effort, well-orchestrated, to 

 2    discredit the enumeration and that he was concerned that 

 3    Mosbacher might feel compelled to adjust because of 

 4    perceived deficiencies in the enumeration.  

 5               Mr. Mosbacher says that there was no input of any 

 6    sort from the White House but John Sununu is expressing his 

 7    skepticism, to put it gently, about adjustment and the 

 8    administrative record so-called is silent about the whole 

 9    affair.  

10               Sununu later stated we are told that adjustment 

11    would occasion major disruption in the apportionment.  

12               Not only are Sununu's overtures significant here, 

13    both for what they are and for the complete absence of this 

14    in the administrative record, but the mindset of the 

15    Commerce Department is also telling.  When John Sununu 

16    called Willkie, who reported the conversation to Mosbacher 

17    but doesn't remember what Mr. Mosbacher said about it, what 

18    did Willkie say to Sununu.  He told Sununu not to worry, 

19    that adjustment was no panacea, that people could see from 

20    the data published that adjustment had its limitations.  

21    This is March '91, when the Secretary is supposed to be 

22    approaching the matter with an open mind and without 

23    prejudgment.  

24               More important, this is at a date, Willkie is 

25    saying these things about a date, when there was no PES data 
                                                              33

 1    at all.  

 2               Both the revealed hostility to adjustment as well 

 3    as the repeated invocation of unsustainable cover stories 

 4    suggests that all of the careful analysis performed by the 

 5    Bureau was doomed to fall on unhearing ears, just as it did 

 6    in 1987.  

 7               We will show that the White House involvement did 

 8    not stop in 1991, although I would have thought Sununu's 

 9    comments at the time were sufficient.  We will show that on 

10    July 5, while everyone is helter-skelter busy getting the 

11    decision out, Mr. Willkie is busily calling, frantically 

12    calling White House General Counsel Boyden Gray to enlist 

13    Boyden Gray's help in deterring White House involvement in 

14    the merits of whether or not to adjust apart from legal 

15    issues.  

16               Why was such a deterrence necessary?  We will 

17    show that Willkie wanted to keep Sununu out of the issue.  

18    So apparently Mr. Willkie was sufficiently concerned that 

19    Mr. Sununu was still making himself felt and heard.  

20               We will also show that there were improper 

21    contacts by the White House counsel's office with members of 

22    the Special Advisory Panel.  

23               Now, the defendants explain that the reason for 

24    that contact was that so that the White House would advise 

25    counsel about legal issues relating to the adjustment 
                                                              34

 1    question.  This is another one of those stories that doesn't 

 2    hold water.  

 3               The Commerce Department has its own legal 

 4    department.  It has ample access to the Justice Department.  

 5    The White House is not where one normally goes in Washington 

 6    to get legal advice.  They certainly didn't have to call the 

 7    White House for legal advice on a constitutional question 

 8    and it is very hard to suppose that where a lawyer at the 

 9    White House would go for advice on legal questions would be 

10    to statisticians on the Special Advisory Panel who are not 

11    normally thought to be repositories of information about the 

12    statutory and constitutional provisions that govern.  

13               This episode, too, is whitewashed by Willkie who 

14    redrafted the very letter he was to receive from the White 

15    House explaining what had happened in all these events.  

16               The Secretary, in announcing his decision said, 

17    if you look at the parties, it's pretty hard to see there's 

18    any measure of gain for either party.  Your Honor, that is 

19    no truer than his statement about White House involvement.  

20               We will show that the impact of adjustment is not 

21    so much at the congressional level, although two 

22    Congressional seats are at stake, but at the level of state 

23    houses and state legislatures.  We will show that the 

24    president himself has for years been intensely interested in 

25    the redistricting of state legislatures, a point that could 
                                                              35

 1    hardly have been lost on the gentleman who now serves as his 

 2    campaign manager.  

 3               And there, in the state legislatures, in the 

 4    words of John Sununu, adjustment would occasion a major 

 5    disruption, a partisan disruption.  We will prove that the 

 6    uncorrected census disproportionately dilutes the vote of 

 7    racial and ethic minorities who overwhelmingly vote 

 8    democratic or more important live in areas that vote 

 9    democratic.  

10               Your Honor, this is not an occasion for an 

11    argument about the law, and I shall not try to make one.  

12    But I do think it appropriate to note one principal point of 

13    agreement with a thought expressed rather emphatically in 

14    defendant's trial brief.  

15               They assert again and again that special 

16    deference is due agency choices involving scientific 

17    disputes that are in the agency's province of expertise.  I 

18    wholeheartedly agree, that the expert agency here is the 

19    Bureau of the Census.  

20               No one seriously asserts the scientific or 

21    statistical expertise of Robert Mosbacher.  He claims none.  

22    I submit the core issue here is the deference he should have 

23    paid to the Bureau.  

24               The Secretary has formally delegated to the 

25    Director of the Bureau "Authority to perform the functions 
                                                              36

 1    vested in the Secretary of Commerce under Title 13 U.S.C."  

 2               Assuming that the Secretary could revoke that 

 3    delegation as to this one piece of the entire census 

 4    operation, assuming that somehow he did revoke it, I submit 

 5    that he could not do so while continuing to wrap himself in 

 6    the mantle of an expert agency.  Neither common sense nor so 

 7    far as I'm aware any judicial precedent provides any support 

 8    for the notion that some heightened difference is due an 

 9    inexpert cabinet officer who overturns an expert agency on a 

10    matter of scientific complexity.  

11               We will show that the Secretary's decision 

12    violates the Constitutional command that the census be as 

13    accurate as practicable, we will show that the decision was 

14    arbitrary and capricious, we will show that the decision 

15    violates the stipulations and orders requirement that it be 

16    reached with an open mind and without prejudgment.  And all 

17    the other fortifications that I find erected around the 

18    decision in defendant's trial brief I submit are so much 

19    heaping Ossa on Pelion.  

20               One has to wonder about a decision that occurs 

21    behind such walls of presumption after presumption, as my 

22    distinguished colleagues seek to erect.  One feels that it 

23    is going to look like an anemic and paltry thing when it is 

24    exposed to the light of day after those walls have crumbled.  

25    Many of those fortifications, as your Honor knows, have been 
                                                              37

 1    previously knocked down by counsel, including 

 2    nonjusticiability, nonreviewability, nondiscoverability, no 

 3    evidentiary hearing are badly undermined.  But the rest are 

 4    also very badly undermined.  They were undermined when the 

 5    Secretary took the decision into his own hands while 

 6    expressing grave concern about the appearance of 

 7    partisanship if the matter were left to the professional 

 8    civil service.  

 9               They were undermined when the Secretary adopted 

10    guidelines which the court held constitute adequate but 

11    minimal performance to satisfy their obligations under the 

12    stipulation and order, and warned that defendants incurred a 

13    heavier burden to explain why no adjustment was made in the 

14    event the Secretary elected to proceed with the actual 

15    enumeration.  

16               The Secretary undermined them himself when he 

17    permitted the decision to be enwrapped in a so-called 

18    administrative record put together after the fact as a 

19    litigation tool under the direction of his trial counsel.  

20               Your Honor, I am keenly aware that you have been 

21    sensitive to the importance of this case.  Of course all 

22    litigants feel their case is important.  But I would feel 

23    derelict if I didn't observe in concluding what's at issue 

24    here is the mechanism by which we apportion power in the 

25    United States, and thus, in the last analysis, the 
                                                              38

 1    legitimacy of our representative institutions.  

 2               The sense, justified sense, that there is a thumb 

 3    on the scale, a thumb disadvantageing the most vulnerable 

 4    portions of our population, is profoundly upsetting.  If it 

 5    is not dealt with in a responsible and professional fashion, 

 6    I very much fear it will be dealt with, it will be taken up 

 7    by, it will be pursued by those who do not proceed in a 

 8    responsibly professional manner and one has only to look at 

 9    what has happened in Los Angeles in the last couple of weeks 

10    to reflect on what happens when people lose confidence in 

11    the impartialty of our institutions.  

12               Your Honor, all cases involve justice.  But I 

13    submit that this case involves justice in its classic, most 

14    classic sense.  As Plato tout us in The Republic "Justice 

15    entails rendering every man his due."  That is what we seek.  

16               Thank you.

17               THE COURT:  Mr. Rifkind, thank you.  Let's take a 

18    ten minute break before we hear the government's opening 

19    statement.  

20               (Recess)

21               THE COURT:  Now we'll hear from the government.

22               MR. SITCOV:  May it please the court, I am 

23    Michael Sitcov, an attorney at the Department of Justice in 

24    Washington, D.C.  Along with my colleagues, Tom Millet, 

25    Judry Subar, Jason Baron and David Glass, I have the 
                                                              39

 1    privilege of representing the defendants in this case.  

 2               Before I get to my prepared remarks, your Honor, 

 3    I would like to express my outrage at Mr. Rifkind's attempt 

 4    to reraise for the umpteenth time the issue of bad faith on 

 5    the part of government employees involved in the 

 6    decision-making process.  Twice the plaintiffs have filed 

 7    briefs in this case in which they have expressly alleged 

 8    that the decision-making process was infected by bad faith.  

 9    And each time, each time those attempts were rebuffed by the 

10    court.  

11               And the reason that those attempts were rejected 

12    by the court, is the reason that you ought not to pay any 

13    attention to them now.  And that is because they're 

14    completely baseless.  

15               We have finally reached the point, your Honor, 

16    where the plaintiffs are going to have to put up or shut up 

17    on that score.  And at the end of this trial, I don't think 

18    you're going to be hearing from the plaintiffs on that score 

19    any more.  

20               Now, because this case is one under the APA, the 

21    Secretary's decision cannot be disturbed unless it is shown 

22    to be arbitrary and capricious.  In other words, even if the 

23    court determines that Secretary Mosbacher's decision was 

24    wrong, the decision must be affirmed unless the court finds 

25    that it was totally unreasonable as well.  And because this 
                                                              40

 1    is a suit under the APA, the only remedy in the event that 

 2    the court finds that the Secretary's decision was 

 3    unreasonable, is a remand to the Secretary with instructions 

 4    to reconsider his decision.  

 5               Under the APA, the court is not empowered to 

 6    decide the adjustment issue itself.  

 7               Now, the precise question before the court is 

 8    whether the Secretary acted in an arbitrary and capricious 

 9    manner when he decided that the proposed adjustment did not 

10    satisfy the criteria established in the guidelines that you 

11    approved three years ago.  

12               It's interesting that the plaintiffs are 

13    complaining now that the Secretary made the decision and 

14    suggested there is some impropriety about that when the 

15    guidelines that they agreed to expressly state that the 

16    decision was the Secretary's.  If that was a problem for 

17    them, three years ago was the time to make that known.  

18               Now, these guidelines also provide that the 

19    census is to be considered more accurate than an adjustment 

20    unless it can be shown that adjusting the 1990 decennial 

21    census will result in a more accurate picture of the 

22    proportional distribution of the population of the United 

23    States on a state by state basis.  

24               That's the same test Judge Sprizzo applied in 

25    Cuomo v. Badridge.  That was the suit some of these 
                                                              41

 1    plaintiffs brought to force an adjustment of the 1980 

 2    census.  It's worth mentioning, at least, that a number of 

 3    the witnesses in this case were witnesses on that case 

 4    although a couple have changed sides.  

 5               One thing that hasn't changed is that the case 

 6    for adjustment is as weak today as it was then.  Our 

 7    evidence will make it clear that the Secretary was right to 

 8    decide that adjustment fails Judge Sprizzo's test and the 

 9    Secretary's decision was not tainted in any way.  

10               First, I'm going to summarize what the evidence 

11    will show about how adjustment is done.  And I think that's 

12    important, especially in this case, your Honor, because in 

13    some ways, adjustment is kind of like salami.  You might 

14    like the way salami tastes, but you might not eat it if you 

15    saw how it was made.  

16               Adjustment uses a process called dual system 

17    estimation or capture recapture.  The principal use of dual 

18    system estimation is for estimating animal populations in 

19    the wild and it works like this.  

20               Let's suppose you wanted to know how many fish 

21    there are in a lake but you don't want to drain the lack 

22    because then all the fish will be dead.  Instead you capture 

23    a thousand fish let's say and you tag them and you throw 

24    them back.  That's the capture.  

25               Then you catch another hundred fish.  That's the 
                                                              42

 1    recapture.  And it turns out 90 of them are tagged.  That's 

 2    suggests that 90 percent fish in the lake are tagged.  To 

 3    compute the total population of fish in the lake you decide 

 4    1,000 by .90 and you add the result to the number you 

 5    originally captured.  

 6               Now, in the situation of census adjustment the 

 7    capture is the census and the recapture is the 

 8    postenumeration survey.  The census was taken as of April 1, 

 9    1990.  And the PES, this postenumeration survey, began 

10    several months later.  To take the PES, the Bureau sent 

11    census enumerators door to door in a random sample of 5300 

12    block clusters consisting of about 12,000 of the nation's 

13    almost 6 million blocks.  Enumerators attempt to identify 

14    everyone who is living in the sample blocks and collect 

15    information about their address as of April 1, 1990.  

16               The evidence is going to show that the Bureau 

17    attempted to match the PES questionnaires with the census 

18    questionnaires for the PES sample blocks.  In the vast 

19    majority of cases, people counted in the PES had also been 

20    counted in the census.  In some cases, people are found in 

21    the PES who are not found in the census.  These people are 

22    called gross omissions.  

23               In still other cases, people are counted in the 

24    census who shouldn't have been counted.  These are called 

25    erroneous enumerations.  An erroneous enumeration would 
                                                              43

 1    include children who were born after April 1, 1990, people 

 2    who died before April 1, 1990 but somehow made it on to a 

 3    census form, a college student who was enumerated at his or 

 4    her parent's address instead of his or her college or people 

 5    counted twice for any reason.  

 6               The undercount is the difference between the 

 7    gross omissions rate and the erroneous enumerations rate.  

 8               To calculate population estimates, the population 

 9    was divided into 1392 demographic groups called poststrata.  

10    Every person in the United States fits into one and only one 

11    poststratum.  Poststrata are defined by census division, 

12    which is a geographical location, for example the Pacific 

13    division.  The type of place of residence, is it a home or 

14    an apartment.  The tenure of residence, do the people who 

15    live there rent or own the house, race, Hispanic, ethnicity, 

16    sex and age.  One poststratum, 9032124, consists of 

17    non-black Hispanic females ages 30 to 44 living in rental 

18    housing in cities in the Pacific division.  

19               By using this dual system estimation process, the 

20    PES results in the census results for the blocks that were 

21    included in the PES are compared and a raw adjustment factor 

22    was developed for each poststratum.  Theoretically, that raw 

23    factor estimated the percentage by which members of the 

24    poststratum were undercounted or overcounted in the census.  

25               Now, tagging fish is easier than identifying the 
                                                              44

 1    same people twice in two different surveys taken months 

 2    apart.  The problem becomes especially acute if you add to 

 3    the mix the additional problem of missing data.  Believe it 

 4    or not, not everybody responds completely or even honestly 

 5    to the census and there are some neighborhoods where when 

 6    you knock on the door and you say hi I'm from the government 

 7    and I'm here to help, you don't get cooperation.  

 8               The result is missing data and that can make it 

 9    impossible to decide whether someone in the PES is the same 

10    person as someone in the census.  To overcome this problem, 

11    we're going to show that the Bureau had to use a model, a 

12    statistical model, that relies on unverified assumptions to 

13    fill in missing characteristics needed to make a match.  

14               Now, even after this process is complete, and 

15    you've got complete enough data to supposedly match everyone 

16    in the PES to someone in the census, in the sample blocks, 

17    there are still people in the PES for whom it is impossible 

18    to determine whether or not they match a census person or 

19    whether the person in the census really exists.  

20               An entirely different set of models is used to 

21    impute match status and enumeration status.  We will show 

22    that these models were never tested before they were used in 

23    the 1990 census.  And we will also show that the Census 

24    Bureau evaluation studies that considered the accuracy of 

25    the models strongly suggest that the model used to impute 
                                                              45

 1    match status doesn't work.  

 2               Now, since the PES was a sample of the nation's 

 3    total population, the adjustment factors were subject to 

 4    sampling error.  Now the plaintiffs often point to the fact 

 5    that the PES was a huge sample as sample surveys go, and 

 6    that's true but the problem is the PES was a mile wide and 

 7    only about an inch deep.  There were 400,000 people in the 

 8    PES and they were divided into 1392 poststrata.  So on 

 9    average, each poststratum only included about 300 people.  

10               And the numbers of gross omissions and erroneous 

11    enumerations were small.  Generally less than two dozen per 

12    poststratum.  For that reason, the sampling errors in the 

13    raw adjustment factors were so high that the resulting 

14    undercount estimates were too unstable to use.  

15               Here I think it might be helpful if I could 

16    explain for a moment what sampling error is.  Like any 

17    sample, the PES is only part of the population.  And when a 

18    sample is used to estimate a numerical characteristic of the 

19    entire population, the estimate will differ from the exact 

20    value if only because the sample is not a perfect microcosm 

21    of the whole.  

22               Sampling error is the difference between the the 

23    exact value of something and the estimate of that thing 

24    based on the results of the sample.  The smaller the sample 

25    the bigger the sampling error and the less stable the 
                                                              46

 1    estimate.  

 2               Sampling errors are a big problem in the raw 

 3    adjustment counts because as I just mentioned the number of 

 4    nonmatches and erroneous enumerations for poststratum is so 

 5    small.  

 6               We will show that in an effort to reduce sampling 

 7    error, the Census Bureau employed a statistical procedure to 

 8    develop so-called smooth adjustment factors.  The smoothing 

 9    process uses regression models to predict adjustment factors 

10    on the basis of what are called explanatory variables.  At 

11    least in theory, the explanatory variables were assumed to 

12    be related to undercounting.  

13               However, we will show quite soon that the 

14    evidence from the PES refutes that notion.  In any event, a 

15    regression model was used to predict values for the 

16    adjustment factors from the explanatory variables.  The 

17    predicted values were then averaged with the raw adjustment 

18    factors to produce smooth adjustment factors.  

19               Now, we're going to explain exactly how the 

20    smoothing process worked and I think it's important for the 

21    court to understand this.  In brief it goes like this.  

22               There are two steps in smoothing.  First the 

23    variances of the raw adjustment factors were smoothed using 

24    a statistical model.  Now the variance of an adjustment 

25    factor is a measure of its sampling error.  The variances 
                                                              47

 1    control the extent to which the smooth factor are more like 

 2    the predicted values or more like the raw values.  

 3               After the variances were smoothed, the smoothed 

 4    variances were used to smooth the adjustment factors 

 5    themselves.  We will show that the Census Bureau's 

 6    evaluation of adjustment did not include a single study 

 7    evaluating the smoothing process or any of the assumptions 

 8    upon which it is based.  

 9               (Continued on next page) 

10    

11    

12    

13    

14    

15    

16    

17    

18    

19    

20    

21    

22    

23    

24    

25    
                                                              48

 1    

 2    

 3               (MR. SITCOV:)  This is particularly unfortunate, 

 4    in light of the fact that smoothing in this context is 

 5    controversial, to say the least, and presmoothing has only 

 6    rarely ever been used outside the 1990 PES.  We are going to 

 7    show quite clearly that the smoothing and presmoothing 

 8    processes didn't do what they were supposed to do.

 9               Once final adjustment factors or computed block 

10    level census counts are adjusted synthetic, that is, the 

11    number of people in each poststratum, in a block, was 

12    multiplied by the appropriate adjustment factor for that 

13    poststratum.

14               Adjusted counts for larger areas -- states, 

15    cities, counties, congressional districts, and so forth -- 

16    were obtained by adding up the adjusted counts, the blocks, 

17    within them.

18               To illustrate exactly how this synthetic 

19    adjustment operates, let's think about the poststratum I 

20    referred to earlier, that's Hispanic females, age 30 to 44, 

21    living in rental housing in cities in the Pacific division.

22               Suppose the adjustment factor for this 

23    poststratum is 1.10 -- in other words, the PES has 

24    determined that people in this poststratum were undercounted 

25    in the census.  If a block in Oregon was found in the 
                                                              49

 1    enumeration to have 10 people that fit into that poststratum 

 2    an adjustment would have required the defendants to multiply 

 3    that number by 1.1.  In the example I've given you, that 

 4    would require an upward adjustments of a block total by 1.  

 5    This would be accomplished by randomly selecting one of the 

 6    ten real census records for someone in this poststratum in 

 7    that block and duplicating it.

 8               Now, there's another problem.  Since that person 

 9    wasn't found in the block, you've got to put them someplace.  

10    So the Census Bureau creates a residence on that block that 

11    puts that adjusted person into the residence; and if there 

12    are more than one poststratum being adjusted upward in a 

13    block, then they'll all go into this group residence.

14               This process is repeated for every block in the 

15    United States in which members of this poststratum had been 

16    counted in the census without regard to whether there had 

17    been an undercount of the poststratum in any particular 

18    block.  The result, theoretically, is that the poststratum 

19    count would be increased nationwide by 10 percent.

20               The adjustment process would also be applied to 

21    poststrata that, according to the PES, were overcounted.  In 

22    these cases, the adjustment factors would be something less 

23    than one.  And in a case of that type, an adjustment would 

24    delete a representative of that poststratum from every block 

25    in which they are represented.  Even if they had been 
                                                              50

 1    properly counted there in the first place.

 2               We are going to show that behind all this 

 3    arithmetic there is a critical assumption, that is, that the 

 4    undercount rate for a poststratum is the same for every 

 5    single block in the country.

 6               For example, using this assumption, a black woman 

 7    who is a partner at a major New York law firm and who lives 

 8    in an apartment on the Upper East Side of Manhattan is 

 9    considered to be no more or less likely to have been 

10    captured in the census than an unemployed black woman of the 

11    same age living in rental housing in a poor neighborhood in 

12    New York.  This homogeneity assumption drives a small area 

13    adjustment.  If this assumption is wrong, as we will show it 

14    is, so are the adjusted counts for small areas.

15               That's sort of a thumbnail sketch of how 

16    adjustment works.  And as you can see, there are assumptions 

17    at virtually every step of the adjustment process.  If the 

18    assumptions fail, then the whole thing goes to pot in a 

19    hurry.  We are going to show that the evidence in favor of 

20    these assumptions, especially the critical ones involving 

21    smoothing, and so forth, is essentially nonexistent.

22               Now, here in the detail, in some more detail, 

23    anyway, is how we are going to do it.  First, you will hear 

24    from Mr. Peter Bounpane.  Mr. Bounpane is a career 

25    professional at the Census Bureau.  He's going to testify 
                                                              51

 1    about how the census is actually done.  He'll describe the 

 2    enormous efforts that the Census Bureau made to get the best 

 3    count of the entire population and the special efforts the 

 4    Census Bureau made to count the traditionally undercounted 

 5    groups.  He's going to show that the Census Bureau did a 

 6    great job of actually counting the population.  And he's 

 7    also going to testify that a large part of the measured 

 8    undercount is due to problems in the measuring process.

 9               Next, you will hear from Dr. Robert Fay, a senior 

10    statistician at the Census Bureau.  Like Mr. Bounpane, Dr. 

11    Fay was a member of the Bureau's Undercount Steering 

12    Committee.  That's the committee to which Mr. Rifkind 

13    referred that recommended by a split vote in favor of 

14    adjustment.  The committee was composed of senior career 

15    officials at the Census Bureau who monitor the census and 

16    the PES results and the results of the Bureau's evaluation 

17    program.  Dr. Fay was one of the majority of the Undercount 

18    Steering Committee members who favored adjustment before 

19    July 15th, 1991.  He will testify in court that his opinion 

20    reflected the fact that he lacked the time to do as complete 

21    a statistical analysis as he would have liked to do because 

22    of the July 15th deadline imposed by the stipulation and 

23    order.

24               THE COURT:  Was Dr. Bryant the chairperson in 

25    that committee?  
                                                              52

 1               MR. SITCOV:  Dr. Bryant?  No.

 2               THE COURT:  She was on it.  

 3               MR. SITCOV:  No, I believe she would have been -- 

 4    she could have attended, but she was not on it.  It was 

 5    career census professionals.

 6               THE COURT:  All right.  

 7               MR. SITCOV:  Since then, Dr. Fay has had time to 

 8    do additional research, and that additional research has 

 9    persuaded him that the decision to adjust the census on July 

10    15th, 1991, would have been wrong.

11               After Dr. Fay you will hear from Dr. Paul Meier.  

12    He is a professional of statistics at the University of 

13    Chicago, and he is one of the most distinguished of academic 

14    statisticians in the United States.  Dr. Meier is going to 

15    discuss what a census should be, that is, an important 

16    national ceremony that provides the best available count of 

17    the population.  He will discuss statistical adjustment of 

18    the census in the abstract and the limits of statistics in 

19    resolving lawsuits.  In this regard, it's worth noting that 

20    Dr. Meier also is a lecturer at the University of Chicago 

21    Law School.

22               Finally, he will testify that failures and key 

23    assumptions in the dual system estimation process make the 

24    PES results unusable for adjustment.

25               Following Dr. Meier, you will hear from Professor 
                                                              53

 1    Leo Breiman, who is a Professor of Statistics at the 

 2    University of California Berkeley.  He's going to testify 

 3    about nonsampling error.  It's also referred to as bias, 

 4    postinformation survey.

 5               As you are going to see, adjustments involves a 

 6    whole lot more than just high statistical theory and fancy 

 7    computer work.  For example, everyone in the P sample is 

 8    interviewed during the PES, and then an attempt is made to 

 9    match them to people in the census.  This is not rocket 

10    scientists, your Honor.  This is a bunch of people with a 

11    bunch of complicated forms that often are filled out legally 

12    or correctly and they are trying to do the matching in maybe 

13    cases by hand.  It's a error-prone process.

14               For instance, people who do the PES interviews 

15    make mistakes.  As I suggested earlier, people interviewed 

16    in the census sometimes aren't honest.  And people who 

17    matched the PES results to the census forms sometimes make 

18    mistakes.  Because the PES is just a tiny fraction of the 

19    total U.S. population, each person in the PES represents 

20    many people in the total population.  Consequently, clerical 

21    mistakes in the PES can be quite costly in terms of 

22    accuracy.

23               For example, Dr. Breiman will testify about a 

24    single group of PES records that was miscoded by census 

25    clerks.  This mistake improperly inflated the undercount 
                                                              54

 1    estimate by almost three quarters of a million people.  Dr. 

 2    Breiman has studied the nonsampling errors the Census Bureau 

 3    considered and some they didn't.  And he'll testify that 80 

 4    percent of the net undercount, that the PES supposedly 

 5    found, can be attributed directly to these types of 

 6    problems.

 7               Demographic analysis, as Mr. Rifkind suggested, 

 8    is a way of estimating the population at the national level.  

 9    It considers birth and deaths, immigration and out- 

10    immigration on the basis of administrative records.  As Mr. 

11    Rifkind quite eloquently stated, one of the plaintiff's 

12    arguments in favor of adjustment is that the PES agrees with 

13    demographic analysis at the national level.

14               On this score, you are going to hear from Dr. 

15    Kenneth Wachter, a Professor of Demography and Statistics at 

16    the University of California Berkeley.  Dr. Wachter has 

17    undertaken his own study of the demographic analysis totals 

18    for the U.S. population, and he is going to testify, with 

19    plausible rates of undercount, that, based on demographic 

20    analysis, can be significantly lower than what the Bureau 

21    found.  That makes the population totals derived from 

22    demographic analysis on the one hand and the 

23    post-enumeration survey on the other farther apart than the 

24    plaintiffs would like you to believe they are.  

25               Dr. Wachter is also going to testify that other 
                                                              55

 1    original research he has done for the defendants 

 2    demonstrates critical failures in the synthetic assumption, 

 3    which I discussed earlier. 

 4               Finally, you will hear from professor David 

 5    Friedman.   Professor Freedman is a Professor of Statistics 

 6    and Mathematics at the University of California Berkeley, 

 7    and he is the author of one of the standard college level 

 8    statistics texts.  Dr. Freedman is going to testify about 

 9    sampling error and loss function analysis.  Loss function 

10    analysis is particularly important to the plaintiffs because 

11    it is the only proof that they are going to offer for the 

12    proposition that adjustment satisfies the test of 

13    distributional accuracy.

14               As I explained earlier, sampling error is a big 

15    problem because the number of matches -- I'm sorry -- 

16    because the number of nonmatches per poststratum is so 

17    small.  The plaintiffs' counter to this problem is that it 

18    is alleviated by smoothing; and even if it isn't, loss 

19    function doesn't save the day.  

20               Dr. Freedman will testify that the magnitude of 

21    the standard error and the raw adjustment factor renders 

22    them unusable, that smooth and presmoothing don't fix the 

23    problem, and that loss function analysis does not provide a 

24    way around the sampling error problem.

25               Now, plaintiffs' case is pretty easy to state, 
                                                              56

 1    and I think cutting through some of what Mr. Rifkind said, 

 2    it's going to sound kind of like this:  There are problems 

 3    in the census, the PES fixed them, census professionals 

 4    favor adjustment and the Commerce Department overruled the 

 5    census for political reasons.

 6               Now, there is no evidence in this case to support 

 7    any of these contentions.  First, the issue is not whether 

 8    the census is flawed.  The issue is:  Does adjustment 

 9    improve the distribution of the population in the United 

10    States on a state-by-state basis?  That's what plaintiffs 

11    have to prove and that's what they cannot prove.

12               Second.  It's not overstating the matter to 

13    describe plaintiffs' evidence in favor of adjustments as 

14    unconscionably weak.  On many of the important statistical 

15    issues in the case, plaintiffs' experts will have to testify 

16    that they haven't looked at the actual PES data, that they 

17    haven't done any independent analyses of it, and that, 

18    instead, they just looked at the work of others and gave it 

19    their blessing.

20               Other of their experts will testify that they 

21    have no opinions about whether the critical assumptions in 

22    the adjustment process were valid, and one is going to 

23    flatly contradict prior sworn testimony he's already given 

24    in this case.

25               Third.  As I noted earlier, Dr. Fay will testify 
                                                              57

 1    that inquiring minds at the Bureau have changed on the issue 

 2    of adjustment.  And I'm sure that it was an oversight by Mr. 

 3    Rifkind when he was reading that lengthy passage from Dr. 

 4    Bryant's recommendation to the secretary that he neglected 

 5    to point out that she specifically said "reasonable minds 

 6    can differ on the issue of adjustment."

 7               Fourth.  As I mentioned earlier, the plaintiffs 

 8    have repeatedly attempted to paint the decision process as 

 9    one which was politically motivated, and each time the Court 

10    has rejected that argument flat.  The reason they failed 

11    before is the reason they are going to fail every time they 

12    try another tactic.  The decision was based -- was made on 

13    its merits, and the evidence will show that it was made 

14    correctly.

15               With respect to Mr. Willkie, we have submitted 

16    large portions of the deposition transcript of Mr. Willkie, 

17    and they completely refute the notion that there was any 

18    political machinations involving the white house in the 

19    adjustment decision.  There is no evidence of that.  

20    Plaintiffs have failed every time they have tried to make 

21    such a showing before the Court.

22               Finally, your Honor, we have argued and we 

23    continue to believe that trial is not only unwarranted in 

24    this case but that it's improper.  The APA does not have one 

25    set of rules for important cases and another set for less 
                                                              58

 1    important claims.  The Supreme Court's most recent decisions 

 2    on the APA make it clear that except in circumstances that 

 3    aren't present here, judicial challenges and agency action 

 4    are to be resolved on the administrative record.  The record 

 5    is insufficient if it's not supportive of the agency's 

 6    decision and the remedy is to send the decision back.

 7               We submitted an administrative record of more 

 8    than 18,000 pages in support of the Secretary's decision.  

 9    The Court has never suggested that it's incomplete, that 

10    it's confusing or that it doesn't support the Secretary's 

11    decision.

12               In this situation, there is no basis for a trial 

13    in this case.

14               Thank you.

15               THE COURT:  Thank you.

16               And about "no basis for the trial," we'll start 

17    by calling the first witness.  

18               MR. ZIMROTH:  Your Honor, the plaintiffs call 

19    Eugene Ericksen. 

20               MR. SITCOV:  Your Honor, might I ask that the 

21    podium be moved a little bit?  I can't see the witness at 

22    all.  

23               MR. ZIMROTH:  I'm wondering, your Honor, if we 

24    can't move that small table here so I can put my exhibits on 

25    it.
                                                              59

 1               (Pause)

 2    EUGENE P. ERICKSEN, 

 3         called as a witness, having been 

 4         duly sworn, testified as follows:

 5               THE COURT:  Mr. Zimroth.  

 6    DIRECT EXAMINATION 

 7    BY MR. ZIMROTH:

 8         Q.    Mr. Ericksen, can you tell His Honor where you 

 9    teach?

10         A.    I teach at Temple University.

11         Q.    And in what departments do you teach?

12         A.    I'm a Professor of Sociology and Statistics.

13         Q.    I'm sorry, Professor Ericksen?

14         A.    I'm a Professor of Sociology.

15         Q.    I have a problem in hearing in one ear.

16         A.    In statistics.  

17         Q.    Excuse me?

18         A.    Sociology and statistics.

19         Q.    Thank you.  

20               What are the areas in which you concentrate your 

21    work?

22         A.    Well, the area that I studied the longest is in 

23    the taking of sample surveys.

24         Q.    And the other areas?

25         A.    Second area is how you calculate estimates for 
                                                              60

 1    areas when the Decennial census is not available.  I'm 

 2    talking about local areas like states and counties and 

 3    cities, and the kinds of variables that I've worked on or 

 4    things like estimating the population size for states in the 

 5    year like 1994, for example.

 6         Q.    And do you also do any work in urban sociology?

 7         A.    Yes.  I've also done work in the field of urban 

 8    sociology in particular, and in the field of demography and 

 9    statistics in general.

10         Q.    What do you mean by when you say "urban 

11    sociology"?

12         A.    By urban sociology, it's the manner -- we study 

13    the manner in which people live in cities, we study the 

14    relationships between groups in cities, and we also study 

15    many of the social and economic problems in cities.

16         Q.    And you mentioned "demography."  Can you tell 

17    Judge McLaughlin what demography is?

18         A.    Yes.  Demography is the standard of 

19    characteristics of populations, how populations grow, how 

20    they become smaller.  Demographers are particularly 

21    interested in data such as births, deaths, migration -- to 

22    give you some examples.

23         Q.    You mentioned that the area that you've been 

24    studying the longest is sample surveys.  How long have you 

25    been working in that area?
                                                              61

 1         A.    Over 25 years.

 2         Q.    Professor Ericksen, in your chosen field, have 

 3    you recently received an honor that you'd like to tell the 

 4    Court about?

 5         A.    Yes.  I was recently named to be a Fellow of the 

 6    American Statistical Association.

 7               THE COURT:  Your life's ambition.

 8               THE WITNESS:  Take them when they come.  

 9    BY MR. ZIMROTH:

10         Q.    When did you first begin doing research in issues 

11    concerning census data?

12         A.    Somewhere around 1968 I wrote the proposal for my 

13    doctoral dissertation.  I wrote a research proposal to the 

14    Bureau of Census in which I proposed to attempt to develop a 

15    new method by which, what we call the postcensal estimates 

16    of counting populations can be estimated.

17         Q.    Why would it be necessary to make estimates of 

18    those?

19         A.    Well, there are many uses of population data that 

20    are required for years that the single census data aren't 

21    available.  For example, there might be -- when there was 

22    revenue sharing, dollars being allocated to states and 

23    counties and cities, one of the elements that was necessary 

24    to allocate those dollars was how many people live in each 

25    of the states and counties and cities that were receiving 
                                                              62

 1    the money.  So, for example, if you wanted to know what the 

 2    population was in 1985, it's necessary to estimate that 

 3    population.

 4         Q.    And how did you go about doing that?

 5         A.    Well, basically, in the method that I developed, 

 6    one would take data from what's called the current 

 7    population survey.  The current population survey is the 

 8    survey that's used at the Census Bureau to estimate things 

 9    like the unemployment rate, and it's a rather large survey, 

10    but for local areas, which are called primary sampling 

11    units, the amount of local data was rather sparse.

12               So what -- the way the method works is, let's 

13    say, if we have a local county estimate for 1985, and we 

14    would divide it by what the census count had been in 1980, 

15    and that would be an estimate of the population growth.  

16    Then you would use a model called a regression model in 

17    which the independent variables would be ratios of things 

18    like change in school enrollment, change in tax exemption, 

19    change in births and deaths, and you'd use this regression 

20    model to estimate what the change in population had been in 

21    all the areas of the country.

22         Q.    And that was the work that you started with your 

23    doctoral dissertation?

24         A.    Yes.

25         Q.    When did you get your doctorate?
                                                              63

 1         A.    In April 1971 is when my dissertation was 

 2    finished.

 3         Q.    Did you continue working after your doctorate?

 4         A.    After my doctorate I continued my work in the 

 5    Census Bureau through a series of what is called joint 

 6    statistical contracts until some time around 1975 or 1976.

 7         Q.    You did this work in conjunction with the Census 

 8    Bureau?

 9         A.    Yes, I did.

10         Q.    Did there come a time when you began to work 

11    specifically on issues of the undercount?

12         A.    Yes.

13         Q.    And when was that?

14         A.    Some time around 1979, a statistician at the 

15    Census Bureau asked me if I thought it might be possible to 

16    use the model that I had developed for estimating postcensal 

17    population change for local areas to apply it to the problem 

18    of estimating what the undercounts were for local areas.

19               I agreed to consider that question and I was 

20    asked by the Census Bureau to write a paper and to present 

21    the results of this consideration at a conference that was 

22    held in February of 1980.

23         Q.    Did you do so?

24         A.    Yes, I did.

25         Q.    Did you continue your work concerning the 
                                                              64

 1    undercount after that conference?

 2         A.    Yes, I did.

 3         Q.    What did you do?

 4         A.    Well, in October of 1980, I was contacted by 

 5    attorneys for the plaintiffs in the New York lawsuit, and I 

 6    offered testimony in that case.

 7         Q.    And did you continue publishing articles -- did 

 8    you publish articles in this area?

 9         A.    Yes.  Since then I had continued to do research 

10    on the undercount, and I published a series of articles on 

11    that subject and continued to work on that and conducted 

12    research up until this very day.

13         Q.    Professor Ericksen, if you will look in your 

14    binder, it's binder No. 2, PX 449.

15         A.    Four-forty what?

16         Q.    9.  What is that document, please?

17         A.    That's a copy of my curriculum vitae.  

18               MR. ZIMROTH:  Your Honor, I move its admission.  

19               MR. SITCOV:  No objection.

20               THE COURT:  449 is admitted.  

21               (Exhibit PX 449 for identification was received 

22    in evidence)

23    BY MR. ZIMROTH:

24         Q.    You mentioned that you testified in 1984 

25    concerning the undercount.  
                                                              65

 1               Would you tell us the instances in which you 

 2    testified concerning the undercount?

 3         A.    Yes.  I also testified in November of 1980 in the 

 4    lawsuit brought by the City of Philadelphia.  I testified in 

 5    the remand of the original New York trial in 1984.  I was 

 6    prepared to testify in this case in July of 1989.

 7         Q.    Have you ever testified on behalf of the United 

 8    States of America?

 9         A.    Yes, I have.

10         Q.    And can you tell us what those instances were.

11         A.    Yes.  In 1987, I was retained by the U.S. 

12    Department of Justice in the case U.S. versus Yonkers 

13    Contracting Company, and in the second case called U.S. --

14         Q.    Tell the Judge what you testified about in that 

15    case.

16         A.    Oh, I was asked to analyze the composition of 

17    juries in the Southern District of New York, and to give an 

18    opinion as to whether the Hispanic population was under- 

19    represented in juries.

20         Q.    Okay.  And other cases?

21         A.    Yes.  About that same time, I did a similar 

22    analysis in the case of U.S. versus Biaggi.  I've also --

23         Q.    And what did you do in that case?

24         A.    I did the same thing.

25         Q.    Okay.
                                                              66

 1         A.    I also testified in a case called U.S. versus 

 2    Shapiro, which was a criminal case involving the mailing of 

 3    obscene materials.

 4         Q.    What was your function in that case?

 5         A.    I was asked to analyze a survey proffered by the 

 6    defendants concerning whether obscene materials were 

 7    consistent with community standards of obscenity.

 8         Q.    Okay.  And are there any other cases in which you 

 9    testified on behalf of the United States?

10         A.    Not that I recall at this time.

11         Q.    Okay.  On behalf of the EEOC?

12         A.    Oh, yes.  I did testify on behalf of the Equal 

13    Employment Opportunities Commission in an employment 

14    discrimination case against Sears Roebuck.

15         Q.    I'm sorry.  I may have misheard you.  Did you 

16    mention United States against Biaggi.

17               THE COURT:  He did.

18         A.    Yes, I did.

19         Q.    You did.  I'm sorry.

20               Have you ever testified on behalf of the 

21    Republican party?

22         A.    Yes, I have.

23         Q.    Would you tell us about when that was?

24         A.    This was a case known as Harrison versus the 

25    Pennsylvania Legislative Reapportionment Committee.  I was 
                                                              67

 1    retained by the Philadelphia Republican Committee in a case 

 2    involving Section II of the Voting Rights Act.

 3         Q.    In September of 1989, were you appointed by the 

 4    then Secretary of Commerce to be a co-chair for the Special 

 5    Advisory Panel?

 6         A.    Yes, I was.

 7         Q.    Of the Special Advisory Panel?

 8         A.    Yes, I was.

 9         Q.    Is that panel still in existence?

10         A.    Yes, it is.

11         Q.    Is there a document that you are aware of that 

12    defines your duties?

13         A.    Yes, there is.

14         Q.    Will you take a look at PX 450, which is also in 

15    Volume 2.

16         A.    (Complies)

17         Q.    Is that the document that you were referring to?

18         A.    Yes, it is.  

19               MR. ZIMROTH:  Your Honor, this is the stipulation 

20    and order which is already part of the record in this case.  

21    I don't think it's necessary for me to move its admission 

22    now.  But if it is, I do.

23               THE COURT:  We don't need it.  

24               MR. ZIMROTH:  All right.  

25    BY MR. ZIMROTH:
                                                              68

 1         Q.    What is your understanding of what some of your 

 2    duties were?

 3         A.    Well, the first thing that we were supposed to do 

 4    was to give advice on the guidelines for adjustment that 

 5    were to be proffered by the Department of Commerce.  More 

 6    generally, our job was to advise the Secretary as to whether 

 7    or not the 1990 census should be adjusted.  Our duties 

 8    included monitoring the progress of the census, 

 9    understanding how the census worked, understanding the 

10    strength and weaknesses of the original enumeration, 

11    understanding the way in which the Bureau proposed to adjust 

12    the census, understanding the strengths and the weaknesses 

13    of the adjustment method, to proffer advice when asked for, 

14    and, finally, to render a recommendation in July 1991 as to 

15    whether or not the 1990 census should be adjusted.

16         Q.    Could you give the Court just a brief overview of 

17    how you went about fulfilling your obligations.

18         A.    Well, in no particular order, one of the things 

19    that we did was read massive amounts of Census Bureau 

20    documents.  We also held meetings among the panel, sometimes 

21    subgroups of the panel, usually with people from the Census 

22    Bureau there.  Occasionally people from the Department of 

23    Commerce came.  

24               We conducted independent research.  To further 

25    that independent research, we retained consultants and 
                                                              69

 1    commissioned them to do additional research.  I hired -- I 

 2    personally hired one full-time staff person, had some 

 3    part-time people working with me.  It was a major research 

 4    project.

 5         Q.    From September of 1989 until July the 15th, 1991, 

 6    can you estimate how much time you personally spent on this 

 7    project?

 8         A.    Well, I did not keep records.  And the amount of 

 9    time that I worked varied depending on when it was, but I 

10    would estimate on average between 20 and 25 hours per week.  

11               MR. ZIMROTH:  Your Honor, I want to proffer 

12    Professor Ericksen as an expert in this matter at this time.

13               THE COURT:  Mr. Sitcov?  

14               MR. SITCOV:  Could I have exactly what he's being 

15    proffered as an expert in?

16               THE COURT:  I assume statistics, sociology, -- 

17               MR. ZIMROTH:  He is an expert in the nature of 

18    the undercount and his works on the Special Advisory Panel, 

19    and as you saw from our statements.  

20               MR. SITCOV:  Nothing.

21               THE COURT:  Very well.  You may testify as to 

22    your expert opinion.  

23    BY MR. ZIMROTH:

24         Q.    Since September of 1989, Professor Ericksen, when 

25    you were appointed by the Secretary, have you ever been 
                                                              70

 1    called by the plaintiffs as an expert?

 2         A.    No, I have not.

 3         Q.    Up to and including today.

 4         A.    No, I have not.

 5         Q.    And do you have any expectation or understanding 

 6    that you will be?

 7         A.    No, I do not.

 8         Q.    Will you look at paragraph 8.

 9               THE COURT:  Of what?  

10               MR. ZIMROTH:  Excuse me.

11         Q.    Of PX 450.  The stipulation.  Do you have that in 

12    front of you, Professor Ericksen?

13         A.    Yes, I do.  

14               MR. SITCOV:  What exhibit? 

15               MR. ZIMROTH:  PX 450, paragraph 8.

16               THE COURT:  On page 5.

17         Q.    I just wanted to read you one sentence and ask 

18    you about that.  

19               "The members of the panel shall be entitled to 

20    full cooperation and shall receive full cooperation from the 

21    defendants including access to all necessary or appropriate 

22    information, and the opportunity to consult with any 

23    employee of the employee of the Bureau." 

24               Professor Ericksen, did you get that cooperation 

25    throughout the period that you served as co-chair?
                                                              71

 1         A.    No, I didn't.  

 2               MR. SITCOV:  Object, your Honor.

 3               THE COURT:  Overruled.  

 4               Go ahead.  

 5               MR. SITCOV:  I'm sorry?

 6               THE COURT:  You started.  I interrupted you.  I'm 

 7    sorry.  

 8               MR. SITCOV:  This was the subject -- this type of 

 9    thing was the subject of plaintiffs' motion in April of 

10    1990.  It was resolved by the Court's decision of June 7th 

11    or 8th of that year.  I don't know why we are reliving that 

12    now.  It has nothing to do with the correctness of the 

13    adjustment decision.  If he had complaints about his work on 

14    the panel, that was the time to decide them.

15               THE COURT:  Mr. Zimroth.  

16               MR. ZIMROTH:  Your Honor has already ruled on 

17    this issue.  I think that the defendants made a motion in 

18    limine on this subject to exclude this testimony, and I 

19    think your Honor ruled in our favor.  I don't expect to 

20    spend a lot of time on this point, your Honor.

21               THE COURT:  How much?  

22               MR. ZIMROTH:  Five minutes exact.

23               THE COURT:  Overruled.  

24    BY MR. ZIMROTH:

25         Q.    I think -- 
                                                              72

 1               MR. ZIMROTH:  Can you read back the question, 

 2    please.

 3               THE COURT:  Did you get cooperation?

 4         Q.    Throughout the period.

 5         A.    No.

 6         Q.    Tell the Judge about it.

 7         A.    In the beginning parts of the existence of the 

 8    panel, it was -- I was stopped from convening meetings I 

 9    wanted to hold.  It was very difficult to get documents.  It 

10    was very difficult to, in short, get the information that we 

11    needed to conduct the affairs of the panel.

12               On top of that, the very first day that we met 

13    with officials from the Department of Commerce, we were 

14    informed that over a hundred thousand dollars had been cut 

15    from our budget.

16         Q.    Did you file an affidavit to the Court giving in 

17    more detail what you have just summarized?

18         A.    Yes, I did.

19         Q.    Will you look at PX 451 and 452 in the same 

20    binder.

21               (Pause)

22         Q.    Are those the affidavits you filed?

23         A.    They are.

24         Q.    Did they accurately reflect the state of affairs 

25    at the time that the affidavits were filed?
                                                              73

 1         A.    They do.  

 2               MR. ZIMROTH:  Your Honor, I would move the 

 3    admission of those affidavits.  I think they are already a 

 4    part of the record in this case.  But just to formalize it, 

 5    I will do that now.  

 6               MR. SITCOV:  Well, they are in the Court's 

 7    document, your Honor.

 8               THE COURT:  No need to mark them again then.  

 9               MR. ZIMROTH:  Excuse me?

10               THE COURT:  There is no need to move them 

11    formally again.  

12               MR. ZIMROTH:  Thank you, your Honor.

13    BY MR. ZIMROTH:

14         Q.    You are aware that the Judge rendered a decision 

15    on the subject matter of what you are writing in your 

16    affidavit?

17         A.    That's correct.

18         Q.    And after the Judge's decision, did things get 

19    better?

20         A.    Yes, they did.

21         Q.    Tell us about that.

22         A.    We normally would receive documents generally 

23    within five working days of asking for them.  In general, we 

24    were able to have the meetings that we required of the 

25    Census Bureau officials.  When all was said and done, we got 
                                                              74

 1    the information we needed, although it was a struggle.  And 

 2    by "the information we needed," I mean the information we 

 3    needed to proffer our recommendations to the Secretary of 

 4    Commerce.

 5         Q.    And what happened on July 15th with respect to 

 6    the your cooperation with the Commerce Department?

 7         A.    On July 15th, or some time -- I don't know if it 

 8    was exactly July 15th, but some time thereabouts, I was 

 9    informed that we would have not be receiving additional 

10    documents from the Census Bureau, that there would be no 

11    more money spent on any panel activities, although there was 

12    money left in the budget.

13               The access that I had to Census Bureau data that 

14    I had been analyzing by remote computer connection at the 

15    Census Bureau was cut off.  For all practical purposes, we 

16    received no more cooperation from the Department of Commerce 

17    with regard to panel activities.

18         Q.    On the basis of the information that you received 

19    before July the 15th, did you submit a recommendation to the 

20    Secretary of Commerce?

21         A.    Yes, I did.

22         Q.    And did you also submit a report detailing 

23    reasons for your recommendation?

24         A.    Well, together with some of my other colleagues 

25    on the panel I submitted a joint report.
                                                              75

 1         Q.    Okay.  Will you please look at in binder No. 2, 

 2    PX 453, and also -- well, start with that one.

 3               (Pause)

 4         Q.    Is that your recommendation?

 5         A.    Yes, it is.

 6         Q.    And will you look at PX -- perhaps you can just 

 7    leave that one open because I want to ask you some questions 

 8    about it in a minute.  

 9               But before that, could you just take a look at PX 

10    142, which is in the same binder.

11               (Pause) 

12               No, I think it's not in the same binder.

13         A.    Yes, it is.

14               (Pause) 

15         Q.    What is 142?

16         A.    142 is the second letter that I wrote jointly 

17    with Professor John Tukey, another panel member, 

18    supplementing my original recommendation to the Secretary of 

19    Commerce.

20         Q.    Okay.  And will you look at 195, which is not -- 

21               MR. ZIMROTH:  Your Honor, that's not in any of 

22    the binders.  That's a separate document.

23         Q.    Would you tell the Judge what that document is, 

24    please?

25         A.    Yes.  This is the joint report that I submitted 
                                                              76

 1    together with three of my colleagues on the Special Advisory 

 2    Panel.

 3         Q.    Can you tell the Judge just how this report is 

 4    organized.

 5         A.    Yes.  After the acknowledgement, we have what I 

 6    call the report itself, which is a 24-page document 

 7    summarizing what we found.  That is, 24 pages of text and 

 8    several tables.

 9               Then Appendix A which follows that --

10         Q.    Just one minute, Professor Ericksen.  

11               MR. ZIMROTH:  Your Honor, for my colleagues, some 

12    of these that are tagged here were not part of the original 

13    report.  We are doing that just for ease of reference just 

14    for the testimony.

15         Q.    Go ahead.  You just mentioned the 24-page report 

16    section, and then what?

17         A.    And then Appendix A was our discussion on the 

18    census process, the strengths and weaknesses thereof.

19               Appendix B was a report that was submitted by a 

20    consultant to the panel, demographer, formerly chief 

21    demographer, I believe, at the Census Bureau, Professor 

22    Passel, on the demographic analysis for 1990.

23               Then Appendix C was our evaluation of the quality 

24    of the PES or post-enumeration survey.

25               And Appendix D is a statement written by another 
                                                              77

 1    of our consultants, Professor Joseph B. Kadane, on the 

 2    problem of correlation bias.

 3               Appendix E is a report that was written by 

 4    additional consultants David Hoaglin and Mark Glickman of 

 5    Harvard University, on what happens when you vary some of 

 6    the assumptions of the smoothing and presmoothing models 

 7    that were used in the adjustment.  

 8               Appendix F is a paper that was written by Kirk 

 9    Wolter, one of the panel members, and a colleague of his 

10    who's still at the Census Bureau concerning how you can 

11    evaluate the quality of adjusted estimates for local areas. 

12               I'm sorry.  That's appendix G.  I skipped 

13    appendix F.

14               Appendix F has to do -- a theoretical matter of 

15    how good you can expect a second measurement.  If you take 

16    two measurements of the same thing and calculate the 

17    average, it makes a point that the average will improve 

18    either of the estimates in general, but you cannot expect it 

19    to improve every time.

20               That was put in because some people expected the 

21    PES to make every area better, and it's important to let 

22    people know that it couldn't do that.

23               And in Appendix H, we --

24         Q.    I'm sorry.  G.  Could you repeat what you said 

25    about G.
                                                              78

 1         A.    Okay.  Appendix G is a paper by my colleague on 

 2    the panel, Kirk Wolter, and his former colleague on the 

 3    Census Bureau, Beverly Causey, on the question of how you 

 4    can evaluate the quality of adjusted counts for small areas.

 5               And Appendix H is a second statement on the 

 6    question of:  If you know how big the errors are for larger 

 7    areas, what can you say about the likely size of error in 

 8    the smaller areas that constitute the large area.  For 

 9    example, if a large area is the state, what might you say 

10    about the errors for counties?  

11               MR. ZIMROTH:  Your Honor, at this time, I move 

12    the admission of 453, 195 and 142.          

13               THE COURT:  453 

14               MR. ZIMROTH:  That's the June 21st letter of 

15    recommendation, in binder No. 2.

16               (Discussion off the record)

17               THE WITNESS:   142?  

18               MR. ZIMROTH:  142 is the July 11th letter.  Did I 

19    neglect to ask you what the July 11th letter was?

20               THE WITNESS:   No, we found it.

21               THE COURT:  Did you testify about that?

22               MR. ZIMROTH:  Did you testify -- why don't you 

23    repeat it if you have.  What is the July 11th --

24               THE COURT:  About this July letter.

25               THE WITNESS:  Yes, I said that that was the 
                                                              79

 1    second letter I wrote to Secretary of Commerce supplementing 

 2    my original recommendation on adjustment.  And it was 

 3    coauthored by a second panel member, John Tukey.

 4               MR. ZIMROTH:  And 195 is the report that 

 5    Professor Ericksen just described.  

 6               MR. SITCOV:  Well, your Honor, as you can see 

 7    from the bates stamp numbers at the bottoms of all these 

 8    pages, these are documents contained in the administrative 

 9    record.  I don't know that they need to be admitted 

10    separately from that.  

11               MR. ZIMROTH:  Your Honor, we --

12               THE COURT:  You want them?  

13               MR. ZIMROTH:  We want them marked separately, 

14    absolutely.

15               THE COURT:  They will be admitted.

16               (Exhibits PX 453, 195 and 142 for identification 

17    were received in evidence) 

18    BY MR. ZIMROTH:  

19         Q.    What was your recommendation to the secretary?

20         A.    My recommendation was that the 1990 census should 

21    be adjusted.

22         Q.    Now, could you take a look at the first page of 

23    453.  In the second full paragraph, and just could you read 

24    that to yourself for a minute.  I want to ask you a few 

25    questions based on that.
                                                              80

 1               (Pause)

 2         A.    Okay.  

 3               MR. ZIMROTH:  Your Honor?

 4               THE COURT:  Yes.

 5         Q.    Do you notice in the second sentence you have a 

 6    phrase there "error in the census"?

 7         A.    Yes.

 8         Q.    What kinds of errors are you talking with?  What 

 9    kind of error are you talking about there?

10         A.    There I was specifically talking about omissions.  

11    Omissions are people who should have been counted but were 

12    not.  There are many millions of those in the Decennial 

13    census.

14         Q.    Can you give the Judge an idea of what types of 

15    omissions?

16         A.    Well, omissions in general will be divided into 

17    two categories.  Sometimes an entire housing unit can be 

18    left off the census.  This could occur, for example, if an 

19    address doesn't get on the address list which is used to 

20    decide where the form should be mailed.  It can also occur 

21    when a house is really occupied, is enumerated as vacant.  

22    Those are the main two ways in which that happens.  So when 

23    every single person in the household is left off the census, 

24    we generally refer to that as a whole household omission.

25         Q.    Is there another kind of omission?
                                                              81

 1         A.    Yes.  The second kind of omission occurs when the 

 2    census form, perhaps, is mailed back in.  But some people 

 3    are left off.  That often happens when you've got people 

 4    living in a house, for example, who are not strongly related 

 5    to the head of household.

 6               For example, if a family has moved to New York 

 7    from the Dominican Republic and has lived in New York for 

 8    two years, and a cousin comes -- let's say the cousin came 

 9    on March 15th, and intended to stay in that house, let's 

10    say, for three or four weeks until he found a job.  That 

11    family might consider the cousin to be a visitor and not an 

12    usual resident of that household.

13               Now, this cousin could have got a job around the 

14    1st of May and moved out the 1st of July and then 

15    established a different resident.  But that person should 

16    have been counted at the April 1st address.  And that would 

17    be considered to be within a household omission because 

18    everybody else there would have been counted and that person 

19    would have been left off.

20         Q.    Now, are there other categories of errors in the 

21    census other than omissions?

22         A.    Yes.

23         Q.    What are they?  Or what is it?

24         A.    Well, the other major category is called 

25    erroneous enumerations.  And the erroneous enumerations get 
                                                              82

 1    less publicity than omissions, but they are very important.

 2               Erroneous enumerations are counts that have 

 3    occurred that should not have.

 4         Q.    Can you give us some examples of how that might 

 5    happen?

 6         A.    Yes.  One example can occur if, let's say, 

 7    there's a family that has a house in Manhattan, and they 

 8    have a weekend house in Connecticut.  It's very likely that 

 9    both of those addresses -- indeed, the Census Bureau intends 

10    that both of those addresses would be on the address list as 

11    used to mail the form.  

12               A form would go to each house.  And somebody, 

13    perhaps the wife, would dutifully fill out the form in New 

14    York, and the husband, who hadn't been paying attention, 

15    would fill out the one in Connecticut.  Two census forms 

16    would go in for the one family.  That would -- I call that a 

17    duplication; in the PES lingo, that would be called someone 

18    enumerated at the wrong address in Connecticut.  Either way 

19    it is an erroneous enumeration.

20               A second category of erroneous enumeration is a 

21    fictitious count.  Sometimes census workers, particularly in 

22    hard-to-count neighborhoods, despair of ever getting the 

23    interview done, and they just go fill out the form.  The 

24    colloquial term for that practice is called curbstoning, 

25    though I've never actually heard of a Census Bureau 
                                                              83

 1    individual sitting on the curb to do it.  Nonetheless, the 

 2    data are fictitious.

 3         Q.    Those are called in the lingo, I think, 

 4    fabrications, right?

 5         A.    They are called fabrications, yes.

 6               There are other kinds -- there is a third 

 7    category that I should mention.  And there are cases where 

 8    -- this especially happens out in rural areas where it's 

 9    hard to construct the address list.  Sometimes the same 

10    house can be counted twice.  That's known in the lingo as a 

11    duplication.

12         Q.    Are you aware that the general accounting office 

13    estimated the number of errors, both omissions and erroneous 

14    enumerations, in the 1990 Decennial census?

15         A.    Yes, I am.

16         Q.    And do you know what that estimate was?  If you 

17    cannot remember offhand, you might look at PX 16, which is 

18    in binder No. 1.

19               (Pause)

20         Q.    You might look at page 1, Mr. Ericksen.

21               Yes.  The question is:  What did the GAO 

22    estimate?  

23         A.    The GAO estimated that the 1990 census contained 

24    a minimum of 14.1 million gross errors.  And by "gross 

25    errors" they mean the sum of omissions and erroneous 
                                                              84

 1    enumerations and perhaps as many as 25.7 million of these 

 2    errors, depending on how broadly a census error is defined.

 3         Q.    Could you tell the Court what that means, 

 4    "depending upon how broadly a census error is defined."  

 5    What would be included in the 25 plus million that's not 

 6    included in the 14.1 million?

 7         A.    Well, in the lower estimate, it is assumed that 

 8    there are only 4.4 million erroneous enumerations.  And the 

 9    4.4 million erroneous enumerations includes fabrications and 

10    it also includes people who are counted twice in the same 

11    local area.

12               It's my understanding that usually means that the 

13    house itself is counted twice.  Then if you add to the 4.4 

14    million, since the net undercount is 5.3 million, that gives 

15    you an estimate of 9.7 million omissions.  And the sum of 

16    9.7 and 4.4 is 14.1 million.

17               Now, one could make the argument that if the 

18    family that I told you about a few minutes ago was counted 

19    in Connecticut but was not counted in New York, then those 

20    two errors balance out.

21         Q.    And is that the assumption upon which the 14.1 

22    million is made?

23         A.    Yes, it is.

24               Now, what the Census Bureau knows from the post 

25    enumeration survey is whether or not somebody got counted in 
                                                              85

 1    a second home.  And the way in which the post-enumeration 

 2    surveys are put together, they are not able to include in 

 3    the estimate knowledge of whether they are counted in the 

 4    principal residence.

 5               Now, sometimes if the family is counted in 

 6    Connecticut and it wasn't counted in New York, by one 

 7    argument you call that an offsetting error.  However, from 

 8    my point of view, since the census is concerned with the 

 9    issue of distribution of population, I could argue that 

10    would count double.

11               On the other hand, if --

12         Q.    In other words, if there was a one-too-many 

13    person or one-too-many households in Connecticut and 

14    one-too-few in New York?

15         A.    That's right.  

16               MR. SITCOV:  Object.  Leading.

17               THE COURT:  Overruled.

18         A.    Now, if the family was counted in Connecticut and 

19    also was counted in New York, I don't think anybody would 

20    disagree that that was an error.

21               So including those, that takes us up to an 

22    estimate of about 10 million erroneous enumerations and 15 

23    million omissions.

24         Q.    Do you know whether or not the Census Bureau 

25    agreed with the general conclusions of the GAO report?
                                                              86

 1         A.    Well, on page 8, I find the statement: 

 2               "We discussed our findings with responsible 

 3    Bureau of Census officials.  They generally agree with the 

 4    facts and analyses presented and suggested several technical 

 5    clarifications will incorporate any clarifications where 

 6    appropriate."  

 7               MR. ZIMROTH:  Your Honor, I move the admission of 

 8    PX 16.  

 9               MR. SITCOV:  We oppose it because there's been no 

10    attempt to authenticate this document.  

11               MR. ZIMROTH:  This is a self-authenticated 

12    document, your Honor.  It is a government report.  

13               MR. SITCOV:  It certainly is not.  It is not a 

14    sealed document of any kind.  There is nothing self- 

15    authenticating about this document.  Plus it's also nothing 

16    but hearsay on the part of this witness.

17               THE COURT:  Well, as to the hearsay objection, I 

18    don't have any problem because I'm sure it's an official 

19    record.  It will fit under 803.

20               As to the authentication problem on it I'm a 

21    little fuzzy.  Can you enlighten me there, Mr. Zimroth?  

22               MR. ZIMROTH:  Yes, I will try, your Honor.

23               Self-authentication, extrinsic evidence of 

24    authenticity as a -- 

25               THE COURT:  What are reading from?
                                                              87

 1               MR. ZIMROTH:  Rule 902.

 2               THE COURT:  All right.  

 3               "Extrinsic evidence of authenticity as a 

 4    condition precedent to admissibility is not required with 

 5    respect to the following: 

 6               "(5)  Official publications.  Books, pamphlets or 

 7    other publications purporting to be issued by public 

 8    authority."  

 9               MR. SITCOV:  Your Honor, I think the appropriate 

10    paragraph would be paragraph 2 of 902.  It says: 

11               "Domestic public documents not under seal.  A 

12    document purporting to bear the signature in the official 

13    capacity of an officer or employee of any entity, included 

14    in paragraph (1) hereof, having no seal, if a public officer 

15    having a seal and having official duties in the district or 

16    political subdivision of the officer or employee certifies 

17    under seal that the signer has the official capacity and 

18    that the signature is genuine."

19               THE COURT:  All right.  Let me reserve on your 

20    proffer.  I'll rule on it first thing after lunch.

21               Just so we are clear, it's an offer of 

22    Plaintiff's Exhibit 16.  

23               MR. ZIMROTH:  Yes, your Honor.

24               THE COURT:  Being challenged on authentication 

25    grounds.           
                                                              88

 1               MR. ZIMROTH:  Just one word on that, your Honor.  

 2    I think the section that Mr. Sitcov read doesn't apply to a 

 3    publication.  This is a publication.

 4               THE COURT:  Your cite was 902 sub --

 5               MR. ZIMROTH:  5.

 6               THE COURT:  -- (5).  Okay.  

 7               MR. ZIMROTH:  And, your Honor, I mean, if Mr. 

 8    Sitcov is going to insist that we get originals for these 

 9    public documents, we'll be here all year, and we'll get it.  

10    I mean, we'll subpoena the GAO and we'll get the original up 

11    here.

12               THE COURT:  Oh, I'm sure Mr. Sitcov isn't going 

13    to play hardball of that variety.  Are you, Mr. Sitcov?  

14               MR. SITCOV:  No, I'm going to play hardball of 

15    the plaintiff's variety, your Honor.

16               THE COURT:  Let's move on.  

17               MR. ZIMROTH:  Unbeknownst to me, your Honor, I 

18    did not know this; we have the original of this particular 

19    document.  But I will say we don't have originals of a lot 

20    of other publications.

21               THE COURT:  I have now been handed up what 

22    appears to be an original of Plaintiff's Exhibit 16.  

23               MR. SITCOV:  I see that, your Honor, but still it 

24    does not bear the seal, as I understand it.

25               THE COURT:  Let me reserve on it.  Let's move on 
                                                              89

 1    to something else.  

 2               MR. ZIMROTH:  Very well.  

 3    BY MR. ZIMROTH:

 4         Q.    Now, Professor Ericksen, you said that sometimes 

 5    errors, omission errors cancel out against erroneous 

 6    enumeration errors, is that correct?

 7         A.    That is correct.

 8         Q.    Sometimes that happens.

 9         A.    Yes.

10         Q.    Why then are the problems of omissions and 

11    erroneous enumerations -- do you think we still then have a 

12    problem with omissions and erroneous enumerations?

13         A.    Oh, yes.  And I think it's important to recognize 

14    that the example that I gave of the fabrication as an 

15    erroneous enumeration and the person sitting inside would be 

16    classified as an omission by the PES, those errors would 

17    offset.

18               The problem though is when you take all the 

19    omissions together, and take all the erroneous enumerations 

20    together, their distributions do not offset.  So the areas 

21    where you tend to find omissions more numerous will not 

22    necessarily be the areas where you find erroneous 

23    enumerations more numerous.  And vice versa.  In areas where 

24    omissions are less numerous may be areas where erroneous 

25    enumerations tend to be more numerous.
                                                              90

 1         Q.    Looking at this issue from the perspective of the 

 2    nation as a whole, do the omissions cancel out the erroneous 

 3    enumerations?

 4         A.    No, they do not.

 5         Q.    Is that what we call the net undercount?

 6         A.    Yes.  The net undercount, in essence, is the 

 7    difference between the number of omissions and the number of 

 8    erroneous enumerations.

 9         Q.    And we are talking now at the national level?

10         A.    At the national level, that's right.

11         Q.    Now, I want to focus on the national level for a 

12    moment and ask you whether or not it is possible to have a 

13    net undercount, a very low net undercount, or indeed a net 

14    undercount of zero, and still have a problem with the count?

15         A.    Yes.

16         Q.    And how would that be?

17         A.    Well, just to give a homely example, if you've 

18    got 10,000 people left out of the census in the city, but 

19    you've got 10,000 duplications and people counted in second 

20    homes, and all that sort of thing, in a resort area, then 

21    the net would be zero.  But there would be too few people in 

22    the city and too many people in the resort area.

23               It's quite possible that there would be a census 

24    taken if current methods were used some time in the future 

25    where the number of erroneous enumerations could number in 
                                                              91

 1    the millions, and could be equal to the number of omissions, 

 2    would also number in the millions.  And that net would be 

 3    zero.  But omissions would be concentrated in certain times 

 4    of areas.  Erroneous enumerations would be concentrated in 

 5    other types of areas, and the distribution of population 

 6    across the states, counties and cities, would be an error.

 7         Q.    Professor Ericksen, in looking back again on your 

 8    June 21st letter, which is PX 453.

 9         A.    I have it.

10         Q.    Okay.  I was asking you about the first half of 

11    that sentence which you used the phrase, "error in the 

12    census."  Now I'm going to ask you about the Second Avenue 

13    which says:  "correct the differential undercut."  

14               Can you explain to the Court what you mean by 

15    differential undercount?

16         A.    Yes.  I explained what the net undercount is as 

17    the difference between omissions and erroneous enumerations.  

18    Now, we can consider two groups.  One group could be blacks 

19    and other minority populations, and the other group could be 

20    whites.  And for blacks and other minorities, we would have 

21    an estimate of omissions, an estimate of erroneous 

22    enumerations, and we would calculate a net undercount for 

23    the minority population.

24               We could also calculate the net undercount for 

25    the white population.  And by the differential undercount, 
                                                              92

 1    we need the difference between the net undercount of the one 

 2    population and the net undercount for the other population.

 3         Q.    Is there such a differential undercount?

 4         A.    You mean between the minority population and the 

 5    white population?  Yes, there is.

 6         Q.    Is there also a geographic component to the 

 7    differential undercount that you were talking about?

 8         A.    Yes, there is.

 9         Q.    And why do you say that?

10         A.    Well, because there's a differential undercount 

11    between the minority and white populations.  It is also a 

12    fact of life in the United States that minority populations 

13    tend to be geographically concentrated.  In any city, such 

14    as New York, it's easy to identify where minority 

15    populations live; it's easy to identify where white 

16    populations live; and it's especially easy to identify the 

17    fact that white populations live in suburbs.

18               So to the extent that there's a higher undercount 

19    for minorities and a lower undercount for white people, 

20    areas where minority is concentrated will have a high 

21    undercount, areas where white people are concentrated will 

22    have a low net undercount, and there will be a differential 

23    undercount between areas.

24         Q.    Based on Census Bureau data, do you calculate the 

25    percentage of minorities in various states of the union?
                                                              93

 1         A.    Yes, I did.

 2         Q.    Can you take a look at Plaintiff's Exhibit 706.  

 3    Excuse me.  

 4               And also did you calculate the difference between 

 5    cities and the rest of the particular states in a few 

 6    examples?

 7         A.    Yes, I did.

 8         Q.    Would you look at 706?

 9         A.    Yes.

10         Q.    Does that represent your calculations?

11         A.    Yes, they do.

12         Q.    Just before you describe what it is, what were 

13    the sources from which you made this calculation?

14         A.    One source was a table giving the populations for 

15    eight groups: blacks, whites, Asians and others who are 

16    non-Hispanic and the same for groups among those who were 

17    Hispanic.

18         Q.    I believe that's PX 489.  Take a look at it.           

19    It's in volume No. 3.

20         A.    Right.  That's the Census Bureau tabulation.

21         Q.    Okay.  And the other second source?

22         A.    The statistical abstract for 1991, which is also 

23    a Census Bureau publication.

24         Q.    Okay.  And, just briefly, could you just 

25    summarize what you found in these areas of calculations?
                                                              94

 1         A.    Yes.  Now, I defined a minority as anyone other 

 2    than non-Hispanic whites.

 3               Now, in discussing this, it is important to 

 4    recognize that question 4 of the Decennial census asks 

 5    people:  What is your race?  And you can be classified as 

 6    white, black, Asian or Pacific Islander, native America-- 

 7    would be four important examples.

 8               Question 7 asks people whether or not you are of 

 9    Hispanic origin.  So a person of Hispanic origin could be of 

10    any race.  So you have white Hispanics and white 

11    non-Hispanics.  And I've considered white Hispanics to be 

12    part of the minority population here.

13               So in the United States, according to the Census 

14    Bureau tabulation -- which, by the way, is unadjusted for 

15    the undercount -- 24.4 percent of America's population was 

16    minority.  It was 72.6 percent in the District of Columbia.  

17    It was 49.6 percent in New Mexico, 42.8 percent in 

18    California, and 39.4 percent in Texas, to give three large 

19    examples of states.

20               On the low end, it was 12.3 percent in my home 

21    State of Pennsylvania, 8.7 percent in Wisconsin, and 1.9 

22    percent in Vermont.

23               I also selected three states which have very 

24    large cities in them:  New York, Illinois and Michigan.

25               Now, for cities, I was not able to find the data 
                                                              95

 1    that I needed for all minorities, so I simply calculated the 

 2    percent black.  In New York City it was now 28.7 percent.  

 3    In the remainder of New York State, excluding New York City, 

 4    it was 7.1 percent.  In Chicago, it was 39.1 percent; in the 

 5    remainder of Illinois, it was 7 percent; in Detroit, it was 

 6    75.7 percent; in the remainder of Michigan, it was 6.2 

 7    percent.

 8               The corresponding portions for Hispanic were 24.4 

 9    in New York City, 4.0 in the remainder of the state; 19.6 

10    percent in Chicago, 4.1 percent in the remainder of 

11    Illinois; 2.8 percent in Detroit, and 2.1 percent in the 

12    remainder of Michigan.

13         Q.    I think you --

14               THE COURT:  Are you going to change exhibits?

15         Q.    No, I just want to ask -- I neglected to ask the 

16    witness to identify the statistical abstract that he got one 

17    source of these figures from.

18               I think that's what number?  26?  Your Honor, in 

19    your binder -- 

20               MR. ZIMROTH:  I'm going to move the admission --

21         Q.    Well, first, is that the statistical abstract 

22    that you were referring to?

23         A.    Yes, it is.  

24               MR. ZIMROTH:  Your Honor, at this time I want to 

25    move the admission of PX 706, which is the summary about 
                                                              96

 1    which Professor Ericksen testified, as well as the source 

 2    documents: PX 26, the statistical abstract, and PX 489, the 

 3    Census Bureau documents.  

 4               MR. SITCOV:  Your Honor, we object to 706.  It's 

 5    not on the list of plaintiff's trial exhibits, which they 

 6    submitted to us, 701.  And I'm informed we just got this 

 7    some time a couple of days ago.  I've never seen it before 

 8    just now.  

 9               We'll make a similar objection with respect to 

10    the statistical abstract as we made with respect to the GAO 

11    report.

12               And I'm not sure -- what was the other one?  

13               MR. ZIMROTH:  489.

14               THE COURT:  This strikes me as a good point to 

15    break for lunch.  

16               MR. ZIMROTH:  Fine, your Honor.  May I just make 

17    one response to Mr. Sitcov?  

18               I'm sorry, did you object to 489 as well?  That 

19    was in what you called the administrative record.  

20               MR. SITCOV:  No, I do not object to 489.  

21               MR. ZIMROTH:  Just to clarify this point, your 

22    Honor, I mean, as far as the Exhibit 706 is concerned, that 

23    is nothing more than a compilation of exhibits which were in 

24    fact given to Mr. Sitcov earlier in our exhibit list.

25               Second of all, Professor Ericksen has testified 
                                                              97

 1    about it.  Third of all, on the authentication issue, I just 

 2    want to cite your Honor to another rule, which also supports 

 3    the authentication, which is rule 901(b)(1).  And this 

 4    witness, I hope you will agree, has knowledge of what he 

 5    speaks of, and I think he's competent to testify that the 

 6    GAO report in fact is the GAO report, and that the 

 7    statistical abstract is in fact the statistical abstract. 

 8               So that's just an additional grounds for 

 9    authentication, in addition to the fact that both of these 

10    documents are self-authenticating.

11               And with that, that's it.

12               THE COURT:  All right.  We'll break now for 

13    lunch.  Thank you.

14               (Luncheon recess) 

15    

16    

17    

18    

19    

20    

21    

22    

23    

24    

25    
                                                              98

 1    

 2    

 3    

 4    

 5    

 6               AFTERNOON SESSION

 7               2:05 p.m.

 8               THE COURT:  Good afternoon.  Please be seated.  

 9               I believe when we adjourned we were concerned 

10    over the admissibility of three documents.  I am prepared to 

11    rule.  Does anybody wish to be heard before I rule?

12               MR. SITCOV:  Yes, I do, your Honor.  

13               We will withdraw our objection as to 26 and 16.  

14    But we will continue to maintain our objection to 706.  And 

15    while we think that 26 and 16 are technically objectionable, 

16    we're willing to waive our objection to that because those 

17    two documents are not particularly difficult for us.  

18               But the problem is that the plaintiffs have 

19    identified a great many exhibits that there will be many 

20    authentication problems about, those two were just the 

21    first.  For example, there is Plaintiffs' Exhibit 525, 526, 

22    528.  Those are all in binders that have been given to you 

23    already.  

24               Each of those is stamped on them on every page 

25    draft.  Now, we're certainly not going to agree to the 
                                                              99

 1    admission of documents of that type.  

 2               Then there are literally dozens of documents like 

 3    89, 90 and 91 that are not in the volumes that you already 

 4    have.  I'm looking at 89 now.  It is a document entitled 

 5    draft.  

 6               There are a bunch of unsigned documents such as 

 7    90, 91.  There's no way for us to be able to authenticate 

 8    them.  We got 701 separate documents in no particular order 

 9    last Saturday.  

10               There certainly is a way under the rules to have 

11    these things authenticated and the plaintiffs chose not to 

12    do that sort of thing and there's no way that we're going to 

13    agree that documents we don't know anything about are 

14    authentic.

15               THE COURT:  Let me first, for the record, rule on 

16    the three documents that are outstanding.  And then I'll 

17    address your concerns.  

18               Plaintiffs' Exhibits 26 and 16 are admitted in 

19    the absence of objection.  

20               (Plaintiffs' Exhibits 16 and 26 for 

21    identification were received in evidence)

22               THE COURT:  Plaintiffs' Exhibit 706 is excluded 

23    on grounds that it does not appear on the document list and 

24    it would be unfair at this juncture to offer it, coming on 

25    the blind side as it does against the government.
                                                              100

 1               MR. ZIMROTH:  Excuse me, your Honor, just for 

 2    clarification.  That document was in fact provided to the 

 3    government prior to today.  It was not on the exhibit list, 

 4    that is true, that we gave them last week.  But it was sent 

 5    to them prior to today and I don't think Mr. Sitcov said to 

 6    the contrary.  If he did, he misspoke.

 7               THE COURT:  Mr. Sitcov, does that alter your 

 8    perceptions?

 9               MR. SITCOV:  No.  If I said I just got it today, 

10    I didn't mean to say that.  We just got it in the last 

11    couple of days and it doesn't alter anything about our 

12    position with respect to that document.

13               THE COURT:  It doesn't appear on the plaintiffs' 

14    exhibit list.

15               MR. SITCOV:  It does not.  Their numbers end at 

16    701.

17               MR. ZIMROTH:  We got, as I'm sure you know, 

18    towards the very end of this discovery process, a ton of 

19    documents that we were analyzing up until the last minute.  

20    With respect to one or two documents that were not on the 

21    exhibit list, we sent the government notification of that 

22    subsequent to last week.  

23               But with respect to this particular document 706, 

24    I think that the situation is entirely different.  That 

25    document is nothing more nor less than an aid for the 
                                                              101

 1    witness to help explain what is already in another document 

 2    that is in testimony.  

 3               And so with respect to a document like that, I 

 4    frankly did not understand and don't understand that that 

 5    has to be given to the government in advance.  But we 

 6    nonetheless gave it to the government in advance.  They had 

 7    it, I don't know when, two days ago.

 8               THE COURT:  Well, I have a suggestion that may 

 9    moot the whole thing.  If that's the purpose for which you 

10    are using it, why not just hand it to the witness and let 

11    him refresh his recollection?

12               MR. ZIMROTH:  What we can do with respect to that 

13    particular document, we can just read the number into the 

14    record.

15               THE COURT:  It is a highly technical objection, I 

16    agree.

17               MR. ZIMROTH:  And with respect, for example, to 

18    another chart that you will see later on today, the witness 

19    -- there are tables that are already, we've shown to Mr. 

20    Sitcov.  And what these charts do, in graphic terms, is show 

21    what the tables already say.  

22               Now, it seems to us that that kind of document is 

23    not required to be turned over to them in advance but we did 

24    it anyway.  We may not have had all the charts completed by 

25    last Monday, chartmakers being what they, but we gave it to 
                                                              102

 1    them as soon as it was ready.

 2               THE COURT:  We'll admit them but they're not 

 3    going to be evidence in chief.  They'll just be ancillary to 

 4    oral testimony as a medical chart would be by a doctor.

 5               MR. ZIMROTH:  There may be one or two documents 

 6    in different categories, if so we'll deal with them when 

 7    they come up.

 8               THE COURT:  I will change my ruling on 706 and 

 9    admit it upon the expressed understanding that the document 

10    is not evidence in chief, it is not evidence of anything 

11    contained therein but it is used merely as an adjunct of 

12    oral testimony to demonstrate or to make more articulate 

13    this information for the witness.  

14               That still leaves the basic problem about the 

15    drafts.  

16               (Plaintiffs' Exhibit 706 for identification was 

17    received in evidence)

18               MR. ZIMROTH:  Those have not been offered.  I 

19    would say that those drafts will be drafts that were 

20    supplied to us by the government so in terms of 

21    authenticity, I don't see how he can say they're not 

22    authentic.

23               THE COURT:  I see a hearsay problem.

24               MR. ZIMROTH:  The ones he's talking about are 

25    documents commissioned by the Census Bureau.  These are 
                                                              103

 1    studies done that experts like Dr. Ericksen relied on in the 

 2    course of their educating themselves about the problem and 

 3    they were studies particularly that the Census Bureau 

 4    commissioned specifically for the purpose of casting some 

 5    light on the accuracy of the postenumeration survey.  It may 

 6    very well be a question of how much weight your Honor wants 

 7    to give to them, that's another story.  But these are the 

 8    most recent documents that the Census Bureau has on the 

 9    subject and if they're not then we want to know why not 

10    because we've asked for the most recent documents.  

11               (Pause)

12               THE COURT:  That said, I think we can resume with 

13    Dr. Ericksen if he would be so kind as to resume the stand.  

14    EUGENE ERICKSEN, resumed.

15               MR. ZIMROTH:  I don't believe there's an 

16    objection to 489?

17               MR. SITCOV:  No.

18               THE COURT:  Admitted.  

19               (Plaintiffs' Exhibit 489 for identification was 

20    received in evidence) 

21    DIRECT EXAMINATION (Cont'd) 

22    BY MR. ZIMROTH:

23         Q.    Prove Ericksen, could you look at Plaintiffs' 

24    Exhibit 706.

25         A.    I have it.
                                                              104

 1         Q.    Would you once again tell us what calculations 

 2    you made concerning the geographical distribution of 

 3    minorities in the United States?

 4         A.    Yes.  First of all, a minority is someone who 

 5    either was Hispanic or was a black nonHispanic, an Asian 

 6    nonHispanic or in a category referring to other races 

 7    nonHispanic.  

 8               For the nation, 24.4 percent of the population is 

 9    defined as minority by that definition.  

10               Now, I gave some examples here.  In the District 

11    of Columbia, 72.6 of the population was minority by that 

12    definition.  The corresponding proportions were 49.6 for New 

13    Mexico, 42.8 for California, 39.4 for Texas, 12.3 for 

14    Pennsylvania, 8.7 for Wisconsin and 1.9 for Vermont.  

15               Now, I did not have the same data available to me 

16    for cities.  So for cities I relied on tables in the 

17    statistical abstract and calculated the proportion black.  

18    And of course the proportion black includes those who are 

19    Hispanic, there's a small proportion of blacks who are also 

20    Hispanic, as well as blacks nonHispanics.  And I subtracted 

21    the relevant numbers from the state totals to estimate the 

22    same proportion for that part of New York State outside of 

23    New York City.  

24               In New York City the population was 28.7 percent 

25    black and 24.4 percent Hispanic.
                                                              105

 1         Q.    Excuse me, before you go on.  In New York City, 

 2    the black population you say was 28.7?

 3         A.    Yes.

 4         Q.    What is the corresponding figure for the rest of 

 5    New York State?

 6         A.    7.1 percent.

 7         Q.    Okay.  And for Hispanics?

 8         A.    For Hispanics, New York City's population was 

 9    24.4 percent.

10         Q.    And the corresponding figure for the remainder of 

11    New York State?

12         A.    4.0 percent.  In Chicago the population was 39.1 

13    percent black.  For the remainder of Illinois, the 

14    corresponding figure was 7.0 percent black.  

15               In Chicago, the population was 19.6 percent 

16    Hispanic.  For the remainder of Illinois, the corresponding 

17    proportion was 4.1 percent Hispanic.  

18               In Detroit the population was 75.7 percent black.  

19    For the remainder of Michigan the proportion was 6.2 percent 

20    black.  

21               In Detroit the population was 2.8 percent 

22    Hispanic.  For the remainder of Michigan the corresponding 

23    proportion was 2.1 percent.

24         Q.    Professor Ericksen, in your opinion does the 

25    census do less well for certain groups in our country than 
                                                              106

 1    it does for others?

 2         A.    In my opinion the census counts minorities and 

 3    other poor groups less well than it counts white people, 

 4    especially those of average or above average economic means.

 5         Q.    In your opinion, what is the reason for that?

 6         A.    The reason for that is that the Census Bureau of 

 7    necessity tries to apply a uniform counting procedure 

 8    throughout the country.  The problem is that a uniform 

 9    procedure works less well in some areas than it does in 

10    others.  It works less well for those people who have lower 

11    levels of education, who have family structures that are 

12    other than the usual mother father and young children.  It 

13    works less well for people who do not live in detached 

14    houses with yards, who do not receive mail on a reliable 

15    basis.  It works less well for people who speak a foreign 

16    language, to give you some examples.

17         Q.    I want you to look at this poster that we have up 

18    which is a reproduction blow-up of Plaintiffs' Exhibit 454.  

19    Could you just tell us what that represents.

20         A.    Those are the major steps in the census process 

21    as it is carried out over most of the country.

22         Q.    In some parts of the country are there some 

23    differences?

24         A.    Well, there are two major, two exceptions that I 

25    should point out.  In the most rural parts of the America 
                                                              107

 1    where the population is very sparse, this is examples of 

 2    where this happens would be out in the middle of Wyoming.  

 3    It's very expensive, it would be very expensive for the 

 4    Census Bureau to compile an address list in advance.  So in 

 5    those areas which are known as list enumerate areas, where 

 6    about 6 percent of the population live, instead of 

 7    constructing an address list in advance, they send 

 8    enumerators out to the area and they list the addresses and 

 9    drop off the census forms at the same time.

10         Q.    In effect the census enumerator acts as a 

11    mailman?

12         A.    That's correct.  That is about 6 percent of the 

13    country and ten percent of the country use a procedure known 

14    as update leave.  They have an address list compiled in 

15    advance but instead of mailing out the forms, they send an 

16    enumerator out to the addresses and they drop them off.  

17    That's about ten percent of the population.  

18               But for everywhere else, that is basicly the way 

19    it works.

20         Q.    Now, I'd like you to first, I'd like to talk 

21    first about the first step in the process which you 

22    described as the construction of the address list.  Could 

23    you describe for the judge how that happens?

24         A.    Sure.  In the urban parts of America, the first 

25    step is to compile commercial address lists.  These are the 
                                                              108

 1    same lists that are compiled by companies who sell addresses 

 2    to advertisers and other kinds of companies that want to do 

 3    mass mailings.  

 4               In areas of generally less densely populated 

 5    suburban areas and some rural areas where these lists were 

 6    not available, the Bureau sends enumerators out to the field 

 7    and the enumerators make the first list.  That's called feed 

 8    list areas.  After that the list goes through various stages 

 9    of updating and checking.  In the second stage generally the 

10    post office will check the list that's obtained from the 

11    commercial companies and add addresses that they know about.  

12               After that, the census sends enumerators out with 

13    the addresses that have been obtained so far and they go out 

14    and look in person to see what other addresses they can find 

15    and they add those to the list, at the same time, trying to 

16    delete duplicated addresses.

17         Q.    And at the end of the process, there is an 

18    address list.

19         A.    Yes.

20         Q.    And in your opinion, are certain groups in our 

21    society more likely to be left off that address list than 

22    others?

23         A.    Yes, they are.

24         Q.    And can you tell us who they are and why that 

25    would be?
                                                              109

 1         A.    Well, the kinds of people who are left off are 

 2    the kinds of people who can afford housing that's not 

 3    particularly good.

 4         Q.    Who cannot what?

 5         A.    Who only can afford housing that's not very good.  

 6    I can give you some important examples of the kinds of 

 7    housing which is left off.  

 8               In Philadelphia, there is a term called the 

 9    Philadelphia row house.  And a Philadelphia row house, I'm 

10    sure you have the same thing in Brooklyn, it's a house 

11    that's got three or four stories that was initially intended 

12    for single family occupancy.  Back in the days when the 

13    neighborhood was occupied by your typical middle class 

14    family, that's who lived there.  

15               Since then the population that lives in that area 

16    is less well off.  The house could be subdivided into 

17    apartments for students.  The house could be subdivided into 

18    apartments for poor people.  Where once there was one house, 

19    there are now three.

20         Q.    One address?

21         A.    One address.

22         Q.    One housing unit?

23         A.    Where once there was one housing unit there is 

24    now three.  The difficulty is that from the outside you 

25    can't tell that anything has changed.  There is still one 
                                                              110

 1    street address, one mail box, one doorbell.  Neither the 

 2    Postal Service nor the commercial address company nor a 

 3    census enumerator walking down the street will have any idea 

 4    that that has occurred.  

 5               A second example of what I'm talking about was 

 6    found by ethnographers who were commissioned by the Census 

 7    Bureau to do an intensive study in Los Angeles county, 

 8    California.  There people living in garages, people living 

 9    in little huts in backyards were found.  This is not the 

10    regular kind of housing that people would expect to find.  

11    And those would not go on an address list either.  

12               A third example occurs in areas where poor people 

13    crowd into apartments and informally subdivide them.  So 

14    basically it's still one apartment, but you have different 

15    people occupying different parts of it and they consider 

16    themselves to be different housing units.

17         Q.    Is there anything about the commercial lists that 

18    you think makes it more likely or less likely that certain 

19    groups would be included?

20         A.    Well, first of all I just explained how a certain 

21    type of housing would not show up on the list.

22         Q.    Even on a commercial list?

23         A.    That's right.  Secondly, the lists are going to 

24    be a better quality in those areas where they are used most 

25    frequently.
                                                              111

 1         Q.    Why is that?

 2         A.    Because that's where they do their business.  In 

 3    areas that can afford to buy the products that the 

 4    advertisers are interested in selling, that's where the list 

 5    will be used.  That's where it will be of better quality.

 6         Q.    What about the post office, they're part of the 

 7    update?

 8         A.    Yes, they are.

 9         Q.    And aren't they able to fix up all these 

10    problems?

11         A.    Well, they will add addresses that they know 

12    about, but in many of the kinds of subdivided housing that 

13    I'm telling you about, the post office is not aware of extra 

14    addresses living there.  They just deliver the mail, they 

15    observe that they might have different names but they're not 

16    aware that the people living together have different names 

17    or are living separately.

18         Q.    Are you aware of studies either by the Census 

19    Bureau or by the GAO concerning coverage of housing in prior 

20    censuses?

21         A.    Yes, I am.

22         Q.    Would you tell us about those.

23         A.    Well, in the 1970 census the Census Bureau did a 

24    study trying to identify the types of housing that were most 

25    likely to be left off the address list.  What they learned 
                                                              112

 1    was that housing that was located in what the term-of-art is 

 2    small unit multi-unit structures.  What that is is a 

 3    structure that has two, three or four units in it, like a 

 4    subdivided row house.  And what they found was that the 

 5    omission of housing units was especially great there.  

 6               This was particularly true in buildings that were 

 7    old.  So in the older parts of cities where you've got 

 8    subdivided row houses, is where in 1970 the Census Bureau 

 9    found addresses left off.

10         Q.    What about in 1980?

11         A.    In 1980 the GAO studied the quality of the Census 

12    Bureau address list that was obtained from the commercial 

13    firms and found that they were particularly deficient in 

14    urban areas.

15         Q.    What about the Census Bureau study in 1980?

16         A.    The Census Bureau study in 1980 found that the 

17    whole household omissions that I talked about this morning 

18    were particularly great among minority populations who as we 

19    know are concentrated in those kinds of neighborhoods in 

20    cities that we've been discussing.

21         Q.    Professor Ericksen, could you please look at 

22    Plaintiffs' Exhibit 21, 425 and 705 and I'll ask you whether 

23    those are the three studies to which you've been referring.  

24               (Pause)

25         A.    Exhibit 21 is a document on which I relied for 
                                                              113

 1    the comments I made about the 1970 census.

 2         Q.    What is that document?

 3         A.    This is a GAO document entitled "Problems in 

 4    Developing the 1980 Census Mail List."

 5         Q.    That's 425.  I just wanted you to identify the 

 6    previous one.

 7         A.    You mean number 21?

 8         Q.    Yes.

 9         A.    "Coverage of housing in the 1970 census PHCE 5 of 

10    the 1970 census, the population and housing."  It's a Census 

11    Bureau publication.

12         Q.    Now, you were in the midst of identifying 425.

13         A.    It says "By the Comptroller General.  Report to 

14    the Subcommittee on Census and Populations, Committee on 

15    Post Office and Civil Service, House of Representatives of 

16    the United States.  Problems in developing the 1980 census 

17    mail list, March 31, 1980."  It's from the General 

18    Accounting Office.

19         Q.    You've seen that document before?

20         A.    Yes.

21         Q.    And 705.

22         A.    705 is dated March 10, 1988, "Preliminary 

23    results, memorandum number 117, evaluation of census 

24    coverage for the 1980 postenumeration program.  Census 

25    omissions as measured by the P sample, prepared by Robert E.  
                                                              114

 1    Fay."

 2               MR. ZIMROTH:  I move the admission of those three 

 3    documents, 21, 425 and 705.

 4               MR. SITCOV:  We object to 705.  Another document 

 5    not on the plaintiff's witness list.  I don't know when we 

 6    got it but I assume it may be as recently as two days ago.

 7               THE COURT:  705?

 8               MR. SITCOV:  I would also note that it identifies 

 9    itself as some preliminary document.

10               MR. ZIMROTH:  Your Honor, if you look at these 

11    evaluation reports, they're all identified as preliminary.  

12    This is a document done in 1988 about the 1980 census.  They 

13    never changed the name from preliminary.  They're called 

14    PRENS.  Preliminary research and evaluation something or 

15    another.  

16               But be that as it may, your Honor, this is a 

17    document which we, this is one document that I told you 

18    about before which we did discover that we wanted to use 

19    late in the game.  We sent it to them last week, I believe 

20    on the 8th.  And I just ask your Honor's indulgence on that 

21    score.  There are not very many such documents.  I don't 

22    know of really more than a handful.  The amount of discovery 

23    we got was voluminous.  We had to look through lots of 

24    documents, a lot of them pouring in at the last minute and 

25    we had to make choices.  
                                                              115

 1               There were also a lot of different teams of 

 2    lawyers and I think we did well in terms of coordination but 

 3    this is one that slipped through the cracks.

 4               MR. SITCOV:  Your Honor, we received from the 

 5    plaintiffs four large boxes of documents that were not in 

 6    any particular order.  We had no way to tell which document 

 7    would go with which witness and as is clear so far from this 

 8    witness, it would be impossible to tell from looking at any 

 9    document that it would belong to a particular witness.  So 

10    on top of any other problem we might have had with these 

11    documents, it would have been impossible for us to determine 

12    what the documents would be used for.  

13               The plaintiffs have had access to these documents 

14    for months.  As Mr. Zimroth just said they had hundreds of 

15    thousands of them.  We just got this document on Saturday.  

16    We came here yesterday.  

17               It's simply, it simply would not be possible for 

18    us to have any kind of reasoned consideration of this 

19    document in that circumstance.

20               THE COURT:  Do I sense that your basic objection 

21    is that you feel you can't adequately cross-examine because 

22    you haven't had the document long enough to study it?

23               MR. SITCOV:  Yes.  And we haven't known, we 

24    didn't know until Saturday which documents of these 700 

25    applied to which witness.
                                                              116

 1               THE COURT:  I will overrule the objection but 

 2    given the government's quandry, I will indicate in good 

 3    faith that if you need further time to prepare to 

 4    cross-examine the witness on these last minute documents I 

 5    will give you all the time you need.

 6               MR. SITCOV:  Thank you.  

 7               (Plaintiffs' Exhibits 21, 425 and 705 for 

 8    identification were received in evidence) 

 9         Q.    I believe you mentioned earlier that there are 

10    also ethnographic reports that make similar points about 

11    housing coverage, is that correct?

12         A.    Yes, there are.

13         Q.    Could you tell the judge what ethnographic 

14    studies are?

15         A.    Sure.  Ethnographers are social scientists, some 

16    are sociologists, some are anthropologists, who go into 

17    details and make detailed studies.  The method of the 

18    ethnographer is to become very familiar with the people who 

19    live in a particular, usually a small neighborhood.  Often 

20    times these ethnographers will actually go and live there 

21    and in this case, I'm sorry, let me back up a minute.  What 

22    the ethnographers try to study are the living conditions of 

23    the people and they try to understand why the people act in 

24    the way that they do.  

25               And the ethnographic information is generally 
                                                              117

 1    preferred to survey information in the sense that it 

 2    provides very good depth to understanding what's going on.  

 3    And so if you want to learn about a small area in depth, 

 4    then you would commission an ethnographic study.

 5         Q.    Has the Census Bureau commissioned ethnographic 

 6    studies?

 7         A.    Yes.  The Census Bureau, following as I 

 8    understand it a recommendation of a panel of the National 

 9    Academy of Sciences, began to commission ethnographic 

10    studies as part of the 1986 test census in east Los Angeles 

11    county.  That was a heavily Hispanic suburban area of Los 

12    Angeles.  

13               They also conducted ethnographic studies in the 

14    1988 test censuses in St. Louis, East Central Missouri and 

15    Washington State and they commissioned a longer list of 

16    ethnographic studies in various sites around the country as 

17    part of the 1990 census.

18         Q.    Would you please look at Plaintiffs' Exhibit 536.  

19    Do you see that?

20         A.    Yes, I do.

21         Q.    What is that document?

22         A.    This is a document entitled "Results of 1988 

23    ethnographic studies of census coverage and plans for 1990."  

24    The authors are Elizabeth Martin, Leslie Braunwig and Robert 

25    Fay.  It's dated October 1990 and it's my understanding that 
                                                              118

 1    Elizabeth Martin was a Census Bureau person who was in 

 2    charge of the ethnographic studies that they commissioned 

 3    the ethnographers to do.

 4         Q.    Does this document set forth the plans for the 

 5    ethnographies for 1990?

 6         A.    Yes, it does.

 7         Q.    Were the ethnographic studies supposed to play 

 8    any role in the evaluation of the PES?

 9         A.    Yes.  For example, the ethnographic studies were 

10    intended to help the Census Bureau understand why omissions 

11    and erroneous enumerations occurred.  More gently they were 

12    intended to evaluate the results of the postenumeration 

13    survey.

14         Q.    Look at the exhibit page 30 and perhaps you can 

15    just read the conclusion.

16         A.    Finally and more important for the 1990 census, 

17    the results of the ethnographic study provide an independent 

18    source of corroborating evidence which, while not based on 

19    the statistically defensible sample, can be used to comment 

20    on and shed light on the results of the PES.  

21               And the PES is the postenumeration survey.

22               MR. ZIMROTH:  At this time I move the admission 

23    of 536.

24               MR. SITCOV:  We object to the admission.  This 

25    stuff is all hearsay.  He certainly can rely on it as a 
                                                              119

 1    foundation for his opinion.  But it is hearsay nonetheless.

 2               MR. ZIMROTH:  Your Honor, these are the Census 

 3    Bureau plans for the 1990 ethnographies that were supposed 

 4    to shed light on the PES.  It is a Census Bureau document 

 5    which we got from the Census Bureau itself.  It's clearly 

 6    not hearsay.

 7               MR. SITCOV:  It clearly is hearsay, your Honor.  

 8    Dr. Ericksen is not the author of this nor is Mr. Zimroth.

 9               MR. ZIMROTH:  But the Census Bureau is.

10               THE COURT:  Why is it not a public document?

11               MR. ZIMROTH:  And it is also a public document 

12    and there are several other documents as well.

13               THE COURT:  Dying declaration?

14               MR. SITCOV:  This is an internal memorandum.  Not 

15    every single thing that is done in a public agency is a 

16    public document.

17               THE COURT:  I will admit it.  

18               (Plaintiffs' Exhibit 536 for identification was 

19    received in evidence)

20         Q.    I'm I'll ask you to look at 455, 456, 501, 502, 

21    504, 507, 512, 521, 522, 525, 526, 528 and 534.  I hope 

22    they're all in the same binder.  

23               What I'd like you to do is one at a time identify 

24    first by the number and then what it is.

25         A.    The Exhibit 455 is a report to the Census Bureau 
                                                              120

 1    by James Diego Vigil.  It's entitled "El Jardin, an 

 2    ethnographic enumeration of a barrio in greater East Los 

 3    Angeles" and it's dated April 1987.

 4         Q.    For you -- you've read these, haven't you?

 5         A.    Yes, I have.

 6         Q.    Just tell the judge what they are as you go 

 7    through it.

 8         A.    Well, this is an ethnographic study in the area 

 9    of East Los Angeles that I told you about before, trying to 

10    understand why -- I believe that there was a substantial 

11    undercount of that population, this is a study about the 

12    reasons why.

13         Q.    Okay.

14         A.    456 is a study of the same area done at the same 

15    time.  The author is John M. Long.  The title is "Comparison 

16    and analysis of ethnographic and census matchings in East 

17    Los Angeles."

18         Q.    Okay.  After those two we should be in the same 

19    binder.

20         A.    The rest of these, I believe, are part of the 

21    1990 ethnographic studies.  If they're not, I'll say so.  

22               501 is entitled ethnographic evaluation of the 

23    1990 decennial census report number 2, the bicentennial 

24    census in a mixed neighborhood in Chicago.  The date is 

25    December 1991.  The author is Ann Terry Strauss.  
                                                              121

 1               502 is report number 3 of the same series.  

 2    Investigating census coverage --

 3         Q.    I don't think you need to read the titles.  Just 

 4    tell us what it is.

 5         A.    501 was an ethnography --

 6               THE COURT:  Start with 3 or 2.

 7         Q.    Start with 502.

 8         A.    All right.  This is a study of an apartment 

 9    complex in Houston, Texas where there are substantial 

10    numbers of immigrants from south of the border.

11         Q.    Okay.

12         A.    504 is a study of a rural community in Langston, 

13    Oklahoma.  

14               507 is the same kind of study in an area of 

15    Manhattan known as Chinatown.  

16               512 is the same kind of study in, as I recall, an 

17    African American section of Hartford, Connecticut.  

18               521 is a study out in Long Island of an area 

19    where undocumented Salvadoreans were concentrated.  

20               522 is a study in an area of the South Bronx 

21    where Mexicans, among other Hispanic groups, lived.  

22               525 is a study in a neighborhood of San Francisco 

23    known as the mission district.  

24               526 is a study of St. Louis Missouri, I believe 

25    the dominant group there were Laotians.  
                                                              122

 1               528 is a study of Hatians in Miami.  

 2               And 534 is actually not part of the 1990 

 3    ethnographic series.  It was what I would call a think piece 

 4    written by a social scientist and it deals with issues of 

 5    why poor people, especially blacks, are not likely to be 

 6    counted.  And it also deals with issues of counting 

 7    Hispanics.

 8               MR. ZIMROTH:  Your Honor, I move the admission of 

 9    Plaintiffs' Exhibits 455, 456, 501, 502, 504, 507, 512, 521, 

10    522, 525, 526, 528 and 534.

11               MR. SITCOV:  We object to all of them for hearsay 

12    and lack of foundation.  I can take them from the top.

13               THE COURT:  Let me find out from Mr. Zimroth why 

14    he is offering them.  For what purpose?

15               MR. ZIMROTH:  These documents, your Honor -- 

16    several purposes.  These documents, most of them were 

17    commissioned by the Census Bureau to look at the causes of 

18    undercount and to see whether or not the undercount -- and 

19    partly also to see whether or not there was any possible 

20    corroboration for the PES.  

21               You will see if you read these documents, I 

22    believe, that the ethnographers found not in a statistical 

23    way that the PES was, but in a very intense way for certain 

24    blocks, that precisely the kinds of things that the PES 

25    measured was seen to be true by the ethnographers.  
                                                              123

 1               These documents were part of the plan that the 

 2    Census Bureau itself had as evidenced by the prior exhibit 

 3    that has been admitted, 536, as part of their plan for 

 4    evaluating the PES.  These are Census Bureau documents.

 5               THE COURT:  They're Census Bureau commissioned 

 6    documents.

 7               MR. ZIMROTH:  Yes.  They're Census Bureau 

 8    commissioned documents and they are documents which are 

 9    reports to the Census Bureau by people that they 

10    commissioned and that they swore in as Census Bureau 

11    employees.  

12               I have here a letter which perhaps I can mark as 

13    a Court's Exhibit but let me read it first.  From one of the 

14    defendant's counsel about these ethnographic studies which 

15    says, basically it says that if we want to contact any of 

16    these ethnographers we should please be sensitive to the 

17    fact that these individuals were sworn as confidential 

18    Census Bureau employees in order to carry out their 

19    responsibilities relating to their agreements with the 

20    agency.  They have therefore received and may continue to 

21    possess Title 13 confidential materials that cannot be 

22    disclosed to plaintiffs or any other noncensus Bureau 

23    employees.  

24               This was part of the Census Bureau plans to 

25    evaluate the PES.
                                                              124

 1               MR. SITCOV:  Your Honor, perhaps Mr. Zimroth 

 2    should take the witness stand so he can describe what the 

 3    documents mean because it appears to me that he is providing 

 4    all the testimony about them.  The fact of the matter is 

 5    that each of the documents starting with 501 bear the legend 

 6    that these, the data or opinions contained therein should 

 7    not be construed as an official Bureau of the Census opinion 

 8    policy or decision unless so supported by other official 

 9    documentation and there isn't any such official 

10    documentation.  Each one of them bears that.

11               THE COURT:  It seems to me that one of the issues 

12    in the case will be the state of mind that the officials in 

13    the Census Bureau, whether they acted reasonably, whether 

14    they were in good faith, etc., and to prove that the 

15    plaintiffs would be entitled to offer these documents not as 

16    any exception to the hearsay rule proving what they contain, 

17    but simply as documents that would have a relevant impact on 

18    the state of mind of the people in the Census Bureau who 

19    read them.

20               MR. SITCOV:  But there is nothing from Dr. 

21    Ericksen that could even remotely address what the people in 

22    the Census Bureau had in the way of state of mind on the 

23    basis of these.

24               THE COURT:  As trier of the fact, I look at these 

25    and I can reasonably infer that this document would cause 
                                                              125

 1    that state of mind.  For those reasons I will admit all of 

 2    the documents starting with 455 and running up to 534.  

 3               (Plaintiffs' Exhibits 455, 456, 501, 502, 504, 

 4    507, 512, 521, 522, 525, 526, 528, 54 for identification 

 5    were received in evidence)

 6               MR. SITCOV:  If you look at 501, 502, 504, 507, 

 7    you see that they postdate the decision.  512 is clearly 

 8    identified as a draft.

 9               MR. ZIMROTH:  If Mr. Sitcov wants to stipulate 

10    that no post July 15th document will be admitted in this 

11    case that's one thing.  If you look at the document that you 

12    have already admitted into evidence, which talks about the 

13    plans, the plans show, I could find you the page, the author 

14    of whom was Elizabeth Martin, I believe that this witness 

15    testified was in charge of the ethnographic studies or 

16    planned for the ethnographic studies, these ethnographies 

17    were due to be done in final at the end of May of 1991.  

18               The information in these reports were clearly 

19    before the Census Bureau before July 15th.  Why they waited 

20    until whenever they waited before putting on the final stamp 

21    I don't know.  But the information was clearly there.  You 

22    can infer that from the fact that in Ms. Martin's memorandum 

23    it says that these drafts were due in April and final were 

24    supposed to be in April of 1991.

25               MR. SITCOV:  That demonstrates why plaintiff's 
                                                              126

 1    failure to do anything to authenticate these documents is 

 2    the reason they ought not to be admitted.  We are to have 

 3    these documents admitted because Mr. Zimroth says a letter 

 4    from someone who is not in the courtroom infers that the 

 5    dates on these documents don't mean what they say.  

 6               But everything in the documents is supposed to 

 7    not only mean what they say, but supposed to be the opinions 

 8    or the knowledge of the Census Bureau and the Secretary 

 9    prior to July 15, 1991.  They simply can't have it both 

10    ways.

11               THE COURT:  Mr. Zimroth, can you in any way 

12    establish to my satisfaction that the information contained 

13    in these postdated documents was known to the right people 

14    in the Census Bureau before July of 1991?

15               MR. ZIMROTH:  I can read to you the plans for 

16    these ethnographies and if you will just bear with me, I 

17    hope I can find it.  I'm looking at page 29, your Honor, now 

18    of Plaintiffs' Exhibit 536.  

19               (Continued on next page) 

20    

21    

22    

23    

24    

25    
                                                              127

 1    

 2    

 3    

 4    

 5    

 6    

 7    

 8               (Pause) 

 9               MR. ZIMROTH:  Your Honor, not only are these 

10    plans here, but the results of the ethnographic studies 

11    which was presented to the Bureau orally in June --

12               THE COURT:  That's the crucial one.  Show me 

13    that.  

14               MR. ZIMROTH:  We'll get you the document, your 

15    Honor.  

16               MR. SITCOV:  Your Honor, we have a strong reason 

17    to believe that the administrative record shows that these 

18    documents were not ready even by November of last year, and 

19    were not submitted to the Bureau until November.  

20               Your Honor, apparently, the committee meeting 

21    that occurred in November of 1990, that memorialized the 

22    fact that these ethnographic studies were not going to be 

23    ready.

24               THE COURT:  Were not going to -- 

25               MR. SITCOV:  -- be ready in time for the July 
                                                              128

 1    1991 decision.

 2               THE COURT:  Well, the formal written document may 

 3    not have been ready and yet they still may have had notice 

 4    of what was coming.  I think that's the point.  

 5               MR. SITCOV:  Well, if they had notice of what was 

 6    coming, then what they heard might be one thing, but I don't 

 7    know that that's what was in the document.

 8               THE COURT:  Well, that it seems to me would 

 9    affect the weight not the admissibility of the document.

10               If there's no evidence that they had any 

11    knowledge of this before the Secretary made the decision, 

12    then the documents would be irrelevant.  

13               MR. ZIMROTH:  May I ask Professor Ericksen one or 

14    two more questions on this point?

15               THE COURT:  All right.

16    BY MR. ZIMROTH:

17         Q.    Professor Ericksen, were you at a meeting in June 

18    of 1991 in which ethnographic reports were summarized and 

19    presented?

20         A.    No.  I think you might be referring to a 

21    conference that took place in June of 1991 where a movie was 

22    taken, but I did not attend that conference.

23         Q.    What was the movie?

24         A.    As I understand, it's a videotape of what all the 

25    ethnographers had to say about their studies.  
                                                              129

 1               MR. SITCOV:  That's what we are trying to find, 

 2    your Honor.  --

 3         Q.    And did you try to get a copy of that tape from 

 4    the Census Bureau?

 5         A.    Yes, I did.  I was informed of the existence of 

 6    that tape, and I wrote to Mr. Chuck St. Laurence, and asked 

 7    him for a copy of it.  I don't remember what his exact 

 8    response was, but to the best of my recollection, he told me 

 9    it would be coming, but it never came.

10               THE COURT:  Let me try to resolve the impasse 

11    this way.  I will overrule the objection for the moment and 

12    admit it upon state of mind subject to your later showing 

13    that it did have a relevant effect on state of mind because 

14    the right people knew it before July; and that if that proof 

15    is not forthcoming, I will then strike the documents.

16               (Exhibit PX 536 for identification was received 

17    in evidence) 

18    BY MR. ZIMROTH:

19         Q.    Professor Ericksen, are these ethnographic 

20    reports that were marked of the kind that you rely upon in 

21    forming your opinion as an expert in this matter?

22         A.    Yes, they are.

23         Q.    And did you read these ethnographic reports?

24         A.    Yes, I did.

25         Q.    And did it form at least part of the basis of 
                                                              130

 1    which you are testifying to?

 2         A.    Yes.  

 3               MR. SITCOV:  I think that's another independent 

 4    grounds for admission, your Honor.

 5               Let's go on.  

 6    BY MR. ZIMROTH:

 7         Q.    This construction of the mail list, is that in 

 8    your opinion an important stage of the process?

 9         A.    Yes, it is.

10         Q.    Why is that?

11         A.    Because the Census Bureau can only mail the 

12    census form out to those addresses that it knows about.

13               If it all the addresses it has it puts on the 

14    list, that's where it mails the forms.  If you are not on 

15    the list, you don't get a form.

16         Q.    Now, let's look at the second stage of the chart 

17    that's in front of you, the mailout/mailback stage.

18               Just briefly, what is that stage?   

19               It may seem too obvious to say, but say it 

20    anyway.

21         A.    Well, I just -- in the main, for 84 percent of 

22    the country, according to Census Bureau reports, the Census 

23    Bureau mails the census form out to the address.  And the 10 

24    percent of date leave -- an enumerator drops the form off at 

25    the address.  This is a stage in which the Census Bureau 
                                                              131

 1    gets the census form to the people.

 2               In the listed enumerator areas, it drops the form 

 3    off to the address.

 4         Q.    And then it's mailed back?

 5         A.    And the idea is that somebody in the household 

 6    will fill out the form and mail it back.

 7         Q.    Is this an important stage in your opinion in the 

 8    process?

 9         A.    Sure.  If people can't get the form, and if they 

10    can't figure out how to figure it out, the Census Bureau 

11    will not get the information that it needs.

12         Q.    Does the success or failure of this stage have an 

13    impact on the rest of the process?

14         A.    Yes.  To the extent the mailout/mailback doesn't 

15    work, it becomes exceedingly difficult to take an extra 

16    census, if not impossible.

17         Q.    Do you know whether or not the Census Bureau 

18    bases budgets on predictions about how well the 

19    mailout/mailback stage will succeed?

20         A.    Yes.  Because the parts of the census that follow 

21    the mailout/mailback phase are so vast, the Bureau would 

22    have to make a budget expectation in order to operate.

23               In 1990 --

24         Q.    Let's go through 1980 first.

25         A.    Okay.
                                                              132

 1         Q.    What was the mail response rate in 1980?

 2         A.    The mail response rate is the proportion of the 

 3    forms that went out that came back.  In 1980, the mail 

 4    response rate was 75 percent.

 5         Q.    And did the Census Bureau make a prediction about 

 6    1990 upon which it based its budget?

 7         A.    Yes.  It predicted that it would be 70 percent in 

 8    1990.  That's implying 5 percentage points.

 9         Q.    In your opinion, is that an important difference?

10         A.    Yes, it is.  Because in terms of the Census 

11    Bureau plans, it's not the mailback rate that matters; it's 

12    what you might call the mail non-back rate that matters.  

13    And in 1980, that was 25 percent.  And in 1990, they 

14    expected it would be 30 percent.  And if you make those 

15    numbers into a ratio, that means that 30 divided by 25 is 

16    120 percent.  So they expected to have to do 20 percent more 

17    work in 1990 than they did in 1980.

18         Q.    What actually happened in 1990?

19         A.    The mailback rate was 65 percent, which means 

20    that the mail non-back rate was 35 percent, and they had an 

21    additional 5 percentage points of work to do.

22         Q.    Will you please look at PX 458, which is in 

23    Volume No. 2.

24         Q.    Do you have that in front of you?

25         A.    I do.
                                                              133

 1         Q.    Could you tell us what that is?

 2         A.    Yes.  This is the final version of what was 

 3    called the check-in rate report.  As a member of the special 

 4    advisory panel, I received these on a weekly basis from the 

 5    Census Bureau.  And these reports told us how well the 

 6    mailout/mailback phase of the census was going.  And this is 

 7    the final report.

 8         Q.    And what did it show?  What did this monitor?  

 9    Did it show the Census Bureau was sick?  What does it show?

10         A.    I don't particularly go for that metaphor.

11         Q.    I'm sorry.  I'm sorry.

12         A.    Look, if you want a metaphor, it's like a 

13    predicted high with 70, and it only got to 65.

14               THE COURT:  That's a great metaphor.

15               THE WITNESS:  Thank you.

16               THE COURT:  That's a statistician's metaphor.  

17               MR. ZIMROTH:  What? 

18               THE WITNESS:  I once took a course on poetry.

19               THE COURT:  Did you pass?

20               THE WITNESS:  Oh, sure.  But they wouldn't let me 

21    in for post-graduate work.

22         Q.    I take back the question.

23               What does it show, Professor Ericksen?

24         A.    What it shows is that the budgeted mail response 

25    rate was 75 percent, and it got up to 65 percent.  
                                                              134

 1               MR. ZIMROTH:  Your Honor, I move the admission of 

 2    this document, 458.  

 3               MR. ZIMROTH:  We don't object.

 4               THE COURT:  458 is admitted.  With gratitude.

 5               (Exhibit PX 458 for identification was received 

 6    in evidence) 

 7    BY MR. ZIMROTH:

 8         Q.    Was the increase in workload -- now I'm talking 

 9    about the increase over what was budgeted, okay?  Was that 

10    distributed equally throughout the United States?

11         A.    No.

12         Q.    Could you tell us how it was distributed, and 

13    perhaps you can point the Judge to the particular chart or 

14    page that shows that?

15         A.    I turned over three pages.  The page numbering in 

16    this document is not easy, but the page that I'm looking at 

17    is entitled "Executive Daily Check-In Rate Report." 

18               THE COURT:  I have it.

19         A.    And in the third paragraph, it indicates three 

20    types of district offices.  

21               Type 1 district offices are those located in 

22    central cities, generally in the type of area where the 

23    Census Bureau expected the taking of the census to be more 

24    difficult.

25               Bedford Stuyvesant is a type 16 district office.  
                                                              135

 1    There's also Type 1 district offices in areas such as 

 2    Harlem.

 3               Type 2 district offices are those offices in 

 4    cities where the census is not so hard to take, but they are 

 5    predominantly in suburban areas.

 6               Type 3 district offices are those district 

 7    offices that are out in rural areas.  The Census Bureau 

 8    divided the country into 447 district offices, and this is 

 9    the administrative unit that they use to administer the 

10    taking of the census.  So every district office is either 

11    type 1, type 2 or type 3.

12               In the type 1 district offices, the mailback 

13    rate, or the mail response rate was 60 percent, leaving 40 

14    percent left to be done; in type 2, it was 66 percent, 

15    leaving 34 percent left to be done; and in type 3, it was 64 

16    percent, leaving 36 percent left to be done.

17         Q.    Does this report also indicate the rates for 

18    cities and states?

19         A.    Yes, it does.

20         Q.    Would you point out where that is for the 

21    Judge?

22         A.    Yes.  I turned over nine more pages to a page 

23    that has the number 1 on it, and it's entitled "Daily 

24    Check-In Rates for States."  And you can see that in the 

25    District of Columbia, for example, the mailback rate was 
                                                              136

 1    54.8 percent; in the state of Florida, it was 59.2 

 2    percent, --

 3         Q.    You needn't read anymore.  Just as long as the 

 4    Judge knows where it is.

 5         A.    Okay.

 6         Q.    And what about the cities?

 7         A.    Yes.  On the very next page, which also says 1 --

 8               MR. ZIMROTH:  This is a Census Bureau document, 

 9    your Honor.

10         A.    This was done under a great time pressure.  

11               It gives the same rates for cities, and it goes 

12    down to -- I think 47.4 percent for Boston is the lowest; 

13    for Cleveland it was 61.4 percent; for New York City it was 

14    53.2 percent; for Seattle, 67.5 percent.  

15               In general, these portions are lower than they 

16    were for the entire states.

17         Q.    Now, this second stage process, the 

18    mailout/mailback stage, in your opinion, does that work 

19    better -- let me put it another way.  

20               Does that present more difficulties for certain 

21    groups in the society than it does for others?

22         A.    Yes.

23         Q.    Perhaps you can take us through each of the -- 

24    let's start with the first step in the processes, the 

25    delivery of the mail, of the forms, and perhaps you can tell 
                                                              137

 1    us why in your opinion it has that effect.

 2         A.    Okay.  First I testified that I there are certain 

 3    houses -- types of housing which are more likely to be left 

 4    off the census form.

 5               Now, even where the addresses are on the census 

 6    form, there are difficulties in delivery.

 7               One of the ethnographic studies -- if necessary, 

 8    I can look up exactly which exhibit it is -- the one that 

 9    was done in the South Bronx encountered an apartment 

10    building where the mailboxes had been destroyed I believe by 

11    fire.  And in that particular apartment building, the census 

12    -- the mail is generally just left in a pile and people go 

13    and get it.

14               Now, the problem is that when the Census Bureau 

15    mails the forms out to people, they don't have names to send 

16    the forms to.  So when there is a pile of forms addressed to 

17    "resident," then people don't necessarily know that it's for 

18    them.

19               There are other examples of where people don't 

20    get separate mail.  Also discovered by the ethnographers, 

21    for example, in the Mission District of San Francisco, the 

22    apartment supervisor got mail for everyone.

23               In the study of the Haitian neighborhood in 

24    Miami, where houses have been subdivided, there was still 

25    just one form sent.  
                                                              138

 1               So oftentimes the mail just doesn't get through.

 2               The next stage is when the form is received, it 

 3    has to be opened.

 4         Q.    Before you get to the opening of the form, I 

 5    think this might be a good place to introduce the forms.  So 

 6    let me do that first, before you are now going to start 

 7    talking about the forms.

 8         A.    Okay.  

 9               MR. ZIMROTH:  Your Honor, at this time I want to 

10    introduce PX 460 and 461, which are the long -- well, let me 

11    -- I hope -- well, maybe there will be an objection, I don't 

12    know, but we'll see.  I want the Judge to have those 

13    originals.  Tell the Judge what those are.

14         A.    Those are census forms.  I think the first one 

15    460 is the short form, and I think 461 is the long form.  

16               MR. ZIMROTH:  And I move their admission, your 

17    Honor.

18               THE COURT:  We have copies in various books.  

19               MR. SITCOV:  No objection.

20               THE COURT:  460 and 461 are admitted.

21               (Exhibits PX 460 and 461 for identification were 

22    received in evidence) 

23               MR. ZIMROTH:  And I think we have some blowups 

24    here.  Let's just start with that one.  

25    BY MR. ZIMROTH:
                                                              139

 1         Q.    I'm sorry, Professor Ericksen.  I interrupted you 

 2    when you started to talk about -- assuming that the mail 

 3    does get delivered, now it has to be opened?

 4         A.    That's right.

 5         Q.    Okay.  So I left you at that stage, and perhaps 

 6    you can continue.

 7         A.    Okay.  So the cover of the census form is quite, 

 8    I believe, an intimidating document.  You notice that there 

 9    are a lot of words on the page -- 

10               MR. SITCOV:  I object, your Honor.  I don't 

11    believe that this witness has been qualified to give this 

12    type of expert witness testimony.

13               THE COURT:  I'm inclined to agree, Mr. Zimroth.  

14               MR. ZIMROTH:  Your Honor, this witness has been 

15    studying the causes of undercount for many, many years.

16               THE COURT:  From a statistical point of view.  

17               MR. ZIMROTH:  No, not only from a statistical 

18    point of view, from an ethnographic point of view, from a 

19    sociologic point of view as well.

20               THE COURT:  All right.  I'll permit it, but I 

21    will take into account that you are wondering beyond 

22    statistics what's he's trained in.  

23               MR. ZIMROTH:  I can go back over that area, your 

24    Honor, but I think you'll see this witness has tremendous 

25    training in this very area.
                                                              140

 1               THE COURT:  Go ahead.

 2    BY MR. ZIMROTH:

 3         Q.    Go ahead, Professor Ericksen.

 4         A.    I would like to say it's general knowledge in the 

 5    field of survey research that mail back forms are especially 

 6    difficult for certain groups of the population to deal with.

 7               If I may, I'd like to read a passage from my 

 8    report to the -- from the report I jointly sent to the 

 9    Secretary with three other colleagues on the panel.           

10               MR. SITCOV:  Is this in response to a question?

11               THE COURT:  No, this is a stream of 

12    consciousness.  

13               MR. SITCOV:  Well then, I'll object.

14               THE COURT:  Sustained.

15               Go ahead, Mr. Zimroth.

16    BY MR. ZIMROTH:

17         Q.    You were talking about what the difficulties were 

18    in terms of why it is that this form may have a different 

19    effect on certain people in society than others?

20         A.    Right.  And the point I'm trying to make is that 

21    people who were less well-educated find a form with all 

22    these words on a page.  I used the word intimidating.  Maybe 

23    I should have said that they are hard to deal with.  People 

24    like that get nervous.  

25               One of the things that we know in the survey 
                                                              141

 1    business is that people whose educational level is low are 

 2    less likely to return these forms.  That's the point that I 

 3    was trying to make.

 4         Q.    How long have you been involved in doing and 

 5    evaluating surveys?

 6         A.    25 years.

 7         Q.    And is this conclusion that you just gave in your 

 8    opinion well-known in the world of surveys?

 9         A.    Yes.  And that was documented in my report to the 

10    Secretary.

11         Q.    So.

12         A.    Okay.  The next point that I would like to make 

13    is that down at the bottom, it is written in Spanish.  This 

14    whole form is written in English.  But if someone's native 

15    language is Spanish, it gives people instructions about what 

16    number to call.  Sometimes that works, sometimes it didn't. 

17               There are examples of ethnographic studies of 

18    people who did try to call that number and had a difficult 

19    time.

20               Another point that the ethnographers made, that a 

21    number of people who speak Spanish are not especially 

22    literate in Spanish.  Which is another problem.

23               Finally, people whose native language is neither 

24    English nor Spanish would find the beginning of this form 

25    difficult.
                                                              142

 1         Q.    Do you know how many languages are spoken in the 

 2    City of New York?

 3         A.    Something over 100.  

 4               MR. ZIMROTH:  Your Honor, would you take a look 

 5    at PX 47897, please.

 6               (Pause) 

 7               You'll find it in Volume 4.

 8               (Pause)

 9         Q.    Could you just tell the Judge what that document 

10    is?

11         A.    Yes.  This is a document that was run by the 

12    computer programmers at the New York City, Department of 

13    City Planning based on a Census Bureau data file, describing 

14    the population of New York City from the 1980 census.

15         Q.    And in terms of the language you spoke, and the 

16    number of people --

17         A.    Yes.  This table describes the numbers of people 

18    who speak the various listed languages at home.

19               There's about 2.4 million people who speak a 

20    language other than English.  If you look on the second 

21    page, you see that the top of the table, about 1.26 million 

22    of them speak Spanish, but the rest of them speak other 

23    languages.  I believe I counted this once, about 120 or so 

24    different languages listed here.  

25               MR. SITCOV:  Okay 
                                                              143

 1               MR. ZIMROTH:  Okay.  Your Honor, I move the 

 2    admission of 497.  

 3               MR. SITCOV:  Your Honor, I'm not sure from the 

 4    description of the document if I've got the right one.

 5               (Discussion off the record) 

 6               MR. SITCOV:  So this is not a Census Bureau 

 7    document?  

 8               MR. ZIMROTH:  It's a computer run from a Census 

 9    Bureau data source.  

10               MR. SITCOV:  Then we object to this introduction, 

11    your Honor.  We don't know anything about it.  It's not a 

12    Census Bureau document.  

13               MR. ZIMROTH:  Excuse me, your Honor.  This is an 

14    official document of the New York City Planning Department, 

15    and I have a -- I'll show it to you, first.

16               (Pause)

17               THE COURT:  All it does is show the numbers of 

18    people who speak all these various languages.  What's the 

19    big deal?  

20               MR. SITCOV:  As I understand it, your Honor, --

21               THE COURT:  Even a New York City bureaucrat can 

22    do that.  

23               MR. SITCOV:  As I understand the certificate, New 

24    York is authenticating the Census Bureau document.  I don't 

25    think they can do that.
                                                              144

 1               THE COURT:  The objection is overruled.

 2               497 is admitted.

 3               (Exhibit PX 479 for identification was received 

 4    in evidence)

 5    BY MR. ZIMROTH:

 6         Q.    Okay.  Now --

 7               THE COURT:  All right.  Would this be a good time 

 8    to take a mid-afternoon break?  

 9               MR. ZIMROTH:  Yes.

10               THE COURT:  Let's take 15 minutes, ladies and 

11    gentlemen.

12               (Recess)

13               (In open court) 

14    EUGENE P. ERICKSEN, resumed

15               THE COURT:  All right.  Mr. Zimroth, you may 

16    resume.

17    DIRECT EXAMINATION CONTINUED 

18    BY MR. ZIMROTH:

19         Q.    Professor Ericksen, we put up the second panel 

20    here, the construction page.  Assuming that the person opens 

21    the forms, is this what he or she sees?

22         A.    Yes.  This is the first page of the census form.

23         Q.    In your opinion, does that form present 

24    particular difficulties for certain groups in our societies 

25    more than others?
                                                              145

 1         A.    Yes.  And what I'd like to say at this stage is 

 2    that I think that the job that the Census Bureau has in 

 3    getting people to understand who should be included is 

 4    formidable.

 5               There are, in the United States, different people 

 6    with different ideas about whose really part of their 

 7    family.

 8               For example, one of the ethnographic studies, I 

 9    believe it was in San Francisco, someone who was visiting, 

10    looking for a job, they found that they would were not 

11    considered to be part of the family, even though it was 

12    really the only residence they had in the United States.  An 

13    ethnographic study of Laotians in St. Louis found that even 

14    among people who have been living together for quite some 

15    time, their definition of what was the family did not 

16    include relatives outside of the father and mother and the 

17    young children.

18               So that one problem is that there are different 

19    cultural definitions of what constitutes family.  Even among 

20    people who are native born in the United States, there are 

21    segments of our society where people don't really have just 

22    one place where they live and eat all the time.  For various 

23    reasons, they may shuttle from one place to another.  It's 

24    not clear what their usual residence is either.

25               Studies of the homeless population indicate that 
                                                              146

 1    many people who are homeless are not always homeless.  They 

 2    go in and out of actually living somewhere, and there are 

 3    other people who while they may never be homeless have 

 4    different places where they live and sleep some of the time.

 5               Another example of a problem is if you think of a 

 6    young black mother who, say, is in her late 20s or early 

 7    30s, could have a child, and she has work hours that makes 

 8    day care a problem.  And this child could stay with a 

 9    relative most days of the week, but go home and stay with 

10    her on the weekend.  That mother would consider the child to 

11    live with her.  However, if such a person read this "usual 

12    residence" literally, might not include such a child.  And 

13    the grandmother might not include the child either.  And the 

14    omission of black children, under the age of 10, has been 

15    shown by demographic analysis to be a serious problem of 

16    census enumeration.

17               So the different definitions of the family are 

18    one problem.  A second problem is just that this is a 

19    difficult instruction.  And one of the problems that we in 

20    the survey business have with mail questionnaires is it's 

21    difficult to get people to read long instructions.  It's a 

22    very difficult problem the Bureau has because the definition 

23    they use of residence are a complicated one and it's 

24    difficult to get people to read these instructions.  It's 

25    not clear what they should do.
                                                              147

 1         Q.    Does that have an impact, in your opinion, on the 

 2    likelihood that someone would return the form at all?

 3         A.    Yes.  One of the things that we know is that the 

 4    people who find a questionnaire difficult to deal with will 

 5    sometimes become frustrated and simply fail to return the 

 6    form at all.

 7         Q.    Are you aware of any studies indicating how long 

 8    it takes to fill out the census form?

 9         A.    Well, I haven't seen studies that I recall.  My 

10    recollection is the Census Bureau expectation was that it 

11    would take 15 minutes on average to fill out the short form.

12         Q.    And the long form?

13         A.    The long form had more questions.  It would take 

14    about three times as long.  As I recall, the expected length 

15    of time was 43 minutes.

16               And in telling you that, I should explain what I 

17    mean "on average."  The typical family might have two or 

18    three people in it.  If the family had six people in it, 

19    then, obviously, it would take longer to fill out the form.

20         Q.    And that's what, the quote-unquote average 

21    household?

22         A.    Yes.

23         Q.    Do you know, and in your opinion, would it take 

24    longer for certain households to fill out the form than 

25    quote-unquote average?
                                                              148

 1         A.    Certainly.  Because you need to fill out 

 2    information for every single person.  The more people you 

 3    have, the more information there is to fill in.  After that, 

 4    the problems of literacy and language difficulty that I've 

 5    talked about would make filling out the form even longer.

 6         Q.    Do you know whether or not there was as many long 

 7    forms or the same percentage of long forms were returned as 

 8    short forms?

 9         A.    No.  The mailback rate for long forms was lower 

10    than the mailback rate for short forms.

11               As I recall, the mailback rate for short forms 

12    was 66 percent; the mailback rate for long forms was 60 

13    percent.  In type 1 district offices, which is where the 

14    urban bureau offices, the mailback rate for short forms was 

15    60 percent, the mailback rate for long forms was 53 percent.

16         Q.    Now, you talked about the difficulty that the 

17    Census Bureau has in describing the residency rules.  And I 

18    take it that you are not by any means being critical of 

19    them, just that it's a difficult task.

20         A.    Yes.  If the Census Bureau asked me to come in 

21    and tell them how to write these rules, I'm not sure I could 

22    do it any better.  It's a very difficult set of instructions 

23    to write.

24         Q.    In your opinion, is confusion about residence 

25    instructions important in terms of the undercount?
                                                              149

 1         A.    Yes, it is.  I've already indicated how confusion 

 2    about the form and perception of the form is difficult, 

 3    would make people less likely to mail it back in.

 4               But the confusion about who to put down causes a 

 5    second problem which may be just as important.

 6         Q.    Would you tell the Judge what that problem is?

 7         A.    Yes.  Census Bureau studies have shown -- and 

 8    this has been corroborated by some of the ethnographic 

 9    studies, that -- well, let me define what sociologists mean 

10    by the nuclear family is: the father and the mother and the 

11    young children.

12               And the people in the nuclear family who fills 

13    out the census form are less likely to be left out -- 

14    substantially less likely to be left out than distant 

15    relatives or people who are not related.  And that seems 

16    directly attributable to the difficulty people have in 

17    dealing with these instructions.

18         Q.    You've described three categories now: the 

19    nuclear family, the distant relations and the non- 

20    relations.

21         A.    That's right.

22         Q.    Do you know of any studies that show that -- 

23    well, you've mentioned studies that show that the further 

24    the relationship of the person from the head of the 

25    household, meaning the person filling out the form, the more 
                                                              150

 1    likely that person is to be missed?

 2         A.    Yes.  The Census Bureau has demonstrated that on 

 3    at least two occasions that I know of.

 4         Q.    What are those occasions?

 5         A.    Well, the test census -- there's a Census Bureau 

 6    study from the 1980 census that showed that non-relatives 

 7    were most likely to be left off, distant relatives were also 

 8    very -- much more likely to be left off, and close 

 9    relatives, as I have described, were much less likely to be 

10    left out.

11               There was another study following the 1986 test 

12    census in Los Angeles County which obtained the same result.

13         Q.    Will you please look at PX 464.

14               Tell the Judge what that is.

15         A.    This is a Census Bureau paper entitled: "An 

16    analysis of Within-household Undercoverage in the Current 

17    Population Survey," written by Robert E. Fay.

18         Q.    And what did it show?

19         A.    This paper shows that distant relations and 

20    non-relatives are much more difficult to count.

21         Q.    And would you tell the Judge what 465 -- and 

22    that's a study of the 1980 census, is that correct?

23         A.    Yes.

24         Q.    And would you tell the Judge what 465 is.

25         A.    465 is entitled "Towards A Theory of Census 
                                                              151

 1    Undercount: An Exploratory Assessment of Data From The 1986 

 2    Los Angeles Test Census" by David J. Fein and Kirsten K. 

 3    West. 

 4               They found the same result in the 1986 test 

 5    census in Los Angeles County.  

 6               MR. ZIMROTH:  Your Honor, I move the admission of 

 7    464 and 465.  

 8               MR. SITCOV:  We don't object.

 9               THE COURT:  You don't.  464 and 465 admitted.  

10               (Exhibits PX 464 and 465 for identification were 

11    received in evidence) 

12    BY MR. ZIMROTH:

13         Q.    Do you know of any studies that deal with the 

14    question whether or not distant relations are more likely to 

15    be living in minority households?

16         A.    Yes, I do.

17         Q.    Would you tell us about that?

18         A.    Well, first of all, I believe there's a -- many 

19    ethnographic studies found that.  I made those same 

20    calculations myself on the basis of 1980 census data.

21         Q.    Would you take a look at 466.

22               (Pause)

23         Q.    Do you have it?

24         A.    Yes, I do.

25         Q.    Is that one such study?
                                                              152

 1         A.    Yes.  This is a study describing ethnographic 

 2    research that had been done up to 1988 on the subject I 

 3    described.

 4               MR. ZIMROTH:  I move its admission, your Honor.  

 5               MR. SITCOV:  Object.  The witness hasn't 

 6    established the foundation, and I notice this document 

 7    begins at page 513.  I don't know what it's of.  And it ends 

 8    somewhere in the higher 500s.

 9               THE COURT:  Yours has page numbers on it?  

10               MR. SITCOV:  Yes, it does, in the upper right- 

11    hand corner.

12               THE COURT:  You are talking about 466.  

13               MR. SITCOV:  Yes.

14               THE COURT:  Mine doesn't have any numbers. 

15               MR. SITCOV:  In the upper right-hand corner the 

16    first one is page 513, and it goes on.  This doesn't appear 

17    to be a government document.  And as I say, the witness has 

18    made no foundation for its admission.

19               THE COURT:  Mr. Zimroth.  

20               MR. ZIMROTH:  Can you identify who Catherine 

21    Hines is, Elizabeth Martin and Gary Shapiro?

22         A.    Elizabeth Martin is the person at the Census 

23    Bureau.  I met her at a meeting of the Special Advisory 

24    Panel.  She was the person in charge of the ethnographic 

25    studies of the 1990 census.  Gary Shapiro was a statistician 
                                                              153

 1    at the Census Bureau.  Catherine Hines is another Census 

 2    Bureau employee.  Peter Hainer of Curry College is a 

 3    consultant of the Census Bureau. 

 4         Q.    Do you know where this document comes from?

 5         A.    Yes.  I believe this is from a -- I've forgotten 

 6    exactly what the Bureau calls them, but they hold the 

 7    research conferences every year and this is from the 

 8    proceedings of the 1988 research conferences.

 9         Q.    If I can just show you this document and see if 

10    it refreshes your recollection, if this is the right one.

11         A.    Right.  It's the Fourth Annual Research 

12    Conference.  

13               MR. ZIMROTH:  Your Honor, I move its admission.

14               THE COURT:  Are you going to find out about the 

15    strange pagination?  

16               MR. ZIMROTH:  It's possible that -- well --

17         Q.    Do you know about the pagination?

18         A.    Yes, all the papers are bound in a volume and 

19    this one just happens to be the one that starts on page 513.  

20               MR. SITCOV:  He still hasn't laid a foundation, 

21    so we continue our objection on this ground.

22               THE COURT:  I think the foundation is adequate.  

23    The objection is overruled.  466 is admitted.

24               (Exhibit PX 466 for identification was received 

25    in evidence)
                                                              154

 1         Q.    You mentioned also that you did your own 

 2    calculation, your own study on whether distant or 

 3    non-relations are more likely to be living in a minority 

 4    household, is that correct?

 5         A.    Yes, I did.

 6               MR. ZIMROTH:  Would you put the next chart up, 

 7    please, 467. 

 8               (Pause)

 9               Sorry, your Honor.  I thought we had blown that 

10    up.  But we hadn't.

11         A.    Okay.  I looked up in chapter C of United States 

12    Summary from the 1980 Census and found the Census Bureau 

13    tabulation of the shares of the counted population in 

14    households who were distant or nonrelatives.  And for 

15    whites, if you add the 3.3 and the 2.6, it comes to 5.9 

16    percent; for blacks, if you do the same addition, it's 12.9 

17    percent; for Hispanic, it's 10.6 percent.  So we see that 

18    these distant and non-relatives are much more likely to live 

19    in black or Hispanic houses than they do in white houses.  

20               MR. ZIMROTH:  Your Honor, I move the admission of 

21    467.  

22               MR. SITCOV:  No objection.

23               THE COURT:  Overruled.  

24               MR. SITCOV:  I said no objection.  If you 

25    overrule it, then we'll --
                                                              155

 1               THE COURT:  Admitted.

 2               (Exhibit PX 467 for identification was received 

 3    in evidence) 

 4    BY MR. ZIMROTH:

 5         Q.    Now, Professor Ericksen, do you know what 

 6    Outreach Programs are?

 7         A.    Yes, I do.

 8         Q.    What do they have to do with this second step, 

 9    the mailout/mailback rate, if anything?

10         A.    Well, as part of the 1990 census, the Census 

11    Bureau wanted to motivate and to inform the population as 

12    much as it possibly could to find, fill out and mailback 

13    their census forms.

14               The Bureau conducted an extensive advertising 

15    campaign.  As I understand it, the advertising was donated 

16    by the Advertising Council; and while the 1990 Census was 

17    going on, the Census Bureau was one of the biggest 

18    advertisers in America.  The Outreach Programs intended to 

19    supplement the advertising because the Bureau made special 

20    attempts to reach out to various community groups and 

21    schools; just to give you two examples, to try to reach 

22    people so that they would encourage folks to be counted.  I 

23    recall seeing packages of materials that were given to 

24    school children, presumably urban school children who had 

25    taken it home to their families so they could be counted.
                                                              156

 1         Q.    Now, Professor Ericksen, I want you to -- are you 

 2    also aware that certain states participated in or 

 3    supplemented those programs and had their own average 

 4    programs?

 5         A.    I am aware that there are states that did that.

 6         Q.    Would you take a look at PX 47.

 7         A.    47?

 8         Q.    Yes.  47.

 9         Q.    Now, just for identification purposes, that is 

10    the recommendation to the Secretary for Michael R. Darvey, 

11    Secretary for Economic Affairs and Administrator for the 

12    Economics and Statistics Administration.

13               And I'd like you to turn to page 40, and I would 

14    like to read into the record that and ask you a question 

15    about it -- several questions about it.

16               "Indeed, the fact that uncounted persons would be 

17    added through an adjustment may actually create a 

18    disincentive to respond to and encourage response to the 

19    census.  A stark example of this disincentive is the case of 

20    Wisconsin, which undertook a $750,000 statewide public 

21    awareness campaign and targeted Outreach Programs to 

22    encourage its residents to participate in the census.  These 

23    efforts resulted in the highest census mail response for any 

24    state in the nation, an accomplishment that was rewarded by 

25    the Census Bureau, yet with constant chance to lose a 
                                                              157

 1    Congressional seat and a portion of its share of federal 

 2    funds, precisely the cost of its low estimated undercount 

 3    relative to other states.  State officials acknowledged in 

 4    public comments that it is doubtful that such support would 

 5    have been provided had they known that adjustment would be 

 6    used to accommodate an undercount.  This example will likely 

 7    result in a loss of state and local support for programs 

 8    encouraging participation in future censuses."

 9               Professor Ericksen, did you do a study of the 

10    effects of Wisconsin's Outreach Program?

11         A.    I did a study of what happened in Wisconsin after 

12    the Outreach Program took place, yes.

13               MR. ZIMROTH:  Okay.  Would you show the next 

14    chart, which is 468. 

15         Q.    Professor Ericksen, this blowup here, which is 

16    468, does that represent the results of your work?

17         A.    Yes, it is.

18         Q.    Could you perhaps come up here and -- 

19               MR. ZIMROTH:  Your Honor, would it be permissible 

20    for Professor Ericksen to come up and explain the chart?

21               THE COURT:  Certainly.

22         Q.    Would you first just start at the top and tell us 

23    what mail return rates are.

24         A.    Okay.  Mail return rates are the percentage of 

25    persons who are counted by mailout/mailback.
                                                              158

 1               So if there are 100 people counted in an area, 

 2    and 60 of them are on the mailback forms and 40 of them had 

 3    to be counted by some other method, we call the mail return 

 4    rate 60 percent.

 5               In other words, it could be thought of as a 

 6    measure of the cooperation of the population.

 7               It's not the same as the mailback rate that I was 

 8    talking about before.

 9               Do you want me to continue?

10         Q.    Please.

11         A.    Okay. 

12         Q.    Explain the chart.  Tell the Judge how you came 

13    to those numbers and what you think they mean.

14         A.    In the United States, 72.3 percent of the 

15    population was counted by mail.  Now, this is higher than 

16    the 65 percent I told you a while ago because we eliminate 

17    vacant housing from the denominator.

18               Now, in Wisconsin, the corresponding figure is 

19    82.5 percent.  So in Wisconsin it's clear that people were 

20    much more likely to respond by mail than the rest of the 

21    nation.  That could be, in effect, of the individual 

22    characteristic of people in Wisconsin; it could be an effect 

23    of the Outreach Program.

24               Now, Wisconsin happens to be a state where about 

25    9 percent of the population is minority, compared to about 
                                                              159

 1    24 percent for the nation.  If you look at white, 

 2    non-Hispanics in the nation, 77.1 percent of those people 

 3    were counted on mailback forms.  That's higher than the 

 4    average for everyone.

 5               In Wisconsin, the corresponding percentage was 

 6    84.7 percent.  So white, non-Hispanics living in Wisconsin 

 7    were much more likely to return their forms than white 

 8    non-Hispanics living elsewhere.

 9               Now, if you look at the minority population in 

10    the nation, only 57.7 percent of minority persons in the 

11    United States were counted by mail.  That's about 19.4 

12    percentage points less than 77.1 percent for whites.

13               In the state of Wisconsin, the difference is only 

14    slightly higher.  It's 59.9 percent.

15               So whatever the benefits -- 

16               (Continued on next page) 

17    

18    

19    

20    

21    

22    

23    

24    

25    
                                                              160

 1    

 2    

 3    

 4    

 5    

 6    

 7    

 8    

 9         Q.    You say the difference is slightly higher?

10         A.    59.9 percent is slightly higher than 57.7 

11    percent.  So whatever the benefits of the outreach program 

12    were in Wisconsin, it seems that they were more likely to 

13    apply to the white nonHispanic population than to the 

14    minority population.  

15               In fact, the difference between the minority and 

16    the white mail return rate in Wisconsin was 24.8 percentage 

17    percentage points.  So in fact, in Wisconsin, minorities 

18    were worse off at the end of the mail out mail back phase of 

19    the census relative to whites than they were in the rest of 

20    the nation.

21               MR. ZIMROTH:  I want to move the admission of 

22    468.

23               MR. SITCOV:  No objection.

24               THE COURT:  468 is admitted.  

25               (Plaintiffs' Exhibit 468 for identification was 
                                                              161

 1    received in evidence)

 2         Q.    I want to move to the next stage in the process, 

 3    the nonresponse follow-up stage.  Could you please describe 

 4    for the judge what that stage is, what happens?

 5         A.    Okay.  After a mail out mail back is completed, 

 6    on about April 25th, that was three and a half weeks after 

 7    April 1st, the Census Bureau had, it felt, the mail back 

 8    forms it was going to get.  It then moved to count the 

 9    people living at addresses where the forms had not been 

10    mailed back in that.  

11               That particular phase of the process is called 

12    nonresponse follow-up.  And this is the part of the census 

13    phase where enumerators go to the house and knock on the 

14    door.  

15               Enumerators were instructed to make up to three 

16    visits to every address to try to find the people there.  

17    And they were also instructed to go at different times of 

18    day to increase the chances of finding the people.  

19               They were also instructed to make up to three 

20    telephone attempts if that was possible to try to contact 

21    the people by telephone.  And they would go to the house and 

22    they would go through all the questions on the census form 

23    and get the data that had not been mailed back in.

24         Q.    And what happened if they made the three visits, 

25    the three telephone calls and they couldn't get the 
                                                              162

 1    information?

 2         A.    At that stage the enumerator was permitted to use 

 3    a procedure known as last resort.

 4         Q.    What is that?

 5         A.    Last resort is a procedure where the enumerator 

 6    is told to get the information from a neighbor, from a 

 7    building superintendent or from some knowledgeable person.

 8         Q.    And is the enumerator instructed about how to 

 9    define a knowledgeable person?

10         A.    No.  I reviewed the instructions for last resort 

11    and there's no definition of what a knowledgeable person is.

12         Q.    Now, what is the C there, closeout?

13         A.    What's the exhibit number for that?  Closeout --

14         Q.    I may have jumped ahead just slightly.  Let's 

15    stay with last resort.  And maybe I missed your answer here 

16    and if so please forgive me.  In the last resort stage, is 

17    the enumerator instructed to get a certain number of 

18    characteristics for the person?

19         A.    Yes.  There are seven questions which are to be 

20    obtained for every individual in the household as well as 

21    information about the housing unit itself.  And the 

22    enumerator is told to get three out of four specified 

23    characteristics for each personal living in the house.

24         Q.    And that's in the last resort?

25         A.    In the last resort phase.
                                                              163

 1         Q.    And supposing the enumerator cannot get three out 

 2    of the -- do you remember what those four are?  If not it's 

 3    not so important but if you do --

 4         A.    I believe it's relationship -- the answer to 

 5    question number one, relationship to head of household, age, 

 6    sex and race.

 7         Q.    Supposing the enumerator cannot get three out of 

 8    the four of those characteristics?

 9         A.    In that case the form would not be accepted as 

10    counted by last resort and the address would still be left 

11    open and the next attempt to enumerate would be in what's 

12    called closeout phase.

13         Q.    Tell us about the closeout face?

14         A.    In the closeout phase the enumerator would return 

15    to the address and make one last attempt to get the 

16    information and in this case, the form would be accepted if 

17    there were two characteristics.

18         Q.    And what happens if it can't get two 

19    characteristics?

20         A.    If it can't get two characteristics but you're 

21    able to determine that the house is occupied, then the form 

22    could be accepted with zero or with one characteristic for 

23    each person.

24         Q.    What is that called?

25         A.    Well, people who are included in accepted census 
                                                              164

 1    forms with zero or one characteristic or called nondata 

 2    defined.

 3         Q.    In these latter stages of nonresponse follow-up, 

 4    last resort, closeout and I'm sorry we left out one which is 

 5    nondata defined, in your opinion is the information that is 

 6    obtained from those process, those stages -- do you see a 

 7    problem in terms of the reliability concerning those stages?

 8         A.    Yes, I do.

 9         Q.    Tell us about it.

10         A.    Well, first of all you're already leaving out 

11    characteristics.  Second of all, there are not really very 

12    good controls that the person providing the information is 

13    well-informed.  A lot of errors are made in these phases of 

14    the census and this stands, as a matter of survey research 

15    common sense, that people are going to be left out because 

16    neighbors don't know everybody who lives in the next house.  

17               It's also -- the ethnographer in the Houston 

18    study reported that the apartment manager reported for a lot 

19    of the units in his building and was unaware of a lot of the 

20    people living there and the reason he was unaware is because 

21    the residents were shielding household members from him 

22    because they were violated the terms of the lease.

23         Q.    In your opinion, this stage which we've called 

24    nonresponse follow-up, does the construction of that phase 

25    have a different impact on certain groups in the society 
                                                              165

 1    than on others.

 2         A.    Yes, it does.

 3         Q.    Why is that?

 4         A.    Well, first of all nonresponse follow-up is much 

 5    more difficult to conduct in neighborhoods where the census 

 6    is harder to take.  It's hard to hire enumerators.  

 7    Enumerators are working in very difficult conditions.  A lot 

 8    of the housing is difficult to find.  The crime rate is 

 9    higher.  The neighborhoods are more frightening.  You're 

10    more likely to run into people who don't speak English.  

11    It's a very difficult job.

12         Q.    You mentioned the difficulties of hiring and 

13    keeping enumerators.  Do you know of any Census Bureau 

14    studies on that subject in the 1990 decennial census?

15         A.    Yes.

16         Q.    Could you look at 470 and first tell us what you 

17    know about the study.

18         A.    Well, while on the Special Advisory Panel, we 

19    were given weekly reports on the conduct of the census, and 

20    it was clear that nonresponse follow-up was going much more 

21    slowly in type 1 offices generally but in type 1 offices in 

22    particular cities, like New York much more specifically.  It 

23    just took much longer to collect all those data.

24         Q.    Would you take a look at 470, please.  

25    Specifically attachment D.
                                                              166

 1         A.    Attachment D?

 2         Q.    Yes.  First of all, what is the document?

 3         A.    This is 1990 decennial research, preliminary 

 4    research and evaluation memorandum number 100, prepared by 

 5    Timothy Divine, Peter Sefton, Cloverfield division and the 

 6    subject is staff turnover for the 1990 census.  

 7               This page was presented at the 1991 joint 

 8    statistical meetings in Atlanta, Georgia, in August 1991.

 9         Q.    Could you look at attachment D.

10         A.    D as in dog or B as in boy?

11         Q.    D ace in dog.  You're looking at D now.

12         A.    Right.

13         Q.    Can you tell us what that indicates?

14         A.    Yes.  The title of the chart is turnover rate by 

15    office type.  What the authors mean by turnover rate here is 

16    the added number of enumerators that had to be hired.  So on 

17    a national basis they had to hire 93 percent more 

18    enumerators than they had originally budgeted.  And they 

19    called that the turnover rate.  I would call it the rate of 

20    having to hire extra enumerators, but of course that was 

21    caused because the turnover rate was higher.  

22               In type 1 district offices which are the problem 

23    offices that I talked about earlier, the corresponding rate 

24    was 127 percent.  In type 2 district offices it was 84 

25    percent.  In type 2 A, now type 2 A is a small subset of 
                                                              167

 1    type 2 offices with the update leave method was used, it was 

 2    71 percent.  And in type 3 offices it was 95 percent.  

 3               So the conclusion to be drawn from this table is 

 4    that it was harder to keep enumerators in type 1 district 

 5    offices.

 6         Q.    And will you look at attachment E.

 7         A.    Okay.  What this is about is a nonresponse 

 8    follow-up is intended to end on June 6, 1990.  In some 

 9    district offices, in many district offices nonresponse 

10    follow-up went beyond June 6th.  In some areas it went on 

11    into July.  So these are the days of completion by office 

12    type.  And we can see that national average was 57.1 

13    percent.

14         Q.    Is that percent or days?

15         A.    I'm sorry?

16         Q.    Is that percent or days?

17         A.    Did I say percent?  Yes, it's 57.1 days.  For 

18    type 1 it was 67.4 days.  Type 2 64.5.  Type 2 A 60.3.  Type 

19    3 61.2.  It takes a lot longer to complete nonresponse 

20    follow-up in type 1 district offices.

21         Q.    And what's the significance of that in your 

22    opinion?

23         A.    They were having a harder time collecting the 

24    data.

25               MR. ZIMROTH:  I move the admission of Plaintiffs' 
                                                              168

 1    Exhibit 470.

 2               MR. SITCOV:  Objection on two grounds.  One is 

 3    foundation and two, the date, the front of 470 is December 

 4    9, 1991.  There's a caveat at the bottom of that page that 

 5    says "The conclusions and recommendations contained herein 

 6    essentially reflect the thoughts of certain staff members at 

 7    the time of publication and should not be interpreted as 

 8    statements of Census Bureau position."  

 9               This document on its face says it's not the 

10    Census Bureau decision, it was well after the adjustment 

11    decision was made.

12               THE COURT:  Mr. Zimroth.

13               MR. ZIMROTH:  I don't see -- if Mr. Sitcov wants 

14    to stipulate right now that no documents produced after the 

15    decision was made -- we would be happy to stipulate to that.  

16    In fact we made a motion on this ground.  Your Honor ruled 

17    to the contrary.  I don't understand what this objection is.  

18               This witness is telling us that in his opinion, 

19    the nonresponse, the latter stages of nonresponse follow-up 

20    is much harder in urban centers then it is elsewhere.  

21               This document provides data that shows that.  

22    This document was produced by the defendants in the 

23    discovery in this case.  In fact, it was produced, I 

24    believe, at a deposition of one of the defendant's 

25    witnesses.  There can be no question about the authenticity 
                                                              169

 1    of this document.

 2               MR. SITCOV:  That is not our Bates stamp number, 

 3    your Honor.  If you compare it to, let's take a look at 

 4    Plaintiffs' Exhibit 47 which includes our Bates stamp.  This 

 5    is not a document that is part of the administrative record.

 6               MR. ZIMROTH:  This document is in response to 

 7    requests that we made or demands that we made in connection 

 8    with the deposition of Peter Bounpane who is one of the 

 9    witnesses in this case.  That's why this document has a B in 

10    front of it.

11               THE COURT:  What is the purpose for which you are 

12    offering this document, Mr. Zimroth?

13               MR. ZIMROTH:  This document is to show, to 

14    support the witness' expert opinion that it is in fact 

15    harder to do the nonresponse follow-up stage in inner 

16    cities, and therefore that there is a differential impact in 

17    the way in which the census is constructed.

18               THE COURT:  Well, if it's offered for that 

19    purpose it's hearsay. 

20               MR. ZIMROTH:  Even if it is hearsay, I think it's 

21    admissible, your Honor, because it forms a basis of this 

22    witness' testimony.  And also it is a Census Bureau 

23    document.  There's no denying that.  It is an official 

24    document of a party defendant in the case.  And it would be 

25    an exception to the hearsay rule for that reason.  
                                                              170

 1               It is also --

 2               MR. SITCOV:  While Mr. Zimroth is waiting, I 

 3    would just point out it is not an official document as this 

 4    legend indicates.  And there has been no testimony from the 

 5    witness and indeed there could not be that he relied upon 

 6    this document in making his recommendation.  His 

 7    recommendation was made to the Secretary three months -- 

 8    well, according to this five months in advance.

 9               MR. ZIMROTH:  Your Honor, this document was 

10    turned to us in response to the fact that at the deposition 

11    of Peter Bounpane he said that he was going to rely on post 

12    July 15th documents to support his testimony.  We then asked 

13    for all relevant post July 15th documents.  This document 

14    was produced in response to that request.  That's point 

15    number 1.  

16               Point number 2 is, regardless of whether this 

17    document was available to this witness in June, June 250th, 

18    when he made his recommendation, he's testifying here today 

19    about some of the same subjects.  And it's added support for 

20    his opinion.

21               THE COURT:  Bounpane is going to be called, isn't 

22    he?

23               MR. ZIMROTH:  By the defendants.

24               THE COURT:  Wouldn't it make more sense to put it 

25    in through Bounpane even if necessary to do it on cross?
                                                              171

 1               MR. ZIMROTH:  Your Honor, it seems to me -- 

 2    you're talking about the hearsay exception, now?

 3               THE COURT:  Yes.

 4               MR. ZIMROTH:  Even if it is hearsay, your Honor, 

 5    I think it is admissible because it forms a part of the 

 6    basis of this witness' testimony.

 7               THE COURT:  There's no exception to the hearsay 

 8    rule for that, that's the problem.  The objection is 

 9    sustained.

10         Q.    Professor Ericksen, in your opinion are 

11    minorities more likely to be counted in the latter stages of 

12    nonresponse follow-up?

13               THE COURT:  Being what?

14               MR. ZIMROTH:  Last resort, closeout and nondata 

15    defined.

16               THE COURT:  B, C and D if there was a D?

17               MR. ZIMROTH:  Right.

18         A.    Well, tabulations that I received from the Census 

19    Bureau while I was a Special Advisory Panel member working 

20    on the joint report that I submitted with Estrada, Tukey and 

21    Wolter, provided data to that.  And it shows, yes, that they 

22    were more likely to be counted by one of those methods.

23         Q.    Will you please -- your Honor, may the witness 

24    come up to the chart?

25               THE COURT:  Certainly.
                                                              172

 1               MR. ZIMROTH:  This chart is labled Plaintiffs' 

 2    Exhibit 702.

 3         Q.    Explain how that chart was constructed.

 4         A.    Yes.  We received --

 5         Q.    Wait for the judge.  

 6               (Pause)

 7               THE COURT:  All right.  I'm with you, sorry.

 8         A.    Okay.  Now --

 9         Q.    Why don't you explain what this chart is meant to 

10    show.

11         A.    Okay.  The Census Bureau provided tabulations of 

12    all the 248.7 million people counted in the original 

13    enumeration and sorted them in two ways.  

14               First, it sorted them according to whether they 

15    were white, black, Asian or other race.  And it also sorted 

16    them as to whether or not they were Hispanic.  

17               It took the same people and it also sorted them 

18    out as to whether they were classified as enumerated by last 

19    resort, which meant that they had the three or four 

20    characteristics that I told you about, or if they were 

21    enumerated by closeout or if they were not nondata defined 

22    which meant they only had zero or one characteristic. 

23               So these are the people who were counted at the 

24    end of nonresponse follow-up for whom the characteristics 

25    were missing.  
                                                              173

 1               Now, if we look over here, at the nondata 

 2    defined, what we have here is, the red bar is the proportion 

 3    of nonwhite Hispanics.  And I don't recall the exact number 

 4    but that looks like it's about .6 percent.  Then the green 

 5    bar is the corresponding proportion for black nonHispanics.  

 6    That's about 1.5 percent.  And the blue bar with white 

 7    stripes is the same for Asians, people of other races and 

 8    Hispanics.  In this proportion it's about 1.2 percent and 

 9    then the white bar with blue stripes is the average for the 

10    whole population and this looks like it's about .8 percent.

11         Q.    That chart was constructed from I believe table 

12    11 of appendix A of your report.  So you can look at the 

13    actual numbers if you want.

14         A.    Okay.  

15               (Pause)

16         A.    So the nondata defined was .6 percent for white 

17    nonHispanics, it was 1.5 percent for black nonHispanics, it 

18    was about 1.2 percent for everyone else, and it's .8 percent 

19    for the total.  

20               Now moving over to the closeout phase the 

21    corresponding percentages are .7 percent for white 

22    nonHispanics, 1.5 percent for black nonHispanics, about.8 or 

23    .9 percent for all others and it was .8 percent for the 

24    total population.  

25               For last resort, it was 1.6 percent for white 
                                                              174

 1    nonHispanics, it was 3.4 percent for black nonHispanics, it 

 2    was about 2.1 percent for everyone else and 1.9 percent in 

 3    total.  

 4               So you combine these three groups for the white 

 5    nonHispanics, if you add the .6 and the .7 and the 1.6 you 

 6    get 2.9 percent.  

 7               For blacks nonHispanics you add the 1.5 and the 

 8    1.5 and the 3.4, you get 6.4 percent.  Which is about one 

 9    out of sixteen people.

10         Q.    One out of how many?  What is that percentage 

11    that you just said?

12         A.    6.4 percent, which is about one out of sixteen 

13    people.

14         Q.    Of that category of people that were enumerated 

15    in the entire United States was enumerated by one of these 

16    three methods?

17         A.    Yes.  And for everyone else it's about 4.1 

18    percent.  And for the nation's population, it was 3.5 

19    percent.  And that's 3.5 percent out of a total population 

20    so approximately 8 and one-half million people were counted 

21    by last resort, closeout or were nondata defined.

22         Q.    What is the significance of that chart in terms 

23    of the impact of nonresponse follow-up on certain groups in 

24    the society, minorities in particular?

25         A.    What this shows to me, number one, the Census 
                                                              175

 1    Bureau had a more difficult time counting the minority 

 2    groups and number two, it indicates because of the nature of 

 3    these procedures that these people were counted less well, 

 4    and it makes it very likely that undercounts were 

 5    concentrated in the households were these people lived.

 6               MR. ZIMROTH:  Your Honor, I move the admission of 

 7    that chart, 702.

 8               MR. SITCOV:  No objection.

 9               THE COURT:  Received.  

10               (Plaintiffs' Exhibit 702 for identification was 

11    received in evidence)

12         Q.    Now, Professor Ericksen, did the Census Bureau 

13    recognize that it had a problem in certain district offices, 

14    certain type 1 district offices I focus you on, with respect 

15    to the latter stages of nonresponse follow-up?

16         A.    Yes, it did.

17         Q.    Last resort, closeout, nondata defined.

18         A.    Yes, it did.

19         Q.    And what was the nature of the problem?

20         A.    Well, there were reports that enumerators were 

21    just putting down one person in a household for last resort, 

22    regardless of who might live there.  And apparently these 

23    reports were so persistent that the Census Bureau invented a 

24    new phase of the census called Pop 1.

25         Q.    That phase had not been part of the original 
                                                              176

 1    design?

 2         A.    No.  This was a reasonable response to a problem 

 3    that emerged in the taking of the 1990 census but was not 

 4    part of the original census plan.

 5               THE COURT:  What did say it was called?

 6               THE WITNESS:  I don't remember what its official 

 7    name was but everyone refers to it as Pop 1.  And I think 

 8    pop stands for population and 1 indicates there's just one 

 9    person in the household.

10         Q.    One person was enumerated in the household?

11         A.    Right.  And I call it Pop 1.

12         Q.    And I think everybody else does, too.  

13               Will you describe what the Pop 1 program was?

14         A.    Yes.  There were ten district offices, they're 

15    all type 1 district offices I believe, they were located in 

16    places like Chicago where this problem was said to have 

17    occurred.  

18               The Census Bureau was sufficiently concerned 

19    about this problem --

20         Q.    Could you speak into the microphone.  I'm having 

21    a little problem hearing you.

22         A.    Sure.  Do I need to repeat what I just said?

23         Q.    No.  Not for me.

24               THE COURT:  I heard it.

25         A.    The Census Bureau was sufficiently concerned 
                                                              177

 1    about this problem that it expanded the program to another 

 2    14 district offices where the rate of last resort was 

 3    especially high.  In other words, where substantially more 

 4    than the average proportion of households were counted by 

 5    last resort.

 6         Q.    So there were 24 in all?

 7         A.    There were 24 in all, and there were 128,000 

 8    housing units that it went back and checked as part of the 

 9    Pop 1 program.

10         Q.    Could you once again just tell us what the 

11    criteria were for picking these 24 district offices?

12         A.    Well, I prefer to read it from my report.

13         Q.    If you need to refresh your recollection, please 

14    feel free.  And after you've read what you need to refresh 

15    your recollection, perhaps you could just explain it to the 

16    judge.

17         A.    The first ten offices chosen to be included in 

18    the Pop 1 check, had an exceptionally large number of last 

19    resort one person households completed after June 6th with 

20    scheduled end date for nonresponse follow-up.  And I refer 

21    to a memo from the census official named Chuck St. Lawrence.  

22               Another 14 district offices were selected to be 

23    part of the Pop 1 check, because 30 percent or more of the 

24    one person households enumerated were counted as last resort 

25    cases during nonresponse follow-up.  And this refers to the 
                                                              178

 1    same memo.  Ten of the fourteen offices were in New York 

 2    City, the others were in Boston, Baltimore, Washington, 

 3    D.C., and Oakland, California.

 4         Q.    And Chicago I think you mentioned.

 5         A.    The original ten, there are five in Chicago and 

 6    the others were Los Angeles, Washington, Philadelphia, 

 7    Syracuse and Fairfield, New Jersey.

 8         Q.    So the Census Bureau went back to how many 

 9    households, you say?

10         A.    128,000, I believe.

11         Q.    And what did it find?

12         A.    It found that in 44 percent of, and I believe it 

13    was 42 percent of the first group of ten households -- let 

14    me check that just a moment.

15         Q.    I just need the total.  I don't need them 

16    separated by categories.

17         A.    I found that in 44 percent of the 128,000 

18    households it revisited, the number of people living there 

19    was indeed other than one.  Now, in some cases there was 

20    nobody there.  And in other cases there were two or more 

21    people there.

22         Q.    And what was the net effect?

23         A.    On average there were more people living there.

24         Q.    Than had been previously enumerated.

25         A.    Yes.
                                                              179

 1         Q.    How many more?

 2         A.    They found another 56,000 people, so as a result 

 3    of this check the population for those 128,000 housing units 

 4    was increased by 43.7 percent.

 5         Q.    And the Census Bureau added those 56,000 people 

 6    into the count, isn't that right?

 7         A.    Yes, they did.

 8         Q.    In your opinion, what inferences is it reasonable 

 9    to draw from the recheck, let me put it that way, the Pop 1 

10    recheck of these 24 households in these 24 district offices?

11               MR. SITCOV:  Object.  I would like to know what 

12    opinions he has drawn.

13         Q.    What opinions have you drawn?

14         A.    Well, first of all, these enumerators did not 

15    enumerate only one person households.  So you have to 

16    question the accuracy of the other work that they did in 

17    those district offices.  

18               Secondly, you have to question the accuracy of 

19    last resort households that were counted outside the 24 

20    district offices where the study took place.

21         Q.    In your opinion, does it say anything about the 

22    particular enumerators that enumerated the houses that were 

23    found to be erroneous?

24         A.    That's what I think I just said.  You have to 

25    question the work that those enumerators did when they 
                                                              180

 1    counted housing units with other than one person.

 2         Q.    Now, did the Census Bureau share some of your 

 3    concern?

 4         A.    Yes, they did.

 5         Q.    And what did they do, what did it do, I should 

 6    say?

 7         A.    They then undertook a study of all one person 

 8    households that were counted during nonresponse follow-up.  

 9    In other words, they took a sample survey of all the one 

10    person households that had been counted by enumerators.

11         Q.    Throughout the United States?

12         A.    This was done throughout the United States.  The 

13    only part of the population that was left out was the one 

14    person last resort households and these 24 district offices.  

15    So the Census Bureau study included not only last resort one 

16    person households, but also one person households that were 

17    counted before last resort took place.

18         Q.    And what did the study show?

19         A.    In this sample of 1000 households the Census 

20    Bureau found that between 20 and 25 percent of the count was 

21    in error.  That other one actually lived in those 

22    households.

23         Q.    What did they do with that finding?

24         A.    This was only based on a sample of 1000 

25    households.  They simply noted it, they did not change the 
                                                              181

 1    count in any way based on the results of this study.

 2         Q.    Did they give as one of its reasons that it was 

 3    waiting for the results of the PES?

 4         A.    Yes, according to the report that I read 

 5    describing the results in the Pop 1 check, it said that it 

 6    would wait for the more comprehensive results of the 

 7    postenumeration survey to come in to make a full evaluation 

 8    of this problem.

 9         Q.    Would you look at 473, please.  Is this the 

10    description of the Pop 1 study that you referred to?

11         A.    Yes, it is.

12               MR. ZIMROTH:  I offer it into evidence, your 

13    Honor.

14               MR. SITCOV:  Object again, there's no foundation 

15    and secondly, if you look at the note, your Honor, it begins 

16    at the bottom of the first page.  It carries over.  It says 

17    the conclusions and recommendations contained herein 

18    essentially reflect the thoughts of certain staff members at 

19    the time of publication and should not be interpreted as 

20    statements of Census Bureau position.  In light of that I 

21    move to strike the witness' testimony on this matter.

22               MR. ZIMROTH:  Are you saying there was no Pop 1 

23    program?  We asked for a description of the Pop 1 program, 

24    this is what they turned over to us.  This is a description 

25    of the Pop 1 program.  It happened during the 1990 decennial 
                                                              182

 1    census.

 2               MR. SITCOV:  Your Honor, we turned over 500,000 

 3    pages of documents.  Now, I don't know what documents we 

 4    turned over that describe the Pop 1 study.  This one clearly 

 5    says that it is not to be interpreted as statements of 

 6    Census Bureau position and that it is preliminary in nature.

 7               THE COURT:  What do you offer it for, Mr. 

 8    Zimroth?

 9               MR. ZIMROTH:  Two things, your Honor.  First of 

10    all that it is a description, it is the description that we 

11    have, it is the one that has been supplied by the defendants 

12    in this case.  Therefore it is an admission about what the 

13    Pop 1 program was.  

14               Second of all it was dated prior to, this report 

15    is dated prior to July 15, 1991 and therefore we think it 

16    ought to have been taken into account, for whatever 

17    information is in here, in making the decision.

18               MR. SITCOV:  Your Honor, they did not, as I 

19    recall, ask specific for any information about the Pop 1 

20    survey.  So it certainly is no admission on our part of 

21    anything.

22               THE COURT:  Mr. Zimroth, he says you didn't ask 

23    for anything about the Pop 1.

24               MR. ZIMROTH:  We asked for documents related to 

25    the running of the 1990 decennial census because one of the 
                                                              183

 1    important issues in this case is how well or poorly the 1990 

 2    decennial census did especially with respect to minorities 

 3    in certain parts of the country.  This document was part of 

 4    the discovery.

 5               MR. SITCOV:  Your Honor, if they wanted to know 

 6    about Pop 1, they could have deposed somebody about Pop 1.  

 7    Then they would have testimony about it.  

 8               Here we have another example of a document that 

 9    has a clear disclaimer on it, that says it's not the 

10    Bureau's final position, and it's not the Bureau's position 

11    at all.  It's the views of some authors.

12               THE COURT:  Who worked for the Census Bureau.

13               MR. SITCOV:  But it specifically says that the 

14    statements should not be interpreted as statements of Census 

15    Bureau position.

16               THE COURT:  Perhaps it should have been their 

17    position.  I think that's the argument.  This is admissible 

18    on the state of mind it should have engendered within the 

19    Secretary.

20               MR. SITCOV:  We don't know, your Honor, if there 

21    was another final report or not.  And Mr. Zimroth is 

22    certainly not in the position to tell us whether there was 

23    or wasn't.  That's why the authentication problem is 

24    important.  If there was a final report, this is not 

25    reflective of the state of mind of anyone.
                                                              184

 1               MR. ZIMROTH:  If there was one, we should have 

 2    it, we should have gotten it in discovery.

 3               THE COURT:  If anybody should know whether 

 4    there's a final report, it's you, Mr. Sitcov.

 5               MR. SITCOV:  I don't think that's entirely, 

 6    right.  As I mentioned, they have received hundreds of 

 7    thousands of documents from us.  I don't know if they 

 8    received the final report or not, if there is one.  This 

 9    past weekend they asked us for a document or a tape that we 

10    gave them some months ago.  So we have given them things 

11    they didn't know they had.  

12               If they wanted this document to come into 

13    evidence, there would have been a way to authenticate it or 

14    to determine whether in fact what is clearly identified as 

15    the draft was the final.  And they chose not to do that.

16               THE COURT:  I'm going to overrule the objection 

17    and admit Plaintiffs' Exhibit 473 as relevant to state of 

18    mind of the people in chain of command in the Census Bureau.  

19               (Plaintiffs' Exhibit 473 for identification was 

20    received in evidence) 

21         Q.    I want to move on, Professor Ericksen, to the 

22    fourth page, post census day coverage improvement programs?

23               THE COURT:  Can you finish it in seven minutes or 

24    so.  If not we can finish it tomorrow.  I'm running out of 

25    gas.
                                                              185

 1               MR. ZIMROTH:  Me, too.

 2               THE COURT:  I think we've gotten a productive 

 3    day's work under our belts.  Let's adjourn for the day and 

 4    resume tomorrow morning here at ten.

 5               MR. ZIMROTH:  Thank you, your Honor.

 6               MR. MILLET:  Your Honor, Thomas Millet.  It would 

 7    be useful for all concerned, given that we are exchanging 

 8    documents on a witnessness by witness basis as they appear, 

 9    and it will be helpful to us to know what paper we need to 

10    have in the courtroom.  If we have any understanding from 

11    the plaintiffs as to whether Professor Ericksen will testify 

12    all day tomorrow or whether they will move to the next 

13    witness.

14               THE COURT:  That seems to me a reasonable 

15    request.

16               MR. ZIMROTH:  How long are you going to 

17    cross-examine him for?

18               MR. SITCOV:  I like him, so probably days and 

19    days.  I don't think it would be as long as his direct.  But 

20    I can't say for certain.

21               MR. ZIMROTH:  My guess is that his direct and 

22    cross should be finished by the end of the day tomorrow.  

23    But we will caucus and if by the end of the session, and if 

24    we have a different estimate of that we will call you, if 

25    you give us a phone number and let you know.
                                                              186

 1               MR. MILLET:  I would then propose at this point 

 2    that plaintiffs do provide to defendants the exhibits they 

 3    expect to use for Dr. Wolter who is anticipated to be their 

 4    second witness and we would provide to your Honor our 

 5    anticipated cross-examination witnesses for Dr. Wolter.  

 6               (Pause)

 7               MR. ZIMROTH:  If it's clear, your Honor, that we 

 8    will not be finished with this witness, then we will give 

 9    them the documents sometime before the end of the day 

10    tomorrow.

11               THE COURT:  He wants them tonight.

12               MR. ZIMROTH:  If it's unclear one way or the 

13    other we will give them to him first thing in the morning.

14               MR. MILLET:  Just one more point, just so the 

15    court knows, we have previously provided to the plaintiffs 

16    our exhibits ordered by witness.  We had understood that was 

17    the agreement.  Apparently that's not the plaintiffs' view.  

18    We would like to have documents as soon as possible 

19    organized by witness so we can take care of some of these 

20    problems.

21               MR. ZIMROTH:  We will get them Dr. Wolter's 

22    documents as soon as we have them assembled in one place.

23               THE COURT:  You don't have them assembled yet?

24               MR. ZIMROTH:  I don't know exactly where they 

25    are.  Mr. Solomon is in charge --
                                                              187

 1               MR. MILLET:  This will only exacerbate our 

 2    problems with documents as has been demonstrated today.

 3               THE COURT:  I'm sympathetic.  If you can get them 

 4    to him tonight that will be helpful.  

 5               MR. RIFKIND:  They're being copied as we speak.

 6               THE COURT:  Fine.  See you all tomorrow.  

 7               (Trial adjourned to Tuesday, May 12, 1992; 10:00 

 8    a.m.) 

 9    

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25
          