                                                              1132

 1    UNITED STATES DISTRICT COURT
      EASTERN DISTRICT OF NEW YORK
 2    ------------------------------x
      THE CITY OF NEW YORK, et al.,
 3    
                         Plaintiffs,          88 Civ. 3474 (JMcL)
 4    
                 v.                           
 5    
      UNITED STATES DEPARTMENT OF COMMERCE,
 6    et al., 
      
 7                       Defendants.
      ---------------------------------x
 8    CITY OF ATLANTA, et al.,
                     
 9                       Plaintiffs,
      
10                v.                          92 Civ. 1566 (JMcL)
                                              
11    MOSBACHER, et al.,
      
12                       Defendants.
      ---------------------------------x
13    FLORIDA HOUSE OF REPRESENTATIVES, 
      et al.
14    
                         Plaintiffs,
15                v.                          92 Civ. 2037 (JMcL)
      
16    BARBARA H. FRANKLIN, Secretary of
      Commerce,
17    
                         Defendant.
18    --------------------------------x
      
19    
                                              May 19, 1992
20                                            9:30 a.m.
      
21    
      
22    Before:
      
23               HON. JOSEPH M. McLAUGHLIN,
      
24                                            Circuit Judge
      
25
                                                              1133

 1    FRANKLIN M. FISHER,               resumed.   

 2    CROSS-EXAMINATION (Continued) 

 3    BY MR. ZITCOV:

 4         Q.    It is the case, isn't it, Dr. Fisher, that you 

 5    had not committed any of your analyses to writing as of the 

 6    time of your deposition in this case in April, is that 

 7    right?

 8         A.    That's true.

 9         Q.    And you didn't --  

10         A.    Except for the notes I told you about later in 

11    the deposition.

12         Q.    And at the time of your deposition, you didn't 

13    intend to commit any of your analyses to writing?

14         A.    Well, what you asked me about, Mr. Sitcov, was 

15    the analyses I had then done and I did not intend to commit 

16    them to writing and I haven't.

17         Q.    And I asked you if you were aware whether other 

18    people who might have been retained as experts by the 

19    plaintiffs had done any analyses that you might be relying 

20    on, and you were directed not to respond.

21               Do you recall that?

22         A.    If you will read through to what you, in fact, 

23    ended up asking me, you asked me whether anyone who was 

24    going to be identified to you as an expert or called by the 

25    plaintiffs to testify had done that, and I said I didn't 
                                                              1134

 1    know.

 2         Q.    Do you recall --

 3         A.    Read to the end of the colloquy and the question 

 4    you asked then.

 5         Q.    Do you recall if I asked you on page 15"

 6               "Q.    Are you personally aware whether any other 

 7    person who has been, who you believe may be an expert for 

 8    the plaintiffs in this case, if any such person has done any 

 9    independent analysis of the census or the PES?" 

10               Do you recall Ms. Goldstein, who is sitting at 

11    counsel table, saying, "I'm going to first of all object to 

12    the question as somewhat vague, but second of all, instruct 

13    the witness based on their interpretation that defendants' 

14    counsel had given to work product privilege and instruct Dr. 

15    Fisher not to respond to any information that came in with 

16    them or the net which he he plans to testify about, but any 

17    other work that is being done that he may or may not know 

18    about at the request of counsel by others Dr. Fisher should 

19    not testify about it if he knows about it." 

20               Do you recall that colloquy?

21         A.    I do recall that, but I think you ought to go on 

22    and read the rest of it down through the question that you 

23    repeated and that I answered.  

24               THE COURT:  I think I heard this.  Move on to 

25    something else.  
                                                              1135

 1               MR. ZITCOV:  Okay.

 2         Q.    Dr. Fisher, would you turn to Plaintiff's Exhibit 

 3    684.

 4         A.    Yes, sir.

 5         Q.    These computations had not been done at the time 

 6    of your deposition, had they?

 7         A.    That's true.

 8         Q.    Would you look at the last couple of lines under 

 9    PRODSE.

10               Do you see that?

11         A.    I do.

12         Q.    And you don't know what data files STPES-01.DAT 

13    through STPES-56 DAT residual dot, and so forth, you don't 

14    know what those data files contain, do you?

15         A.    You asked me those yesterday if that was 

16    identification of data files with someone has the census 

17    data and tapes can find out, but I do not personally know 

18    what they refer to.

19         Q.    And you did not personally do the computations?

20         A.    No.

21         Q.    And you did not personally supervise them being 

22    done?

23         A.    That's not true.  I supervised them in the way I 

24    did it yesterday.

25         Q.    You did not check the computations?
                                                              1136

 1         A.    I had them put through the standard checking 

 2    procedure that is done under my supervision at the firm that 

 3    did this.

 4         Q.    But you didn't personally check it?

 5         A.    I did not personally check it.

 6         Q.    Could you turn to 685.

 7               This chart didn't exist at the time of your 

 8    deposition?  

 9               THE COURT:  Wait.  685 is identical to 684. 

10               THE WITNESS:  No, your Honor, it's not.  One is 

11    done with the PRODSE method, the other one is done with the 

12    up UNWEPERS.

13         Q.    This chart did not exist at the time of your 

14    deposition, did it?

15         A.    No, it did not.

16         Q.    And these computations were done as a result of a 

17    request you made after your deposition, isn't that right?

18         A.    Yes.

19         Q.    The same with 684, isn't it, that those 

20    computations were done as a result of a request you made 

21    after your deposition?

22         A.    Well, I had --  there had been some discussion 

23    for both of these whether this could be done.  They had not 

24    been done at the time of my deposition.

25         Q.    And you asked that they be done after your 
                                                              1137

 1    deposition?

 2         A.    I certainly asked that they be done, yes.

 3         Q.    And, again, looking at the source line, you don't 

 4    know what any of those data files are, do you?

 5         A.    The same set of data files --  actually, not 

 6    quite, but the answer is the same.

 7         Q.    And you didn't personally do any of the 

 8    computations on this chart, did you?

 9         A.    That's true.

10         Q.    Turn to 686, please.

11               This is another chart that didn't exist at the 

12    time of your deposition, isn't that right?

13         A.    Yes.

14         Q.    This is another chart that was created at your 

15    request after the your deposition, isn't it?

16         A.    Yes.

17         Q.    If you look at the source line of the PRODSE, 

18    there are some data files there, isn't that right?

19         A.    Yes.

20         Q.    You don't know what those data files are, do you?

21         A.    Not with any specificity.

22         Q.    And you didn't personally do the computations on 

23    this, did you?

24         A.    That's true.

25         Q.    Can you look at the note, please, under the last 
                                                              1138

 1    line of data?

 2         A.    Yes.

 3         Q.    The note says, "The expected number of states 

 4    with improved accuracy is 40"?

 5         A.    Yes.

 6         Q.    Do you see that?

 7         A.    Yes.

 8         Q.    And this is a chart that shows how many states 

 9    have or the expected number of states whose distributive 

10    share would be improved with adjustment, is that right?

11         A.    Yes, among other things.

12         Q.    Do you recall when I took your deposition on 

13    April 16, I asked you if you knew what the expected number 

14    of states will have a distributive share that is made more 

15    accurate by adjustment, and you recall that you responded, 

16    "I don't know"?

17         A.    That's true, and I didn't know then.

18         Q.    Okay.

19               So sometime after your deposition, you had 

20    computations done that would give you that answer, is that 

21    right?

22         A.    Yes, that's correct.

23         Q.    And that was after your deposition?

24         A.    Yes.

25         Q.    Would you turn to Exhibit 687, please.
                                                              1139

 1               Now, this is another chart that didn't exist at 

 2    the time of your deposition, isn't that right?

 3         A.    Yes.

 4         Q.    And this is another chart that was done at your 

 5    request after your deposition?

 6         A.    Yes.

 7         Q.    And you didn't personally do the computations on 

 8    this, did you?

 9         A.    No.

10         Q.    You don't know what the source files are, do you?

11         A.    Not with any specificity.

12         Q.    And this is another chart that proposes to 

13    explain expected number of states with improved accuracy, 

14    that is, distributional accuracy based on adjustment, is 

15    that right?

16         A.    Yes, among other things.

17         Q.    Pardon me?

18         A.    Among other things.

19         Q.    Okay.  And, again, you didn't know at your 

20    deposition the number of states that would have a 

21    distributive share made more accurate by adjustment, did 

22    you?

23         A.    You may have misspoke.  I didn't know that then.  

24    Nobody knows that now.  But what this chart shows is the 

25    expected number of states.
                                                              1140

 1         Q.    And you didn't know the expected number of states 

 2    at that time, did you?

 3         A.    No.

 4         Q.    Would you turn to Plaintiff's Exhibit 690.

 5               This chart didn't exist at the time of your 

 6    deposition, did it?

 7         A.    Not so far as I know.

 8         Q.    And this computation, the computations that this 

 9    chart is based on hadn't been done as of the time of your 

10    deposition, had they?

11         A.    I did not know --  I don't know when these were 

12    done, but I did not know of them at the time of my 

13    deposition.

14         Q.    Okay.

15               What is the source of these data?

16         A.    The data, the source of these data are given, in 

17    fact, the data themselves are given on another exhibit 

18    supplied to you which I believe is Plaintiff's Exhibit 694, 

19    but speaking more simply, what this --  in order to produce 

20    this chart, what is required is to know the adjusted count 

21    by state, the census count by state, and that's all that is 

22    required.

23         Q.    And this --

24         A.    You could put this together in less than an hour.

25         Q.    Okay.  Let's look at Plaintiff's Exhibit 694.
                                                              1141

 1               You didn't do these computations, did you?

 2         A.    No.

 3         Q.    And the source there are the same files that were 

 4    the sources of 684 to 687, isn't that right?

 5         A.    Yes.

 6         Q.    And you don't know --

 7         A.    I'm sorry.  You need one more --  I said 

 8    previously that you need the adjusted count, the census 

 9    count for each state, that's true, but you also need the 

10    percent minority as given by the census for each state.

11         Q.    And you don't know, I think we have established, 

12    what are in those data files, do you?

13         A.    Well, it's hard to imagine what is not in those 

14    data files are the numbers that are here.

15         Q.    But you don't know that?

16         A.    No, I haven't looked specifically at the data 

17    files.

18         Q.    Would you turn to 692, please.

19         A.    Okay.

20         Q.    That is another chart that was prepared after 

21    your deposition, isn't it?

22         A.    Yes.

23         Q.    And you didn't prepare this chart yourself, did 

24    you?

25         A.    No.  It was prepared by Charles River Associates 
                                                              1142

 1    under my direction, as I told you yesterday, when I 

 2    testified yesterday.

 3         Q.    You didn't do the computations yourself, did you?

 4         A.    No.

 5         Q.    And this also refers to Plaintiff's Exhibit 694 

 6    as the source, is that right?

 7         A.    Yes.  The other one probably --  does the other 

 8    one refer to 694 or 695?

 9         Q.    It doesn't refer to any.

10         A.    Just a minute.

11               Oh, this is the same chart.  The difference is 

12    that the source has been added.  It is 694.

13         Q.    Okay.  Then the same answers to my questions 

14    about whether or not you knew what was in the source 

15    materials in 694 would apply here?

16         A.    Yes.  It's not the same chart, but anyway, yes, 

17    the same answers.

18         Q.    Will you turn to 693, please.

19         A.    Okay.

20         Q.    This chart didn't exist at the time of your 

21    deposition, did it?

22         A.    No.

23         Q.    You didn't do the computations that this chart is 

24    based on yourself, did you?

25         A.    No.
                                                              1143

 1         Q.    And this, the source for this is Plaintiff's 

 2    Exhibit 695.

 3               Will you turn to 695.

 4         A.    Okay.

 5         Q.    You didn't do the computations in this chart, did 

 6    you?

 7         A.    No.

 8         Q.    And you don't --

 9         A.    Mr. Sitcov, it has not been my practice for many 

10    years to literally do the computations myself for anything.

11         Q.    Okay.  And you don't know what the files that are 

12    identified as the sources of 695 are, do you?

13         A.    Not specifically.

14         Q.    I think we can now skip over 694 and 695.

15               Could you turn to 696, please.

16               Do you have it?

17         A.    Yes.  So long as you keep going in order it's 

18    easy.

19         Q.    Okay.

20               You don't know what the files are that are 

21    identified in the source?

22         A.    Not with any specificity.

23         Q.    Okay.

24               This chart refers to a standard deviation for an 

25    observed loss function difference, does it not?
                                                              1144

 1         A.    Yes.

 2         Q.    And that standard deviation can also be called 

 3    standard error, can't it?

 4         A.    Yes.

 5         Q.    The standard deviations is computed from a list 

 6    of numbers, isn't it?

 7         A.    It certainly can be.  It doesn't have to be.

 8         Q.    What is the list of numbers that this standard 

 9    deviation was computed from?

10         A.    That is essentially the question you asked me at 

11    the end of yesterday just before we recessed as to how this 

12    is calculated and I will be glad to tell you.

13         Q.    What I want to know is what is the list of 

14    numbers?

15         A.    No, no.  You will get that answer.  It's okay.  

16    Calm yourself, I am answering the question.

17         Q.    I am quite happy to hear whatever else you have 

18    to say, but I want to know what the list of numbers is?

19         A.    I understand.  What I am saying is that those two 

20    answers are essentially the same.

21         Q.    Okay.

22         A.    What the census did for the final loss functions 

23    a thousand simulations.

24               One takes the --  one takes for each simulation 

25    of the target the value of the --  the difference for the 
                                                              1145

 1    squared loss function, if I -- this is for the squared --  

 2    yes, this is for the squared.

 3               One takes for each simulation of the target the 

 4    difference of the squared error loss functions and produces 

 5    --  that produces a thousand numbers for that.  That is the 

 6    list of numbers from which this is calculated.

 7         Q.    Did you do the calculations that resulted in the 

 8    data in this chart yourself?

 9         A.    No.

10         Q.    Will you turn to 697, please.

11               Would you look at the source line?

12         A.    Yes.

13         Q.    It is correct that you don't know what those data 

14    files are?

15         A.    Not with any specificity.

16         Q.    This is another document that came into existence 

17    after I took your deposition, isn't it?

18         A.    Yes.

19         Q.    And you didn't do these computations yourself, 

20    did you?

21         A.    That's true.

22         Q.    For the last exhibit and this one, that is, 696 

23    and 697, were the standard deviations computed by the 

24    bootstrap method?

25         A.    In effect, yes.  The bootstrap method, I 
                                                              1146

 1    described to you the bootstrap method in which this is done 

 2    from the internal simulations.

 3         Q.    And it is your opinion, is it not, that the 

 4    bootstrap method is not the preferred method for 

 5    calculations when you are using the squared error loss, 

 6    isn't that right?

 7         A.    I believe that I --  yes, I said that, but I said 

 8    that in the context of calculating not this variance, but a 

 9    different variance.

10               I don't believe you could possibly calculate this 

11    variance any other way.

12         Q.    Will you turn to 697, please.

13               Do you have it?

14         A.    Yes.

15         Q.    You don't know what the source documents are in 

16    this exhibit, do you?

17         A.    They are the files listed at the bottom, and I 

18    don't know with specificity what those files contain.

19         Q.    This is an exist that was created after your 

20    deposition, isn't it?

21         A.    Yes.

22         Q.    You didn't do the computations that are 

23    represented in this exhibit yourself, did you?

24         A.    That's true.

25         Q.    Turn to 699, please.
                                                              1147

 1               It is also the case that you don't know what the 

 2    source documents are for this exhibit, isn't that right?

 3         A.    They are the source or the files listed in the 

 4    bottom and I do not know with specificity what those are.

 5         Q.    This is another --

 6         A.    What's in them, rather.

 7         Q.    This was another document that was created after 

 8    your deposition, isn't that right?

 9         A.    Yes.

10         Q.    And you didn't do these computations yourself, 

11    did you?

12         A.    No.

13         Q.    Turn to 700.

14               Again, you don't know what --

15         A.    If you are going to ask me the same set of 

16    questions about this you are going to get the same set of 

17    answers.  Is that okay?

18         Q.    Absolutely.  

19               THE COURT:  Maybe you could lump them all into 

20    one question.

21         Q.    All right.  Let me ask you to look at 700, 701 --  

22    700 and 701.  

23               THE COURT:  You should have done this sooner.

24         Q.    You don't know what the source files are for 

25    either of these exhibits, do you?
                                                              1148

 1         A.    Not with --  I don't know what's in those source 

 2    files with any specificity.

 3         Q.    And you didn't do any of the computations that 

 4    are reflected in these exhibits, did you?

 5         A.    That's true.

 6         Q.    And these exhibits weren't done until after your 

 7    deposition, isn't that right?

 8         A.    That's also true.

 9         Q.    Will you turn to 717, please.

10         A.    Okay.

11         Q.    This document didn't exist at the time of your 

12    deposition, did it?

13         A.    That's true.

14         Q.    You didn't do the computations that are reflected 

15    here, did you?

16         A.    Not personally.

17         Q.    You don't know what the data files are that are 

18    listed on for the source, do you?

19         A.    I don't know what's in them with any specificity, 

20    no.  

21               MR. ZITCOV:  Your Honor, at this time we are 

22    going to move to have excluded from evidence the files that 

23    I have just asked Dr. Fisher about and have his testimony 

24    about them struck as I mentioned yesterday.  

25               THE COURT:  Let me reserve on the motion to 
                                                              1149

 1    strike until I have heard the redirect.  

 2    BY MR. ZITCOV:

 3         Q.    Dr. Fisher, do you know if the Bureau's loss 

 4    function analysis depends on the synthetic assumption?

 5         A.    Indirectly it does.

 6               You asked me at my deposition about the synthetic 

 7    assumption and I couldn't remember it by name.

 8         Q.    And now you remember it?

 9         A.    Yes.  I asked you to tell me what it was you had 

10    in mind and you wouldn't, but it was something I was 

11    familiar with although I had forgotten the name.

12         Q.    Would you look at Plaintiff's Exhibit 698, 

13    please.

14               On 698, there is an asterisk and it says, 

15    "Probability that the observed loss function difference in 

16    favor of adjustment is due to chance if the census is really 

17    more accurate is less than one chance in 150 million." 

18               Do you see that?

19         A.    Yes.

20         Q.    Is there only one chance in 150 million of the 

21    failure in the synthetic assumption?

22         A.    No, that's not what is being measured here.

23         Q.    All right.

24               Now, the Bureau's loss function analysis depends 

25    on the co-variance matrix for the smooth adjustment factors, 
                                                              1150

 1    doesn't it?

 2         A.    Yes.

 3         Q.    Is there only one chance in 150 million that the 

 4    co-variance matrix needs to be doubled to be an unbiased 

 5    estimate of the true co-variance matrix?

 6         A.    No.  But I can assure you that if you doubled the 

 7    matrix, you are not going to change the results of the 

 8    hypothesis test except for the numbering involved.

 9         Q.    The Bureau loss function analysis depends on the 

10    total error model, doesn't it?

11         A.    Yes.

12         Q.    And you had reviewed that model, hadn't you?

13         A.    Not in any detail.

14         Q.    Is there only one chance in 150 million of a 

15    failure in the total error model?

16         A.    No, but I can tell you again that the results of 

17    these hypotheses tests are going to be quite robust against 

18    various forms of failure in the total error model.  You have 

19    to increase the variance in those models quite a lot before 

20    you fail to reject the null hypothesis.

21         Q.    The Bureau's loss function analysis depends on 

22    sharing the biases down from the 13 evaluation post-strata 

23    to the 1,392 post-strata, isn't that right?

24         A.    Yes.

25         Q.    When I took your deposition on April 16, you 
                                                              1151

 1    hadn't reviewed the methods used by the Bureau to do that 

 2    sharing down, had you?

 3         A.    You asked me that yesterday and I pointed out to 

 4    you that I had told you what I thought they were and by and 

 5    large I was right.

 6         Q.    Is there only one chance in 150 million of a 

 7    failure in the Bureau's procedure for allocating bias?

 8         A.    No.  But the results are really very robust to at 

 9    least two fairly different assumptions as to how that is 

10    done.

11         Q.    You are aware, aren't you, that the Bureau 

12    conducted a preliminary loss function analysis?

13         A.    Yes.

14         Q.    You didn't testify about that analysis yesterday, 

15    did you?

16         A.    No.

17         Q.    You are aware, aren't you, that Bruce Spencer 

18    advised the Census Bureau on the conduct of the loss 

19    function --  of that loss function analysis, aren't you?

20         A.    I don't know that I am specifically aware of 

21    that, but it certainly wouldn't surprise me.  I know he 

22    advised them about loss function generally.  I don't 

23    remember that specific one.

24         Q.    Could you turn to Exhibit 686.

25               Yesterday you directed me to an article by 
                                                              1152

 1    Spencer and Mulry for information about the formulas used by 

 2    Charles River Associates to compute the probabilities on 

 3    this chart.

 4         A.    That's not quite right.  I directed you to a memo 

 5    from Spencer to Mulry produced by the Census Bureau, I 

 6    believe, in connection with the Bounpane deposition.

 7         Q.    And that is the same Bruce Spencer who is the 

 8    co-author --  I'm sorry --  that is the same Bruce Spencer 

 9    who advised the Bureau on conducting the loss function 

10    analyses?

11         A.    Yes.

12         Q.    You are aware, aren't you, that the preliminary 

13    loss function analysis contained a certain mathematical 

14    blunder, is that right?

15         A.    It contained something that certainly needed to 

16    be corrected, yes.

17         Q.    The loss function used a bias estimate of the 

18    census loss, didn't it?

19         A.    That's one way of putting it.

20               Let me put it precisely.

21               The loss function included, the preliminary loss 

22    function attributed to the census a piece of --  a loss 

23    which was too big by a knowable factor or factor that you 

24    could get to cancel out.

25               I'm sorry, by a knowable factor.
                                                              1153

 1         Q.    And the updated loss function analysis corrected 

 2    that mistake, isn't that right?

 3         A.    Yes.

 4         Q.    The preliminary loss function analysis used as 

 5    the loss estimate defined --  well, you have Plaintiff's 

 6    Exhibit 42 there?

 7         A.    Probably.  Hang on.

 8               (Pause)

 9         Q.    Will you turn to page 3 of Plaintiff's 

10    Exhibit 42.

11         A.    Okay.

12         Q.    That shows, doesn't it, that the preliminary loss 

13    function analysis uses the loss estimate defined at the top 

14    of page 3 of Plaintiff's Exhibit 42, is that right?

15         A.    Probably.

16               Hang on.  I need to check it.

17               (Pause)

18               One of us is confused, possibly me.

19               I think you are confusing what is known as the 

20    preliminary loss function analysis with what is in this 

21    document, which talks about the preliminary results of the 

22    loss function analysis.  That's not the same thing.  This is 

23    the final loss function.

24         Q.    Look at page 2.

25               Do you see the heading on page 2, it says, "Loss 
                                                              1154

 1    function, preliminary estimates"?

 2         A.    Yes.  But that's the point.  This is not the 

 3    preliminary --  this is not what we refer to the thing that 

 4    has the error in it referred to as the preliminary loss 

 5    function analysis, that is a much earlier document.  These 

 6    are preliminary estimates of the final loss function.

 7         Q.    So is it your testimony that this formula that 

 8    appears at the top of page 3 is not the formula that was 

 9    used in the preliminary loss function analysis?

10         A.    It's related to it and I would have to check this 

11    against the preliminary loss function documents, but I am --  

12    I'm sorry.  As I said, one of us is confused and it may be 

13    me.

14               The correction is discussed at page 5.

15         Q.    Okay.  So this is the preliminary loss function 

16    analysis?

17         A.    It certainly begins that way, yes.

18         Q.    Okay.

19               (Continued on the next page) 

20    

21    

22    

23    

24    

25    
                                                              1155

 1         Q.    Now, the preliminary loss function analysis uses 

 2    a loss estimate that is defined at the top of page 3 of 

 3    Plaintiff's Exhibit 42, is that right?

 4         A.    Yes.  It starts that way.

 5         Q.    Could you turn to Plaintiff's Exhibit 653.

 6         A.    Do you want me to keep this one out at the same 

 7    time?

 8         Q.    Yes, please.

 9         A.    Okay.

10         Q.    Do you see the line at the top of page 7 of 

11    Plaintiff's Exhibit 653?

12         A.    Yes.

13         Q.    That is a formula for census loss, isn't it?

14         A.    Yes.

15         Q.    That is a norm la I asked you about yesterday, 

16    isn't it?

17         A.    Yes.

18         Q.    You testified yesterday that nothing could 

19    possibly go wrong with that formula, didn't you?

20         A.    What I testified yesterday was that that formula 

21    as stated here is a definition as to what the author is 

22    going to mean by the use of the term "LCI."  If you are 

23    going to take that to mean that nothing could possibly go 

24    wrong with the use of that formula as it goes on further in 

25    the analysis, I didn't testify to that.  That would be 
                                                              1156

 1    silly.

 2         Q.    You testified yesterday, didn't you, that there 

 3    are 13 evaluation poststrata?

 4         A.    I think that's right.

 5         Q.    The total error model quantifies biases for the 

 6    evaluation poststrata?

 7         A.    Yes.

 8         Q.    You testified yesterday about the PRODSE and 

 9    UNWEPRS method of sharing down the biases from the 

10    evaluation poststrata?

11         A.    Yes.  

12         Q.    You are aware, aren't you, that the census bureau 

13    spread the biases down from the 13 evaluation poststrata to 

14    the 156 age and sex groups which comprised them?

15         A.    Yes, I think that's right.

16         Q.    Suppose one of the defendants' experts developed 

17    an alternative method of sharing the biases down which 

18    assumed independence across the 12 age/sex groups comprising 

19    each of the evaluation poststrata.  Would that be a 

20    reasonable alternative imputation procedure?

21         A.    I don't know what it means.  What do you mean by 

22    independent.

23         Q.    Independence of errors in the nonsampling errors 

24    in the PES.

25         A.    Say it again?
                                                              1157

 1         Q.    What I mean by "independence" is independence of 

 2    the errors in the nonsampling errors in the PES.

 3         A.    I'm sorry.  I'm really having difficulty 

 4    following.  Could you just say it slowly?  The preposition 

 5    in the middle is escaping me.  The independence of?

 6         Q.    Independence of errors or independence in the 

 7    nonsampling errors.

 8         A.    You have to say it slowly.  It's the two little 

 9    words I am not hearing.

10         Q.    Independence.

11         A.    Yes.

12         Q.    Of.

13         A.    Yes.

14         Q.    Errors in the nonsampling errors in the PES.

15         A.    I don't understand what that means.  It probably 

16    means something, but I don't think you are saying it right.

17         Q.    Did the census bureau assume independence of 

18    errors across the 12 age/sex groups comprising each of the 

19    evaluation poststrata?

20         A.    I don't know.  At the level we are talking, I'm 

21    not sure what it means to do that.

22         Q.    Do you recall testifying at your deposition that 

23    you didn't know if that was what the census bureau did?

24         A.    That is what I just said.

25         Q.    You are aware, aren't you, that there are 39,000 
                                                              1158

 1    state and local governments in the United States?

 2         A.    No.  I'm under oath.  I cannot say that I am 

 3    actually aware that.

 4         Q.    Do you know that that is about the number?

 5         A.    No.  I know there are a lot.

 6         Q.    Does that sound about right?

 7         A.    39,000 comes within my definition of a lot.

 8         Q.    You are aware, aren't you, that there are only 

 9    5,000 block clusters in the post-enumeration survey?

10         A.    That sounds right.

11         Q.    So the bureau collected data in at most 5,000 of 

12    those 39,000 jurisdictions, is that right? 

13         A.    I'm not sure that follows, because it is not 

14    totally clear to me that a block cluster couldn't spread 

15    across jurisdictions.

16         Q.    There are at least some jurisdictions in which 

17    the PES collected no data whatsoever, isn't that right?

18         A.    Yes, I'm sure that's true.

19         Q.    Would you say it is at least 20,000?

20         A.    Oh, I'm not going to hazard.  I'm sure there are 

21    what I would still say a lot.

22         Q.    Do you recall that I asked you yesterday about a 

23    hypothetical survey to estimate the average height of the 

24    lawyers in every law firm in New York?  Do you recall that?

25         A.    Yes, I do.
                                                              1159

 1         Q.    You testified that you would need to take a 

 2    sample from each one of those law firms to estimate the 

 3    height by firm?

 4         A.    Yes, except unless you were willing to assume 

 5    that there was something about law firms in New York that 

 6    they had in common as regards the height of lawyers.

 7         Q.    If you are unwilling to do that, then you would 

 8    have to take a sample from every single law firm in order to 

 9    estimate the height by law firm?

10         A.    If you thought that every law firm was likely to 

11    be really different from every other law firm -- I'm not 

12    making jokes now -- in terms of the heights of their 

13    lawyers, and that was in some sense systematic, then you 

14    would have to do that.  That would be to assume that knowing 

15    the heights of lawyers in some of the law firms gives you no 

16    information about the heights of lawyers in other law firms.  

17    I don't think that is a very good assumption.

18         Q.    You could estimate the undercount for every one 

19    of the substantial number of jurisdictions in the U.S. using 

20    data from 5,000 block clusters, is that your opinion?

21         A.    Yes.

22         Q.    You can tell us how many of those jurisdictions 

23    can be expected to be improved by adjustment, is that right?

24         A.    Yes.  

25               MR. SITCOV:  I have no further questions, your 
                                                              1160

 1    Honor.  

 2               THE COURT:  Redirect?  

 3               MR. RIFKIND:  May I consult for just a moment, 

 4    your Honor?  

 5               THE COURT:  Sure.  

 6    REDIRECT EXAMINATION 

 7    BY MR. RIFKIND:

 8         Q.    Dr. Fisher, the large number of charts that Mr. 

 9    Sitcov just took you through showing various data sources at 

10    the bottom where it says "Source," do you have an 

11    understanding as to where those source documents or source 

12    files, source tapes, come from?

13         A.    Of course.  They come from the Bureau of the 

14    Census.

15         Q.    Do you know when CRA was able to obtain those or 

16    when the plaintiffs were able to obtain -- 

17               THE COURT:  Who is CRA?  

18               MR. RIFKIND:  Charles River Associates.  I'm 

19    sorry.

20         Q.    Do you know when the plaintiffs or those 

21    assisting them were able to obtain those data files from the 

22    census bureau?  

23               MR. SITCOV:  Your Honor, unless he knows from his 

24    personal knowledge, the only way he can be answering is 

25    hearsay, and I don't think that would be appropriate here.  
                                                              1161

 1    That is another reason why this stuff ought to be excluded.  

 2               THE COURT:  Let's find out whether he has 

 3    personal knowledge.

 4         A.    You have to explain to me what constitutes 

 5    personal knowledge.  If you mean was I there when they were 

 6    delivered in the mail room, the answer is no.  If you mean 

 7    do I know when work was able to begin, I have some idea 

 8    about that.  That, of course, depends on somebody telling 

 9    me, yes we can now do it.  

10               MR. SITCOV:  Then I renew my objection now.

11         Q.    Was there a time when you were advised that the 

12    data was unavailable?

13         A.    Yes.

14         Q.    Mr. Sitcov, I called your attention yesterday to 

15    Plaintiff's Exhibit 2 at page 282, the National Academy 

16    panel.  Do you remember that?

17         A.    I do.

18         Q.    I ask you to refer now to page 282.  Do you see 

19    word at the bottom of that page, "Recommendation 7.2"?

20         A.    I do.

21         Q.    Would you read that to the Court.

22         A.    "Recommendation 7.2.  In measuring the total loss 

23    associated with an adjustment procedure, we recommend that 

24    the contribution to this loss attributable to a geographic 

25    region should reflect its population size.  Thus, we 
                                                              1162

 1    recommend against loss functions based solely on the number 

 2    of political entities losing or gaining through adjustment."

 3         Q.    Focusing on the first sentence of that 

 4    two-sentence recommendation, what happens, if you know, when 

 5    you measure total loss taking account of the population size 

 6    of the geographic region?

 7         A.    There are at least two ways to do that.  I do 

 8    know.  There are at least two ways to do that.  One way, in 

 9    effect, I described yesterday.  If you wish to do it in 

10    terms of what fraction of the population lives -- actually, 

11    there are at least three ways to do it.  Nobody ever expects 

12    the Spanish Inquisition, if I may quote the classics.  

13               If you ask what fraction of the population lives 

14    in those states which are more likely than not to have their 

15    distributive accuracy improved through adjustment, it is 

16    approximately 90 percent.

17         Q.    That is what you testified to yesterday, is that 

18    right?

19         A.    I did.  The second one I also testified to 

20    yesterday.  If you weight the states by population and then 

21    multiply by probabilities, as it were, getting something, it 

22    doesn't have a good name, but something like the expected 

23    size of the population living in areas that will be made 

24    more accurate by adjustment, that answer is on the order of 

25    85 percent.  
                                                              1163

 1               The third way, not testified yesterday, if you 

 2    take the squared error loss function for distributive shares 

 3    and weight each state by its population, something which, as 

 4    I recall, the National Academy panel actually recommended, 

 5    you get an answer which is even more favorable to adjustment 

 6    than doing it unweighted.

 7         Q.    So is it your testimony doing what the panel is 

 8    here recommending comes out more favorably for adjustment?

 9         A.    That's true.

10         Q.    Focus on the second sentence in that 

11    recommendation.  "We recommend against loss functions based 

12    solely on the number of political entities losing or gaining 

13    through adjustment."  

14               Is that, in substance, the standard or test that 

15    the Secretary was using in the portions of his opinion you 

16    were referring to yesterday?  

17               MR. SITCOV:  I object to the form.  He can ask if 

18    he thinks it is, but he can't ask what the Secretary was 

19    doing.

20         Q.    In your view, is that what the Secretary was 

21    doing?

22         A.    Yes.

23         Q.    In your view, the Secretary was doing what the 

24    panel was recommending against doing?

25         A.    That is true.
                                                              1164

 1         Q.    Putting aside the panel's recommendation and 

 2    putting aside your own view as I take it you expressed 

 3    yesterday that what the Secretary was doing was not a 

 4    rational thing to do, accepting the test the Secretary was 

 5    using, does that favor adjustment or does it favor the 

 6    original enumeration?

 7         A.    It clearly favors adjustment.  That is the 

 8    purpose of calculating the expected number of states that 

 9    will be made better off -- I'm sorry -- will have the 

10    distributive shares made more accurate through adjustment.  

11    The answer to that comes out 40 to 11 in favor of 

12    adjustment.

13         Q.    Is that conclusion based on the standard that the 

14    Secretary's report indicates he was using?

15         A.    Yes.

16         Q.    Robust to variance?

17         A.    It certainly is.  As I testified yesterday, if 

18    you double the variance of the total error model, the answer 

19    goes from 40-to-11 to 35-to-16, still in favor of 

20    adjustment.  For this, 25-1/2 is the right break point.  In 

21    order to tip the scale in favor of the original enumeration, 

22    you have to increase the variance by a factor of of between 

23    7 and 8, far larger than the bureau or, indeed, the 

24    Secretary suggested was a reasonable range.

25         Q.    You took issue yesterday with the way the 
                                                              1165

 1    Secretary calculated the standard he was applying, is that 

 2    correct?

 3         A.    Yes, I did.

 4         Q.    Assuming nonetheless that one accepted the 

 5    Secretary's test and calculated it the way the Secretary 

 6    calculated it, does the outcome favor adjustment or does it 

 7    favor the original enumeration?

 8         A.    It favors adjustment.  You come out with 21 

 9    states whose distributive accuracy will be made worse and 29 

10    states -- I'm counting District of Columbia as a state here 

11    -- 29 states in which it will be made better.  It is not 

12    close, because the appropriate break point for that is not 

13    25 but on the order of 34.

14         Q.    Is that conclusion, that is, accepting the 

15    Secretary's test and calculating it the way the Secretary 

16    calculated it, also robust to changes in variance?

17         A.    It is.  If you double the variance, which is what 

18    the Secretary does in his report, you come out with, I 

19    believe it is, 29 states whose distributive accuracy will be 

20    made worse by adjustment.  But if you understand what you 

21    are doing and you understand what the bureau told you, at 

22    least twice, and if the Secretary had understood that, let's 

23    put it that way, then the Secretary would have realized that 

24    29 is not, in fact, a vote in favor of the original 

25    enumeration.  29 is still well below the break point of 34.  
                                                              1166

 1    That still favors adjustment.  

 2               Look, it is the case that if you use this 

 3    standard and you use it right, there really isn't any doubt.  

 4    If you use the standard and you use it wrong, calculating 

 5    the way the Secretary did, and you take the trouble to 

 6    understand what it is you have done, it is still in favor of 

 7    adjustment.  

 8               MR. RIFKIND:  I have no further questions.  

 9               THE COURT:  Before you leave, I need a little 

10    help.  On the motion which is still sub judice on the 

11    government's part to strike documents and testimony, that 

12    analytically is a motion for a sanction.  Accordingly, the 

13    rules of evidence don't apply to it.  For that reason, I 

14    would ask you to explore in a little more detail than you 

15    did why the underlying data were not available at the time 

16    of the deposition and became available thereafter, leading 

17    to the creation of the documents thereafter.  

18               MR. RIFKIND:  Your Honor, if I may, there is a 

19    somewhat complex story which I am not sure that Professor 

20    Fisher can genuinely speak to.  If you would like, if it 

21    will help in resolving this issue, I can assemble that story 

22    in the form of an affidavit or memorandum and submit it.  

23               THE COURT:  All right.  Why don't you do that 

24    this evening or something.  But get from Dr. Fisher now 

25    whatever it is he can contribute, if anything.  
                                                              1167

 1               MR. RIFKIND:  Let me just consult for a second.  

 2               MR. SITCOV:  Your Honor, will we have an 

 3    opportunity to respond?  

 4               THE COURT:  If you want it, you have it.  

 5               MR. SITCOV:  I don't know that I want it, but I 

 6    may need it.  Thank you.  

 7               MR. RIFKIND:  I am not really sure that there is 

 8    anything further I can elicit from Professor Fisher on the 

 9    subject.  The cutting point is a longish struggle between 

10    plaintiff's counsel and defendant's counsel to obtain data 

11    files from the bureau.  

12               I learned last night, but not from Dr. Fisher, 

13    for example, that as to a key set of studies that we were 

14    doing, that is, all of PX-696 to 701, it was not until April 

15    20th, when we obtained from the government and gave to our 

16    computer people a document that permitted them to generate 

17    PX-696 and 701.  So it couldn't possibly have been done 

18    before April 20.  In fact, I had people staying up sort of 

19    round the clock for ten days in order to be able to get that 

20    work done by May 1, which is when I gave it to the 

21    government, as I was required to.  

22               I might also say that when we gave the material 

23    to the government on May 1, the letter we sent to them at 

24    that time said certain of the exhibits are based on data in 

25    the form of computer tapes, disks, etc., much of which is 
                                                              1168

 1    obtained from defendants, i.e., them.  If you wish to obtain 

 2    any databases from us, please call me.  

 3               I have not yet received that call.  

 4               THE COURT:  Let me put the question to Dr. 

 5    Fisher.  

 6               When you were exploring this a few minutes ago, 

 7    you began to say that CRA, Charles River Associates, which 

 8    compiled the numbers, didn't have the data in time.  Did you 

 9    have any way of knowing that other than what Mr. Rifkind 

10    just said?  

11               THE WITNESS:  Do you mean do I know independently 

12    of Mr. Rifkind telling me?  

13               THE COURT:  Did somebody at CRA tell you?  

14               THE WITNESS:  Oh, yes.  

15               THE COURT:  Tell me that.  

16               THE WITNESS:  Mr. Stoddard informed me.  

17               THE COURT:  Stoddard?  

18               THE WITNESS:  Stoddard.  He is a senior research 

19    associate at CRA and had been of principal assistance in 

20    this work.  He informed me of several events.  I hope I can 

21    get them the right way around.  

22               First, as to whatever exhibits it was that Mr. 

23    Rifkind just mentioned, the ones involving the hypothesis 

24    testing -- 

25               MR. RIFKIND:  696 to 701.  
                                                              1169

 1               THE WITNESS:  Thank you.  He informed me of what 

 2    Mr. Rifkind had just said:  We did not, in fact, receive the 

 3    data files until April 20th, approximately, some days after 

 4    the deposition.  Those exhibits could not have been produced 

 5    before that.  

 6               With respect to the exhibits which described the 

 7    expected number of states, there is an earlier date for 

 8    which we had the data, but we did not have the programs that 

 9    the census bureau had used to generate various things 

10    required for that, and eventually we ended up having to 

11    reconstruct what the census must have done.  

12               That work was much delayed because we could not 

13    secure the program.  We kept thinking we would secure the 

14    program, and that eventually arrived after the work had been 

15    done.  We didn't want to do a lot of the unnecessary work.  

16    As a result, that was also not completed until after the 

17    deposition.  

18               With respect to the charts showing percent 

19    minority by state against percent undercount, those charts 

20    are trivial to produce.  They are trivial for us to produce, 

21    they are trivial for the census to produce.  They can be 

22    done by a relatively low-level employee in approximately an 

23    hour.  They don't have any deep data issues in them.  

24               The only other thing I have to say about this is 

25    that at my deposition Mr. Sitcov asked me whether I had done 
                                                              1170

 1    any analyses, and I said what I had done.  Then he asked me 

 2    whether I had reduced any of them to writing, and I said no.  

 3    Then he asked me whether I intended to, and I said no.  

 4               He didn't ask me if I intended to do any further 

 5    analyses that might be put into writing, and indeed I 

 6    described to him the fact that one would want to calculate 

 7    things like the expected number of states and the hypothesis 

 8    test.  

 9               He also asked me at the end of that colloquy a 

10    question I was permitted to answer, which was, in effect, 

11    did I know of any one identified to the defendants and 

12    likely to be called as a testifying expert who was doing 

13    calculations, and I said I did not know, which was true.  He 

14    did not ask me a question which would have called for me to 

15    reveal what Charles River Associates was thinking of doing.  

16               MR. SITCOV:  Your Honor, may I just respond 

17    briefly to a couple of points that Mr. Rifkind and Dr. 

18    Fisher just made?  

19               THE COURT:  Are you going to recross too?  

20               MR. SITCOV:  Yes, just on a couple of points.  

21               THE COURT:  All right.  

22               MR. SITCOV:  When I took Dr. Fisher's deposition, 

23    the colloquy to which he was referring, I asked him a 

24    question two ways, because the first way I asked it, his 

25    attorney directed him not to respond.  The first way I asked 
                                                              1171

 1    the question was, "Are you personally aware whether any 

 2    person who has been, who you believe may be an expert for 

 3    plaintiffs in this case, if any such person has done any 

 4    independent analyses of the census or the PES?"  That starts 

 5    on line 23 of page 15 of his deposition transcript.  

 6               The answer I got from Ms. Goldstein, who is 

 7    sitting at counsel table, was, "I'm going to first of all 

 8    object to the question as somewhat vague, but second of all 

 9    instruct the witness, based on their interpretation that 

10    defendant's counsel had given to the work product privilege, 

11    and instruct Dr. Fisher not to respond to any information 

12    that came in with them other than that which he will -- he 

13    plans to testify about.  But any other work that is being 

14    done that he may or may not know about at the request of 

15    counsel by others, Dr. Fisher should not testify about if he 

16    knows about it."  

17               I did not ask him about trial witnesses at that 

18    time.  

19               THE COURT:  Go on reading. 

20               MR. SITCOV:  May I finish, your Honor?  

21               THE COURT:  Are you going to continue reading the 

22    deposition?  

23               MR. SITCOV:  Yes, I will.  Then I said, "So you 

24    are directing him not to respond to those things?  

25               Ms. Goldstein said, "I'm in the anticipation of 
                                                              1172

 1    where this line of questioning is going.  That's what I'm 

 2    saying at this point before it gets too far.  You may or may 

 3    not know about anything.  But based on your and your 

 4    co-counsel's instructions to your witnesses not to testify 

 5    about work product, that's my instruction."  

 6               I said, "My question is, are you instructing him 

 7    not to answer the question that's pending?"  

 8               Then Ms. Goldstein, "Let me hear the question 

 9    again."  

10               This time I changed it so that it would not be 

11    the same question.  "Do you know if any other expert who has 

12    been identified to us by plaintiffs as an expert who may 

13    testify at trial in this case has done an independent 

14    analysis of the 1990 census and/or PES?"  

15                   "MS. GOLDSTEIN:  My instruction stands."  

16               So I asked him two ways, your Honor.  

17               THE WITNESS:  That question was answered, Mr. 

18    Sitcov.  Why don't you read the answer.  

19               MR. SITCOV:  If Dr. Fisher would like the answer, 

20    I am picking up now.  

21                   "MS. GOLDSTEIN:" --

22               MR. RIFKIND:  No. 

23                   "MR. SITCOV:  So you're instructing him not 

24    to respond?  

25                   "MS. GOLDSTEIN:  Without disclosing the work 
                                                              1173

 1    product information, yes.  To the extent that he can, then 

 2    he can respond."  

 3               Right?  

 4               MR. SITCOV:  That's what Ms. Goldstein said.  

 5               THE COURT:  That is what you are reading.  

 6               MR. SITCOV:  No.  I said what I said.  

 7               THE WITNESS:  Why don't you read what Dr. Fisher 

 8    said.  

 9               THE COURT:  The hell with Dr. Fisher.  

10               MR. SITCOV:  Then we will drop our motion, if 

11    that is the way you feel about it.  

12               THE COURT:  Why don't you just read verbatim what 

13    is in front of you.  

14               MR. SITCOV:  I am.  

15               Ms. Goldstein says, "My instruction stands."  

16               I said, "So you're instructing him not to 

17    respond?"  

18               Ms. Goldstein said, "To the extent he can't 

19    respond without disclosing that work product information, 

20    yes.  To the extent that he can, then he can respond.  

21                   "THE WITNESS:  Well, you will all be excited 

22    to know that I believe the answer to the question you have 

23    asked is no, I do not know."  

24               As I said, I asked him both ways.  I asked him 

25    first did he know if any expert that the plaintiffs had 
                                                              1174

 1    retained would be doing any work.  He was instructed not to 

 2    respond.  I then asked him had any trial witness done any 

 3    such work.  

 4               THE COURT:  Did he know whether any trial 

 5    witness.  

 6               MR. SITCOV:  Yes, did he know.  He was instructed 

 7    not to respond but said, as to the trial witnesses, I take 

 8    it, that no trial witness would be doing that.  

 9               THE WITNESS:  He said he didn't know.  

10               THE COURT:  Correct.  

11               MR. SITCOV:  He didn't know if any trial witness 

12    would be doing that.  But that still doesn't answer the 

13    question that is posed by exactly these documents.  

14               Obviously, people who were not trial witnesses 

15    did this.  I was not able to discover that.  

16               There is another point, and we will certainly be 

17    happy to address this in more detail once we see the 

18    affidavit that Mr. Rifkind is going to be preparing.  The 

19    magistrate, Magistrate Ross, required the plaintiffs to pay 

20    our costs of duplicating this information when it was 

21    disclosed to them, I guess, late last month because we had 

22    already disclosed it to them before.  They had this 

23    information.  What they did with it or what they chose not 

24    to do with it, I don't know.  But it is not the case that 

25    they did not have this information.  
                                                              1175

 1               I would also request that we be given by 

 2    plaintiff specific identification of files by name and 

 3    specific identification by request of when these files were 

 4    requested in their declaration.  I don't want there to be 

 5    any doubt -- 

 6               THE COURT:  In their?  

 7               MR. SITCOV:  I understood that Mr. Rifkind was 

 8    going to be preparing an affidavit.  We want all the 

 9    specifics possible because we believe that he is wrong.  We 

10    believe that this information was given to them in a very 

11    timely fashion and that they had it for quite some time.  

12               That does not address the other point that Dr. 

13    Fisher had testified that he had not intended to talk about 

14    specific data.  I asked him on page 15, "Have you done any 

15    independent analysis of the census in the PES?"  

16               The answer was, "Well, I've thought about the 

17    loss function at some length and I've come to certain 

18    conclusions about it, and in that sense it's the case I've 

19    done independent analyses.  But those are analyses of the 

20    principles involved rather than independent analyses of the 

21    data themselves."  

22               MR. RIFKIND:  It doesn't say anything about what 

23    he expected to testify about.  It answers your question, 

24    "Have you done any independent analysis," which I take it he 

25    correctly answered, honestly, I haven't, except to the 
                                                              1176

 1    extent that he said.  

 2               THE COURT:  I think we have thoroughly vented 

 3    this problem.  

 4               MR. RIFKIND:  Could I add just one sidelight on 

 5    it?  

 6               MR. SITCOV:  No.  

 7               MR. RIFKIND:  I think it is quite clear, and if 

 8    it isn't we will make it abundantly clear, that Professor 

 9    Fisher meticulously and carefully answered every question 

10    that was put.  Mr. Sitcov raised a question about some 

11    instructions that were being given.  I think, in fact, the 

12    instructions did not block the relevant testimony.  But I 

13    would like to put it for just one second in the context of 

14    what was going on here.  

15               We gave the government a 26(b)(4) identification 

16    of what the witnesses were going to talk about.  We asked 

17    for the same from them and we were told that we weren't 

18    going to get it, we were going to get a witness with a very 

19    general statement, the name of the witness with a very 

20    general statement about his subject.  

21               We then went and took the depositions of their 

22    experts, and they told us that they were not going to 

23    produce the witnesses' documents, the documents that had 

24    already been prepared at the time of the deposition.  In 

25    candor, we asked the Court if we could have them, and we 
                                                              1177

 1    were advised that the Court's view was we would get them on 

 2    May 1 when they submitted their reports and all the rest.  

 3    We understood that to be the law of the case.  In fact, we 

 4    were sufficiently disturbed about that that we asked twice.  

 5    Your Honor, and I am not challenging the ruling now, said 

 6    you will get it on May 1.  

 7               So we got on May 1 the reports of their 

 8    witnesses.  At their depositions we did not get their 

 9    underlying studies.  So we were examining their witnesses in 

10    a vacuum without access to the data that they were then 

11    generating.  

12               It was in that context that Ms. Goldstein 

13    instructed the witness not to divulge anything that 

14    defendant's counsel at that stage had claimed to be work 

15    product.  

16               THE COURT:  Can we agree on some schedule here 

17    for your affidavit and then your response, if any?  

18               MR. SITCOV:  Sure.  

19               MR. RIFKIND:  I think 24 hours.  The day after 

20    tomorrow would be fine, I think it would be safer.  

21               THE COURT:  The day after tomorrow you will give 

22    it to Mr. Sitcov?  

23               MR. RIFKIND:  Yes.  

24               THE COURT:  How long will you need?  

25               MR. SITCOV:  Day after tomorrow will be Thursday.  
                                                              1178

 1    That might put us in a little bit of a bind because we will 

 2    be starting our case right about then.  The next day would 

 3    be -- 

 4               THE COURT:  Pick a time.  

 5               MR. SITCOV:  How about Tuesday, which is the day 

 6    we come back after the weekend?  

 7               THE COURT:  I don't have to decide this 

 8    immediately.  All right.  I don't want to hear any more 

 9    about this particular problem.  I will do it on papers.  

10               Do you have some reacross, Mr. Sitcov?  

11               MR. SITCOV:  Yes, very brief.  

12    RECROSS-EXAMINATION 

13    BY MR. SITCOV:

14         Q.    Do the census bureau's loss function analyses 

15    take account of the population in geographic areas?

16         A.    No, they didn't do it that way.

17         Q.    Could you turn to Plaintiff's Exhibit 654, 

18    please.

19         A.    Okay.

20         Q.    That document has a Bates stamp number on the 

21    first page 7855.  I would like you to turn to 7857.

22         A.    Yes.

23         Q.    Paragraph 4 says, "The census bureau will 

24    generate and analyze lots of data.  The loss functions 

25    should be thought of as organizational tools but not the 
                                                              1179

 1    only set of tools that assist the bureau in its review and 

 2    adjustment decision-making processes."  

 3               Do you agree with that statement?

 4         A.    This was, if you look up what this is, this is a 

 5    list of the issues and concerns raised.  It is not a list of 

 6    analyses.  

 7               Do I agree that the census bureau followed this 

 8    or what, whether the census bureau should have followed 

 9    this?

10         Q.    Yes.  Do you agree that the loss function should 

11    not be the only set of tools? 

12         A.    It depends in part what you mean by "loss 

13    function."  In some sense loss functions just are a way of 

14    describing systematic criteria for dealing with, in this 

15    case, this problem.  In that sense I don't agree with it.  

16               If you if you mean the particular loss function 

17    analysis done by the bureau should not be the only one, then 

18    I'm not sure.  If you were to include in that the 

19    proposition do the loss function, see how robust it is, do 

20    some tests, and so on, then I would say yes, I think that is 

21    the appropriate criteria.

22         Q.    Would it be irrational if the census bureau 

23    decided that loss functions should not be the only set of 

24    tools?

25         A.    I would have to know exactly what they meant by 
                                                              1180

 1    that.

 2         Q.    According to this statement, if they meant it in 

 3    this way, that the census bureau will be generating lots of 

 4    data and that loss functions should be considered 

 5    organizational tools but not the only set of tools?

 6         A.    Yes, but I don't know what they meant they were 

 7    going to do with the rest of the tools.  

 8               MR. SITCOV:  No further questions.  

 9               MR. RIFKIND:  We are done.  

10               THE COURT:  Thank you.  It has been a pleasure 

11    listening to you, Dr. Fisher.  It has been most 

12    enlightening.  

13               THE WITNESS:  Thank you, your Honor.  I have 

14    enjoyed it.  Can we hide the bit about the spot on the tie 

15    from my wife?  

16               THE COURT:  Good luck.  

17               (Witness excused) 

18               (Recess) 

19               (Continued on next page) 

20    

21    

22    

23    

24    

25    
                                                              1181

 1               THE COURT:  Mr. Solomon.  

 2               MR. SOLOMON:  Your Honor, the plaintiffs call 

 3    professor bruise cane.

 4    BRUCE EDWARD CAIN,
                        

 5             called as a witness by the plaintiff, having

 6             first been duly sworn, was examined and

 7             testified as follows: 

 8    DIRECT EXAMINATION 

 9    BY MR. SOLOMON:

10         Q.    Professor Cain, what is your occupation?

11         A.    I am a professor of political science at the 

12    University of California Berkeley, and I'm the associate 

13    director of the Institute of Governmental Studies.

14         Q.    That is also affiliated with the University of 

15    California?

16         A.    Yes, organized research unit of the University of 

17    California Berkeley.

18         Q.    Describe to the court, please, your educational 

19    background?

20         A.    I have a BA from Bowdoin College in 1970, I have 

21    a B. Phil from Trinity College in Oxford where I was a 

22    Rhodes Scholar.  

23               THE COURT:  B. Phil? 

24               THE WITNESS:  Bachelor of Philosophy.  It's a 

25    master's degree, essentially.
                                                              1182

 1         A.    And I have a Ph.D. from Harvard in 1976 in 

 2    political science.

 3         Q.    When you were at Oxford on the Rhodes 

 4    Scholarship, what course of study did you take?

 5         A.    You are required to take two papers, one in 

 6    political theory and the other one in institutions, and then 

 7    I took two other papers in comparitive politics.

 8         Q.    What was the course of study for your Ph.D.?

 9         A.    That consisted of course work, again, including 

10    substantive course work in American politics, course work in 

11    political theory and course work in statistics and political 

12    methodology.

13         Q.    What was the subject of your dissertation?

14         A.    My dissertation as an application of voting 

15    models derived for American policies to British politics, so 

16    it was a study of how those models compared when they were 

17    applied to British politics.

18         Q.    Did your study include statistical analyses?

19         A.    Yes, it did.  It involved modeling, the decision 

20    to vote in British elections using patterns of modeling that 

21    were used in the United States.

22         Q.    When you say modeling, do you mean statistical 

23    modeling?

24         A.    Yes, statistical modeling using regression, 

25    probit logic, standard techniques in political science.
                                                              1183

 1         Q.    In connection with the work you did for your 

 2    Ph.D. or leading up to it, what training or study did you 

 3    have in statistics or mathematics?

 4         A.    I took the --  I took a two year sequence courses 

 5    in statistics and political methodology, including 

 6    introductory statistics course by various regression 

 7    techniques and a specialized course in which we look at 

 8    simultaneous equation modeling and the logic and probit type 

 9    of models.

10         Q.    Since completing your Ph.D. work, have you 

11    continued to be involved in statistical modeling of 

12    political data?

13         A.    Yes.  A fair amount of my work involves the 

14    applications of statistics to American politics.

15         Q.    Can you give the court some sense of how you used 

16    statistics in your work?

17         A.    Well, very typically, for example, an example 

18    would be in a study that we did of the 1988 presidential 

19    elections, we had collected a number of the polls that the 

20    LA Times had from the different states that had primaries 

21    and we analyzed those polls using these techniques to 

22    discover whether the issue concerns raised in the New 

23    Hampshire primary were similar or different than those 

24    raised in the New York primary, and the had standard way to 

25    use this is to try to figure out the relative effect of 
                                                              1184

 1    different variables of interest in understanding, say, 

 2    voting behavior or behavior of representatives.

 3         Q.    Was your dissertation published?

 4         A.    Yes, it was published in two articles, one that 

 5    appeared in --  actually, both of them --  one appeared in 

 6    political --  American Journal of Political Science, which 

 7    was on strategic voting, and the other one appeared in 

 8    Comparative Political Studies and it was a model of issue 

 9    voting in England, the effect of different issues on voting.

10         Q.    After your Ph.D. and before where you are now, 

11    were you offered an academic position?

12         A.    Yes.  I was offered a position as an assistant 

13    professor at the California Institute of Technology.

14         Q.    Did you become a professor at California Tech?

15         A.    Yes, I did.

16         Q.    You move from being a professor at California 

17    Tech to professor at Berkeley, is that right?

18         A.    In 1989 I moved up to Berkeley to become a 

19    professor of political science and the political science 

20    department at Berkeley and assumed the position of associate 

21    director of the IGS.

22         Q.    And IGS is the Institute of --

23         A.    Institute of Governmental Studies.  To make it 

24    easier I will just call it IGS.

25         Q.    What is the IGS?
                                                              1185

 1         A.    The IGS is an organized research unit of the 

 2    University of California, that is to say, it receives state 

 3    money to conduct research, to do conferences, to educate 

 4    students in a setting on the campus.

 5               In addition, we receive private funds and we do 

 6    some contractual work with the state on various problems 

 7    that they may have.

 8         Q.    What courses generally do you teach as a 

 9    professor of political science?

10         A.    Well, I teach the American politics course, which 

11    is a general survey course of executive, legislative and 

12    voting behavior; I teach a course in representation at the 

13    graduate level where we look at various issues of 

14    representation from the redistricting to campaign finance to 

15    bribery laws.

16               I teach a course in California politics in state 

17    and local politics which is a survey of policy and 

18    institutional issues facing state government.

19         Q.    In connection with your course work, do you study 

20    the structure and role of administrative agencies in the 

21    American system of government?

22         A.    Yes.  That's a very important part of the 

23    American government course, is to look at the operations of 

24    the executive branch and to look at different kinds of 

25    models of how decisions were made by the executive branch.
                                                              1186

 1         Q.    In particular, do you study the process by which 

 2    decisions are made by political entities or administrative 

 3    agencies?

 4         A.    Yes.  That would be --  that's another very 

 5    common topic that we would cover in such a course.

 6         Q.    Do you supervise dissertations as part of the 

 7    work you do at California?

 8         A.    Yes.  Both at California Tech and at the 

 9    University of California Berkeley I have supervised in the 

10    past and continue to supervise now graduate students working 

11    in the area of quantitative American politics.

12         Q.    Are you a member of professional organizations or 

13    associations?

14         A.    I am a member of the American Political Science 

15    Association and the Western Political Science Association 

16    and off and on depending on the year I sometimes belong to 

17    some of the regional associations, depending on whether I 

18    attend their meetings in a given year.

19         Q.    Are you active in scholarly journals?

20         A.    Yes.  I have been on a couple of editorial boards 

21    and currently serve on one.

22               I also regularly referee for almost every major 

23    journal in the political science field, and I would say on 

24    average I referee five or six articles a month on topics, 

25    again, in the area of quantitative American politics.
                                                              1187

 1         Q.    When you say quantitative, does that include the 

 2    use of statistics?

 3         A.    Yes.  The term quantitative American politics 

 4    means the application of statistics and quantitative data to 

 5    understanding trends or institutions in American politics.

 6         Q.    Have you won any awards for teaching?

 7         A.    Yes, I did.  I won an award I believe in 1988, 

 8    but I may be wrong, it mate be 1989, but I won an award 

 9    given by the students, the associated students of the 

10    California Institute of Technology, which is designated in 

11    my curriculum vitae as the ASCIT Teaching Prize.

12         Q.    Have you won any awards for your published work?

13         A.    Yes.  I wrote a book called Personal Vote, along 

14    with John Ferejohn and Morris Fiorina, the book was called 

15    Tt Personal Vote, and it was a comparison of constituency 

16    representation in Great Britain and the United States and 

17    that book won the Ferna Prize, which is the prize given by 

18    the legislative study subcommittee of the American Political 

19    Science Association.

20         Q.    What are the primary fields of research that you 

21    engage in?

22         A.    I do American voting behavior, I have written on 

23    Congress, state legislatures, redistricting, campaigning 

24    finance, areas of democratic theory, institutional design.  

25    These are some of the areas that I have written on.
                                                              1188

 1         Q.    Have you published books on any of those 

 2    subjects, in addition to the book you told us about?

 3         A.    Yes.  In addition to the Personal Vote, I have 

 4    written a book, two books on reapportionment and 

 5    redistricting.  The first one was called the Reapportionment 

 6    Puzzle, and the second one was called Congessional 

 7    Redistricting, which I co-authored with David Butler from 

 8    Oxford University.

 9               In addition, I edited a book called Developments 

10    in American Politics, which is a survey of topics related to 

11    American government.

12         Q.    You also publish articles on the fields of your 

13    research?

14         A.    Yes, that's correct.

15         Q.    In addition to the academic work that you do, do 

16    you do consulting work?

17         A.    Yes, I do some consulting work.

18         Q.    Describe briefly the subject matter of your 

19    consultations generally to the court.

20         A.    Most typically they involve voting rights 

21    litigation in which I'm called in to do empirical 

22    assessments that are related to the so-called jingles 

23    criteria in voting rights cases, which is a setup of sort of 

24    empirical test to determine whether or not a particular 

25    minority group has a claim under Section 2 of the Voting 
                                                              1189

 1    Rights Act to representation.

 2         Q.    In connection with the consultation work that you 

 3    have done, have you represented, consulted for the United 

 4    States Department of Justice?

 5         A.    That's correct.  In a very important case which 

 6    is Garza versus the County of Los Angeles, I was a 

 7    consultant to the Justice Department in which I was 

 8    responsible, I was the principal contractor for providing 

 9    data to the Justice Department, data which they used both to 

10    assess the nature of the claim under the jingles criteria 

11    and also to look at the prospects of creating districts that 

12    might give Latinos equal opportunity to elect a 

13    representation of their choice.

14         Q.    Did the work that was done in connection with 

15    that data include statistical modeling and work?

16         A.    There was statistical modeling done, but it was 

17    not done by me.  I was, I was the --  I helped bring 

18    together the data using people that had worked for me in the 

19    past.  But I did not do the statistical analysis of that 

20    particular case because I was moving up to Berkeley at the 

21    time and I just didn't have the time to do it, so it was 

22    given over to other people.

23         Q.    Are you currently involved in any redistricting 

24    consulting?

25         A.    I am currently a consultant to the Los Angeles 
                                                              1190

 1    City Council and their redistricting.

 2         Q.    Have you consulted for any state or local 

 3    government outside of California and the Department of 

 4    Justice?

 5         A.    Yes.  I consulted now twice for the Attorney 

 6    General's office of the State of Massachusetts, again, 

 7    related to potential voting rights litigation.

 8         Q.    Have you been qualified to testify as an expert 

 9    in political science at any time prior to today?

10         A.    Yes.  I was, I was qualified as an expert witness 

11    in the so-called Prop 73 trial, which was the service, the 

12    Service Employees Union trial that is designated in my 

13    vitae.

14         Q.    In connection with the redistricting work, do you 

15    use census data?

16         A.    Yes, quite frequently.

17         Q.    In connection with the research and teaching that 

18    you do, have you studied the role of the Census Bureau?

19         A.    Yes, in connection to writing the book on 

20    Congressional Redistricting, we have a section in the book 

21    about the Census Bureau and the dispute over adjustment.

22         Q.    Would you look, please, at what we have marked 

23    for identification as PX 568 in the binder.  

24               MR. SOLOMON:  For the record, that is the CV of 

25    Professor Cain.
                                                              1191

 1         Q.    Can you identify that for the court, please, 568?

 2         A.    Yes, that is my vitae.  

 3               MR. SOLOMON:  We would offer PX 568, your Honor.  

 4               MR. SUBAR:  No objection.  

 5               THE COURT:  568 is admitted.

 6               (Plaintiff's Exhibit 568 marked for 

 7    identification was received in evidence.) 

 8    BY MR. SOLOMON:

 9         Q.    In connection with the work you have done to 

10    prepare yourself to testify before the court today, 

11    generally describe what you have reviewed or read.

12         A.    I have reviewed the report of Secretary Mosbacher 

13    and statements of Secretary Mosbacher in his press 

14    conference; I have reviewed other reports, such as the one 

15    by Professor Wolter and the undercount steering committee; I 

16    have reviewed the pleadings of the court; I've reviewed 

17    depositions of people such as Wendell Winkle, Lance 

18    Tarrance, William McGehee, Swanson.

19               I have looked at statements by former Director 

20    Barabba, I've looked at some memoranda circulated within the 

21    Department of Commerce, and obviously I've looked at data, 

22    the adjusted and unadjusted figures and their indications.

23         Q.    In connection with that work, have you been paid 

24    for your time?

25         A.    No, I have not.
                                                              1192

 1         Q.    Are you being paid for the time you are spending 

 2    here today?

 3         A.    No, I am not being paid.

 4         Q.    Do you have any promise, has any promise been 

 5    made to you or is there any expectation that you will 

 6    receive any remuneration in connection with this work?

 7         A.    No, unfortunately.  

 8               MR. SOLOMON:  We proffer Professor Cain on the 

 9    subjects of political science, including statistical 

10    analysis of political and sensus data and redistricting.  

11               MR. SUBAR:  May I have a few questions on the 

12    voir dire, your Honor?  

13               THE COURT:  Yes

14    VOIR DIRE EXAMINATION

15    BY MR. SUBAR:

16         Q.    Dr. Cain, you have never written concerning 

17    census matters, have you?

18         A.    As I said, Mr. Subar, I have written in my book 

19    on Congressional Redistricting about the census as it 

20    applies to redistricting and apportionment.

21         Q.    Was that book published, been published since I 

22    took your deposition about a month and a half ago?

23         A.    It came out this spring, but I'm not sure of the 

24    timing of that.  It was right around the time of the 

25    deposition.
                                                              1193

 1         Q.    If I suggested, Dr. Cain, that at your deposition 

 2    I asked you the following question and you gave me the 

 3    following responses, would that refresh your recollection, 

 4    at page 10, line 16.

 5               "Q.    Do you know when the CV that has been 

 6    marked as Exhibit 1 was prepared?

 7               "A.    A couple of months ago.

 8               "Q.    And you prepared another one since?

 9               "A.    I think there may be a later version, for 

10    example, the book, redistricting, with David Butler McMillan 

11    forthcoming has been published, so it has a 1992 publication 

12    date." 

13               Does that refresh your recollection, Dr. Cain, 

14    that you --

15         A.    Yes.  I mean, it depends on what you mean by 

16    "published." 

17               It was sitting --  I didn't know whether it was 

18    sitting in a warehouse or whether it actually had --  I 

19    mean, I had seen the copy edited stuff and I had sent it 

20    back at that point.  Whether I actually had physical copies 

21    of the book or not was probably what I had in mind.

22         Q.    It certainly had been written before the 

23    deposition?

24         A.    It certainly had been written, no question about 

25    that.
                                                              1194

 1         Q.    And do you recall, Dr. Cain, that at your 

 2    deposition I asked you the following question and you gave 

 3    me the following response, at page 53, line 20.

 4               "Q.    But you have never written concerning 

 5    census matters?

 6               "A.    I've never written concerning census 

 7    matters." 

 8               Does that refresh your recollection, Dr. Cain?

 9         A.    It does, but at the time I had forgotten that I 

10    had.

11         Q.    Dr. Cain, you haven't studied concerning census 

12    matters in enough detail that you would want to come before 

13    the court as an expert, have you?  

14               THE COURT:  On what?

15         A.    Precisely that.  There are some issues --  you 

16    will not --  the judge will be relieved to know that I have 

17    absolutely nothing to say about smoothing, absolutely 

18    nothing to say about many of the technical matters that seem 

19    to have put lots of people to sleep the last day or so, so 

20    you are right, Mr. Subar, I have nothing to say about the 

21    detailed process by which the Census Bureau has conducted 

22    its PES and I intend not to say anything about them.

23         Q.    But as of the time of your deposition, you had 

24    not studied concerning census matters in enough detail that 

25    you would want to come before the court as an expert, is 
                                                              1195

 1    that correct?

 2         A.    Mr. Subar, at the time that I spoke to you, I had 

 3    worked extensively with the census data, I knew about and 

 4    had looked at and we discussed during the deposition the 

 5    implications of using adjusted and unadjusted data in 

 6    California, so there are certain issues about the census 

 7    which I am intimately familiar with from a decade of working 

 8    on apportionment and voting rights cases.

 9               It seems to me as long as I am talking about 

10    that, you and I have an understanding about what I know.

11               If I try to tell you about the statistical issues 

12    related to the PES, then I'm out of my league and you'll 

13    know immediately.  

14               MR. SUBAR:  Your Honor, subject to that 

15    understanding that I think the court now has with Dr. Cain, 

16    we have no objection to his qualification as an expert, with 

17    one exception:

18               Some of the questions that Mr. Solomon asked and 

19    some of the exhibits that were given to us in connection 

20    with distribution testimony suggested that Dr. Cain might be 

21    testifying, if the court permits, with regard to the 

22    structure and role of the Census Bureau, and that is an area 

23    that falls outside the scope of Dr. Cain's witness statement 

24    that was given to us in response to an interrogatory that we 

25    propounded to the plaintiffs.  
                                                              1196

 1               THE COURT:  I am not sure in abstract that I can 

 2    take a position on any of that, but in general I will 

 3    qualify Dr. Cain as expert in political science and the 

 4    political implications or ramifications of the census.  

 5               MR. SUBAR:  Thank you, your Honor.  

 6    BY MR. SOLOMON:

 7         Q.    Professor Cain, is the accuracy of census data 

 8    relevant to the redistricting process?

 9         A.    Yes, it is.  It's very important.

10         Q.    Tell the court why.

11         A.    Well, the aim of redistricting is to make 

12    district populations as equal as possible and it's very 

13    important that the data be reliable and unbiased so that 

14    when we go through the process of adjusting the populations 

15    with the new data that we can be sure that we have districts 

16    that are equally populated and not systematically 

17    underpopulated in any way.  

18               THE COURT:  Are you talking about Congressional 

19    districts or all districts? 

20               THE WITNESS:  Well, somebody mentioned that there 

21    are 39,000 jurisdictions and one of the things that we have 

22    to bear in mind is that the redistricting occurs with all 

23    jurisdictions that have any kind of general governmental 

24    responsibility and, as a result, we have redistricting not 

25    only of Congress, but of the Assembly and State Senate and 
                                                              1197

 1    Houses, of city councils, boards of supervisors, water 

 2    boards, boards of education, so everything we are talking 

 3    about applies up and down the same rules of redistricting 

 4    essentially apply to anything that has districts and has a 

 5    general governmental purpose.

 6         Q.    Are you familiar with the term differential 

 7    undercounting?

 8         A.    Yes, I am.

 9         Q.    What is your understanding of that term?

10         A.    Well, differential undercounting refers to the 

11    fact that the census undercounted certain racial and ethnic 

12    groups, in particular, Latinos and African-Americans, by 

13    greater amounts than other groups in the population.

14         Q.    Do you believe there was a differential 

15    undercount in the 1990 census?

16         A.    Yes, I do believe that.

17         Q.    What is the basis for that belief?

18         A.    The basis for that is the reports that were 

19    issued at the time of the decision, the reports of Secretary 

20    Mosbacher and reports of others as well.

21         Q.    In connection with the book that you wrote, did 

22    you, in fact, study the subject of differential 

23    undercounting?

24         A.    In the Congressional redistricting book we do 

25    refer to this problem and write about it.
                                                              1198

 1         Q.    In your opinion, does the differential undercount 

 2    that you made reference to have an effect on the voting 

 3    power of racial and ethnic minorities?

 4         A.    In my opinion, it does and it was foreseeable.

 5         Q.    Tell the court why you believe it does.

 6         A.    Well, because of the differential undercount, 

 7    areas with high concentrations of Latinos and 

 8    African-Americans are going to be credited with less 

 9    population than they actually have and, as a result, when 

10    the district boundaries are drawn, they will receive less 

11    representation than they otherwise should receive, or to put 

12    it another way, their districts will be substantially 

13    overpopulated relative to the ideal population, because we 

14    will not know the true number, we won't have the missing 

15    people included in the census figures.

16         Q.    Most of us, I think, and familiar to the court, 

17    explain very quickly the ideal district.

18         A.    Well, what happens in a redistricting is that you 

19    take the state population or the jurisdiction population if 

20    it's a city or a county, you divide it by the number of 

21    seats and that gives you the ideal population, and then the 

22    goal is to try to get the districts as nearly as practicable 

23    to that ideal population.

24         Q.    Do you believe that there is a malapportionment 

25    imbedded in the 1990 census uncorrected by the PES?
                                                              1199

 1         A.    Yes, I do.

 2         Q.    Have you, in connection with your work that you 

 3    are going to testify to here today, done any numerical 

 4    analysis about the level of malapportionment that is 

 5    imbedded in the uncorrected 1990 census data?

 6         A.    Yes.  Although a team, we went ahead and I have 

 7    taken a look at some figures that will confirm this.

 8         Q.    Tell the court generally what was done?

 9         A.    Well, what we did --  what I did was I wanted to 

10    see what would happen if we looked at congressional and 

11    state legislative districts with two sets of data; first 

12    looked at their population using the unadjusted data and 

13    secondly look at their populations with the adjusted data 

14    and then look at the difference between the two populations 

15    and see if, in fact, what seems obvious is true, namely, do 

16    we find that the districts that have substantial 

17    concentrations of minorities have substantial amounts of 

18    undercount.

19         Q.    Look, please, at PX 591 marked for 

20    identification, a document entitled, "Percent undercount 

21    approved redistricting plan total population of California," 

22    behind tab 591 in your notebook.

23         A.    Yes, I've got that.

24         Q.    Can you identify that document?

25         A.    Well, this is essentially the output or the table 
                                                              1200

 1    showing the output of what happens when you at that time 

 2    approved, state approved congressional districts and there 

 3    are 352 of them because California receives some additional 

 4    seats.  

 5               For each one of the districts you have the 

 6    districts identification number, you have its population 

 7    with the unadjusted figures, you then have the population 

 8    with the adjusted figures and then you have the percent of 

 9    undercount which is the difference between the adjusted and 

10    unadjusted figures divided by the unadjusted population to 

11    give you a base, and that is expressed as a percentage.

12               Then what we have done here is ranked the data in 

13    terms of the degree or percent of undercount so that the 

14    first district, which is the 35th Congressional District, 

15    has the largest degree of undercount, it is 8.7 percent.

16         Q.    Tell the court what you mean by approved 

17    redistricting plan?

18         A.    Well, California attempted to redistrict through 

19    the legislature.  The negotiations dragged on and, 

20    unfortunately, never came to a resolution with the 

21    government.

22               This went to court masters and the court drew 

23    these lines and they are now in place and will be used for 

24    the 1992 election.

25         Q.    So the census POP column that shows the numbers 
                                                              1201

 1    that are roughly similar, that is the population that has 

 2    actually been approved, is that right?

 3         A.    That is correct.  When the --  these are the 

 4    figures that the masters believe they have put into the 

 5    districts.

 6         Q.    And the adjusted POP column shows that those are 

 7    the number of people estimated by the PES to be in the 

 8    districts that have just been approved?

 9         A.    Well, using the PES to adjust the census data, 

10    that is the figure that you come up with.

11         Q.    Who prepared this chart?

12         A.    This was prepared by the Charles River 

13    Associates.

14         Q.    Tell the court how it was prepared.

15         A.    Well, essentially there are seven files that need 

16    to be combined in order to prepare this data.  There are the 

17    two census files, the file containing the unadjusted data 

18    and the file containing the adjusted data; there are three 

19    files that contain descriptions of the congressional 

20    assembly and senate plans which were obtained from the State 

21    of California; there was a file which, again, was obtained 

22    from the State of California or from the law firm for the 

23    State of California which has the political data that you 

24    need in order to merge in order to get the registrations 

25    that go with the census blocks; and then there is an 
                                                              1202

 1    identifying file for the blocks.

 2         Q.    Did you satisfy yourself that CRA properly did 

 3    the numerical analysis to confirm the view that you asked 

 4    them to?

 5         A.    Yes, I did.  

 6               MR. SOLOMON:  We would offer PX 591, your Honor.  

 7               MR. SUBAR:  No objection.  

 8               THE COURT:  591 is admitted.  

 9               (Plaintiff's Exhibit 591 marked for 

10    identification was received in evidence.)

11    BY MR. SOLOMON:

12         Q.    The first district, the district with the highest 

13    undercount and I think in your opinion the district with the 

14    most malapportion is District 35.  Tell the court what is 

15    District 35.

16         A.    District 35 is the district represented by Maxine 

17    Water.  It's an African-American district.

18               Maxine Water, you will recall, was the person, 

19    probably the pivotal person in the LA riots from the 

20    African-American community.  Her district was one of the 

21    most heavily devastated in the riots.  

22               MR. SUBAR:  Objection.

23               Your Honor, I would move to strike the reference 

24    to the LA riots and to what role Congresswoman Water might 

25    have played in that situation.  It's entirely irrelevant to 
                                                              1203

 1    this case.  

 2               THE COURT:  The motion is denied.

 3         Q.    The second most heavily undercounted district --

 4         A.    That is the districts that contains Watts and 

 5    Compton.

 6         Q.    The districts at the bottom of the chart, can you 

 7    just generally describe those for the court?

 8         A.    District 10 is in Contra Costa County and it is 

 9    predominantly white, suburban, middle-class district.

10               District 4 is in the so-called motherload, which 

11    is the gold country.

12         Q.    Gold country?

13         A.    Gold as in G O L D.

14         Q.    Okay.

15               Did you ask CRA to do any other numerical 

16    analysis to confirm your view of the malapportionment that 

17    is imbedded in the uncorrected data?

18         A.    Yes, I did.

19         Q.    Would you turn, please, to tab 587 where we have 

20    marked for identification as PX 587 a chart and just some of 

21    it's backup entitled, "Percent undercount by percent 

22    minority." 

23               I would think, Professor Cain, that you could 

24    explain to the court 587 and 588 and 589 and 590 at the same 

25    time.
                                                              1204

 1         A.    Yes.  They are of a piece.  They vary only in 

 2    that we are looking at different kinds of districts.

 3               What you see is a scatter plot which visually 

 4    gives you a representation of the different congressional 

 5    districts in terms of two variables:

 6               One is the degree of undercount or the percent 

 7    undercount that that congressional district has and the 

 8    other is the percent minority.  So for each point you have 

 9    its position with respect to both of those variables.

10               And what we have done is classified the points 

11    into different shapes for your visual aid.

12               The triangles represent the points that are in 

13    the lowest fifth with respect to percent minority, the dots 

14    are the points that are in the upper 20 percent and the 

15    squares are the ones that are in the middle.

16              In addition, you will see a line and that line is 

17    essentially a fitted regression line.  What that is is a 

18    standard statistical technique which minimizes the distance 

19    of the points to the line so it's the best fit for that 

20    scatter points.  

21               And from that you can read the estimation of the 

22    slope, which tells you the magnitude of the relationship 

23    between the two variables and you get an estimate of the 

24    standard error, which tells you how much confidence you can 

25    have that this relationship is statistical significant, that 
                                                              1205

 1    is to say, it wasn't produced by chance.

 2               Those statistics are reported in the somewhat 

 3    hard to read because it's a small square box at the bottom 

 4    that says, "Test of significance."

 5         Q.    Am I correct that PX 587 shows the results of the 

 6    malapportionment analysis, if you will, for the California 

 7    congressional districts?

 8         A.    That's correct.

 9         Q.    And then 588 shows the results of the 

10    malapportionment within the California senatorial districts?

11         A.    That's correct.

12         Q.    And 589 shows the results of the malapportionment 

13    of the California assembly districts?

14         A.    That's correct.

15         Q.    Explain to the court what PX 590 is?

16         A.    590 is slightly different, because here what we 

17    have done is grouped blocks into 100 points of about 298 

18    each which represent, if you start at the 100 percent 

19    minority points, it represents the 298 most minority blocks 

20    grouped together and then the next point is the next 298.

21               And so basically it does pretty much the same 

22    thing, only instead of working with districts we now have 

23    equal numbers of blocks and, again, showing the location of 

24    these aggregations of blocks with respect to percent 

25    minority and percent undercount.
                                                              1206

 1         Q.    You mentioned that some tests of significance of 

 2    the data were done.

 3               What conclusion do you draw whether the data, 

 4    before you explain what it is, whether the data here is 

 5    significant?

 6         A.    We conducted a number of tests of significance.

 7               We conducted the tests of significance that are 

 8    related to the regression analysis, we conducted analysis of 

 9    variance testing whether differences between means were 

10    significant.

11               In addition, we explored different kinds of 

12    cutoff points, what if instead of 20 percent we cut it off 

13    at 30 percent, and we tested the significance of that.

14               Then in addition we ran a number of other tests 

15    which I don't rely as heavily on because they are not 

16    standard tests in political science, but they sort of 

17    further explore assumptions about the tests.  That is, 

18    instead of using the actual values, they rank order the 

19    districts in terms of their relationship to the median or 

20    their --  it breaks down --  some of the tests are under the 

21    assumption that the distribution is normal and some of them 

22    that it is exponential, et cetera.  

23               The main point of that is that for every test 

24    conceivable that the computer could yield, do we discover 

25    that we get a showing of significance regardless of what we 
                                                              1207

 1    throw at it, and the basic conclusion is that in all cases 

 2    we find that the results are statistical significant by 

 3    extremely large margins, so it doesn't seem to be dependent 

 4    on any known convention or premise in the test.  

 5               MR. SOLOMON:  Just for your information, your 

 6    Honor, the test that the witness was referring to, the 

 7    results of those tests have been produced to the government.  

 8    They are not in the book here because he is not relying on 

 9    them.

10         Q.    What conclusions do you draw from these analyses?

11         A.    Well, if you look at the scatter points, you can 

12    see that it tilts up to the right, which indicates that as 

13    the districts become more heavily minority, the degree of 

14    undercount or the percent undercount increases.

15               So what that tells you is that the districts 

16    drawn by the court using the unadjusted data, when we look 

17    at it with the adjusted data, these districts are going to 

18    be substantially overpopulated, that is, if we were to 

19    recompute the ideal population and we were to look at the 

20    deviations from the ideal population, you would discover 

21    that the most heavily minority and most heavily undercounted 

22    districts would be overpopulated by between four and five 

23    percent and you would discover that some of the districts 

24    that are least minority are underpopulated on the order of 

25    two to three percent, which is far beyond the constitutional 
                                                              1208

 1    standard for congressional districts.  

 2               When the masters, for example, did their 

 3    redistricting, they were aiming for a total variance of less 

 4    than .5, that is, one half of one percent.  That was the 

 5    goal.  And many people criticize that redistricting for 

 6    having what they regarded as an unconstitutionally large 

 7    variance, but nonetheless the court felt that .5 was enough.

 8               What we are talking about are populations 

 9    variations that are several times that .5 total deviation, 

10    we are talking about seven or eight percent total population 

11    deviation.

12         Q.    And the deviations you are talking about, is that 

13    the malapportionment that you made reference to earlier?

14         A.    That's right.  The basic notion is that a 

15    district is malapportioned if it is not equally populated 

16    with the other districts, so a district which is 

17    substantially overpopulated relative to other districts.

18               Now, this is, in essence, what we have done, what 

19    we see here is something very much akin to the 

20    malapportionment that was involved in the original Baker 

21    versus Carr.

22               It's kind of ironic.  We sort of perpetuated this 

23    as a result of the undercount.  

24               MR. SUBAR:  Objection.  Dr. Cain is beginning to 

25    testify about legal conclusions. 
                                                              1209

 1               THE WITNESS:  No, I'm merely --  

 2               THE COURT:  Overruled.

 3               You may continue.

 4         A.    I'm sorry.

 5               The point I'm simply making is that because the 

 6    malapportionment follows racial lines and because the racial 

 7    groups live in urban areas, the urban areas are 

 8    overpopulated and the rural and suburban areas are 

 9    underpopulated and it is caused because of the racial 

10    alignment.  I'm saying that is very much analogous to the 

11    situation that prompted the initial malapportionment cases.  

12               MR. SUBAR:  Your Honor, if so qualified, we 

13    certainly don't rely on the objection that I just made a 

14    moment ago, but I would ask the court strike the witness' 

15    testimony for another reason.

16               Dr. Cain is testifying about the relationship 

17    between what he understands to be the undercount and what he 

18    understands to be the minority population of various areas 

19    in the State of California.  That is outside the scope of 

20    his witness statement, and if the point of distribution 

21    testimony is to establish that there was a differential 

22    undercount, that is an issue that is not in this case 

23    because Secretary Mosbacher in his decision document 

24    acknowledged that there is a differential undercount by 

25    race.  
                                                              1210

 1               THE COURT:  If I can cut through all the 

 2    underbrush, I think what he is saying is, why can't we 

 3    stipulate to most of what Dr. Cain is now telling us?  We 

 4    know this and no one disputes it.  

 5               MR. SOLOMON:  If the government is prepared to 

 6    stipulate what the professor has already said, I will simply 

 7    offer these documents and move on.  

 8               THE COURT:  I was going to suggest this ten 

 9    minutes ago, but I didn't want to intrude.  

10               MR. SUBAR:  Your Honor, we can certainly 

11    stipulate as measured by the post-enumeration survey results 

12    reported last July 15, there was a differential undercount. 

13               Beyond that at this point we are not prepared to 

14    stipulate anything further.  If that does the trick, that 

15    does the trick.  

16               MR. SOLOMON:  Your Honor, the point of this was 

17    to show you in a graphic way with respect to a redistricting 

18    plan --  

19               THE COURT:  That's just what I don't need is 

20    another graph.  

21               MR. SOLOMON:  But with respect to redistricting 

22    plans that have, in fact, will be been approved that there 

23    is a differential and it malapportioned these various ways 

24    of cutting up California.

25               If your Honor has heard enough, I would like to 
                                                              1211

 1    offer on behalf of plaintiffs PX 587, 88, 89 and 90, and if 

 2    those come in, I move on.  

 3               MR. SUBAR:  Subject to my earlier objection, I 

 4    guess it was, no other objection.  

 5               THE COURT:  They come in subject to that.

 6               (Plaintiff's Exhibits 587, 588, 589 and 590, 

 7    respectively, marked for identification were received in 

 8    evidence.) 

 9    BY MR. SOLOMON:

10         Q.    Dr. Cain, I would like to point you to a couple 

11    of statements in Secretary Mosbacher's report and ask you 

12    with respect to it as a political scientist to comment on 

13    them.  

14               THE COURT:  That is Plaintiff's Exhibit 9?  

15               MR. SOLOMON:  Yes, your Honor.

16               For the court's convenience, we have excerpted 

17    some of those pages in tab 9 of the book.  

18    BY MR. SOLOMON:

19         Q.    Look, please, at page 2-61.

20               The Secretary on page 2-61 says, "In the 

21    conclusion, the third sentence, "I am concerned that an 

22    adjustment would reduce state and local support for future 

23    censuses, adversely affect the department's ability to 

24    obtain appropriate funding for future censuses, adversely 

25    affect the quality of the work done in the future by 
                                                              1212

 1    temporary census enumerators," and what I would like you to 

 2    focus on is the statement of the Secretary that, "An 

 3    adjustment would reduce or adversely affect the department's 

 4    ability to obtain appropriate funding for future censuses." 

 5               That, by the way, Professor, is the same 

 6    statement that is made on 2-60 where the Secretary says, 

 7    "Tarrance and Wachter expressed concern that an adjustment 

 8    would adversely affect the department's ability to obtain 

 9    sufficient funding for future censuses.  I share this 

10    concern." 

11               Please comment on that.  

12               MR. SUBAR:  Objection, your Honor.  Again, this 

13    one really is considerably more substantive.  This line of 

14    questioning is beyond the scope of Dr. Cain's witness 

15    statement that we were provided in discovery.  

16               THE COURT:  I don't have that statement in front 

17    of me.  

18               MR. SOLOMON:  Your Honor, I can cut through that 

19    easily.

20               There is a statement, among other things, says 

21    that Professor Cain is going to testify to the political 

22    consequences of the Secretary's decision and that they were 

23    knowable prior to July 15, 1991.

24               After this, his deposition was also taken and he 

25    wasn't instructed not to answer any questions, and he was 
                                                              1213

 1    able to ask any question that Mr. Subar wanted.  

 2               The fact of the matter is we are laying the 

 3    predicate with these questions to demonstrate that, in fact, 

 4    a political consequence of his decisions were known and 

 5    knowable and, in fact, taken into account.  

 6               MR. SUBAR:  Your Honor, first of all, to correct 

 7    Mr. Solomon, there was at least one question I was not 

 8    permitted to get the answer to at Dr. Cain's deposition, I 

 9    don't recall if there were more.  

10               But the political consequences --  for your 

11    Honor's information, I happen to have an extra copy of Dr. 

12    Cain's witness statement.  I would hand it to your clerk if 

13    that would help matters.

14               (Handing to the court)

15               (Pause)

16               You your Honor will notice that there are three 

17    paragraphs that discuss the expected subject matter of Dr. 

18    Cain testimony.

19               The first reads that the census 

20    disproportionately undercounts people in areas in which 

21    voters are predominantly Democratic as compared to areas 

22    where voters are predominantly Republican.

23               Second is the decision not to correct the 1990 

24    decennial census results in underrepresentation in areas 

25    known to be Democratic.
                                                              1214

 1               The third isn't that Dr. Cain will testify about 

 2    the political consequence, the other political consequences 

 3    of the Secretary's decision, it's only that the political 

 4    consequences of the Secretary's decision were knowable prior 

 5    to July 15, 1991.

 6               There is no indication in here that Dr. Cain 

 7    would be testifying about any other political consequences 

 8    of the decision beyond those mentioned in the first two 

 9    paragraphs which focus strictly on party politics.  

10               THE COURT:  The motion, if that's what it is, is 

11    denied, and if it is an objection, it's overruled.  

12               MR. SUBAR:  Thank you.  

13    BY MR. SOLOMON:

14         Q.    Do you have the question, sir?

15         A.    Yes.  It's been a while.  

16               THE COURT:  What was it?

17         Q.    Does anybody else?

18         A.    The question was --  

19               THE COURT:  You tell us.

20         A.    The question was did I have anything to say about 

21    the Secretary's comment that the adjustment decision would 

22    reduce state and local support for future censuses, 

23    adversely affect the department's ability to obtain 

24    appropriate funding for future censuses, et cetera.

25         Q.    That's what I would like you to focus on?
                                                              1215

 1         A.    Yes.

 2               I will say that is purely speculation.  If you 

 3    were classifying things into what was knowable and what I 

 4    showed you before was part of the common knowledge of 

 5    anybody who had been following it, this would not be 

 6    knowable in anyway.  There is no political science, no 

 7    demonstrated empirical work that you could rely on, it is 

 8    simply a conjecture about how people are going to react some 

 9    time in the future, and --  

10               THE COURT:  I suppose the argument is, as a 

11    layman, which I assume the Secretary is, if you are going to 

12    start adjusting the initial raw counts don't mean as much as 

13    they used to, people aren't going to be as particular in 

14    funding those things because they know down the line it will 

15    all come out in the wash. 

16               THE WITNESS:  Well, that was not the experience 

17    in California.  In the City of Los Angeles, for example, 

18    they felt that it was extremely important to get the best 

19    enumeration they could, because that would be the baseline, 

20    and then they would argue about the adjustment.  

21               So that wasn't --  we have had an experience mof 

22    this.  We went through the 1990s with the expectation, up 

23    until 1987, when the Department of Commerce overruled the 

24    Census Bureau, there was an expectation that there would be 

25    a PES, that it would be seriously considered, and yet it did 
                                                              1216

 1    not get in the way of peoples' efforts to do the best they 

 2    could in outreach.  Insofar as we have any evidence on that 

 3    proposition, it goes against the statement of the Secretary.  

 4               MR. SUBAR:  Your Honor, I will have to move to 

 5    strike Dr. Cain's testimony regarding what peoples' 

 6    expectations were.

 7               If he is talking about the expectations of people 

 8    who were involved in planning for the 1990 census up to 

 9    1987, there is no foundation that Dr. Cain knows that.

10               If he is talking about the expectations of the 

11    quarter of a billion people in this country who the census 

12    attempted to count as they prepared for the 1990 census, 

13    which is what the Secretary's decision documents refer to 

14    here, there could be no foundation that Dr. Cain knows the 

15    expectations of all of us.  

16               MR. SOLOMON:  May I respond?  

17               THE COURT:  I have no idea of what you just told 

18    me.  

19               MR. SUBAR:  Your Honor, if I could clarify, then, 

20    maybe I will try to restate it.

21               Dr. Cain testified just a moment ago with regard 

22    to his understanding of the expectations of people as they 

23    prepared for the 1990 census.  He didn't make clear whether 

24    he was talking about the expectations of political leaders 

25    who were preparing to take and to help take the census --  
                                                              1217

 1               THE COURT:  That's what I think he is talking 

 2    about.

 3               Is it? 

 4               THE WITNESS:  Yes.  

 5               THE COURT:  People who are going to do the work 

 6    on the census? 

 7               THE WITNESS:  That's right.  

 8               THE COURT:  Not citizens? 

 9               THE WITNESS:  The general public we know from 

10    York and Gallup studies has very little knowledge of what is 

11    going on in the census.  

12               MR. SUBAR:  Then, your Honor, there is no 

13    foundation that Dr. Cain knows what all of the expectations 

14    of that political leadership might have been, and 

15    furthermore, Dr. Cain limited his testimony to the 

16    expectations of people up until 1987.  

17               THE COURT:  Did you? 

18               THE WITNESS:  No.  I mean, I haven't used that as 

19    an example.  

20               MR. SOLOMON:  Can I ask a few questions?  

21               THE COURT:  Yes, you might get us on the track.  

22               MR. SOLOMON:  Mr. Subar is welcome to 

23    cross-examine this expert.  We don't want to get to the 

24    point where we are just interfering with the witness' 

25    ability to give your Honor whatever information he has here.  
                                                              1218

 1               MR. SUBAR:  That wasn't my intent, your Honor?

 2         A.    I apologize.  

 3               THE COURT:  I'm sure not.  It was general 

 4    ineptness.  

 5               MR. SUBAR:  Thank you, your Honor.  

 6               MR. SOLOMON:  I am not going to comment on that 

 7    one.  

 8    BY MR. SOLOMON:

 9         Q.    Okay.

10               I think you have completed your answer with 

11    respect to the first quote that I gave you.

12               I would like you to turn to the second quote, 

13    which is just to get back to the subject that Mr. Subar 

14    wants you to talk about, apparently.  It's on page 2-61 as 

15    well.

16               I am reading at the top of the page, the second 

17    full sentence, "As Wachter recognized, adjustment may pose 

18    significant risk to the technical independence of the Census 

19    Bureau professionals who have traditionally been free from 

20    external influence in the implementation of their mission." 

21               Now, have you studied the Census Bureau's role 

22    within the Department of Commerce?

23         A.    Yes, that's correct, I have looked at how the 

24    decision was made.

25         Q.    And in connection with the work that you have 
                                                              1219

 1    done here, have you reviewed testimony and documents 

 2    concerning the Census Bureau's role?

 3         A.    That's correct, I have.

 4         Q.    The Secretary relies on Professor Wachter as a 

 5    demographer about the adjustment posing risks to the 

 6    technical independence of the Bureau's professionalism who 

 7    have traditionally been free from external influence.

 8               Do you agree with the Secretary?

 9         A.    No, that doesn't correspond with the facts.

10               I mean, the fact is that the Bureau had advised 

11    both in 87 and in 91 to go ahead with PES and it was, it was 

12    a decision by the Commerce Department, the Secretary, to 

13    stop that, so it was --  traditionally it has been the 

14    Census Bureau that made the decision.  The Census Bureau 

15    made the decision in 1980.

16               So I think he's got it wrong.  I think that the 

17    technical independence as outlined by Vincent Barabba, who 

18    is a former director of the census, is preserved through the 

19    doctrine delegation of allowing the Census Bureau to make 

20    the decision on purely technical and statistical grounds 

21    rather than letting the Commerce Department intervene and 

22    overturn its decision.

23         Q.    The Secretary makes reference to the mission of 

24    the Bureau of the census.

25               As a political scientist studying organizational 
                                                              1220

 1    structure of our government, do you have an understanding of 

 2    the mission of the Census Bureau?

 3         A.    I do.  

 4               MR. SUBAR:  Objection.  Your Honor, I --  

 5               THE COURT:  Sustained.

 6               It's about time you got one.

 7         Q.    You made a reference to Vincent Barabba in your 

 8    last answer.

 9               Who was Mr. Barabba?

10         A.    He is a former director of the Census Bureau.  In 

11    fact, twice director.

12         Q.    I'm sorry, you said twice?

13         A.    Yes.

14         Q.    In connection with the opinion that you just 

15    offered to the court, did you review anything by Director 

16    Barabba?

17         A.    Yes.  I have reviewed things that he has written, 

18    contributions that he has written and his testimony before 

19    Congress.

20         Q.    Would you look, please, at what we have marked 

21    for identification only as PX 726 in tab 726 in your binder.

22               If you would identify that document for us.

23         A.    That is a statement of Vincent Barabba before the 

24    U.S. House of Representatives Subcommittee on Census and 

25    Population Hearing.
                                                              1221

 1         Q.    Please describe to the court generally what 

 2    reliance you place on the statement of Director Barabba?

 3         A.    Well, in his testimony, indeed, in almost all his 

 4    writings, Vincent Barabba emphasizes that it's very 

 5    important if you are going to produce data which is reliable 

 6    and impartial and perceived as such that the Census Bureau 

 7    have an independence and an integrity and professionalism 

 8    and reputation for that and that is something he outlines in 

 9    his discussion.

10               And then he outlines in some detail how, starting 

11    in 1980, when this controversy over the undercount started 

12    to heat up, how they took steps to insure the independence 

13    and the perception that the decision was being made on 

14    statistical grounds, and so he describes how they did that 

15    and the features of that; that they had a very open process 

16    in which they worked with experts on the outside, had 

17    undercount workshops, consulted them, exchanged informations 

18    with them; they relied heavily on information and data about 

19    the reliability of the adjustment, that is, that it was 

20    something that was an empirical assessment or a scientific 

21    assessment, that is, they would have ideas and then they 

22    would go out and test those; and that it would be evaluated 

23    as a statistical and technical issue by the Census Bureau 

24    and that these were the key features.  

25               MR. SUBAR:  Your Honor, I am going to have to 
                                                              1222

 1    move to have this response stricken.

 2               The question was to what degree did Dr. Cain rely 

 3    on this document.

 4               We don't have an objection to that question and 

 5    we wouldn't have an objection to the answer had the question 

 6    been answered.

 7               The reason that I am moving to strike is because 

 8    Dr. Cain has been testifying as to the contents of the 

 9    document.  The document hasn't been authenticated and it 

10    contains hearsay, and in that respect we object to the 

11    introduction of the substance of the document as opposed to 

12    testimony concerning distribution reliance on it.  

13               MR. SOLOMON:  May I respond, your Honor?  

14               THE COURT:  Well, let me do this backwards.

15               Why don't we get the document into evidence and, 

16    thus, moot his objection.  

17               MR. SUBAR:  Your Honor, I will object to the 

18    introduction of the document.

19               Dr. Cain might be able to testify about his 

20    reliance on it, but the document itself is hearsay and 

21    unauthenticated and, therefore, objectionable.  

22               THE COURT:  If it's taken even for his reliance 

23    on it I have to know the contents thereof and the contents 

24    thereof are precisely what he is telling us.  

25               MR. SUBAR:  If that is your ruling, to which we 
                                                              1223

 1    take an exception to it, that is your ruling, your Honor.  

 2               THE COURT:  If it's an exception it's overruled.

 3               Go ahead.  

 4    BY MR. SOLOMON:

 5         Q.    As a user of census data yourself, do you have an 

 6    opinion about whether the role played by the Census Bureau, 

 7    the openness, the professionalism, the scientific approach 

 8    that was taken that you just described your reliance on Dr. 

 9    Barabba concerning that, is that relevant to you as a user 

10    of census data?

11         A.    Oh, absolutely, as a user of census data who 

12    intends to publish articles and referee journals and my 

13    reputation would depend on those articles, as a person who 

14    sometimes works with the media in developing stories about 

15    what is happening inside California using the census data, 

16    as a person who uses this data to try to advice school 

17    districts and cities and counties as to how to handle the 

18    very difficult problems of minority representation, it's 

19    absolutely essential that the data be perceived as objective 

20    and unbiased and comes from a source which is objective and 

21    unbiased.

22         Q.    Since the events of 1987 that you did make a 

23    brief reference to, has your view concerning the data coming 

24    from the Bureau of the Census changed?

25         A.    Yes.
                                                              1224

 1               The outline that Vincent Barabba had of how to 

 2    have a model of an agency that could produce data that we 

 3    could believe in was a very good one.  It's very 

 4    consistently with scholarly principles.  We do the same 

 5    thing with political science.  

 6               There are a number of people who produce polls in 

 7    political science who are associated with campaigns or 

 8    associated with particular causes and the political science 

 9    profession conducts through the University of Michigan a 

10    national election study every two years and it follows many 

11    of the same procedures that Mr. Barabba points out, namely, 

12    that it has open consultations as to how to collect that 

13    data, it is done on a scientific basis, it is not done to 

14    prove a particular point for the Democratic party or the 

15    Republican party or for particular groups.

16               That model of how to do social science and then 

17    kind of unperturbed, unbiased objective fashion is the model 

18    that Vincent Barreber outlined, and since that time we have 

19    a departure from that outline, we no longer observe the 

20    principle of delegation to the Census Bureau, the decision 

21    is being made in ways which is more mysterious and appears 

22    to be more politicized.

23         Q.    In your last answer you say that the process 

24    appeared to be more politicized.

25               Are you aware of the special advisory panel that 
                                                              1225

 1    was appointed pursuant to a stipulation and order entered by 

 2    this court?

 3         A.    Yes, I am.

 4         Q.    Are you aware of of the personnel that were 

 5    appointed by the Secretary on to that panel?

 6         A.    Yes, I am aware.

 7         Q.    Was one of the people Lance Tarrance?

 8         A.    Yes, that's correct.

 9         Q.    Tell the court who Mr. Tarrance is, please.

10         A.    Mr. Tarrance is a Republican polster.  

11               MR. SUBAR:  Objection; relevant.  

12               THE COURT:  Overruled.

13         Q.    What do you mean by a polster?

14         A.    Well, again, I was alluding to this before.

15               There are a number of people who collect polls 

16    and there is a class of individuals that collect polls for 

17    candidates or for specific causes, and in that world they 

18    can only work for one particular party because there would 

19    be no trust of them if they tried to work both sides of the 

20    aisle, so to speak, and Lance Tarrance is one who polls 

21    primarily for Republican candidates.

22         Q.    I would like to read to you from page 47 of the 

23    deposition of Mr. Tarrance and see whether that doesn't 

24    confirm your view.  

25               MR. SUBAR:  Your Honor, this is hearsay.  
                                                              1226

 1    Moreover, the depositions of a number of people, including 

 2    Mr. Tarrance, are going to be introduced into evidence 

 3    pursuant to an understanding between the parties and the 

 4    court that one side have an opportunity to present what they 

 5    want in terms of the deposition excerpts and the other side 

 6    will have an opportunity to cross-examine.  

 7               THE COURT:  The longer you talk the more confused 

 8    I get.

 9               I think you were right when you started, so stop 

10    there.  

11               MR. SUBAR:  Then I will stop.  

12               THE COURT:  Sustained.  

13               MR. SUBAR:  Thank you.  

14               THE COURT:  I don't want him commenting on 

15    somebody else's deposition.  I can read the deposition.  

16    Whether he believes it or not, how he makes it out to be, 

17    it's my business, not his.   

18               No offense. 

19               THE WITNESS:  No offense taken.  

20               MR. SOLOMON:  It is not so much for his view of 

21    the deposition, but, rather, one of the sources of his 

22    opinion that I was going to use this for, but I will move 

23    on.  

24    BY MR. SOLOMON:

25         Q.    Do you know whether Mr. McGehee is also appointed 
                                                              1227

 1    by the Secretary as one of the four special advisory panel 

 2    members?

 3         A.    That's correct.

 4         Q.    Tell the court who Mr. McGehee is?

 5         A.    Well, Mr. McGehee is a graduate of the Claremont, 

 6    now called Claremont McKenna College, which is a school that 

 7    specializes in social science, economics and politics, and 

 8    is home to the Rose Institute, which is one of the premier 

 9    redistricting units in the country.

10         Q.    Do you know whether Mr. McGehee has done any work 

11    for any party other than the Republican party?

12         A.    I know of --  I know that the Rose Institute 

13    essentially works for the Republican party, and so it's 

14    quite consistent with what the fact that he would not have 

15    worked with anybody other than Republicans.

16         Q.    To a political scientist, what did the 

17    appointment by the Secretary of Messrs. Tarrance and McGehee 

18    signify to you, if anything?

19         A.    Well, it signifies a departure from the notion 

20    that this would be a decision made on statistical grounds.  

21    These are not individuals who profess to have any 

22    statistical expertise.  So if nothing else, it means that 

23    these are people that are going to make the judgment on some 

24    other basis.

25         Q.    I would like you to turn to what I hope is in 
                                                              1228

 1    your binder as tab 678 where we have marked for 

 2    identification PX 678.  It is a computer printout, for the 

 3    record, of the press conference and the press statement made 

 4    by Secretary Mosbacher when he announced his decision.

 5               Did you hear the press conference, Professor 

 6    Cain?

 7         A.    Yes, I did.

 8         Q.    Have you since listened to a tape of it and tried 

 9    to compare whether some of the statements that are in here 

10    are the same as what Mr. Mosbacher said?

11         A.    Yes, I did.

12         Q.    I would like you to turn to page 6.  I would like 

13    to read you a question and answer and ask you to comment on 

14    it.  It's in the middle of the page.  

15               The question is, "Some of the people who favored 

16    adjustment are bound to say that politics did play a part in 

17    your decision, because if you adjust you would be counting 

18    more minorities and Hispanics who tend to vote Democratic.

19               What is your response to that criticism? 

20               "Secretary Mosbacher:  First of all, it was not 

21    political.  If you look at the parties, it's pretty hard to 

22    see whether there is any measure gain for either party.  

23               You know, I've heard arguments that perhaps 

24    Arizona and California, which were the two that, under the 

25    adjustment, the figures we saw, would have gained and 
                                                              1229

 1    Pennsylvania and Wisconson would have lost, certainly in 

 2    Arizona there is no doubt it is higher percentile 

 3    Republicans, but politics really isn't what it's about and, 

 4    unfortunately, it coming out so closely even it didn't even 

 5    appear important." 

 6               I would like your to begin with respect to the 

 7    statement by the Secretary that it's pretty hard to see 

 8    where there is any measure gain for either party?

 9         A.    Well, for the same reason that we were almost 

10    embarrassed by the obviousness of tables that we presented, 

11    it would be obvious to the Secretary that the gains go far 

12    beyond the apportionment of the last two congessional 

13    districts.

14               As we said before, there are 39,000 jurisdictions 

15    out there, not all of them do elections, but many of them 

16    do, and as we saw from the table, there were real effects 

17    with respect to the representation of minority groups within 

18    states, within cities, within counties, within water 

19    districts, within special districts; within all those 

20    jurisdictions there were real effects, so they are 

21    multipliers, if you like, by all the number of districts.

22               So merely focusing on the apportionment of the 

23    last two districts, surely it would be part of the common 

24    knowledge of anybody with a moderate degree of 

25    sophistication that because the differential undercount 
                                                              1230

 1    there would be differential malapportionment, that having 

 2    more people is better than having less people with respect 

 3    to redistricting.

 4               That's all the Secretary would have to know to 

 5    realize the affect is far greater than the apportionment 

 6    than to one state as opposed to another. 

 7    

 8    

 9               (Continued on the next page) 

10    

11    

12    

13    

14    

15    

16    

17    

18    

19    

20    

21    

22    

23    

24    

25    
                                                              1231

 1         Q.    You spoke in your last answer of apportionment 

 2    and malapportionment.  My question is whether it was equally 

 3    well known that there was a partisan aspect to the decision.

 4         A.    Yes.  

 5               MR. SUBAR:  Objection to equally well known.  

 6               THE COURT:  Let him clarify that.

 7         Q.    You said it was very well known to anybody with a 

 8    moderate degree of political sophistication, I think is what 

 9    you said.  Let's do that.  In your opinion, did Secretary 

10    Mosbacher have a moderate degree of political 

11    sophistication?  

12               MR. SUBAR:  Objection.  Speculative.

13               THE COURT:  Sustained.

14         Q.    Tell the Court who Secretary Mosbacher is?

15         A.    Secretary Mosbacher had previously been involved 

16    in President Bush's presidential compaign and is now 

17    currently heading his presidential reelection campaign.  So 

18    this is a person who would know a fair amount about politics 

19    and in particular the political facts that would connect the 

20    racial malapportionment to the partisan reapportionment.

21         Q.    The question that I would then pose to you is to 

22    someone with the degree of political sophistication that you 

23    have just described Secretary Mosbacher having, would it 

24    have been obvious that the adjustment decision would have a 

25    measure of gain for one party or another?  
                                                              1232

 1               MR. SUBAR:  Objection.  That is still 

 2    speculative.  

 3               THE COURT:  Overruled.  You may answer. 

 4         A.    Yes.  Because the groups we are speaking of, 

 5    Latino and African-American communities, are widely known to 

 6    be disproportionately associated with the Democratic Party 

 7    rather than the Republican Party.

 8         Q.    What is the basis for the statement that you just 

 9    made?

10         A.    There are many.  Casual observation, distribution 

11    of Latino and black elected officials is one confirmation.  

12    Voting behavior in the districts, Congressional districts, 

13    and assembly and Senate districts which are overwhelming 

14    minority would be a second.  

15               Academic studies.  If you ever had a political 

16    science course and you were exposed to the American voter or 

17    the changing American voter, which is a classic in college, 

18    would have taught you that there is a strong relationship 

19    between the African-American community and the Democratic 

20    Party.  

21               Commercial polls, field polls, any number of 

22    studies of the primaries that are going on right now.  All 

23    these things show that there is a partisan predisposition on 

24    the part of these groups towards the Democratic Party.

25         Q.    Do you have an estimate that you feel comfortable 
                                                              1233

 1    with of the percentage of Democratic voters among blacks?

 2         A.    Typically, we find about an 8 or 9 to 1 ratio.  

 3    That is to say, in black precincts or African-American 

 4    precincts, you are finding registration rates or voting 

 5    rates of about 80 or 90 percent for the Democratic Party.

 6         Q.    80 or 90 ratio meaning 8 to 1 or 9 to 1?

 7         A.    Right.

 8         Q.    You talked about subdivisions of political 

 9    jurisdictions other than the states.  Do you have any 

10    information concerning whether it has been a stated purpose 

11    of the Republican Party to enlarge its presence with respect 

12    to substate and legislative houses?

13         A.    There is no question that -- 

14               MR. SUBAR:  Objection.  Hearsay.  

15               THE COURT:  Overruled.

16         A.    There is a community of scholars and people 

17    involved in redistricting that travel around the country.  

18    At all of those conferences the representatives from the 

19    Republican National Committee, they have a subsection for 

20    redistricting, have talked about how they feel that the 

21    progress of the Republican Party in state legislatures in 

22    gaining control has been slow and held back by previous 

23    redistricting, how getting control of the state legislatures 

24    and getting control of the governorships is absolutely 

25    critical if they were going to turn around their fortunes in 
                                                              1234

 1    Congressional races.  

 2               That would be essential if they were going to end 

 3    the divided government situation that they found themselves 

 4    in for the last quarter century or so.  So there is no 

 5    question that a primary goal of the Republican Party was to 

 6    get the best redistricting they could at the state level, 

 7    because redistricting is done, after the apportionment 

 8    phase, the redistricting is done at the state or local 

 9    level.

10         Q.    Was that fact that you just testified to 

11    something that would have been known to someone of moderate 

12    political sophistication?

13         A.    Unquestionably.

14         Q.    Have you heard Vice President Quayle make 

15    speeches concerning the importance of capturing state 

16    houses?

17         A.    Yes.  Particularly in California, where the 

18    redistricting was considered very critical, there were a 

19    number of fund raising events in the spring preceding the 

20    redistricting in which be both President Bush and Vice 

21    President Quayle articulated the need for a better 

22    apportionment on the part of the Republican Party.

23         Q.    Did you do any quantitative analysis to test the 

24    proposition that the Secretary's statements that there was 

25    no measure of gain for either party was true or false?
                                                              1235

 1         A.    Yes.  Once again, although it seems obvious, we 

 2    did conduct such a study.

 3         Q.    Could you explain at the Court how it was done?

 4         A.    It was essentially analogous to the analysis that 

 5    we did previously with the percent minority.  We aligned the 

 6    districts with respect to their party registration and their 

 7    undercount, and we analyzed the relationship between the 

 8    percent Democrat and the percent undercount to see whether 

 9    the Democratic areas were more substantially undercounted 

10    than the Republican areas.

11         Q.    Did CRA perform those analyses for you?

12         A.    That's correct.

13         Q.    Would you look, please, at what we have marked as 

14    PX-583, 584, 585, and 586, and explain those to the Judge.  

15    These are the last charts.

16         A.    Yes.  Unfortunately, more charts.  I apologize, 

17    but they are in color.  

18               Once again, we have produced scatter plots where 

19    the points on the plots represent the different 

20    Congressional, assembly or state senate districts.  We again 

21    classified them with triangles, dots, and squares, with the 

22    triangles representing those that have the lowest percent 

23    Democratic registration, the dots representing the highest 

24    20 percent Democratic registration, and the squares 

25    representing those in the intermediate ranges.  
                                                              1236

 1               Once again, we have fitted a regression line and 

 2    report statistics.  Once again, the statistics of that 

 3    regression and also tested the differences in the means 

 4    between the top and the bottom group indicate a high degree 

 5    of significance, which is to say that we can be fairly 

 6    confident that this relationship was not -- not fairly -- 

 7    extremely confident that this relationship would not have 

 8    produced by chance.

 9         Q.    Does the conclusion that you just drew, and I 

10    just want to make sure we understand each of the charts, 

11    apply to the Congressional districts in California?  That is 

12    583; and the California assembly districts, and that is what 

13    you are showing in 584; the senatorial districts, that is 

14    585; and then the percentile block analysis that you 

15    described to the Court with respect to race and ethnicity, 

16    for all blocks in California?

17         A.    That is correct.

18         Q.    Did you review this work by CRA to satisfy 

19    yourself that it was done correctly?

20         A.    Yes, I did.  

21               MR. SOLOMON:  We would offer PX-583, 584, 585, 

22    and 586.  

23               MR. SUBAR:  No objection, your Honor.  

24               THE COURT:  These are the last charts?  

25               THE WITNESS:  Not quite.  
                                                              1237

 1               THE COURT:  They are admitted. 

 2               (Plaintiff's Exhibits 583, 584, 585, and 586 for 

 3    identification were received in evidence).

 4         Q.    Is the phenomenon that you have just described, 

 5    that is, the obvious relationship between percent undercount 

 6    and percent Democrat, unique to California?

 7         A.    No, it is not.

 8         Q.    What numerical work did you have done to test 

 9    that view of yours?

10         A.    Unfortunately, we couldn't find at the time 

11    approved plans for some of the states we wanted to look at.  

12    So as an alternative what we did was take the 1980 

13    Congressional districts in New York and California and 

14    compute the percent undercount of those districts.  So we 

15    are applying 1990 adjusted and unadjusted data within the 

16    boundaries of the 1980 Congressional districts for 

17    California and New York.

18         Q.    In your view, was that an appropriate way to try 

19    to test your hypothesis?

20         A.    Once again, we can see whether the proportionate 

21    increase in undercount is greater in the districts that are 

22    more Democratic.  In this particular case we didn't have 

23    party registration data, so we take advantage of that to 

24    provide a second kind of measure of Democraticness versus 

25    Republicanness.  We classified it by the party of the 
                                                              1238

 1    representative who happens to be in that seat.

 2         Q.    That is a second way of testing the hypothesis of 

 3    Democratic versus Republican?

 4         A.    Right.  Instead of using party registration, look 

 5    at the partisanship of the person who is in office.

 6         Q.    What states was that done for?

 7         A.    That was done for New York and California.

 8         Q.    Did the results of that study confirm your view?

 9         A.    Yes.  Once again, disproportionately, the 

10    districts that had the highest degree of undercount were 

11    represented by Democrats, and the districts that had the 

12    lowest degree of undercount were represented in Congress by 

13    Republicans.

14         Q.    Are you aware, Professor Cain, that the 

15    Department of Commerce refused to release the 100 percent 

16    data, the adjusted data, for the use by census bureau data 

17    users?

18         A.    Yes, I am aware of that.

19         Q.    Is that fact relevant to your opinion that there 

20    is a partisan aspect to the Secretary's decision?

21         A.    It is certainly consistent with the hypothesis 

22    that if you didn't want states to use data that might 

23    potentially be damaging to the Democratic Party, they 

24    wouldn't have it.

25         Q.    I would like you to turn back to PX-678, which is 
                                                              1239

 1    the press conference, and I would like to read you another 

 2    question and answer and ask you some questions.  This 

 3    appears on page 7 of the press statement PX-678.

 4         A.    Did you say page 7?

 5         Q.    Page 7 to 8, right.  

 6               "Question:  Mr. Secretary, Mike Posner of 

 7    Reuters.  Did the White House play any role in this decision 

 8    at all?  You mentioned all the people you discussed the 

 9    situation with.  You failed to mention anything about the 

10    White House.  Did you contact the White House?  Did they 

11    approve it or veto the decision?  

12               "Secretary Mosbacher:  No, I did not contact the 

13    White House.  The White House had no input on this.  And I 

14    neither talked to the President nor anyone else at the White 

15    House during this decision-making period about this at all.  

16    Nor did they seek to have any input in it."  

17               Do you see that, sir?

18         A.    Yes, I do.

19         Q.    On the basis of what you know, is that a true 

20    statement?  

21               MR. SUBAR:  Objection.  Foundation.  

22               THE COURT:  On the basis of what you know.  The 

23    objection is overruled.

24         A.    First of all, it is so obvious you didn't need to 

25    be contacted by the White House.  But on the basis of what 
                                                              1240

 1    we know, there were contacts between the White House and the 

 2    commerce department in both directions.  

 3               MR. SOLOMON:  This is the subject of deposition 

 4    testimony which we will submit not through this witness but 

 5    at an appropriate time, your Honor.  I would like to ask 

 6    this witness some questions about that deposition testimony.  

 7    It would be very brief.  

 8         Q.    Who is Governor Sununu?  

 9               MR. SUBAR:  Objection, your Honor.

10               MR. SOLOMON:  You haven't heard the question yet.  

11               THE COURT:  Who is Governor Sununu.  He is a 

12    little -- 

13               MR. SUBAR:  Although he hadn't asked the question 

14    by the time I said "objection," I am objecting to this line 

15    of questioning on the same grounds that I objected with 

16    regard to deposition matters earlier.  

17               THE COURT:  Overruled.  The question is, who is 

18    Mr. Sununu?

19         A.    He was a former governor of New Hampshire who at 

20    the time was chief of staff for Bush in the White House.

21         Q.    Mr. Sununu was at the White House?

22         A.    That's correct.

23         Q.    Are you aware of whether Governor Sununu has any 

24    particular expertise in statistics?

25         A.    I do not know of any expertise that he has.
                                                              1241

 1         Q.    Are you aware as a political scientist of anyone 

 2    contacting Governor Sununu or being contacted by Governor 

 3    Sununu for purposes of securing advice on matters of 

 4    statistics?

 5         A.    His name doesn't come to mind, no.

 6         Q.    If I asked you to assume from the deposition 

 7    testimony that Governor Sununu told Mr. Willkie, the third 

 8    highest ranking officer at the commerce department, that in 

 9    his view adjustment would cause a major disruption in 

10    reapportionment, would that be consistent with your view of 

11    the obviousness of the proposition that politics was 

12    involved?  

13               MR. SUBAR:  Objection. 

14         A.    He would certainly be qualified -- 

15               THE COURT:  So obvious that I don't need to hear 

16    it.  Sustained.  

17               MR. SUBAR:  Thank you.

18         Q.    I would like to have you focus on two more 

19    comments by the Secretary, and I think we are then done.  If 

20    you turn back to Tab 9, which was the report of the 

21    Secretary ostensibly, pages 2-61 and 2-62.  This is the 

22    paragraph that I was reading to you before.  In part, the 

23    Secretary there says, "The evidence indicates that the 

24    controversy over adjustment is likely to have a negative 

25    effect on future censuses, regardless of the outcome of the 
                                                              1242

 1    adjustment decision."  

 2               From the standpoint of a political scientist, do 

 3    you agree with that?

 4         A.    The evidence that the controversy itself is going 

 5    to have a negative effect on future censuses is essentially 

 6    there is no strong evidence.  If you look at the opinion 

 7    polls that have been conducted, they indicated, if anything, 

 8    there is a slight preference on the public to adjust as 

 9    opposed to not adjust the count.  But when it comes right 

10    down to it, nobody is willing to say that this is likely to 

11    have much of an effect in the future on participation.  

12               One thing I do note is that the studies do show 

13    that there is a much stronger preference for adjustment on 

14    the part of the undercounted groups.  So what we would have 

15    to be careful of is if we don't adjust, we may compound the 

16    problem of getting participation from the -- we might 

17    compound it.  There might be an extra cynicism about the 

18    process if there isn't an adjustment.  But all of this is 

19    highly speculative.  The study that York and Gallup shows is 

20    that the preferences about the census are fairly weak.

21         Q.    Are you aware of any evidence at all that 

22    supports the Secretary's position?

23         A.    No.  

24               MR. SOLOMON:  I think the word the witness used 

25    was "participation," your Honor.  
                                                              1243

 1               THE COURT:  Right.  I figured that out.

 2         Q.    I would like you next to turn to the first page 

 3    of these excerpts that I have given you in Tab 9 here, page 

 4    1-6 of the Secretary's opinion.  The Secretary says -- 

 5               MR. SUBAR:  I'm sorry.  Page 1-6 at Tab 9?  It is 

 6    not in my volume unless it is out of order.  

 7               THE COURT:  There are no 1s in there.  

 8               MR. SOLOMON:  Let me just read it.  We can save 

 9    time.

10         Q.    The second says, "An adjusted set of numbers will 

11    certainly disrupt the political process and may create 

12    paralysis in the states that are working on redistricting or 

13    have completed it."  

14               THE COURT:  What was that word, create what?  

15               MR. SOLOMON:  "Paralysis in the states that are 

16    working on redistricting or have completed it."

17         Q.    You spend a great deal of time on the 

18    redistricting process, don't you, sir? 

19         A.    Yes, that is correct.

20         Q.    You, in fact, perform the computer analyses from 

21    time to time or direct people to perform the computer 

22    analyses that are necessary for redistricting, correct?

23         A.    That's correct.

24         Q.    Do you agree with the Secretary's statement?

25         A.    No.  The fact of the matter is that because we 
                                                              1244

 1    live in the era of modern computers, and everything is on 

 2    the computer, it is very easy to make adjustments in plans 

 3    in a short period of time.  To take an example, we have 

 4    close to four million people in Los Angeles city.  We are 

 5    currently doing redistricting alternatives.  

 6               We can do a redistricting alternative from 

 7    scratch in an evening.  We can make changes between dinner 

 8    breaks, a dinner break and going out for dessert.  So it is 

 9    not a technical difficulty.  You could easily make these 

10    adjustments in a timely period.  

11               Moreover, the extent of change that we are 

12    talking about is substantially less than, say, the kinds of 

13    malapportionment that states faced correcting in the 1960s.  

14    Certainly, California had a much more substantial problem, 

15    bigger discrepancies that they had to deal with, where you 

16    had a state senate that was based totally on counties.  

17               So there is an understanding in the redistricting 

18    world that you have to make changes.  Quite often there are 

19    threats of litigation.  So disruption is kind of the norm.  

20    Dealing with disruption is something that these people 

21    thrive on and can easily handle.  

22               MR. SOLOMON:  Thank you very much.  No further 

23    questions.  

24               THE COURT:  Mr. Subar?  

25               MR. SUBAR:  Thank you, your Honor.    Your Honor, 
                                                              1245

 1    perhaps I could take just a moment to hand you a binder that 

 2    I may be asking you to refer to, and to hand one to the 

 3    witness as well.  

 4    CROSS-EXAMINATION 
                        

 5    BY MR. SUBAR:

 6         Q.    Professor Cain, as a California resident, you 

 7    feel that you have a personal stake in the adjustment 

 8    controversy, don't you?

 9         A.    I as a resident of California, I would only want 

10    the seat if we properly do deserve the seat.  If the data is 

11    biased, then I feel, as a resident of the state of 

12    California, the data should be corrected.  And if we deserve 

13    to have the allocation, then we should get the allocation.  

14    I don't intend to represent that seat, so there is no 

15    personal stake.  I think only an idiot would run for 

16    Congress.

17         Q.    That is your opinion as a political scientist?

18         A.    I think it is an opinion that most Americans have 

19    these days.

20         Q.    You don't know, do you, that Secretary Mosbacher 

21    considered the political implications of the adjustment as 

22    an actual fact?

23               THE COURT:  It is an inference on your part?

24         A.    It is an inference based on the fact that these 

25    things are obvious, that is correct.
                                                              1246

 1         Q.    For all you know, he made efforts to isolate 

 2    himself from political forces, is that correct?

 3         A.    I don't know.  Once you know something, I don't 

 4    know how you unknow it.  But he may have tried to unknow 

 5    things.

 6         Q.    Professor Cain, are you aware of the fact that 

 7    the director of the census bureau is a political appointee?

 8         A.    That is correct.  

 9               THE COURT:  What does that mean?  

10               MR. SUBAR:  That she is a political appointee?  

11               THE COURT:  Yes.  

12               MR. SUBAR:  It means -- your Honor could refer to 

13    Defendant's Exhibit 701, which is at Tab 1 in the notebook 

14    that I handed the Court and the witness, which is 13 U.S. 

15    Code Section 21.  That provides, your Honor, that the bureau 

16    shall be headed by a Director of the Census appointed by the 

17    President by and with the advice and consent of the Senate.  

18               THE COURT:  That is all you mean by "political 

19    appointee"?  

20               MR. SUBAR:  That is all I mean by "political 

21    appointee."

22         Q.    Dr. Cain, are you aware of the fact that the 

23    director of the census bureau Barbara Bryant is a 

24    Republican?

25         A.    I am not aware of her party identification.
                                                              1247

 1         Q.    Could you flip to Tab 2 of the notebook in front 

 2    of you.  

 3               MR. SUBAR:  Your Honor, that document is an 

 4    excerpt from a periodical publication called Who's Who of 

 5    American Women, which we offer into evidence.  

 6               THE COURT:  Which one is she on?  

 7               MR. SUBAR:  She is on the second page of the 

 8    exhibit, second column, third full paragraph down, Barbara 

 9    Everett Bryant.  

10               THE COURT:  It is admitted as Defendant's Exhibit 

11    702. 

12               (Defendant's Exhibit 702 for identification was 

13    received in evidence) 

14               MR. SUBAR:  Thank you.  

15               MR. SOLOMON:  Your Honor, if we are going to talk 

16    about Director Bryant, my first request is why don't we 

17    bring her here to testify.  

18               THE COURT:  I take it he is offering this to show 

19    that she is a Republican.  

20               MR. SUBAR:  That is correct, your Honor.  

21               MR. SOLOMON:  Your Honor, I don't have a problem 

22    with the authenticity of this document.  I have a problem 

23    that that is what the document says.  Whether or not the 

24    document is right, this witness can't know, Mr. Subar can't 

25    know.  Director Bryant can know.  
                                                              1248

 1               THE COURT:  Cannot?  

 2               MR. SOLOMON:  Can know.  I would object to the 

 3    introduction of this document for just that reason.  

 4               THE COURT:  The objection is overruled.  It is a 

 5    commercial list.  It comes right under the classic 

 6    illustration of a hearsay exception.  Martindale Hubbell is 

 7    admissible, a telephone book is admissible.  Who's Who in 

 8    America is admissible.  

 9               MR. SOLOMON:  I am not about to try to teach your 

10    Honor any evidence.  I would simply observe that when those 

11    come in, they come in for what is written in them, not for 

12    the truth of the statements.  

13               THE COURT:  They come in for the truth because 

14    they are constantly relied upon by people in the 

15    professions.  

16               MR. SUBAR:  Thank you, your Honor.  

17               THE COURT:  I'll send you a bill for that 

18    lecture.

19               MR. ZIMROTH:  And we'll pay you just as we have 

20    everyone else.

21         Q.    Dr. Cain, are you aware of the fact that the 

22    plaintiffs in this case stipulated that the Secretary of 

23    Commerce could make the decision as to whether or not the 

24    results of the 1990 census would be adjusted?

25         A.    That's always true.  It just historically usually 
                                                              1249

 1    has never been true.  That is, historically it has been 

 2    delegated, for the reasons I outlined.

 3         Q.    Certainly, there is no question, is there, that 

 4    the Secretary of Commerce is a political appointee, correct?

 5         A.    That's correct.  And in a more expanded sense, 

 6    too.

 7         Q.    So any Secretary of Commerce who might get 

 8    involved in the decision-making process with respect to 

 9    adjustment could be subject to political pressure, whether 

10    that person be a Democrat or a Republican, is that right?

11         A.    Unquestionably, yes.

12         Q.    You are not a demographer, is that correct?

13         A.    I am a political scientist who uses demographic 

14    information.

15         Q.    You are not an expert on some of the -- many of 

16    the statistical aspects of the adjustment controversy, is 

17    that correct?

18         A.    I am not an expert on the PES, no.

19         Q.    Your opinion in this case that you expressed in 

20    your direct testimony is based on the assumption that the 

21    adjustment process embodied in the post-enumeration survey 

22    makes sense, isn't that correct?

23         A.    Well, I put it differently.  I said that the 

24    process as set up by Vincent Barabba and Courtney Slater in 

25    the 1980s was a process designed to insulate it from 
                                                              1250

 1    appearance and reality and politicization, through the 

 2    doctrine of delegation, the doctrine of openness, and the 

 3    doctrine of expertise.

 4         Q.    You don't know, though, whether adjustment makes 

 5    sense as a technical matter, do you?

 6         A.    I know that adjustment does correct the 

 7    historical bias in the data.

 8         Q.    But you don't know whether the process embodied 

 9    in the post-enumeration survey makes sense as a statistical 

10    matter, do you?

11         A.    I can't comment on smoothing or loss functions or 

12    any of these things, no.

13         Q.    You don't know whether the numerical results of 

14    the post-enumeration survey are closer to the truth than the 

15    results of the census enumeration, do you?

16         A.    We know that the aggregate numerical numbers are 

17    closer to the truth because almost everybody, indeed 

18    everybody, has conceded what the differential undercount is 

19    in the aggregate.

20         Q.    But you don't know whether the lower level 

21    numbers are closer to the truth?

22         A.    Actually, I do have something that I can offer in 

23    assistance to you.  That is that if you look at our charts, 

24    if it were the case that this distributional error that they 

25    are talking about was introducing an enormous amount of 
                                                              1251

 1    randomness in the variance, then you would see that scatter 

 2    of points.  It would be much wider than you see.  That is, 

 3    you would see some districts that didn't have a large 

 4    percentage of minority that had high degrees of undercount 

 5    or vice versa because of the random noise that was 

 6    introduced.  

 7               So what I can tell you is that from the aggregate 

 8    level, looking at the representational effects when all the 

 9    blocks net themselves out, that the data looks entirely 

10    plausible.

11         Q.    We will get to that later, Dr. Cain.  For right 

12    now, don't you recall giving the following responses to 

13    questions that I asked you at your deposition?  This is at 

14    page 67, line 14.

15              "Q.    Okay.  But do you have any opinion 

16    whatsoever as to which set of numbers, the adjusted numbers 

17    or the unadjusted numbers, is closer to the truth?

18              "A.    I'm not an expert in adjustment.

19              "Q.    Okay.  Do I take it that you don't have an 

20    opinion on that?

21              "A.    I'm not comfortable giving an opinion on 

22    that."  

23               Does that refresh your recollection?

24         A.    Yes, you have refreshed my recollection.

25         Q.    If adjustment doesn't make sense as a technical 
                                                              1252

 1    matter, your views that you expressed during your direct 

 2    testimony would change, isn't that correct?

 3         A.    If the vast majority of the panels and everybody 

 4    that examined it of experts concluded that adjustment didn't 

 5    make sense, then I would be inclined to believe that it 

 6    didn't make sense.

 7         Q.    The point of your direct testimony was, was it 

 8    not, that the adjustment decision has political implications 

 9    that are knowable?  Is that a fair statement?

10         A.    Yes.  These political implications were knowable 

11    and obvious.

12         Q.    In other words, what you are saying is the groups 

13    that tended to vote Democratic were undercounted, in your 

14    view, is that correct?

15         A.    Among other political effects.  The most 

16    important one, of course, is that the groups themselves are 

17    malapportioned.  But, in addition, the Democrats were 

18    malapportioned, and now it is probably the motive for 

19    malapportioning of minority groups.

20         Q.    Certainly political groups, the Democratic Party, 

21    Republican Party, are nonracial groups; you would agree with 

22    that, wouldn't you?

23         A.    No.  I mean it is precisely because the 

24    Democratic and Republican Party tends to be aligned along 

25    racial lines and ethnic lines that you observe the pattern 
                                                              1253

 1    in the data that you do.  The two points are intimately 

 2    connected.

 3         Q.    But certainly the Democratic Party per se 

 4    wouldn't be defined as a racial group; it's a political 

 5    party, right?

 6         A.    Actually, the Democratic Party has been 

 7    struggling with precisely that issue.  That is exactly why, 

 8    when Paul Kirk took over the Democratic Party, he tried to 

 9    abolish the Latino caucus and the African-American caucus 

10    and the Women's caucus.  It was precisely to loosen that 

11    identification between the groups.

12         Q.    In your opinion, it is your opinion, isn't it, 

13    that it is the job of the political system, not the courts, 

14    to accommodate the demands that nonracial groups have for 

15    fair representation?

16         A.    No.  My point was that in issues of political 

17    gerrymandering the court has to be very careful because both 

18    the Congress and the court acknowledge that there is no 

19    right to proportional representation.  Denied that standard, 

20    we as political scientists cannot provide another objective, 

21    simple standard of what the fair share is for the Democratic 

22    and Republican Party.  So my argument is that we can't help 

23    define what that right is to the court, and therefore the 

24    court has to be very wary when it intervenes in that area.

25         Q.    So you are saying that that is not your opinion, 
                                                              1254

 1    it is not your opinion that it is the job of the political 

 2    system, not the courts, to accommodate the demands that 

 3    nonracial groups have for fair representation?

 4         A.    No.  I was elaborating my opinion.  I was saying 

 5    that because of this difficulty of defining what the fair 

 6    share of representation is for parties, that we are better 

 7    off in a situation of allowing the political process to try 

 8    to come to an agreement, and I have come out in favor of 

 9    institutional arrangements which produce greater degrees of 

10    consensus between the parties in making these decisions.

11         Q.    In your direct testimony, Dr. Cain, you used the 

12    term "redistricting" a number of times.  I believe you also 

13    used the term "apportionment."  Could you explain to the 

14    Court what the difference between those two terms is?

15         A.    I hope this isn't too obvious too.  

16               THE COURT:  It isn't.  I thought they were the 

17    same.

18         A.    "Apportionment" refers to the allocation of 

19    Congressional districts upon the completion of the census so 

20    that there is a sort a proportions formula that says that if 

21    you get a certain amount of population, then you get a 

22    certain number of seats.  So California, for example, got 

23    additional seats, seven new seats.  

24               Once that allocation has been made to the states, 

25    it is then up to the states to do the redistricting.  The 
                                                              1255

 1    redistricting is not done at the federal level.  So the 

 2    first phase of that, the allocation of the districts is 

 3    called "apportionment."  

 4               The second phase, which is the adjustment of the 

 5    district boundaries in order to make the populations equal, 

 6    that phase is referred to as "redistricting."  

 7               Now, to confuse things a little bit, but why not, 

 8    the term "reapportionment" is sometimes also used.  People 

 9    object to that use of the term, but actually it is a very 

10    good term, because it turns out that the difference between 

11    the two stages is not so clear.  When you are drawing 

12    district boundaries, you are also apportioning 

13    representation over cities and over counties and over areas 

14    and over neighborhoods.  So, in fact, the process of 

15    redistricting is also a form of apportionment, but it is not 

16    an apportionment across the states, which is what we mean by 

17    Congressional apportionment.

18         Q.    When it comes to gubernatorial elections, neither 

19    apportionment nor redistricting is relevant, is that 

20    correct?

21         A.    Well, that depends on what you mean by 

22    "relevant."  In the last gubernatorial race in California, 

23    many would say that that race was about apportionment, 

24    because the Republican Party regarded it as absolutely 

25    essential that a Republican become a governor of California 
                                                              1256

 1    in order to veto any redistricting plan that the legislature 

 2    might send to it.  That is a reason why, as we referred to 

 3    before, George Bush and Dan Quayle were sent to California, 

 4    to raise money and to roust the troops on this issue, 

 5    because it was a very central issue in the gubernatorial 

 6    race.  

 7               But it certainly is true that the governor's race 

 8    is not run in districts.  So in that sense that's correct.

 9         Q.    Let's look at some of the charts for a moment, 

10    Dr. Cain, that you prepared.  You can pick any of the charts 

11    from 583 through 590, or you can look at all of them as a 

12    group.

13         A.    583?

14         Q.    583 through 590.

15         A.    Right.

16         Q.    The Y axis, the vertical axis, on all of these 

17    charts is labeled "Percent undercount."

18         A.    Yes.

19         Q.    But you don't really know exactly what the 

20    percent undercount is, do you?

21         A.    Yes.  The percent undercount is the adjusted 

22    minus the unadjusted over the unadjusted.

23         Q.    That is assuming that the adjustment is correct?

24         A.    That's correct.

25         Q.    The exhibits show, in your view, that more 
                                                              1257

 1    minorities than nonminorities were undercounted and that 

 2    more Democrats than non-Democrats were undercounted, is that 

 3    a fair statement?

 4         A.    Districts, yes.

 5         Q.    But they don't show what sort of redistricting 

 6    there would be if adjusted numbers were used in a 

 7    redistricting process, do they?

 8         A.    This is an argument not about gerrymandering, it 

 9    is an argument about malapportionment.  The argument is that 

10    no matter who draws the lines and no matter what their 

11    intent is, when they get done, if you use the unadjusted 

12    data and then you go back through and you apply the adjusted 

13    data, you will discover that the districts are 

14    malapportioned.  So it doesn't matter what the intent was, 

15    whether it was partisan or not; the fact is that if you use 

16    the unadjusted data, you will be underrepresenting minority 

17    areas.

18         Q.    In fact, the process of redistricting is a very, 

19    very complex process, isn't it?

20         A.    That is correct.

21         Q.    It works differently in different states?

22         A.    I wouldn't overdo that point.

23         Q.    The degree to which the redistricting process is 

24    politicized certainly varies from state to state, doesn't 

25    it?
                                                              1258

 1         A.    If we are back onto the gerrymandering issue, the 

 2    probabilities of gerrymandering do vary from state to state.

 3         Q.    And the degree to which the process is partisan 

 4    depends on a number of factors, including whether there is 

 5    single party control over the state legislature, is that 

 6    correct?

 7         A.    That's correct.

 8         Q.    The nature of the voting rule within the state 

 9    legislature as applied to the redistricting process?

10         A.    Yes.  This sounds very familiar.

11         Q.    The degree of political competitiveness within a 

12    state legislature?

13         A.    This is obviously insightful research.  I must 

14    believe it.

15         Q.    It is not my research.  It's yours.

16         A.    Oh, I thought so.  Yes.  Then I agree to every 

17    word.

18         Q.    In the redistricting context, political 

19    scientists speak in terms of districts that are safe for one 

20    party or another and districts that are less than safe for 

21    one party or another, is that not correct?

22         A.    Yes.

23         Q.    We don't know, do we, if Republicans or Democrats 

24    would have come out ahead in, let's say, California in terms 

25    of the number of safe seats for Democrats or Republicans had 
                                                              1259

 1    adjusted numbers been used in redistricting as opposed to 

 2    unadjusted numbers?

 3         A.    This is not an argument about the likely plans 

 4    that would be drawn as a result of using one set of figures 

 5    rather than another.  It says that regardless of who draws 

 6    the plans, it results in malapportionment, if you use one 

 7    dataset versus the other.

 8         Q.    But the point is that we don't know how any of 

 9    those plans in any of the states would have come out, do we?

10         A.    We know that whatever these plans are, the 

11    minority areas will receive less representation, both in 

12    actual fact and psychologically.  When people are thinking 

13    about whether or not they have an obligation, a voting 

14    rights obligation, to create a Latino seat, which is a 

15    situation that we are in in Los Angeles right now, the 

16    benchmark figures, that is, is there a 65 percent population 

17    that you can draw a majority seat in, it makes a difference 

18    whether the number is 60 or 65.  

19               So if the population is not properly enumerated, 

20    you could come to very different decision rules about how 

21    you are going to handle the redistricting, regardless of who 

22    is in charge.

23         Q.    Again, you don't know whether the process would --

24         A.    No, that's right.  This is not an argument that 

25    using one set of figures and the other affects the 
