Tribunal Criminal Tribunal for the Former Yugoslavia

Page 28128

 1                           Thursday, 26 April 2012

 2                           [Open session]

 3                           [The accused entered court]

 4                           [The witness takes the stand]

 5                           --- Upon commencing at 9.19 a.m.

 6             JUDGE KWON:  Good morning, everyone.

 7             Due to some technical difficulties, we moved courtroom -- we've

 8     moved to Courtroom III and the start of the trial has been delayed for

 9     15 minutes.

10             Yes, where were we?  Yes, Ms. West, please continue.

11             MS. WEST:  Thank you, Mr. President, good morning.  Good morning,

12     Your Honours.

13             JUDGE KWON:  I'm sorry, but I will note that we'll continue to

14     sit pursuant to Rule 15 bis in the absence of Judge Morrison.

15                           WITNESS:  EWA TABEAU [Resumed]

16                           Examination by Ms. West: [Continued]

17             MS. WEST:  Mr. Registrar, may I have 65 ter 11218, and may I have

18     page 37 in e-court.

19        Q.   Good morning, Dr. Tabeau.

20        A.   Good morning.

21        Q.   So I just want to remind you, and myself as well, that we both

22     need to speak rather slowly.

23        A.   Of course.

24        Q.   And after you give an answer, I'm going to make a pause so the

25     transcript can catch up with us.


Page 28129

 1             Yesterday, when we ended, we were talking about sources that you

 2     used in your municipality report.  And you mentioned that you used the

 3     census from 1991 and that you also used voters' registrations records

 4     from 1997.

 5             Can you remind us of the number of persons who registered to vote

 6     in 1997?

 7        A.   Yes.  The number of registered voters was high.  In absolute

 8     terms, it was 2.5 million who registered.  And it was a high

 9     participation rate of almost 90 per cent.

10        Q.   Can you also tell us how many people registered -- or were

11     registered in the census in 1991?

12        A.   The population registered in the census was 4.4 million people.

13        Q.   Did you do a comparison between the 1991 census and the 1997

14     voters' registration?

15        A.   Yes.  This comparison was done by, first of all, linking all

16     individual records, one record representing one person, and in this way

17     to trace those reported in 1991 in the census in a post-conflict source

18     on surviving population.  That is the voters.

19        Q.   And what percentage of the records did you find were linked?

20        A.   It was approximately 80 per cent of the voters who registered

21     that were confirmed in the population census.

22        Q.   Was this the methodology that you used in your analysis?

23        A.   The linking of records is the core of the methodology we used.

24     The linking gives us the possibility to compare pieces of information

25     about the same persons at two different moments of time - in 1991 and in


Page 28130

 1     1997.

 2        Q.   Okay.  We're going to talk about that in detail.  And on the

 3     screen in front of you, you should have Annex C.  It's page 226 in

 4     Annex C, and this talks about your methodology.

 5             MR. ROBINSON:  Excuse me, Mr. President.  We don't have the

 6     e-court working on our side here, so we're not able to follow.

 7                           [Trial Chamber and Registrar confer]

 8             JUDGE KWON:  Yes.  All I can see is the documents in private --

 9     or personal computer, but it's not working on the common drive.

10             MS. WEST:  Mr. President, I am able to cite the page of the

11     report itself.  So if everyone has their hard copy, we can still

12     continue.

13             JUDGE KWON:  Did you bring the hard copy with you?

14             MR. ROBINSON:  We don't have that, no.

15             JUDGE KWON:  We didn't bring it either.  Given the --

16             Shall we take a short break so that the technicians could look

17     into the matter and fix it?

18             MS. WEST:  Yes, please.

19             JUDGE KWON:  I consult the Registrar.

20                           [Trial Chamber and Registrar confer]

21             JUDGE KWON:  Yes, the Chamber will rise for ten minutes, and

22     hopefully the problem can be solved in that time.

23             We rise for ten minutes.

24                            --- Break taken at 9.25 a.m.

25                           --- On resuming at 9.37 a.m.


Page 28131

 1             JUDGE KWON:  Yes, Ms. West, please continue.

 2             MS. WEST:  Thank you.

 3        Q.   And so we are going to speak now specifically about the

 4     methodology which we can see is part of Annex C.

 5             Dr. Tabeau, what were the three variables in which you were

 6     interested for your analysis?

 7        A.   I was interested in place of residence of individuals reported in

 8     1991 census, on one hand, and on the other hand, in a post-war source,

 9     that is, voters' register.  These were the two of the three variables.

10     And the third variable was ethnicity.  Ethnicity as reported in the

11     open-ended question in the 1991 census.

12        Q.   And why were you particularly interested in these three

13     variables?

14        A.   These three variables make it possible to analyse the population

15     movements between these two moments of time, 1991 versus 1997, for every

16     ethnic group separately.  When I am saying "every group separately," I

17     mean for every main ethnic group, like Serbs, Muslims, Croats, and all

18     other ethnicities jointly as "others."

19        Q.   You mentioned that in the census in 1991 place where the person

20     lived was mentioned.

21        A.   Yes.  This -- we even have a detailed address for every person,

22     but, as well, we have information about the so-called settlements and

23     larger areas, these are municipalities.

24        Q.   In 1997, did the voter registration records specifically indicate

25     where that person lived in 1997?


Page 28132

 1        A.   Well, there is a variable, a self-reported place of residence,

 2     but we rather use another variable from the voters' register, this is the

 3     place of registration to vote.  So this is a very good proxy for the

 4     place of residence, at that time.

 5        Q.   Now, in a general way, can you tell us about the process you

 6     employed to figure out which persons were on the 1991 census and were

 7     also on the 1997 voters' registration?

 8        A.   In this particular case, the matching could be done using the

 9     unique personal identification number, the JMBG, as the same people

10     reported in the census and in the voters' register had the same unique

11     number.

12             So this was the first step in our matching process.  Of course,

13     there was a group of records with -- for which it wasn't achieved for

14     various reasons, so in the next step we did our matching using

15     descriptive data items, such as names - first, father's name, surname -

16     date of birth, place of birth.  This matching is more complex than the

17     first type.  The first is based on one numeric data item.  It's very,

18     very easy to do this kind of matching.  Names are more complex.  Still, a

19     lot of success can be achieved.

20        Q.   Now once you had a set of linked records, what did that tell you

21     about the identified people?

22        A.   These matched records was the group for -- of whom we had a lot

23     of information for both years, for 1991 and 1997.  Basically we were

24     interested in comparing places of residence in both these years, and

25     ethnicity.  So having these lots of information, we selected three


Page 28133

 1     variables and studied in order to compile statistics as the ones

 2     presented in the municipalities report.

 3        Q.   Okay.  So let me ask you some questions about that.

 4             Did the linking, the successful link, tell you that the person

 5     survived the war?

 6        A.   Yes.

 7        Q.   Did the successful link tell you what the ethnicity of the person

 8     was?

 9        A.   Yes.

10        Q.   Did it tell you where they lived in 1991?

11        A.   Yes.

12        Q.   And did it also tell you where they registered to vote in 1997?

13        A.   Yes.

14        Q.   What was the significance for you of a person who lived in one

15     place in 1991 but lived in a different place in 1997?

16        A.   It was an indication of a movement which we considered to be a

17     displacement compared with the original 1991 place of residence.  So

18     basically we used a statistical definition to decide whether a person can

19     be considered a displaced person or a refugee, or the person didn't move.

20        Q.   All right.  You mentioned earlier in your testimony that there

21     were groups of individuals who were not included in the 1997

22     registration.  You said that these were people who didn't want to

23     register, some maybe too old, too ill, and others perhaps too young.

24             Was anyone who was born in 1980 and afterwards included in your

25     analysis?


Page 28134

 1        A.   No.  This the first criterion to -- for being eligible to vote.

 2     It is the age criteria.  Those who are eligible have to be 18 years of

 3     age at time of elections.  So these people who were born after 1980 plus

 4     other later years are not part of the voters' register.

 5        Q.   And as a result of that, did that make your numbers larger or

 6     smaller?

 7        A.   Of course smaller, because obviously also among children below

 8     18 years of age they were displaced persons who obviously moved with

 9     their parents.

10        Q.   This methodology that you've spoken about and these sources, are

11     these all included in the annex that we see on the screen now?

12        A.   Yes, they are.

13             MS. WEST:  Mr. President, I would tender 65 ter 11218.  This is

14     the annex.

15             JUDGE KWON:  How many reports, expert reports, are you going to

16     tender through Dr. Tabeau, Ms. West?

17             MS. WEST:  Mr. President, there's a total of six.  For the

18     municipalities, there's just one plus this annex that you see.  For the

19     Sarajevo portion of the case, there are three.  And for the Srebrenica

20     portion, there are two.

21             JUDGE KWON:  In total, 11 items.

22             MS. WEST:  I think the extra items are just addendums to some of

23     the reports.

24             JUDGE KWON:  But you are tendering them as well?

25             MS. WEST:  I will.


Page 28135

 1             JUDGE KWON:  Are there any objections to the admission of any of

 2     those items?

 3             MR. ROBINSON:  No, Mr. President.

 4             JUDGE KWON:  They are all admitted.

 5             MS. WEST:  Thank you very much.

 6             JUDGE KWON:  Shall I -- the exhibit number will be informed by

 7     the Registrar in due course.

 8             MS. WEST:  Thank you.

 9             JUDGE KWON:  Let's move on.

10             MS. WEST:  May we have 65 ter 10723; e-court page 28.

11        Q.   Dr. Tabeau, we're now going to talk about estimates, statistical

12     estimates, that you made in regard to these reports.

13             Before we do so, I'd like to look at your report, the

14     Mr. Karadzic report on municipalities, to review how it's divided.

15             First off, in 1991 how many municipalities were in the BiH?

16        A.   109.

17        Q.   After the war ended, how many municipalities were there?

18        A.   189 codes were assigned to municipal areas, but a number of these

19     codes were empty.  So in total it was 149.

20        Q.   Why were there more after than before the war?

21        A.   This is because of how the country changed during the war.  The

22     Dayton line actually split a number of municipalities in two areas, one

23     belonging to Republika Srpska and one to the Federation of Bosnia and

24     Herzegovina.  That is the main reason.  But there were also a low number

25     of new municipalities, very small areas cut off from the old


Page 28136

 1     municipalities.

 2        Q.   Were you able to account for the split in your results?

 3        A.   Yes.  We analysed municipalities as available from the

 4     post-Dayton classification, official classification, which number is

 5     higher than the pre-war municipalities number.

 6        Q.   Now, the annex to your report, this is what we see on the screen,

 7     Annex A, is lengthy, and I just want to go through it briefly.

 8             This covers your results for internally displaced persons for all

 9     the 27 municipalities relevant in this indictment.  In regards to

10     Table 1, can you tell us what's included in Table 1?

11        A.   Table 1 reports on the changes in the ethnic composition.  There

12     is the population size in 1991 and of the complete population and the

13     size of the sample population from the voters' register, 1997.

14        Q.   And you noted earlier that ethnicities were reported in the 1991

15     census.  How many ethnicities did you include in your tables?

16        A.   In all tables, only four major groups are included.  The category

17     "others" comprises the remaining category different than Serbs, Muslims,

18     and Croats.

19        Q.   Okay.  So we have a total of four.  Does that mean there are four

20     tables for each municipality in this annex?

21        A.   Yes.  Each ethnic group has a separate table.  So ethnic --

22     change in the ethnic composition is reported separately for Muslims,

23     Serbs, Croats, others, and each of these tables will be listing all the

24     27 municipalities included in the indictment of this case.

25             MS. WEST:  And, for the record, that's e-court pages 0964 through


Page 28137

 1     0975.

 2        Q.   Now let's go to Table 2.  Can you tell us what Table 2 regards?

 3        A.   Table 2 presents the minimum number of internally displaced

 4     persons and refugees obtained from our linking methodology.  These are

 5     minimum numbers, which means only the numbers that were obtained from the

 6     matched records of voters backed with the census.  They are called,

 7     therefore, minimum, as the voters' register is a sample, very large, but

 8     sample, so we have the minimum number.  They can be documented with lists

 9     of all these displaced.

10        Q.   Again, is this four tables as well for each municipality for

11     all 27 -- excuse me, for each ethnicity for all 27 municipalities?

12        A.   Yes, correct.

13             MS. WEST:  For the record it's e-court page 0976 to 0987.

14        Q.   If we can look at Table 3.  Tell us what those tables regard.

15        A.   Table --

16             MR. ROBINSON:  Excuse me, Mr. President, I would ask the

17     Prosecutor if there is going to be any kind of redactions to reflect that

18     20 municipalities are charged pursuant to the Trial Chamber's decision

19     under Rule 73 bis (D) as opposed to the 27 that were charged at the time

20     she did her analysis.

21             JUDGE KWON:  Ms. West.

22             MS. WEST:  Your Honour, I understand the elimination of those

23     seven, but I do believe, like your observations about evidence on Zepa,

24     the information on those other seven do still go to the overarching JCE,

25     and I would include them.


Page 28138

 1             JUDGE KWON:  And then we can calculate on our own based upon the

 2     tables that follow.

 3             MS. WEST:  Correct.  So it's easy to take out the seven

 4     municipalities because the tables are split up.

 5             JUDGE KWON:  Mr. Robinson.

 6             MR. ROBINSON:  Well, Mr. President, our position is that the

 7     Chamber shouldn't consider any evidence with respect to the seven

 8     municipalities which have been stricken.  You ordered the Prosecution not

 9     to lead evidence of the seven municipalities.  And in the municipalities

10     portion of the case, we have had rulings by the Chamber with respect to

11     Brcko and possibly another municipality where you found that the material

12     of the 20 municipalities that were before you were sufficient and that it

13     was not necessary in this component of the case to receive evidence of

14     other municipalities to show a pattern.

15             So, consistent with that ruling, I don't believe that the Court

16     should consider any of the seven municipalities that have been stricken

17     from the indictment.

18             JUDGE KWON:  However, as a matter of practicality, I don't think

19     it's necessary, absolutely necessary, to ask the doctor to calculate

20     again, as long as we can do it on our own.

21             MR. ROBINSON:  Yes --

22             JUDGE KWON:  Why don't you leave it in the hands of the Chamber

23     later.

24             MR. ROBINSON:  Very well.

25             JUDGE KWON:  Thank you.


Page 28139

 1             Let's proceed.

 2             MS. WEST:

 3        Q.   Dr. Tabeau, tell us about Table 3.

 4        A.   Table 3 is a sort of follow-up on Table 2.  It presents a more

 5     complete number of internally displaced persons and refugees.  The more

 6     complete number could be obtained as an estimate.  It is a number that is

 7     produced using standard statistical techniques for estimation of unknown

 8     numbers of certain events in the population.  And this estimated number

 9     of internally displaced persons and refugees is presented in Table 3.  As

10     with any estimates, statistical estimates, also some uncertainty around

11     this number is expressed in this table.

12        Q.   Now, you mentioned the term "refugee."  Can you tell us what the

13     statistical definition is of "refugee"?

14        A.   Well, a refugee is a person who left his or her home and moved to

15     another country different than Bosnia and Herzegovina.  Unlike the

16     definition of the internally displaced person who moved only within

17     Bosnia-Herzegovina and didn't go to another country.

18        Q.   Do you understand there to be a legal definition of "refugee"

19     that is different than the statistical definition?

20        A.   Yes, of course, I am aware of such definition.  And, as a matter

21     of fact, what we discussed yesterday, among sources we had this one

22     source called DDPR, which is the official register of internally

23     displaced persons and refugee in Bosnia and Herzegovina.  Official

24     register was made by UNHCR together with local authorities in Bosnia and

25     Herzegovina.  So this data, this source, represents IDPs and refugees as


Page 28140

 1     presented according to the legal definition of a refugee and an

 2     internally displaced person, but cross-referencing this with our

 3     statistics, I can say there is a huge consistency between the two.

 4        Q.   And lastly --

 5             JUDGE KWON:  Ms. West, going back to the issue that Mr. Robinson

 6     raised just now, just in case, i.e., for the -- purely for the reference

 7     of the Chamber, I would like to produce the separate table, or number or

 8     numbers or ratio, in consultation with Dr. Tabeau after her evidence is

 9     over that would include the municipalities that have been taken out from

10     the indictment, but not now.

11             Yes, please proceed.

12             MS. WEST:  Thank you, Mr. President.  We will do that.

13        Q.   Dr. Tabeau, looking at the last portion of the tables,

14     Tables 1 BH, 2 BH, 3 BH, tell us about those.

15        A.   These three tables are exactly the same as Table 1, 2, and 3,

16     only that they are presenting one large area, Bosnia and Herzegovina, as

17     a whole, for just reference purposes, no more than that.

18        Q.   So this includes all the municipalities; is that correct?

19        A.   Yes, it includes all municipalities in Bosnia.

20        Q.   The findings that you made in the table, are those findings also

21     conveyed in writing elsewhere in your report?

22        A.   Yes.  Findings presented very specifically in a detailed way in

23     the Annex A are discussed more generally in the body text of the report.

24     Again, separately for every ethnic group.

25        Q.   Dr. Tabeau, I do not want to go through all these tables, but I


Page 28141

 1     think the best way to convey this information is by giving one example.

 2     And you've prepared a handout showing one example; is that right?

 3        A.   Yes, correct.

 4             MS. WEST:  May we have 65 ter 23719.  And for the parties, this

 5     is the handout that has been distributed.

 6             JUDGE KWON:  Which is titled as "courtroom presentation."

 7             MS. WEST:  Thank you.  Yes.

 8        Q.   We'll go to page 1 of that.  Tell us what the source of the

 9     material was for this handout.

10        A.   This comes from my report from Tables 1, 2, 3 in the annex.

11     Annex A, of course.

12        Q.   And we see now Table 1, and this specifically regards Prijedor;

13     is this correct?

14        A.   Correct.

15        Q.   Now, can you tell us, going across the rows, we see at the top

16     that you have the ethnicities.  And then you have 1991 and a number of --

17     numbers follow.  What are those?

18        A.   Well, this table generally reports on the change in the ethnic

19     composition between 1991 and 1997.  Numbers in the first part of the

20     table under 1991 and 1997 are just absolute numbers of those individuals

21     based -- for whom we study this issue.

22        Q.   Okay.  So when you say absolute numbers, when we look at row

23     1991, are those the actual numbers reported by the census?

24        A.   Yes.  So, for instance, for all ethnicities, there were 94.028

25     individuals that were born before 1980 reported in the census.


Page 28142

 1        Q.   And then we look at 1997, that row, are those the actual absolute

 2     numbers reported from the voters' registration?

 3        A.   That is correct.  This is the sample of registered voters linked

 4     back with the records of the population census.

 5        Q.   Now let's go down to the percentage section and the 1991 row.  Is

 6     that the percentages those ethnicities made up in the entire population

 7     of the country in 1991?

 8        A.   Well, this example is just for one municipality.

 9        Q.   Excuse me.

10        A.   So we are discussing the municipality of Prijedor in this

11     example.  Indeed in the second part of the table our reading should go in

12     rows.  In the first row there is the ethnic composition expressed in

13     percentages for those living in the municipality of Prijedor.  And in the

14     second row, there is ethnic composition, again in relative terms, in

15     percentages, for 1997.  These are comparable figures because of the

16     relative nature, unlike the absolute numbers, because here there is a

17     problem that one -- 1991 is representing a complete population, and 1997

18     is a sample.

19             So comparisons are always possible and meaningful if we look at

20     the percentages.  The percentages represent very well the entire

21     population in both cases.

22        Q.   Let's look at what happened to the Serb population in Prijedor

23     from 1991 to 1997.

24        A.   The Serb population increased or the share of Serbs in the

25     municipality of Prijedor increased from 43.4 per cent in 1991 to


Page 28143

 1     89 per cent in 1997.  It is an increase of more than 100 per cent.

 2        Q.   Let's look at what happened to the Muslim population.

 3        A.   The share of Muslims in 1991 in Prijedor was 42.6 per cent, and

 4     in 1997, just 1 per cent.  It is a very significant decrease by 97.6

 5     per cent.  That is the reading, how these numbers should be looked at in

 6     my report.

 7        Q.   Now let's move to Figure 1.  Is this just a visualisation of what

 8     those numbers would look like?

 9        A.   Yes.  These are exactly same numbers as in Table 1 presented

10     graphically in Figure 1.  The colours, red is associated with the Serbs,

11     and the green, Muslims.  These are the two groups, major groups, we have

12     just discussed.

13        Q.   And if we move to Table 2, this is the overall number of IDPs and

14     refugees in Prijedor.  Is this table comprised of information from

15     several of the tables in your annex?

16        A.   Yes.  Most importantly, this table reports on population

17     displacement.  But there are pieces of information from all three tables

18     in Annex A, one -- Table 1, Table 2, and 3 put next to each other and

19     making the story more complete than just if you read it table by table.

20        Q.   Can you go over the column labels and tell us which each of them

21     mean?

22        A.   Of course, first column is self-explanatory.  It is just ethnic

23     groups listed.

24             The next one reports on the 1991 census population.  Of course,

25     those born before 1980.  Perhaps a short explanation, that we work with


Page 28144

 1     this segment of the population is because we want to compare the

 2     comparable, that means the same segments in both years, 1991 and 1997,

 3     and we don't want to mix up the whole population which is generally

 4     available from 1991 census with a sample of ages 18-plus of the eligible

 5     voters.

 6             Well, going back to the table, so 1991 is included -- 1991 census

 7     is included for reference purposes.

 8             So, next column, the second column, is giving us the minimum

 9     number of internally displaced persons and refugees in Prijedor, or,

10     rather, from Prijedor.

11        Q.   And when you say "minimum number," is this the number that you

12     actually counted from the 1997 voters' registration?

13        A.   Yes, yes, that is correct.  It's the minimum number.  Those

14     records matched back with the census for whom we were able to compare the

15     place of residence.

16        Q.   And the next column is an estimate.

17        A.   Yes.

18        Q.   How did you get to that?

19        A.   This estimate is based on a sample proportion of IDPs and

20     refugees, that is, the displaced people.  In the voters's sample, this

21     fraction is, in statistical terms, a sampling proportion.  We take this

22     proportion and apply it to the exposed population, that is, the 1991

23     census, and in this way we produce an estimate which is -- which should

24     be seen as some kind of producing a more complete number, like an

25     extrapolation of the sample results over the complete population.  That's


Page 28145

 1     the estimate, estimated number.

 2             JUDGE KWON:  Ms. West, doctor confirmed that all these numbers

 3     were taken from the tables that exist in her report.  For example,

 4     sixty -- no, 642.  So as far as this table is concerned, this Table 2, so

 5     with the specific numbers, I'd like to be shown each table that shows the

 6     number here.  For example, where can we see 642?

 7             Do you follow?

 8             MS. WEST:  Yes, I do.

 9             JUDGE KWON:  Thank you.

10             MS. WEST:  And I think Dr. Tabeau has indicated that for you on

11     the smaller table below the table.

12        Q.   Dr. Tabeau, maybe you can tell us about the smaller table.

13             JUDGE KWON:  So, for our understanding, could we go back to that

14     table and coming back to this table.  That will be much easier for us to

15     follow.

16             MS. WEST:  If I can just clarify:  Do you want me to do that

17     during the presentation itself?

18             JUDGE KWON:  Now.

19             MS. WEST:  Okay.

20        Q.   So, for example, let's do 642.

21        A.   Yes.

22        Q.   That particular number, 642 --

23        A.   Yes.

24        Q.   -- can you tell us exactly on what page of the annex we can find

25     that?


Page 28146

 1        A.   Yes.  From the table, little table at the bottom, we can read

 2     that the number for Serbs is taken from the Table 2 with the letter S.

 3     This table is available at page 51.  So we have to go to page 51,

 4     Table 2, Serbs.

 5             MS. WEST:  And that would be the ERN ending in 0986.

 6             And for e-court purposes, this is 65 ter 10723.

 7             JUDGE KWON:  Yes, Table 2 as continued.

 8             THE WITNESS:  So, Your Honour, can I take you to this right place

 9     in the table?

10             JUDGE KWON:  Yes.  E-court will follow.

11             THE WITNESS:  Okay.

12             JUDGE KWON:  I found it, but ...

13             THE WITNESS:  Yes.

14             So that is how we read the references at the bottom of the table.

15             JUDGE KWON:  Yes.  I would like you to explain how that 242 --

16     two-hundred -- I'm sorry, 642 is reached in this table.

17             So, shall we upload it?

18             MS. WEST:  Please.  Again, that's --

19             JUDGE KWON:  It's hard photo -- hard copy page number 51.

20             MS. WEST:  Mm-hm.

21             JUDGE KWON:  ERN ending with 986.

22             MS. WEST:  Thank you.

23             JUDGE KWON:  It's in the 65 ter number 10723.

24             Next page.

25             THE WITNESS:  Next page.  Yes.


Page 28147

 1             If we go to row 14, which is Prijedor, and move to the right of

 2     this -- of this row -- I'm not sure ... oh, I did something wrong.  I

 3     actually shouldn't probably do that.  Or should I?

 4             JUDGE KWON:  Yes, we can see it.  642.

 5             THE WITNESS:  Yeah, 642 is in the column called IDPs and

 6     refugees.

 7             JUDGE KWON:  How was that number reached in this table?

 8             THE WITNESS:  Yes.  In this table, we only report the tabulation

 9     based on our matching analysis.  These 642 records or persons were

10     identified in the voters' register and linked back with their records of

11     the 1991 census, and we compared for these 642 their place of residence

12     in 1991 versus the place of registration to vote, which is an

13     approximation of their residence in 1997, and concluded these places were

14     different, and we, therefore, labelled them as displaced individuals.

15             Different means these persons moved from the pre-war municipality

16     where they used to live, which was Prijedor, to all kinds of other

17     municipalities in Bosnia and Herzegovina, and also possibly, and likely,

18     some of them moved abroad.

19             JUDGE KWON:  Thank you.

20             On this occasion, Doctor, could be kind so as to explain to us

21     all the numbers in that row, i.e., 14 Prijedor.

22             THE WITNESS:  Yes.

23             First number, first column, 49.019, is the total population

24     identified in 1997.  This is the population who originally resided in

25     Prijedor, which was reported in 1991 census.  It is basically this


Page 28148

 1     population which constitutes our sample of voters for Prijedor.

 2             For these 49.000, we are able to say how many of them became

 3     displaced and how many are still residing in their original municipality,

 4     that is, Prijedor.

 5             JUDGE KWON:  Just a second.  At the outset you stated 49.019 is

 6     the total population identified in 1997.

 7             THE WITNESS:  Yes.  But this is the table for Serbs.  That means

 8     that we are talking about the Serbs from Prijedor.

 9             JUDGE KWON:  Mm-hm.  But also you said this is the population who

10     originally resided in Prijedor which was reported in 1991 census.

11             THE WITNESS:  Yes.  Well, this table -- first of all, it's good

12     to understand.  This table is made according to the original place of

13     residence in 1991.  Original place of residence in 1991.  We look at

14     those who originally lived in 1991 in Prijedor, all ethnicities, and we

15     followed them in the voters' register.  Yes?

16             So in the voters' register we have a total population for

17     Prijedor from Prijedor originally in 1991 equal 49.019.  Okay?

18             JUDGE KWON:  Very well.

19             THE WITNESS:  Of this 40.000, there is 24.070 -- 24.070 of Serbs.

20     Now we are moving to the ethnicity presented in this table.  Okay?

21             JUDGE KWON:  Yes.  Are you going to explain the next number,

22     21.964?

23             THE WITNESS:  21.964 is the number out of 49.019 who became

24     displaced in 1997.

25             JUDGE KWON:  Yes.  Yes.


Page 28149

 1             THE WITNESS:  These are two reference numbers for Prijedor;

 2     right?  As the whole municipality.  We move now to the next panel of the

 3     table.  And we see 24.070 in the first column for Serbs.  That means we

 4     had 24.070 Serbs identified among the 49.019.  Okay?

 5             JUDGE KWON:  Yes.

 6             THE WITNESS:  And then the 642 are reported as displaced, as

 7     displaced Serbs.

 8             So the 642 should be referred to in relation to 24.070 in our

 9     sample.

10             Obviously it is a small per cent of 2.7 per cent.

11             JUDGE KWON:  I'm sorry, forgive me, I'm still -- it's unclear to

12     me how we can reach that number of 642.

13             THE WITNESS:  This table --

14             JUDGE KWON:  It's coming from the separate investigation or

15     examination.

16             THE WITNESS:  This number cannot be calculated from this table.

17     This is not possible.  This number was obtained in the matching process.

18     And from the matching process, we can tabulate the matching results

19     according to what is presented in this table.

20             JUDGE KWON:  So where can we find that analysis in your report?

21             THE WITNESS:  This, how I obtained numbers like 642, is explained

22     in -- in -- common language, say, but it is not that I give you the

23     actual data from which you can count these numbers.  I can compile such

24     lists.  So you could be receiving from me a list of, as you see, for

25     Prijedor, the minimum number is 21.964, so it will be a list which would


Page 28150

 1     contain twenty -- almost 22.000 names.  These are -- and from that list

 2     you can count 642 displaced Serbs.

 3             MS. WEST:

 4        Q.   And when you --

 5             MS. WEST:  Mr. President, if I may.

 6             JUDGE KWON:  Please, yes.

 7             MS. WEST:

 8        Q.   In order to count those 642, are the people you include those

 9     who, in 1991, those Serbs in 1991, lived in Prijedor but those -- those

10     Serbs who, in 1997, lived somewhere else?

11        A.   These 642 are Serbs who lived in 1997 somewhere else.  That's the

12     meaning of this number.

13             MS. WEST:  Mr. President, if you would be satisfied we can

14     provide you a list of those 642.

15             JUDGE KWON:  Has the Defence been disclosed with such material?

16             MS. WEST:  The Defence has been disclosed with the 1991 census --

17             THE WITNESS:  Voters --

18             MS. WEST:  -- I believe the OSCE voters' registration as well.

19     If they wanted to make their own calculations, they can do the same.

20             JUDGE KWON:  Could you be kind so as to show us the part, in

21     common language as doctor has referred to, that deals with this 642?

22             THE WITNESS:  Your Honour, in the report, municipalities are not

23     discussed one by one because there are far too many to do so.  But what

24     is done in the report, there are summaries of these kind of results

25     presented for the entire area of 27 municipalities, and then for each


Page 28151

 1     group separately.  Serbs are one of these groups.  And we can go to the

 2     report and I can tell where it is discussed generally for Serbs for the

 3     entire area in the report.

 4             JUDGE KWON:  Thank you.  Yes.  But if you could deal, with the

 5     doctor, as to the methodology how she counted this number.

 6             MS. WEST:  I will do so.  Thank you.

 7             JUDGE KWON:  Okay.

 8             MS. WEST:

 9        Q.   So, Dr. Tabeau, we're going to focus now on how you got to 642

10     and the comparison between the census and the OSCE records.

11             When you made this comparison, did you -- did you compare names?

12        A.   Of course.  Well, we -- Your Honour, just to make clear what we

13     are talking about:  In this analysis, on one hand, I have a database with

14     4.4 million rows.  These are my records.  That is the census.  That is my

15     starting point.

16             On the other hand, there is another data table which comprising

17     2.5 million records.  Okay?  Rows in that table.  So my purpose is to

18     make a connection between these two for as many persons as possible.

19             For a majority of them, it is easily possible because I do it

20     using the JMBG, the unique identification number.

21             So, in the next step, I improve it by running a number of

22     additional criteria to match as many as possible.  Once I am done with

23     the matching, for every person who is connected between these two tables,

24     I can look at the full record of information and compare many things,

25     including the place of residence.  So if I take one row from Table A and


Page 28152

 1     one row from Table B, Table A being 1991 census, Table B being the 1997

 2     voters' register, and I look at Mr. Whoever, Cekic, for example, I look

 3     and compare his first name, father's name, surname, date of birth, and

 4     place where, on one hand, he lived in 1991 and where he registered to

 5     vote in 1997.

 6             Okay, if the place is the same, I mark this other variable in my

 7     working file.  If the person became displaced, I also mark it.  So I

 8     create a data item, a variable, which tells me for every linked person

 9     what happened, whether he or she moved or not.

10             So once I'm done with this kind of flagging of records, then I

11     move to the next step.  I make my only tabulation of the marking of

12     displacement.  And, of course, I also have information about persons,

13     about their ethnicity and age and many, many other things.  So, in the

14     next step I can create tables like this one by counting, by aggregating,

15     the values on the newly created variables into categories like 642 of

16     displaced Serbs from Prijedor.

17             MS. WEST:

18        Q.   And that's -- Dr. Tabeau, that's 642 that we see.  Those are

19     Serbs who you saw from the census in 1991 who lived in Prijedor, and

20     those are the same Serbs you linked to 1997 who did not live in Prijedor.

21        A.   Correct.

22        Q.   And did you do this for all the municipalities and all four

23     ethnicities?

24        A.   Yes.

25             JUDGE KWON:  Thank you, Ms. West.


Page 28153

 1             MS. WEST:  Thank you.

 2             JUDGE KWON:  Shall we go back to the presentation table.

 3             MS. WEST:  Which is 65 ter 23719.  We can go to e-court page 3,

 4     please.

 5             THE INTERPRETER:  Kindly slow down for the sake of the

 6     interpreters, please.  Thank you.

 7             MS. WEST:

 8        Q.   Now, this table 3, on top, regards the ethnic composition --

 9             JUDGE KWON:  Ms. West, were we not still on page 2, e-court

10     page 2, Table 2?

11             MS. WEST:  Yes.  We can go back to that.  Thank you very much.

12     So this would be e-court page 2, Table 2 at the bottom.

13             JUDGE KWON:  So we just discussed 642.

14             MS. WEST:  And now we can discuss the estimate.

15             JUDGE KWON:  Yes.

16             MS. WEST:

17        Q.   Dr. Tabeau, do you see that number, 1.089?

18        A.   Yes.

19        Q.   Can you tell us how you got to that estimate?

20        A.   This is the sampling proportion applied to the 1991 census

21     population, and, therefore, a new number is produced, which obviously is

22     larger than the 642.  642 was a minimum obtained, as I explained.  This

23     number, 1.089, comprises the 642 plus --

24             JUDGE KWON:  Just a second.

25             THE INTERPRETER:  Kindly slow down, please.  Thank you.


Page 28154

 1             JUDGE KWON:  The table refers to 3S, so shall we go --

 2             THE WITNESS:  Yes.

 3             JUDGE KWON:  -- shall we go to that page --

 4             THE WITNESS:  Yes.

 5             JUDGE KWON:  -- and see how it is calculated.  So, for example

 6     purpose, I would like to just take this exercise.  Thank you.

 7             MS. WEST:  Thank you.  Yes.  So for 3S, we're going to go to --

 8     this will be 65 ter number 10723, and it's the ERN ending in 0998,

 9     please.

10             JUDGE KWON:  Do you have a hard copy page number?

11             MS. WEST:  63.

12        Q.   Dr. Tabeau, is that correct?

13        A.   Yes, absolutely.

14        Q.   So now we are talking about the estimate for Prijedor.  So if we

15     focus on line 14, Prijedor.

16        A.   Yes.  Again, in this table, the first three columns refer to the

17     entire population, and three last columns only to Serb population from

18     Prijedor.

19             So if we go to the fourth column --

20             JUDGE KWON:  Why don't we start from the first --

21             THE WITNESS:  Okay.

22             JUDGE KWON:  -- column.  Estimate.

23             THE WITNESS:  Yes.

24             45.688 is the first number in this table for Prijedor.  It's the

25     estimated, more complete number of IDPs and refugees from Prijedor.


Page 28155

 1             Next to it there are two numbers.  45.536 is the first one, and

 2     45.841 is the second one.

 3             JUDGE KWON:  Just pausing there.  The number we saw previously in

 4     the Table 2S, i.e., the total population identified in 1991 who lived in

 5     Prijedor area in 1991 was 49.019.  So how did we come to this 45.000

 6     number?

 7             THE WITNESS:  These two numbers are -- should not be compared.

 8             The first number is just a sample of voters from Prijedor

 9     identified and matched with the 1991 census.

10             JUDGE KWON:  Then how do we reach this number of 45.000?

11             THE WITNESS:  This is a number, estimated number, of internally

12     displaced and refugees, of the displaced population.  Of the displaced

13     population.  This is an estimate.  That means we cannot compare it with

14     any number you mentioned, especially with the sample size, because it is

15     an estimate obtained using sampling proportion and applied to the

16     1991 census population.

17             JUDGE KWON:  Mm-hm.

18             THE WITNESS:  It is a complex statistical analysis that we are

19     presenting in this report.  On the one hand, one part of figures that we

20     produced are numbers that are documented with names and other information

21     about persons.  But estimates are something else.  They are produced

22     using statistical techniques in order to express what part of information

23     is missing from the minimum numbers.

24             So if we go back to the table we just discussed for Prijedor, and

25     if we would like to compare these 45.000 with the right number from the


Page 28156

 1     other table, then it would have to be 21.964, which we have on page 2 in

 2     our presentation.

 3             If we go to page 2 of our presentation, to Table 2, then in the

 4     second column at the bottom, the total of all displaced population from

 5     Prijedor is 21.964.  I repeat: 21.964.

 6             JUDGE KWON:  Yes.

 7             THE WITNESS:  If we move to the next column, the total at the

 8     bottom, we see the number 45.688.  This is exactly the same number that

 9     we have just seen in -- on page 63 in the big table from the annex.

10             So these two numbers can be compared.

11             The estimated number, 45.688, is produced using statistical

12     estimation technique, which is called extrapolation of sampling

13     proportion.  It's described in the report, Slobodan Milosevic report.

14     This number, 45.688, includes the 21.964 identified as a minimum number,

15     but on the top of it, there is a part that cannot be documented with

16     names and lists.  This is the estimated part.  But even though it is not

17     documented, the estimate, which is of a very good statistical quality,

18     and I can tell why, the estimate shows what is missing in the minimum

19     number, what is not there, what cannot be documented for various reasons.

20     That's the difference between the two.

21             JUDGE KWON:  So, Doctor ... I was waiting for the French

22     translation to be completed.

23             As we see in this table, Table 2 in your presentation, the ratio

24     between the minimum number and estimated number is different per

25     ethnicity.  For example, Serbs moved from 600 to 1.000, while Muslims


Page 28157

 1     increased more than two times.

 2             So I'm interested in knowing how you reached these numbers using

 3     statistical method.

 4             THE WITNESS:  Yes.

 5             JUDGE KWON:  So that we laymen could understand.

 6             THE WITNESS:  The estimation was made separately for each ethnic

 7     group and for a good reason, very good reason.

 8             The rates of displacement are different for ethnic groups.  It is

 9     not that there is one uniform rate of displacement observed from the raw

10     data that we have.  So what Your Honour just noted from the table is a

11     very correct observation; that is, that the ratios were different for

12     ethnic groups.  The ratios were estimated from our analysis, from

13     matching of individual records for every person, all ethnicities, with

14     the census.

15             JUDGE KWON:  Did you apply the same ratio as far as same ethnic

16     group is concerned, regardless of the regions or municipalities?

17             THE WITNESS:  Well, ratios were obtained separately for every

18     municipality and every ethnic group.  So it is not that we assumed that

19     there is no differentiation between various regions.  Things were

20     happening differently.

21             So this analysis sticks very closely to what is in the data.

22             JUDGE KWON:  Thank you.

23             And now, Ms. West, I would like to deal with how we come to that

24     ratio.

25             MS. WEST:  Thank you, Mr. President.


Page 28158

 1             JUDGE KWON:  Thank you.

 2             MS. WEST:  Thank you, Mr. President.

 3        Q.   So you just mentioned that you looked at the ratio for each

 4     ethnicity and each municipality.

 5             JUDGE KWON:  But I note the time.

 6             Shall we take the break now?

 7             THE WITNESS:  Mm-hm.

 8             JUDGE KWON:  Yes, we'll break now.  Given the -- some intervals,

 9     we'll take a break only for 25 minutes, if it is okay with the staff.

10             We'll resume at 11.00.

11                           --- Recess taken at 10.36 a.m.

12                           --- On resuming at 11.01 a.m.

13             JUDGE KWON:  Yes, Ms. West, please continue.

14             MS. WEST:  Thank you, Mr. President.  I just point out today,

15     transcript page 24, at line 2, it indicates "numbers."  I think, in fact,

16     I said "names."

17             May I have 65 ter 11218; e-court page 40, please.

18        Q.   Dr. Tabeau, we're now going to talk about how you came to

19     estimates.  And today, transcript page 17, line 11, you said:

20             "This estimate is based on a sample proportion of IDPs and

21     refugees, that is, the displaced people.  In the voters' sample, this

22     fraction, in statistical terms, is a sampling proportion."

23             Can you tell us what you mean by "sampling proportion"?

24        A.   I suggest we move to page 51, table 2S, Serbs.  I'm speaking of

25     the Annex A of the municipalities report prepared for this case.


Page 28159

 1             MS. WEST:  That would be 65 ter 10723.  ERN ending in --

 2             JUDGE KWON:  -- 986.

 3             MS. WEST:  -- 986.  Thank you.

 4             THE WITNESS:  We just discussed this table.  In particular, the

 5     number 642 of displaced Serbs from Prijedor.

 6             So if we go to line 14, Prijedor, then in the third-last column,

 7     there is this number, 642.

 8             Next to this number, there is a per cent, 2.7, which tells us

 9     what was the sampling proportion of displaced Serbs in the identified

10     voters' population of Serb ethnicity.  Basically, if we look to this

11     table, 642 must be divided by 24.070.  24.070 is the size of Serb

12     population from Prijedor identified in the 1997 voters' register.

13             JUDGE KWON:  Doctor, if you could remind us what 24.070 number

14     means.

15             THE WITNESS:  24.070 is my sample of Serbs among the voters,

16     registered voters.  This is my sample for Serbs in 1997: 24.070.

17             Out of 20.070 [sic] of all Serbs in my sample for 1997, I have

18     identified 642 of them as displaced.  This gives me the percentage 2.7 of

19     displaced Serbs.

20             JUDGE KWON:  Could you tell us again the meaning of

21     two-thousand -- 21.964?  The second column.

22             THE WITNESS:  Twenty-one thousand -- yeah, second column.  I see.

23     21.964 is more than the Serbs.  It is all displaced and refugees of all

24     ethnicities in my entire sample for Prijedor.

25             So entire sample means all ethnicities.  Sample of Serbs is the


Page 28160

 1     Serbs out of all ethnicities.

 2             JUDGE KWON:  Then why is the Serb sample, which is 24.000, bigger

 3     than the sample of all -- of all ethnicities?

 4             THE WITNESS:  Yes.  The 24.070 should be compared with 49.000.

 5     49.019.  That are the two numbers that can and may be compared.  This is

 6     the -- 49.019 is all ethnicities in 1997.

 7             JUDGE KWON:  Thank you.

 8             THE WITNESS:  24.000 is the Serbs.  Then further, the displaced

 9     population, 21.964 is all ethnicities.  Of them, 642 were Serbs.

10             So these first three columns in this table are given for

11     reference.  We will see they are the same for every ethnic group.  Only

12     the next three columns are related to a given ethnic group indicated in

13     the title and number of the table.

14             JUDGE KWON:  Does the last percentage number, two-thousand --

15     2.9, have any specific significance?

16             THE WITNESS:  Not for my estimation.  It is just given for,

17     again, reference.  If we want to know what was the fraction of Serb

18     displaced among all displaced, then it would be 2.9.

19             JUDGE KWON:  That's the ratio reached 64 hundred -- if we divide

20     642 by 21.964?

21             THE WITNESS:  Yes.  That is exactly how it is.

22             But for my estimation, the other percentage matters, 2.7.  It's a

23     different percentage.

24             MS. WEST:

25        Q.   And that's the --


Page 28161

 1             JUDGE KWON:  Just a second.

 2             MS. WEST:  Sorry.

 3             JUDGE KWON:  Thank you.  Now you can continue.

 4             MS. WEST:

 5        Q.   And the 2.7 that you calculated for Serbs in Prijedor is unique

 6     to Serbs in Prijedor; correct?

 7        A.   Yes, correct.  So if we would like to see the same per cent for

 8     Muslims, we would have to go to another page.  And this is page 42 --

 9        Q.   Let's not do that right now.  I want to ask you another question.

10        A.   Okay.

11        Q.   So that sampling proportion, that 2.7, is different for each

12     ethnicity in each municipality; is that correct?

13        A.   Yes, correct.  That is how it is.  That is the specificity, the

14     uniqueness, of every of these sampling proportions.

15        Q.   And it is sampling proportion ratio that you used in your

16     estimate to come to a more complete figure?

17        A.   I just used this per cent two and seven -- 2.7 in my estimation.

18     I want to show this.  But if we don't need to go to Muslims, then I

19     suggest we move to the census population, 1991, that is Table 1.  So we

20     just continue talking about the Serbs.  Okay?  So we have the proportion,

21     sampling proportion, 2.7.  I take this proportion and apply it to the

22     census population of Serbs from Prijedor.  We just reported in Table 1,

23     page 39 of Annex A, 39 --

24        Q.   That would be --

25             JUDGE KWON:  I believe, before moving further, just one -- my


Page 28162

 1     last question with respect to this table.

 2             Coming back to the number 642, which you calculated yourself with

 3     your team, and you stated that you flagged the material in -- in terms of

 4     the tables.  Has such material been disclosed to the accused so that they

 5     can verify whether the marking was correctly done or not?

 6             MS. WEST:  Your Honour, I will speak more about this on the

 7     break, but I understand the census and the OSCE records are disclosed to

 8     the accused so they can make their own calculations.  It's unclear to me

 9     whether the flagged table was disclosed, but I will get back to the

10     Court.

11             JUDGE KWON:  Thank you.

12             MS. WEST:

13        Q.   So now we will go to -- it's the same 65 ter number, but it's the

14     ERN ending in 0974.  This -- the table --

15        A.   Page 39, Table 1.  Yes, for Prijedor.

16        Q.   Okay.  Thank you.  What do you want to show us in this table,

17     Doctor?

18        A.   Serbs, yes, it is correct, line 14, it is Prijedor.  We go to the

19     second column for Prijedor.  There is a number there, 40.822.  40.822.

20             This is --

21        Q.   I'm sorry, Dr. Tabeau.  I'm looking on the screen.  At line 14

22     for Prijedor I see 40.825.

23        A.   Oh, sorry.  25, yes.  I was seeing 25 but, sorry.

24        Q.   Yeah, go ahead.

25        A.   So 40.825.  So this is the 1991 census population of Serbs from


Page 28163

 1     Prijedor.  Of course, it is only those who were born before 1980.

 2     Otherwise, I could not compare this number with what I have in the

 3     voters' register.

 4             So 40.825, I multiply this number by 2.7 per cent, that is my

 5     sampling proportion of displaced Serbs from Prijedor.  In result of this

 6     multiplication, we see the number 1.089 in Table 3, of course, and in our

 7     presentation under the estimated number of displaced Serbs from

 8     Prijedor ... if we go to the presentation, page 2.

 9             MS. WEST:  65 ter 23719, e-court page 2, please.

10        A.   Table 2, at the bottom of the page.  For Serbs, estimated number

11     of IDPs and refugees is 1.089.  That is how the method works.

12             So for any other ethnic group, the sampling proportions are

13     different, and in result -- and, of course, the 1991 census populations

14     for each ethnic group are different, the size of them.  So this results

15     in the numbers listed as estimated numbers in this third column, and the

16     total of which for all ethnic groups is 45.688.  That is the estimated

17     more complete number of displaced persons and refugees from Prijedor.

18        Q.   Okay.

19             JUDGE KWON:  Just a second.  We see here -- I'm coming back to

20     the number of 40.825 who lived in Prijedor in 1991.

21             THE WITNESS:  Yes.

22             JUDGE KWON:  And those are Serbs?

23             THE WITNESS:  Yes.

24             JUDGE KWON:  So bearing that in mind, could we go back to the

25     page 52 of the report.


Page 28164

 1             MS. WEST:  65 ter 10723.

 2             JUDGE KWON:  Yes.

 3             MS. WEST:  ERN 09--

 4             JUDGE KWON:  I think it's --

 5             MS. WEST:  I'm --

 6             JUDGE KWON:  I'm coming back to page 51.

 7             MS. WEST:  0986, please.  ERN page 0986.

 8             JUDGE KWON:  Here, total population identified in 1997 who lived

 9     in Prijedor in 1991 was 49.019.

10             How -- could you tell us the difference between these two

11     numbers?

12             THE WITNESS:  Your Honour, the number 40.825, these are

13     1991 Serbs, should be compared with the number 24.070 from Table 2S,

14     page 51.

15             So you mentioned, Your Honour, the number 49.000 from this table.

16     49.019.  Which is a larger number because it covers all ethnicities.  If

17     we want the Serbs, to compare with the Serbs, then we have to move to the

18     right in this table and look at the number 24.070.  24.070 is a sample of

19     those 40.825.  It's a sample, a large sample, very large sample, from

20     those reported in the census, and I use this sample to produce the more

21     complete number of displaced persons.

22             JUDGE KWON:  Then if you could tell us again the meaning of

23     40.825.

24             THE WITNESS:  40.825 is the number of Serbs listed in 1991 census

25     for Prijedor, and they were born before 1980.


Page 28165

 1             JUDGE KWON:  And 24.000?

 2             THE WITNESS:  24.070 is a sample of those Serbs listed in the

 3     census in Prijedor who were, as well, identified as voters registered in

 4     1997 elections.

 5             JUDGE KWON:  So, first, 4.4 million versus 2.5 million?

 6             THE WITNESS:  That is very correct, yes.  And linking persons,

 7     the same -- the same, very same persons in both years.

 8             JUDGE KWON:  Thank you.

 9             Shall we move on?

10             MS. WEST:  Thank you very much.

11             If we can go back to the handout, so it's 23719, e-court page 2.

12     Thank you.

13        Q.   We've -- we were just looking at the Serbs estimated number

14     1.089.  Now, if we continue along that row, we see 95 per cent confidence

15     interval.

16             Dr. Tabeau, is this a type of statistical error that you're

17     accounting for?

18        A.   Yes.  This estimated number is a point estimate.  Point estimates

19     don't tell us with a certainty about what we estimate.  They tell us with

20     a certainty probability or confidence that this unknown estimated number

21     would be equal, what is shown in Table 2.  In this particular case, the

22     confidence interval expresses the range of uncertainty, assuming

23     5 per cent error of the estimation.  So if we go to Serbs and look at the

24     estimated number 1.089 and the numbers -- two numbers next to it, the

25     first one is 1.036, the second one is 1.142, so you can see that point


Page 28166

 1     estimate is somewhat there between these two values, lower end and upper

 2     end of this interval, and the interval itself gives us the 95 per cent

 3     confidence that our value is correct.

 4        Q.   Thank you.  And now if we'll go to the last two columns.  You've

 5     got a minimum number percentage and an estimated number percentage.

 6             Tell us about those two columns.

 7        A.   It is that in these columns some numbers from this particular

 8     table are related to each other.  A simple calculation of minimum number

 9     divided by the 1991 population is reported under 1.6 per cent.  That is

10     the share of the minimum number of displaced in the 1991 census

11     population.  And next to it, we see the familiar 2.7 per cent.  That is

12     the number that relates the estimated unknown number of displacement to

13     the 1991 population.

14        Q.   Thank you.  Doctor, I'm not going to go through the Muslim

15     numbers with you, but just like the estimate we gave for the Serbs in

16     Prijedor who were displaced as 1.089, what is the estimate you give for

17     the Muslims from Prijedor that were displaced?

18        A.   Well, the number of displaced Muslims from Table 2 is 1900 --

19     sorry, 19.290.  That is the second row, second column.

20        Q.   And that's the minimum absolute number; correct?

21        A.   Yes.  It's the minimum absolute number.

22        Q.   What's the estimate?

23        A.   The estimate is 39.322.

24        Q.   I now want to --

25             JUDGE KWON:  And finally, that may be a matter of course, when


Page 28167

 1     discussing the person -- internally displaced persons, the reason for

 2     such displacement was not looked into.  Whether they were forced or they

 3     voluntarily moved, you didn't look into such matters.

 4             THE WITNESS:  No.  Causes -- causes of displacement are not

 5     looked at at this -- in this report.  It is just measurement of

 6     displacement.

 7             JUDGE KWON:  Thank you.

 8             MS. WEST:

 9        Q.   So now, Doctor, I'd like to talk about -- we just talked about

10     Prijedor in particular.  Now I would like to talk about the overall

11     number in the 27 municipalities.

12             If we can go to the next page of the handout, and we're going to

13     look at the bottom table, Table 4, which is the same table we just looked

14     at for Prijedor, but this accounts for all 27.  And, here, let's focus on

15     the Muslim line going across.

16             So if we start at 481.109, tell us what that represents.

17        A.   It's the 1991 census count of Muslims in the 27 municipalities of

18     the Karadzic case area.

19        Q.   And then the 161.047, what number does that -- what does that

20     represent?

21        A.   This number represents the number of -- minimum number of

22     displaced Muslims from these municipalities.

23        Q.   And what was the estimate you came up with, using a sampling

24     proportion for Muslims?

25        A.   It is 284.239.  That is the estimated more complete number of


Page 28168

 1     displaced Muslims from these municipalities.

 2        Q.   Now, the President just asked you whether you analysed the causes

 3     of their displacement, and, of course, your answer to that was no.  But

 4     could you analyse the timing of their displacement, what time between

 5     1991 and 1997 these people moved out?

 6        A.   It is not easy to analyse the timing because generally sources

 7     reporting on timing are not available.  However, we were able to find a

 8     few sources reporting on early 1993 and later early 1995 which we could

 9     use to compare our numbers with.

10        Q.   If we can go to the next page of the handout.  This will be

11     page 4 in e-court.  See on top it says Table 2, population change in some

12     municipalities?

13             And here we see the table begins separated into ethnicities.  It

14     begins in 1991.  It ends in 1997.  But in the middle we have the years

15     1993 and 1995.  What sources did you consult to get informations [sic]

16     from 1993 and 1995?

17        A.   These two sources were compiled by authorities of

18     Republika Srpska, specifically by the Ministry of Interior of

19     Republika Srpska, in the course -- during the conflict, actually.  These

20     are relatively brief sources that cannot be seen comparable with

21     population census made by statistical authorities, yet they report

22     approximate numbers that are very useful for our purposes to compare with

23     our statistics.

24        Q.   And here we only see a listing of nine municipalities.  Why are

25     there only nine and not 27?


Page 28169

 1        A.   This is because of the specificity of these sources.  They only

 2     covered the so-called Banja Luka region, which is different generally

 3     than the 27 municipalities studied here.  However, there is a number of

 4     municipalities, specifically nine, that overlap with -- between the two

 5     sources, the two types of sources: our and RS sources.

 6        Q.   And we will focus particularly on the Muslim table on top.  And

 7     since we just spoke about Prijedor, let's look at Prijedor again.

 8             And, here, for Prijedor, in the column that says 1993 "(moved

 9     out)," we have 42.000.  What does that number represent?

10        A.   This is the approximate number of Muslims from Prijedor that are

11     reported here by RS sources as gone, moved out from this municipality, as

12     of early 1993.

13        Q.   And then, for the column 1995, it says "(residents)" for

14     Prijedor, 3.600.  What does that represent?

15        A.   This number represents the Muslims registered still by RS sources

16     still in the municipality of Prijedor.  It would be 3.600 out of 1991

17     population reported here for Prijedor as 49.351.

18        Q.   And so in regard to timing, what was your conclusion as to when

19     the Muslims in Prijedor moved away?

20        A.   Obviously, already by early 1993, a large number of Muslims from

21     Prijedor is reported by independent sources as moved out.  So that is a

22     good indication of the timing when people were moving and becoming

23     displaced.

24        Q.   These blue boxes we see on these two tables, do these just

25     indicate corrections you made to your report since the last time?


Page 28170

 1        A.   Yes.  There was a mistake for Sanski Most that has been

 2     corrected, and this is highlighted blue.

 3        Q.   If we can move to the next page, please, of the handout --

 4             JUDGE KWON:  While we are on this page --

 5             MS. WEST:  Mm-hm.

 6             JUDGE KWON:  -- can you take the table at the bottom of the page

 7     which refers to Serbs, which should be C, and not B.

 8             THE WITNESS:  Yes.

 9             JUDGE KWON:  Here, in Prijedor, we see the number 47.581 in 1991.

10     Did we see this number before while looking at various tables related to

11     the Prijedor and Serbs?

12             THE WITNESS:  Your Honour, we couldn't see this number because

13     this number includes all ages of Prijedor population.  That means all

14     those, disregarding when they were born.

15             In our analysis, in previously seen tables, we looked at a

16     segment, a large one, but a segment of population, of all those who,

17     later on in 1997, became 18 years of age and could vote and register.  So

18     our population, 1991 population, is obviously smaller than what is

19     reported in this table.

20             JUDGE KWON:  Thank you.

21             MS. WEST:  If we can turn to the next page of the handout.

22        Q.   And here we see two graphs, Figure 14 and 15.  Are these just

23     visualisations of what we saw on the previous page?

24        A.   Yes.  These are the same numbers as on the previous page.

25     Figure 14 represents Muslims.  Figure 15, Serbs.


Page 28171

 1             What is shown in these two figures is the population in 1991

 2     versus 1995.  So the dark brown bars represent 1991.  Green ones, 1995.

 3     For both.  And a per cent is shown instead of absolute numbers, as

 4     per cent is a measure that can be easily compared.

 5             For Muslims, the pattern clearly indicates that in 1995 the

 6     fractions were extremely low compared with 1991, which just further

 7     confirms what we studied and obtained in our report.  And Figure 15 shows

 8     the opposite.  The fractions of Serbs increased considerably in 1995

 9     compared to 1991, and, again, the same result was obtained in our report.

10        Q.   Thank you, Dr. Tabeau.

11             On the following page of the handout, you did a similar analysis

12     of the Srebrenica municipalities; correct?

13        A.   Yes, it is correct.

14        Q.   Now, this is an analysis for IDPs, internally displaced people,

15     not for killed people; right?

16        A.   Yes.  It is displacement, not killed people.

17        Q.   And in this analysis did you apply the same methodology that we

18     talked about for the Prijedor numbers?

19        A.   Yes.  It is the same methodology and the same sources.

20        Q.   Okay.  So I would now like to move --

21             MS. WEST:  Mr. President, I'm going to move now to the Sarajevo

22     reports.

23             JUDGE KWON:  While we are moving, I have to -- I'd like to note

24     that we passed a number of Prosecution Exhibit 5000.  65 ter 12133 has

25     the honour to have the number P -- Exhibit P5000.


Page 28172

 1             MR. ROBINSON:  Our congratulations to the Prosecution.

 2             MS. WEST:  My congratulations to Dr. Tabeau.

 3             THE WITNESS:  I'm the lucky person.

 4             MS. WEST:

 5        Q.   So we'll move on to Sarajevo.  Can you tell us - you've been at

 6     the OTP for several years - what was the first study that you did

 7     regarding the events of Sarajevo?

 8        A.   Chronologically the first report I made was for Galic case.

 9        Q.   Okay.  And that Galic report was for the period of time

10     September 1992 to August 1994.  Was that the period of time when Galic

11     was the Sarajevo-Romanija Corps commander?

12        A.   Yes.  As far as I recall, yes.

13        Q.   What was the next report that you did on Sarajevo?

14        A.   The next one was for the case of Dragomir Milosevic.  And the

15     coverage was September 1994 to November 1995.

16        Q.   Right.  And was that the period of time where Dragomir Milosevic

17     was then the corps commander?

18        A.   Yes.

19        Q.   Okay.  And what was the last report that you wrote for Sarajevo?

20        A.   The last report I made for this case, for the Karadzic case, and

21     it covers the first five months of the siege.  That is, from April 1995

22     to 9 September 1995.

23        Q.   And would that be April of 1992?

24        A.   Sorry.  Of course, 1992.  Yes.  I'm very sorry.

25        Q.   So the Karadzic report covers the period of time not covered by


Page 28173

 1     Galic and the Milosevic reports?

 2        A.   Yes.  These reports are exclusive.  So three of them covered the

 3     entire period from April 1992 to November 1995.

 4        Q.   All right.  And we will talk about all three.  And we're going to

 5     do it chronologically.  But first I want to ask you:  What was your task

 6     for all of these reports?

 7        A.   My task was to come up with a number, reliable number, of victims

 8     of the siege, both civilian victims and military victims.

 9        Q.   And for each of these three reports, did you use the same

10     methodology?

11        A.   Two of these reports were based on the same methodology and the

12     same sources, and the third one, the Dragomir Milosevic report, was

13     slightly different, still within the standard methodology used throughout

14     the years here in this Tribunal.

15             MS. WEST:  May we have 65 ter 12130, please.  12130, yes.

16        Q.   Doctor, this is your report for this case.  It's April 1992 to

17     September of 1992.  What area, what geographical area, does this report

18     cover?

19        A.   The area covered in this report is the so-called

20     within-front-lines area of Sarajevo.  In terms of municipalities, this

21     within-front area covers large parts of six urban municipalities of

22     Sarajevo, and they are all listed in my report.

23        Q.   What was the main source you used in preparation of this report?

24        A.   The main source used to compile statistics, again, minimum

25     numbers, on killed persons and wounded person was the so-called Sarajevo


Page 28174

 1     household survey.  This is a large survey conducted in Sarajevo by local

 2     researchers by mid-1994.  So the front -- within-the-front-lines area

 3     should be seen as the area that was located inside the front lines as of

 4     this moment.

 5        Q.   Okay.  And we will speak in detail about the survey in a moment.

 6             But were there other sources that you consulted peripherally as

 7     well?

 8        A.   Of course.  There was another source reporting on that, the

 9     so-called Bakije funeral home, the largest and the oldest funeral home of

10     Sarajevo that was burying the Muslim population, exclusively Muslims.

11     Well, we used this source mainly to produce an estimate of a more

12     complete number of victims.  This source is not used to compile

13     statistics on killed and wounded population.

14             We also used another source, the 1991 population census, which we

15     actually always use when working on victims in order, first of all, to

16     validate the -- the information about persons listed in a source on

17     victims with how the same person is reported in the census, that is, the

18     validation of victim records and possibly improving them using the census

19     population information.  And secondly, we used the census in order to

20     acquire the ethnicity of victims.  We always use this one definition of

21     ethnicity based on the open-ended question from the 1991 population

22     census.

23             In addition to the census, we also used military lists of all

24     three armies that were engaged in combat in Bosnia, and these lists were

25     used to give us information about who can be considered a military


Page 28175

 1     person, a soldier, and who, a civilian.

 2             In this particular case, in Sarajevo household survey, the

 3     civilian military status was, however, also available from our main

 4     source.

 5        Q.   Okay.  So we are going to talk about the main source, this

 6     household survey.  And, for the record, you wrote in detail about this in

 7     your Galic report, starting at page 10.  We don't need to go there.

 8             But can you tell us when the household survey of Sarajevo was

 9     conducted?

10        A.   It was conducted in mid-1994, as far as I remember, from April to

11     August, or even early September of 1994.

12        Q.   And who conducted it?

13        A.   The group who conducted it and designed the questionnaire were

14     researchers from the Institute for War Crimes and International Law in

15     Sarajevo.  They were able to develop this questionnaire in this difficult

16     time for the importance of this survey to everybody, and also with a lot

17     of assistance from the statistical authority of Bosnia-Herzegovina.

18        Q.   What area was covered by the survey?

19        A.   As I mentioned, the area was within the front lines, six urban

20     municipalities, large parts of these municipalities.  Obviously the

21     number of individuals and households included in this survey must have

22     been high, and it was.  It was a large number of 85.000 households that

23     were interviewed in this survey.  85.000 households are equivalent to

24     approximately 340.000 individuals.

25        Q.   Okay.  And you said there was interviews.  Were they face-to-face


Page 28176

 1     interviews?

 2        A.   Yes.  These were face-to-face interviews conducted by a group of

 3     interviewers that were trained and instructed how to question the

 4     households in this survey.

 5        Q.   You just mentioned that it was 85.000 households equivalent to

 6     approximately 340.000 individuals.  How do you know it was

 7     340.000 individuals?

 8        A.   It is an approximate number.  The exact number, I don't know, and

 9     I don't know because this huge massive material has never been

10     computerised in whole.  Only the part that was needed for our work was

11     computerised, and yet it was a lot of work at that time.

12             This number 340.000 individuals is an estimation based on the

13     assumption that the average size of one household is four persons.

14        Q.   Prior to the war, did you know what the average size of a

15     household was?

16        A.   Yes, it was smaller than four persons.  It was three-something.

17     There is a footnote in my report which tells the exact number, but --

18        Q.   Why did you increase that number?

19        A.   I increased that number because there were newcomers to Sarajevo

20     who were staying with their families who originally lived in the Sarajevo

21     area, so that is one reason.  And in the Galic report, we also calculated

22     demographic rates, that is, the relative measures of killing and

23     wounding.  And if we would keep the population size small, this would

24     affect the rates by increasing them, and we don't want to do that.  We

25     rather keep the rates conservative.  That means we rather include more


Page 28177

 1     population at risk.

 2        Q.   Okay.  Now, you mentioned -- I think you mentioned there was a

 3     questionnaire.  And the actual questionnaire is in the annex to the

 4     Karadzic report.  But can you tell us in particular the questions in

 5     which you were interested?

 6        A.   These were questions related to whether there were family members

 7     killed during the war in this given family, whether there were persons

 8     wounded, whether there were natural deaths.  I think these were the three

 9     major questions that I was interested in.

10        Q.   Okay.  And once the interviews were done -- strike that.

11             The process of computerising and screening all this information,

12     was that a process that you yourself were involved?

13        A.   Yes.  We actually didn't even try to request a database from the

14     institute who -- who made this survey because they didn't have one.  What

15     they had were the original questionnaires.  The questionnaires needed to

16     be computerised and processed, and this is what I was in charge of and in

17     which I was involved with some other colleagues from my unit.  So we

18     actually selected a sample of original questionnaires and copied them and

19     arranged data entry, and all this was done as our project, as the

20     Demographic Unit project.

21        Q.   And out of the 85.000 household questionnaires, how many did you

22     take?

23        A.   Approximately 40.000.  This is the number of questionnaires that

24     reported either on killing or on wounding in the family, or on natural

25     deaths, or combinations thereof.


Page 28178

 1        Q.   Okay.  And the remaining -- the other 45.000, they didn't have

 2     information in which you were interested; is that correct?

 3        A.   That is correct.

 4        Q.   All right.  Now, we've already spoken about the matching or the

 5     linking process.  Was that the same process used here?

 6        A.   Yes, it was the standard procedure, matching procedure, as in

 7     other projects.

 8        Q.   And the 40.000 questionnaires that you had, did you match them to

 9     the census?

10        A.   Not all of them.  There is a high matching rate though, more than

11     81 per cent.

12        Q.   By doing so for the 81 per cent, did you become aware of how that

13     81 per cent reported their ethnicity?

14        A.   Yes.  Through the links, we could learn about how they reported

15     themselves in the 1991 census.

16             MS. WEST:  If we can go to the handout number 23719, looking for

17     e-court page 7, please.  So it will be number -- page 6 in the handout.

18     This is a table on the Karadzic Sarajevo report that we have put in the

19     handout, and this is from page 4 and 5 of the report.

20             Now, I'd like to look at some of the results that you came up

21     with.  And here we see Table 1, an overview of the summary statistics

22     from the household survey.

23             We see under variables, the column labelled "variables," sample

24     size.  Tell us what you mean here by sample size.

25        A.   It sample size should be understood here as a selection of cases,


Page 28179

 1     of records, from the overall computerised Sarajevo household survey

 2     questionnaires.  The sample size should be seen as the size of victims

 3     population.  Records were selected based on the relevance to this

 4     particular indictment period, which is April 1992 to 9 September 1992, so

 5     only these records were selected that were falling within this

 6     time-frame.

 7             So under the sample size in the second column, there is the

 8     overall total.  And under the column "killed," there is a number

 9     reported.  2.028 is the number.  And that's the number, a minimum number,

10     of killed persons selected because of the relevance to this particular

11     indictment period.  All these cases of killing took place within the

12     front lines of Sarajevo as of mid-1994.

13             And the next number, absolute number, in the second-last column,

14     is 7.374.  I repeat the number: 7.374.  This is the number of wounded

15     persons from the same territory and the same time-period.

16        Q.   Thank you.  And included on the questionnaire was information

17     that the person killed or wounded was a civilian or a soldier; is that

18     right?

19        A.   Yes.  There was an indication of the civilian/military status.

20        Q.   All right.  And in addition, in describing the event that caused

21     the killing or the wounding, what type of information was included?

22        A.   We had information about the year of birth, which made it

23     possible for us to study ages of those killed and wounded, and most

24     importantly we also had information about the cause of killing and

25     wounding, a cause that we were able to process according to categories


Page 28180

 1     such as shelling, sniping, fire-arms, other fire-arms, and a number of

 2     additional other causes.

 3        Q.   Okay.  And that information is at the bottom of this table;

 4     correct?

 5        A.   That is correct.

 6        Q.   The second table we see, here you just took out the civilians; is

 7     that right?

 8        A.   That's right.

 9        Q.   So if we turn now to the next page of the handout - page 7 in the

10     handout, page 8 in e-court - this regards causes of death and causes

11     of -- cause of wounding.  And here we see five particular categories.

12             Can you tell the Chamber how it is that you organised these five

13     categories?

14        A.   The causes were reported by respondents without any instructions

15     or standardisation of categories.  People reported causes as they knew

16     them.  So my task was to study these reports and to come up with a

17     meaningful classification that could be done in such a way that every

18     cause would be reported in a consistent way.

19             During our process of standardisation, we kept the originally

20     reported categories as often as possible.  So what see -- what we can see

21     under fire-arms, shelling, sniping, general shooting, are categories that

22     were originally reported by the respondents.

23        Q.   Now, I just want to focus on that category of fire-arms.

24             On the column to the extreme right on each table, you have a

25     percentage.


Page 28181

 1        A.   Yes.

 2        Q.   When you add up the percentages on the first table, Table 2, for

 3     fire-arms, so we begin with 59.7 per cent, 14.1 per cent --

 4        A.   Oh, here.

 5        Q.   -- and down to 5.1 per cent.

 6        A.   Yes.

 7        Q.   -- what is the total percentages of deaths caused by fire-arms?

 8        A.   It is 88.4 per cent, which is a very large per cent of all causes

 9     reported by respondents.

10        Q.   Now, that's for people killed.  Now, if we move over and go to

11     wounded people and do the same thing and add up the percentages for

12     fire-arms, what is the total per cent?

13        A.   This per cent is even higher, and it is equal, 96.8 per cent.

14     Almost 97 per cent.

15        Q.   So far, in this report, the numbers we have spoken about are

16     minimum numbers.  They're direct, actual numbers that you yourself

17     counted.  Were you able to translate these numbers into actual estimates

18     of people killed and wounded?

19        A.   Yes.  The numbers which are included in Table 1 are minimum

20     numbers.  They are minimum numbers because even though our source was

21     very large and reliable, still it is not that one source can report all

22     victims of war.  This is not how the experience shows us things are done

23     during conflict.  So it is certain that Sarajevo household survey is not

24     a complete source.  The estimation that we made included two sources: One

25     is the Sarajevo household survey, and the second one is the Bakije


Page 28182

 1     funeral home.

 2        Q.   Okay.  So in order to give an overall estimate of numbers of

 3     killed and wounded in Sarajevo in this time-frame, what statistical tool

 4     did you use?

 5        A.   We used a simple method which is used particularly for this type

 6     of estimation, the method called capture/recapture.

 7        Q.   And with this method, does it require one independent source or

 8     two?

 9        A.   This method requires at least two independent sources.  It is

10     possible to have three, if exists.  In our case, we had two sources,

11     independent sources, at our disposal, so capture/recapture estimation was

12     completed based on two.

13        Q.   Now, one of those sources we have already spoken about in detail,

14     and that's the household survey report.  And the second source you

15     mentioned was the Bakije funeral home records.  Can you tell us a little

16     bit more about that source?

17        A.   The Bakije funeral home is a Muslim funeral home.  That means

18     from this source we have information about deaths of Muslims exclusively.

19     And the number for the entire war period reported in this source is

20     approximately 12 -- more than 12.000, which compares with other sources,

21     like Sarajevo household survey.  Only that this is just one ethnic group,

22     the Muslims.

23        Q.   All right.  And these records cover a time-frame that is bigger

24     than the time-frame for this report; correct?

25        A.   Correct.  But information about dates of funeral, which are very


Page 28183

 1     close to the dates of death, we can extract the relevant records for our

 2     purpose, for our estimation.

 3        Q.   Now, you've already mentioned the statistical tool called

 4     capture/recapture.  In a very general way, can you explain to us what

 5     that is?

 6        A.   Yes, I will.  One more remark about the Bakije funeral home.  It

 7     covers both killings and natural deaths.  It's one important starting

 8     point for how we applied the capture/recapture.

 9             Capture/recapture is a method that was originally developed to

10     estimate the size of wild populations that cannot be easily measured.  If

11     we think of fish, for instance, then it is not possible to go to the

12     ocean and then count them all.  In such cases, samples are taken of fish

13     from the sea and compared, cross-referenced.  And based on this,

14     estimation can be made which provides us with unknown total.

15             The same method was applied to human populations, like, for

16     instance, HIV patients who are also counted with huge difficulties, were

17     estimated in various studies using capture/recapture.  The method is

18     based on a simple assumption, that the overlap between two samples is a

19     meaningful piece of information that can take us to produce the unknown

20     overall size of population.

21             Well, in order -- in common language:  If we have a sample taken

22     from a population, larger population, if we mark individuals selected for

23     the sample and let them go to mix up with the rest of the population,

24     would be the first step in capture/recapture.

25             In the second step, a second sample is drawn, and we check how


Page 28184

 1     many individuals in the second sample have been marked as also present

 2     previously in the first sample.  The individuals checked to be first in

 3     the first sample and overlap with the second sample is the overlap of the

 4     two samples.

 5             The method is based on a simple assumption, that the proportion

 6     of the first sample size to the unknown population size is exactly the

 7     same as the proportion of recaptured individuals to the size of the

 8     sample 2.

 9             So we have two proportions.  One proportion illustrating the

10     recaptures individual -- individuals, in sample 2, in relation to the

11     size of this sample, and the other proportion is the entire sample 1 to

12     the unknown population --

13        Q.   And --

14        A.   -- size.

15        Q.   -- if you apply that proportion to the entire population, is that

16     how you get your estimate?

17        A.   Yes.  Indeed.  This is, again, like extrapolation of sampling

18     proportion in order to estimate the unknown overall population size.  The

19     details of this method are described in the report in Annex --

20        Q.   The details of this are in the Karadzic report, page 30 and 58.

21             MS. WEST:  If we can go to the next page in the handout, please.

22             So what we'll see here is Table 12 where you applied that

23     estimation.  And we see that it's split up into ethnicities for this

24     period of time in 1992 in this particular geographical location.

25             Under "observed," we see Croat, 53.  What does "observed" mean?


Page 28185

 1        A.   These are exactly same numbers as those reported in Table 1.  So

 2     the overall total in Table 12 is 2.028, and it's exactly the same as the

 3     overall number of killed in Table 1.

 4             So observed are the counts based on Sarajevo household survey.

 5     Ethnicity comes from links of victim records with the 1991 census.  There

 6     is a small category of unknowns, 470; for those we don't have ethnicity,

 7     as they are unmatched.

 8        Q.   Okay.  So the ethnicity that you reported here, this is -- is

 9     this how these people reported themselves in the census?

10        A.   Yes.

11        Q.   And when we go from 53 Croats observed to the estimated 80, to

12     get that number, did you apply the capture/recapture theory?

13        A.   Yes.  But the method it was applied is rather complex, because,

14     as I mentioned, the funeral home, Bakije does only include information

15     about Muslims.  Moreover, it includes both killings and natural deaths.

16             So capture/recapture was in the first step applied to the Muslim

17     population exclusively.  Muslims, both killings and natural deaths, as

18     reported in both sources.  So overlap was studied only for Muslims, both

19     killed and natural deaths, in both sources.

20             Applying capture/recapture to one category is not complex

21     mathematically and statistically.  So once this was done, in the next

22     step we extrapolated these results.  First we split the estimate between

23     killings and natural deaths.  And further, when looking exclusively at

24     killings and saw how much Muslim killed population increased from

25     capture/recapture results, once this was finished, we applied the same


Page 28186

 1     degree of increase to every other ethnic group.

 2             So the method was applied in several steps, in three steps.

 3     Basically what we present here in Table 12 are relevant, 100 per cent

 4     relevant, results related to killings only and to the time-frame of

 5     indictment period.

 6        Q.   Okay.

 7             JUDGE KWON:  Yes, Ms. West.  So let's take the example of

 8     Muslims, which we see in this Table 12.

 9             So could -- could the doctor explain us how 1.811 number could be

10     reached using that capture and recapture estimation?

11             THE WITNESS:  Yes, of course.  Yes, of course.

12             I suggest we move to the annex from the report.

13             MS. WEST:

14        Q.   Doctor, would that be Annex 6?

15        A.   Page 59.

16             MS. WEST:  This is 65 ter 12130.  ERN ending in 8525.  It's

17     almost the end of the report.

18             JUDGE KWON:  Second-last page.

19             MS. WEST:  Thank you.

20             THE WITNESS:  If we could please take a look at Table A6.1 in the

21     middle of this page.

22             This table summarises the input data used for the estimation.  We

23     have numbers of both killed persons and natural deaths shown as available

24     from the two sources.  First one, HSS 1994, is the Household Survey

25     Sarajevo.  In this source, the number of killed Muslims and natural


Page 28187

 1     deaths of Muslims equals 1.707.  The number of the same type of deaths

 2     reported in the Bakije funeral home is 1.576.

 3             From the matching of the two sources at the individual record

 4     level, we obtain the overlap between these two sources.  The overlap

 5     means these are cases reported in both sources.  This number, overlap, is

 6     1.039.

 7             If we go down, there is a formula which tells us the capital N

 8     with a hat, which is the unknown size of the population that we estimate,

 9     can be obtained from a simple operation which multiplies Sample 1 by

10     Sample 2 and divides this multiplication by the overlap.  From this

11     formula we obtain the number reported in the last paragraph at the

12     bottom, which says N-hat equals 2589.251.  This number, 2.589,

13     approximately, is the estimated number of both killed Muslims and natural

14     deaths of Muslims.  That is the first step of our capture/recapture

15     estimation.

16             Starting from here, we apply the proportion of killings and

17     natural deaths as obtained from Sarajevo household survey and distribute

18     this number into two categories.  And one portion obtained from the

19     proportion of killings in Sarajevo household survey gives us the number,

20     estimated number, in 12 -- in Table 12, which equals 1.811.

21             JUDGE KWON:  Thank you.  Just -- just a second.  I have to admit

22     that this is beyond my capability.

23             Mr. Robinson, do you challenge this formula which is related to

24     this purely the capture/recapture method, scientifically?

25             MR. ROBINSON:  I'm going to defer this to Dr. Karadzic because


Page 28188

 1     he's been working with our expert who doesn't speak English, so I myself

 2     don't know.

 3             THE ACCUSED: [Interpretation] Excellencies, the sample with

 4     plants and animals in an ocean, where it's impossible to have a different

 5     register, is absolutely not applicable to a civilised environment --

 6             JUDGE KWON: [Previous translation continues] ... I'm not asking

 7     you as to the applicability of this formula to the Sarajevo case.  But my

 8     question is whether you challenge this formula itself when it applies to

 9     wild animal or whatever.  Otherwise I have to ask the doctor to explain

10     in more detail so that layman could understand.  Or give us some examples

11     of this method.

12             If you do not challenge this formula as such, we can continue.

13             THE ACCUSED: [Interpretation] I challenge the formula.

14                           [Trial Chamber confers]

15             JUDGE KWON:  But if you do so, it will be noted that your expert

16     is challenging this formula itself.

17             Very well.

18             JUDGE BAIRD:  Now, Dr. Tabeau, you have heard that Dr. Karadzic

19     is challenging the formula.  Would it be possible for you to give it to

20     us in a more simplified form so that it could be a little more easily

21     digested?

22             Is it possible at all?

23             THE WITNESS:  Honestly, I thought I was giving it in a very

24     simple form.

25             JUDGE BAIRD:  Where you really?


Page 28189

 1             THE WITNESS:  I thought I were.

 2             JUDGE BAIRD:  I see.

 3             THE WITNESS:  So to think of an even simpler example is hard, but

 4     I can prepare something.  I can do my best and come up with a very simple

 5     example, because, yes, it is a simple method which is based on equality

 6     of two proportions.

 7             JUDGE BAIRD:  But we shall be grateful for that even more

 8     simplified form.

 9             THE WITNESS:  Of course.

10             JUDGE BAIRD:  Extremely grateful.

11             THE WITNESS:  Of course, yes.  One more remark, if I may.  It is

12     not that I invented it.  It is that I have references for the authors and

13     applications.  And it is not what -- that I'm telling that I came up with

14     this formula and all kinds of things like, you know, error estimation and

15     stuff.  There is a huge literature on this.

16             JUDGE BAIRD:  We weren't implying that at all.  We weren't.

17             THE WITNESS:  Okay.

18             MS. WEST:  Mr. President, may I make a suggestion?

19             JUDGE KWON:  The Chamber has a meeting, some internal meeting at

20     12.30, so I was about to suggest to adjourn now for a break.

21             And then can I hear your suggestion, Ms. West.

22             MS. WEST:  My suggestion was going to be that Dr. Tabeau is going

23     to have to come back another day.  And if we could hold off on this

24     example so she can give us a meaningful thought --

25             JUDGE KWON:  Yes.


Page 28190

 1             MS. WEST:  -- maybe we could address it on Tuesday.

 2             JUDGE KWON:  Thank you.

 3             JUDGE BAIRD:  Fair enough.

 4             JUDGE KWON:  And were the parties informed that on Tuesday we'll

 5     start at 11.00 in light of the fact that the accused will have an

 6     interview with a certain witness I don't know?

 7             We'll rise for the break and we'll resume at 1.30.

 8                           --- Luncheon recess taken at 12.21 p.m.

 9                           --- On resuming at 1.34 p.m.

10             JUDGE KWON:  Yes, Ms. West, please continue.

11             MS. WEST:  Thank you, Mr. President.  I would just like to make

12     three corrections to the transcript: Page 9, line 19, it should be

13     "Muslims" and not "multiples."  Page 13, line 6, it should be "Annex A,"

14     not "Annex C."  Page 39, line 11, it should be "causes," not "cautions."

15        Q.   Dr. Tabeau, when we finished at the last session, we were talking

16     about estimates.  And now I want to move on to another part of this

17     report that regards timing.  This is another area that you analysed.  It

18     was the timing of civilian deaths per day versus soldiers killed per day.

19             Can you tell us what the purpose was of this particular analysis?

20        A.   The analysis of timing has the purpose to investigate the siege

21     intensity.  This can be done when analysing the timing of killings and

22     woundings of soldiers.  And in addition to this, the purpose was to

23     investigate what was the timing of killings and wounding of civilians.

24     Whether these two timings were consistently the same or different.

25     Whether the numbers of killed civilians were the same, higher, or lower


Page 28191

 1     compared with the numbers of killed soldiers.

 2        Q.   Before we go into the results in detail, can you just tell us in

 3     general what you found?

 4        A.   Generally, the starting point for this analysis was that when we

 5     look at timing of killings or woundings of civilians in relation to

 6     timing obtained for soldiers, that the killings of civilians should be

 7     insignificantly -- insignificant compared with the numbers of soldiers.

 8     If there are combat activities, our assumption was that in combat

 9     soldiers are killed or wounded, on the first place.  If accidentally

10     civilians are killed, then the numbers of killed civilians should be much

11     lower than those of soldiers.  And this observation should be consistent

12     for the entire siege period.

13             So that was the starting point.

14             The outcome of our analysis confirmed that in case of this

15     episode of the siege of Sarajevo, the expectation has not been confirmed.

16     There were many days in the siege when the numbers of killed civilians

17     were much higher than the numbers of killed soldiers, or the same, or

18     just little lower than the number of killed soldiers.  So this finding is

19     contrary to the initial expectation, and this is what we elaborated in

20     the Karadzic report in the part related to timing.

21             MS. WEST:  If we can have e-court page 18, and I believe we're on

22     65 ter 12130 already.  Thank you.  The top paragraph is what we'd like to

23     look at.

24        Q.   This is the part of your report that regards timing, and I just

25     have a couple questions for you in the top paragraph.


Page 28192

 1             Here you wrote about what you just termed as an assumption:

 2             "The reason for this is a simple belief [sic] that if civilians

 3     were not actively involved in fighting and stayed aside from it, one

 4     should expect that their deaths were accidental and unintended and thus

 5     must be proportionally lower than the number of killed soldiers."

 6             So is this what you meant when you talked about the assumption

 7     you had made?

 8        A.   Yes, this is exactly what I meant.

 9        Q.   And did your support -- did your results support this assumption?

10        A.   No.  The results support the contrary.

11             MS. WEST:  May we have e-court page 28, please.

12        Q.   Just below the box, you're talking your results.  And you say

13     that:

14             [As read] "The above observations lead to the conclusion that two

15     different processes were responsible for the victimisation of civilians

16     and soldiers.  One process was the driving force behind the killings and

17     woundings of civilians, and another process stayed behind the killings

18     and woundings of soldiers."

19             Dr. Tabeau, what do you mean by here by "process"?

20        A.   Yes, the process, I'm referring to in this paragraph is meant as

21     the determining factors that were leading to killings of civilians and

22     soldiers.

23             The determining factors, of course, explicitly not -- are not

24     studied here other than the study of causes of death, of course.  Causes

25     of death reported in this report which were discussed in the earlier


Page 28193

 1     sessions included shelling and sniping, other shooting, indirect causes

 2     related to war activities.  These are expressions of these processes that

 3     were leading to death of civilians and soldiers.

 4        Q.   And could these processes include military activity?

 5        A.   Of course.  Shelling and sniping doesn't occur by itself.  It is,

 6     of course, combat activities, military activities, that are taking us --

 7     leading us to the causes expressed as shelling, sniping, shooting.

 8             MS. WEST:  May we have e-court page 30.  The top paragraph.

 9        Q.   And here you speak about these results again.  And you say:

10             "Based on the results, ... the existence of the two different

11     mechanisms driving the victimisation of civilians and soldiers has been

12     confirmed.  There is no reason to believe that killing and wounding of

13     civilians was a direct consequence of, or related to, fighting of the

14     armies involved in the combat in and around ... Sarajevo."

15             So my question for you regards these mechanisms.  You say that

16     the existence of two different mechanisms.  What do you mean by that?

17        A.   Well, these two different mechanisms -- if we say soldiers who

18     are killed in combat activities, and this is this one mechanism, combat

19     and fighting that kills the soldiers, the question is, What is the other

20     mechanism that kills civilians?  And if it is not combat and fighting,

21     then it has to be targeting of civilians using means that are normally

22     used for combat of armies.  So what I'm trying to say is that in my

23     opinion it has to be -- it had to be the targeting, intentional targeting

24     of civilians that resulted in killing and wounding them in large numbers

25     which are in many instances much, much higher, disproportionately higher,


Page 28194

 1     than numbers of killed soldiers.

 2             MS. WEST:  If we could go to the handout, which 23719, and

 3     e-court page 9.

 4        Q.   We're going to talk about the numbers.  Momentarily we will see a

 5     table.  This is Table 8 taken from your report.

 6             And this regards the top 15 per cent of daily killings for

 7     soldiers.  Can you tell us about this table?

 8        A.   This table is a selection of data on killing of soldiers.  This

 9     selection relates only to dates on which 15 per cent highest killings

10     were noted.  So it is if we would sort all the dates from these siege

11     episode from the top number of killings of soldiers to the lowest number

12     of killings of soldiers, then in this table it is only 15 per cent of the

13     top dates with the highest numbers of killings of soldiers.

14             Further, for this top 15 per cent, we report two variables:  One

15     is the number of killed civilians from all causes, and the second is the

16     number of killed soldiers.  And further, we calculate a ratio which is

17     given in the second-last column in this table, CS ratio.  This ratio

18     expresses the number of killed soldiers in relation to the number of

19     killed civilians.

20             Its value for the first date, 2nd of May, 1992, is 6.  That only

21     means that per one civilian killed, there were six soldiers killed on

22     this day.

23        Q.   And so based on the assumption that you had when you first

24     started this analysis that you told us about, that on days when there

25     would be a high number of soldiers, there would be a less number of


Page 28195

 1     civilians, does this table support that assumption?

 2        A.   Yes.  This expectation I mentioned initially only means that the

 3     CS ratio is higher than one.  If it was one, the two numbers are exactly

 4     the same.  If it is higher than one, and in -- the first row equals six,

 5     that means that soldiers were killed significantly more than civilians.

 6             So because in this table on all dates except for two, indicated

 7     with a "yes" in the last column, so for all other dates, the ratio is,

 8     indeed, higher than one.

 9        Q.   And if we can continue with that assumption, would one then

10     expect to see, on the days with -- when most civilians were killed, even

11     more soldiers killed?

12        A.   Yes.  This is it correctly what we expect.

13             MS. WEST:  If we can turn to the next page of the handout,

14     please, which is Table 6.

15        Q.   Can you tell us about this table.

16        A.   This is a similar table to the one we just discussed.  The

17     difference is that the selection of dates is made based on the number of

18     killed civilians.  This table represents the top 15 per cent of highest

19     killings of civilians.  And for this 15 per cent top daily killings, we

20     calculated again the same measure, CS ratio, and in the last column of

21     the table, with a "yes," I marked the values of the CS ratio, which are

22     lower than one or just equal one.  I think all are lower than one.  This

23     indicates that the number of killed civilians is higher than the number

24     of killed soldiers, which is the opposite to what we have seen for

25     soldiers for the top 15 per cent of killings of soldiers.  It is the


Page 28196

 1     actually complete opposite of it.

 2        Q.   And what significance does that have for you in terms of the

 3     mechanisms that killed soldiers versus civilians?

 4        A.   Well, this is -- my initial result that makes me think these

 5     mechanisms must be different.  These mechanisms are considerably

 6     different, which I further confirmed by several additional analysis

 7     summarised in the report in the section on timing.

 8        Q.   If we just look below this table, what we see --

 9             JUDGE KWON:  Just a couple of questions on the part ...

10             The distinction of civilians and the soldiers was also coming

11     from the two surveys, household surveys and the Bakije material.

12             Am I correct in so understanding?

13             THE WITNESS:  Yes.  This distinction that is presented in this

14     table and analysed in the report comes from one source.  The Household

15     Survey Sarajevo.

16             However, just to double-check that we were correct by using this

17     distinction - the issue is, after all, of -- is essential, I think - we

18     also compared the reporting of the civilian/military status in Sarajevo

19     household survey with an alternative approach.  The alternative being

20     matching of the Sarajevo household survey with the military lists, and

21     for all those marked, taking the military status from the military lists

22     as the one observed, really.

23             We further compared these two approaches, and in the second

24     approach the number of civilians killed was slightly higher than in the

25     first approach.  So in order to be conservative, we decided to use the


Page 28197

 1     reporting from Sarajevo household survey.  But I want to stress that a

 2     code -- a large per cent of records had the same coding.

 3             JUDGE KWON:  Doctor, what did you mean by "military list" when

 4     you said you matched the Sarajevo household survey with military lists?

 5             THE WITNESS:  Earlier today when discussing the sources, I

 6     reported that a number of additional sources were used, not for the list

 7     of victims itself, but for analysis like this one.  The military lists

 8     were the lists I obtained from the ministries of defence of

 9     Republika Srpska and the Federation of Bosnia and Herzegovina.  These are

10     lists of those reported as killed in action or who went missing, being in

11     the army.

12             The lists covered three armies: The VRS, of course; ABiH army;

13     and HVO army.  So, in total, for the entire country, there is about

14     50.000 military casualties reported there.

15             JUDGE KWON:  I was waiting for the French translation.

16             And as you -- as to your comment or conclusion that civilians

17     were deliberately targeted, in relation to that I wonder whether the fact

18     that there are many civilians victims would necessarily lead to such a

19     conclusion?

20             THE WITNESS:  It is not only the fact that there were that many

21     civilian victims in relation to military casualties on the same dates, it

22     is not only this reason that led me to draw this conclusion.  I studied

23     the similarity between the timing of killing of civilians and soldiers

24     statistically and the same I have done for wounding.  I calculated

25     several statistical measures of the similarity of these two different


Page 28198

 1     distributions and consistently my result was that these two patterns are

 2     statistically different and significantly so.

 3             That is one type.

 4             I also studied other techniques to confirm one -- the same thing.

 5             JUDGE KWON:  I noted - please correct me if I'm wrong - that more

 6     people died because of shelling than because of sniping or shooting.

 7             THE WITNESS:  Well, it is probably right.  If we go back to the

 8     table on causes of death, then we can verify this conclusion.  But, of

 9     course, it is a right conclusion.

10             JUDGE KWON:  Then does the location of the army have any impact

11     or effect on such result if the -- for example, if an army is located in

12     the middle of civilian population.  Does it have any effect on that,

13     Doctor?

14             THE WITNESS:  Well, I -- I believe it does matter where the army

15     is located.  Absolutely.  Yeah.  For shelling, there is a separate

16     analysis in the report.  And I compared the victims of shelling for both

17     civilians and soldiers, and the conclusion is not different for this

18     particular category.  So these are two different patterns, I think.

19             As for the location of the army, the VRS was located, of course,

20     on the hills with an excellent view on the town down the valley, so I

21     think it was a perfect location for shelling of any victim.

22             JUDGE KWON:  Thank you.

23             Yes, please proceed.

24             I take it you have been to Sarajevo several times.

25             THE WITNESS:  I've been many times to Sarajevo and had the


Page 28199

 1     opportunity to walk along the actual front lines on the hills.

 2             JUDGE KWON:  Thank you.

 3             MS. WEST:

 4        Q.   If we go to the bottom of this page, there's another figure below

 5     this table.  This looks -- it's a chart, it's a visualisation.

 6             Can you just tell us what this is?

 7        A.   This is a selection of days from this episode of the siege.  In

 8     total there were 162 days in this episode.  On this chart, 51 days are

 9     shown.  These are the days on which the number of killed civilians was

10     higher than the number of killed soldiers.  It is a larger per cent of

11     the overall number of days, 31.5 per cent of days, for which this

12     observation holds.

13        Q.   Okay.  Now, this issue of timing, did you make this same analysis

14     in your second and third report about Sarajevo as well?

15        A.   Yes.

16        Q.   Okay.  We'll eventually get there, but now we're going to --

17     we'll move to the second report, which is the Galic report.

18             MS. WEST:  If we can have 65 ter 12131.

19        Q.   Now, this report picks up the timing -- the time-period where the

20     Karadzic report leaves off; is that right?

21        A.   Yes, that's right.

22        Q.   And what area does this report -- what geographical area does it

23     cover?

24        A.   It is exactly the same area within the front lines area in

25     Sarajevo.  And the same source is used for this report exactly as for the


Page 28200

 1     previous one.

 2        Q.   Okay.  So you answered my next question.  The household survey

 3     was used for this report?

 4        A.   Correct.

 5        Q.   All right.  And the results you obtained for the Galic report,

 6     are those similar to the Karadzic report?

 7        A.   Yes.  But the numbers are higher compared with the first episode,

 8     because the episode itself is longer than the first one.

 9        Q.   Right.  And for the Galic report, the episode is almost about two

10     years; is that correct?

11        A.   Correct.

12             MS. WEST:  If we can -- sorry to keep going back and forth, but

13     may I have the handout again, and now we're looking at page 10, which is

14     e-court page 11.

15        Q.   We're going to review some tables for Galic, and this is going to

16     be quite similar to what we did in Karadzic -- the Karadzic report.

17             So at first we will look at Table 1.

18             MS. WEST:  And we need the handout which is 65 ter 23719, please.

19     E-court page 11.  I think everybody has the handout itself.

20        Q.   So we have Table 1 in front of us.  Can you tell us what this

21     table is?

22        A.   It is the same table as already shown for the first episode in

23     the Karadzic report.  Here, the numbers are higher.  The number of killed

24     persons is 3.798.  And the number of wounded persons is 12.919.

25             Further down in the table, we include basic distributions by sex,


Page 28201

 1     age, civilian/military status, and, of course, cause of killing and

 2     wounding.

 3        Q.   Okay.

 4             JUDGE KWON:  Just one question.

 5             As to the distinction between civilian and soldiers, what are the

 6     examples of unknown status?  Does it mean that they didn't answer the

 7     question?

 8             THE WITNESS:  Yes.  That is exactly the meaning: no answer to

 9     this question.

10             JUDGE KWON:  Thank you.

11             MS. WEST:

12        Q.   And under "status" in this table, we see civilians killed, 1.399;

13     civilians wounded, 5.093.  And then we go to the table below it.

14        A.   Yes.

15        Q.   Is this where we deal specifically with civilians?

16        A.   Yes, that is civilians only.  Yep.

17        Q.   Okay.  And is this separated in the same manner?

18        A.   Yes, it is.

19        Q.   If we can turn the page to page 11.  Again we see another table

20     with which we're familiar.  This regards killed, cause of death, wounded,

21     cause of wounding.

22             Let's work on Table 6, the section up top that involves

23     fire-arms.  If we add all those totals up --

24        A.   These are very high totals for fire-arms.  For killings, it is

25     89.2 per cent; and for wounding, the fire-arms are responsible for 95.4


Page 28202

 1     per cent of all wounding cases.

 2             The remaining causes have much less significance.

 3        Q.   Thank you.

 4             MS. WEST:  If we turn the page --

 5             JUDGE KWON:  Just -- can we stay there.

 6             Second group in the causes of death among direct casualties of

 7     war, and I see "human shield, 4."  How do we come to this conclusion?

 8             THE WITNESS:  There is no -- my conclusion here, it is the data

 9     as reported by the respondents.

10             JUDGE KWON:  So they responded they -- certain individuals died

11     as human shields?

12             THE WITNESS:  That is the meaning, I understand, of this

13     response.

14             There is very little affect of my work here.  As I said, it was

15     for me to clean it up a little bit and regroup and present in a table.

16     It is not that I am recording the human shields from something else to

17     human shields.  It is not what I have done.

18             JUDGE KWON:  Thank you.

19             Yes, Ms. West.

20             MS. WEST:  Thank you, Mr. President.  And if I can pick up on

21     that.  May we go to 65 ter 12130.

22        Q.   I just want to go back to the Karadzic report for just one

23     moment, and we're going to look at the annex, and the reason being is the

24     annex includes the actual questionnaire that was used in the household

25     survey.


Page 28203

 1             MS. WEST:  So 65 ter 12130.  May I have the ERN ending in 8506,

 2     please.  I think it's probably page 40.

 3        Q.   This the English translation of the questionnaire.  And you

 4     mentioned that you were specifically interested in killed and wounded,

 5     and I see here that's 5 and 6.  Tell us about this information that was

 6     asked.

 7        A.   Well, as we can see for killed persons, and the same is later for

 8     wounded - I'm talking about question 5 - so information collected about

 9     those killed included names of these individuals, year of birth, position

10     in the household, date of death, place of death, cause of death, and the

11     civilian/military status.  So all these items were collected explicitly

12     reported on the questionnaire.

13        Q.   So when you received information in your tables, for example, the

14     direct casualties of war, and it was listed - we just saw it a second

15     ago - cold steel, slaughtered, tortured, for example, execution, human

16     shield, disappeared - is that information that is specifically coming

17     from this questionnaire?

18        A.   Yes.  This is coming from the responses to these questions.

19        Q.   And that was the questionnaire in the Karadzic case.  That was

20     the same questionnaire used for the Galic case as well?

21        A.   Yes.  That's the same questionnaire and the same source.

22             MS. WEST:  If we can go to the handout, please, to page 12 of the

23     handout, and we'll continue on.

24        Q.   Now we spoke earlier about the statistical formula of

25     capture/recapture, and I don't want to talk about that again, because


Page 28204

 1     we'll talk about it next time, but I do want to talk about the results

 2     that came from that formula in the Galic case.

 3             And up top we see Table 18.  Can you tell us about the numbers

 4     you found for this --

 5             JUDGE KWON:  Just a second.  It should be next page.

 6             MS. WEST:  Thank you.

 7        Q.   Tell us about these results in this two-year period of time.

 8        A.   These are results in Table 18 of the capture/recapture method,

 9     exactly the same method and the same sources as were applied in the

10     Karadzic report.  The information reported here includes the estimated

11     number for all ethnic groups, including Muslims, with the number 3.437.

12     This was produced as the initial estimate in the first step of the

13     estimation and extrapolated, in the next phases, exactly the same way as

14     it was done for Karadzic.

15        Q.   And so for here we see, under "Muslims," the surveyed during that

16     period of time reported 2.340.  And then what was the estimate?

17        A.   The estimate is 4.352 for all ethnic groups.  The number 2.963 is

18     lower than what we have seen in Table 1 because, here, only records

19     linked are shown.  That makes for the difference, makes up for the

20     difference.

21        Q.   Okay --

22        A.   But it is exactly the same data we're talking about.

23        Q.   When you say "linked," linked from what to what?

24        A.   In this report, in the Galic report, the links between Sarajevo

25     household survey and the Bakije funeral home, the links go to the


Page 28205

 1     population census.  The matched records listed here are those matched

 2     with the population census and therefore are lower than the overall

 3     numbers presented in Table 1.

 4        Q.   Now, we're going to move on to timing.  And you mentioned that

 5     you did the same timing analysis in the Galic report that you did in the

 6     Karadzic report.  And let's look here at Table 13.

 7             Tell us what this table reports.

 8        A.   Table 13 reports on killed civilians and soldiers for days when

 9     only -- when six or more civilians were killed.

10             This table relates only to one year of the two studied.  It's

11     just a selection from the second year.  There is another table in the

12     report for the first year so that the analysis is complete.

13             So also in this table, when we go through the numbers of killed

14     civilians and soldiers, we see that for all but two dates the civilians

15     are killed in higher numbers than soldiers.  And the same is observed for

16     the days selection of high days from the first year of the siege in the

17     table in the report similar as this one.

18        Q.   All right.  When go to Table 14, days when even more civilians

19     were killed, 20 civilians or more were killed, what you did -- excuse me,

20     wounded.

21        A.   Wounded.

22        Q.   What did you find?

23        A.   Well, it is again a table related to very high days, 20 civilians

24     wounded on one day, it's a lot, so, again, the same pattern.  The numbers

25     of wounded civilians much, much higher than the number of wounded


Page 28206

 1     soldiers.  The same is in the report for the first period.

 2        Q.   And what are the significance of these figures to you?

 3        A.   Well, these figures confirm the conclusion I already presented

 4     for the first episode of the siege discussed in the Karadzic report.  I

 5     think my conclusion also holds for this report.

 6             MS. WEST:  I'd like now to move to the third Sarajevo report.  65

 7     ter 12135.  And this is the -- a report for when Dragomir Milosevic was

 8     the corps commander.

 9        Q.   What's the time-frame of this report?

10        A.   This report comprises events from the last year of the siege,

11     August 1994 to November 1995.  The final, final stages of the war.  Yes.

12     And the area, similar but not identical to the previous reports.  This

13     report is compiled using actually different sources.  The same

14     methodology.  But because of the different sources, the area is also

15     slightly different.

16        Q.   Okay.

17        A.   But the core of it remains the same.

18        Q.   Okay.  So this report uses the same methodology; right?

19        A.   Yes.

20        Q.   Why did you use different sources?

21        A.   For the obvious difficulty that Sarajevo household survey

22     actually ended with reports for August/September 1994 and there is no

23     more data in it.  So what I studied in Dragomir is not covered in the

24     Sarajevo household survey.

25        Q.   What sources did you use?


Page 28207

 1        A.   Sources are different than Household Survey Sarajevo but also

 2     very reliable and made by professional statisticians.  I used two -- the

 3     so-called DEM-2 databases, one of Republika Srpska and one of the

 4     Federation.  These DEM-2 databases are called so because they are

 5     electronic expression of what is reported to statistical authority on the

 6     so-called DEM-2 form.  This is a form that is used in official death

 7     notifications that citizens are obliged to make in case of death in their

 8     family to statistical authority.  So these were my major sources --

 9             JUDGE KWON:  Doctor for record, could you spell them two.

10             THE WITNESS:  D-E-M-2, like a number, 2, DEM-2.  It is -- the

11     "dem" comes from "demographic" death notification form.

12             JUDGE KWON:  Thank you.

13             MS. WEST:

14        Q.   And these databases, you had one from the RS and one from the

15     Federation?

16        A.   Yes, that is correct.  These databases were based, as I said, on

17     official death notifications from the war time.  All together for the

18     entire country these databases comprised 140.000 records, cases, death

19     cases.  Information collected by the authorities and processed here

20     included both war-related violent deaths and also natural deaths.  We

21     actually requested the information about natural deaths in order to be

22     able to exclude it from further analysis.  But if we would not have it,

23     it would be to make sure that we worked with categories that are not

24     mixed up.

25        Q.   Okay.  What other sources did you use?


Page 28208

 1        A.   The third source.  In addition to these two, was the ICRC list of

 2     missing persons.  ICRC is International Committee for Red Cross.  They

 3     are mandated to trace missing persons in case of conflict and war.  I

 4     included the list of missing persons by ICRC because missing persons are

 5     not reported in official death notifications.  It is impossible because

 6     death certificate don't exist for these kind of cases.  And secondly, I

 7     choice ICRC for its quality.  A very well-known, high quality of

 8     information.  So these three sources were the basic sources for compiling

 9     the list of victims.

10        Q.   And now let's talk about your sources for your wounded numbers.

11     What were those sources?

12        A.   The wounded persons are taken their records from Sarajevo

13     hospital records.  I was able to use records from three hospitals in

14     Sarajevo.  This is not all hospital records available for this period.

15     There were more hospitals.  There were two war hospitals in Sarajevo.

16     And emergency health stations, these records are not used.

17             So the three hospitals that I used are actually a small sample of

18     all information about wounding process.  It's a very small, limited

19     sample, but it's a very good sample because, first of all, the selection

20     concentrated on civilians.  All men at age 18 to, I think, 65, were

21     excluded.  They are not reported in the sample.  Only persons below age

22     18 and women are included.  Yet it is a small number that was collected.

23             JUDGE KWON:  Just a second.

24             That's the amount of time you should put a pause.  French

25     translation is just now completed.


Page 28209

 1             Please continue, Ms. West.

 2             MS. WEST:  Thank you.

 3        Q.   I'd now like to talk about the methodology you used in getting

 4     the numbers for the killings.

 5             MS. WEST:  And in the reports - for the parties - this starts at

 6     page 11.

 7        Q.   Tell us what your first step was when you looked at all these

 8     sources and had to put it together.

 9        A.   Well, what we always do in the beginning, we study each source,

10     assess it, clean it up a little bit, exclude duplications, and match it

11     with the population census.  We have done this with all our sources on

12     victims.

13             In the second step, we integrated the sources of -- on victims in

14     one large data table.

15        Q.   So when you talk about integration, does that mean you merged the

16     DEM-2 databases and ICRC lists?

17        A.   Yes.  That's the meaning of it.  Merging is increasing the number

18     of cases by putting one source on the top of the other.

19        Q.   What was the next step?

20        A.   Well, one more thing.  So these three sources, DEM-2 databases,

21     ICRC lists, is the core, the most important source, as such, for the

22     Dragomir Milosevic report, but because the whole database was established

23     for actually entire country, so in addition to these three, we also

24     included Sarajevo household survey and Bakije funeral home for earlier

25     periods, not those who were later analysed in Dragomir.  Just


Page 28210

 1     clarification.  But it doesn't change the nature of the data used for

 2     Dragomir.

 3             So having merged all these lists, the next step is to identify

 4     the overlap of the lists.  The overlap means that records of the same

 5     persons are repeated in several sources.  As we can easily imagine, they

 6     are repeated in DEM-2 databases, especially the Federation part of it,

 7     and the Sarajevo household survey.  Moreover, we already know that

 8     household survey Sarajevo overlaps quite well with Bakije funeral home.

 9     So there is a lot of overlap in this merge of lists.

10             But, of course, we are able -- we know -- we have techniques that

11     can be done to identify the overlap and eliminate all records that would

12     duplicate information about victims.

13             So that is what we have done in this phase.

14        Q.   So at this point you have a large database for the entire country

15     for the entire war period; is that right?

16        A.   Right.

17        Q.   So in order to focus solely on the Dragomir Milosevic time-period

18     and the geographical area of Sarajevo, what did you do?

19        A.   In order to be able to -- to focus on this area, we had to

20     extract a portion of this database, the portion related to Sarajevo, and

21     for this portion, eliminate any duplicated information, specifically for

22     this part, to end with a reasonable list, reliable list, from which we

23     could extract relevant records for the report for the Dragomir Milosevic

24     case.

25        Q.   Once you arrived at the correct time-period and the correct area,


Page 28211

 1     how many records did you have?

 2        A.   Approximately 26.000 records.  These twenty-six -- specifically

 3     26.466, this number covers the entire war period, meaning April ninety --

 4     even January 1992 to December 1995, and, actually, not six, but ten

 5     Sarajevo municipalities.  The ten Sarajevo municipalities comprise the

 6     six urban municipalities exactly the same as studied in the Galic report

 7     and Karadzic report, plus four extra municipalities.

 8             This is a broad selection, but we don't make these kind of

 9     databases every day, and there might be always questions that are broader

10     than just the report that we are making.  So the idea of making this kind

11     of database as the basis is actually related to these additional

12     questions that may come up in the future.  But having this database, we

13     have selected from it only the relevant records according to time and

14     area criteria.

15        Q.   Okay.  So now we have a database that regards this time-frame,

16     this area.  Does that database differentiate between soldiers and

17     civilians?

18        A.   In DEM-2 databases and in the ICRC list, there is no information

19     about civilian/military status.

20             So it was not possible for us to take it as reported in our

21     original sources and process it into statistics.

22        Q.   So what did you do?

23        A.   We, in this case, used the military lists that I mentioned

24     earlier today and cross-referenced these lists by matching,

25     individual-level matching, with the records in the Sarajevo database.


Page 28212

 1        Q.   And if you found any reporting that an individual was a soldier,

 2     even if that was inconsistent with other reporting, how did you count

 3     that person?

 4        A.   Well, in the first step we just marked all those confirmed on

 5     military lists as soldiers.  That was step number one.

 6             And, further, we have also Sarajevo household survey in this

 7     database.  So for a number of records, we can compare further the

 8     information obtained from the military lists with the actual reporting in

 9     the Sarajevo household survey.

10             So for many records, the information was consistent.  Actually,

11     for majority of records.  For a small number of records, there was a

12     different reporting of individuals.  So we again opted for the

13     conservative approach, that we coded as soldiers everybody ever reported

14     as soldier in any source.

15             MS. WEST:  If we can turn to the handout and go to page 13.  This

16     is Table 6 on top.  It regards an estimate of military status.

17        Q.   Dr. Tabeau, can you tell us about this table, please?

18        A.   This table relates to the entire Sarajevo database.  We see the

19     overall total of individuals included here listed in the last row as

20     26.466.  That is the general Sarajevo database, entire war period.

21             Above this number, there is a distribution of it into civilians,

22     militaries, or soldiers, and there is a category called unknown.  What

23     interests us at this stage is the category called unknown.  This is the

24     number of cases for which we did not have any information from our

25     matching with military lists.  But we don't want, of course, to have a


Page 28213

 1     large number of unknown status because it -- it hampers a good analysis

 2     of this data.  And moreover, for a large number of the unknowns, a good

 3     information was available from Sarajevo household survey.  So we go back

 4     to Sarajevo household survey and check how these unknowns are reported

 5     there, and obviously we find two components listed in this table under

 6     the column "correction."  1.715 are those who were reported civilians.

 7     So we take them and record them into civilians.  And they are gone from

 8     the unknowns.  And we go and look for soldiers among the unknowns, and we

 9     see 886 reported as soldiers in Sarajevo household survey.  So we record

10     these cases as soldiers.  So they are gone from unknowns and moved to

11     soldiers.  This is one type of correction: Improving of the unknowns.

12     But, as I say, we always compare, always cross-reference information if

13     we have it.

14             So when we compared the coding of the status between Sarajevo

15     household survey and military records, we saw a small number, 608, 608,

16     that were actually civilians in our first approach but soldiers in

17     Sarajevo household survey.  In order not to inflate the number of

18     civilians, we recorded them into soldiers because this number was

19     reported inconsistently between these two sources, and we don't want to

20     keep the number of civilians high.  We want to keep it low, as low as

21     possible, because we want to be conservative.  We don't want to make

22     numbers higher than they should be.

23        Q.   So if we go to the column entitled "final number," we can see

24     that you've altered your figures, but we still see a number in the

25     unknown table.


Page 28214

 1        A.   Yes.  For this number, there is nothing we can do.  And this is

 2     the cases that will stay like that, that we can only hope that the number

 3     of unknowns in the Dragomir Milosevic case area and time span will be

 4     low.

 5             We will see in a second, in the second table, how really high it

 6     was.

 7        Q.   All right.  So let's move down to Table 11.

 8             And we have two tables, and let's start on the left.

 9             Can you tell us about this table?

10        A.   Yes.  To comment on the number of records with unknown

11     civilian/military status, in total there were 571 such records related to

12     the Dragomir Milosevic case area and time span.  This number is mentioned

13     in the table -- in the left table, at the bottom, under "unknown

14     status" --

15        Q.   And just stop there are for a moment.  That number, 571, is

16     different than in Table 6, the unknown, 3454; right?

17        A.   Of course.  Because --

18        Q.   And why?

19        A.   Yes.  Why was it different?  It is, in Table 11 we actually made

20     one extra step between Table 6 and 11.  In Table 11, instead of the

21     entire database, we talk about a selection of cases from the Sarajevo

22     database, the selection that relates to the Dragomir Milosevic case.

23     Except, of course, for cases of natural deaths which are also included.

24     And in a second I will explain how they were excluded.  But for the time

25     being, we have to take a note of the fact that we extracted records, and


Page 28215

 1     speak of Dragomir Milosevic cases only.

 2        Q.   So Table 11 regards only the Dragomir Milosevic time-frame;

 3     correct?

 4        A.   Correct.

 5        Q.   All right.  So continue on explaining Table 11.

 6        A.   The second issue that needed to be sorted out next to the unknown

 7     civilian/military status was the unknown cause of killing, of death.

 8             In our databases, there are two parallel reports on causes of

 9     death.  One is the reporting of medical causes of death.  There is an

10     example in the report of medical causes.  I think it is ... it is

11     Table 12 on page 26.

12        Q.   Go ahead.

13        A.   So medical causes include, for examples, injury, all kinds of --

14     different types of injuries, like crushing injuries, other injuries,

15     open-head wounds, neck fractures.  Well, the cause of death is expressed

16     as a medical category.  A list of the causes is included in 12 -- in

17     Table 12 on page 26.

18        Q.   And if you could stop there.

19             MS. WEST:  This is page 26 of 65 ter 12135 and we won't go there.

20     And it's a number of causes.

21        Q.   But tell us the source of these causes.  Where did you get this

22     information?

23        A.   These causes are reported in DEM-2 databases.  The large

24     databases, they were the core sources for this report.

25             But, as I said, parallel to the medical causes of death, there is


Page 28216

 1     another way of reporting of causes of death.  There's another way

 2     indicated explicitly whether a death case is a war-related violent death

 3     or not.  For us, for our purposes, this information is essential, because

 4     we are not looking for all available medical causes, disregarding how

 5     relevant they might be to our cases, it is important for us to make sure

 6     that we extract violent war-related deaths for this analysis.

 7             And in Table 11, the left part of the table, we can read that

 8     there were 981 cases of violent war-related causes.  981 is the total of

 9     violent war-related deaths.  This number, 981, is an absolute minimum of

10     the cases that we consider to be the victims related to the last episode

11     of the siege, the one related to Dragomir Milosevic report.

12        Q.   Okay.  And, now --

13             JUDGE KWON:  Just can you go back, can we see the page, Table 12,

14     medical causes of violent war-related deaths.

15             MS. WEST:  That is 65 ter 12135, e-court page 26, please.

16             JUDGE KWON:  Is the page ending with 6397?  Yes.

17             MS. WEST:  Mr. President, did you have a question on that?

18             JUDGE KWON:  Yes, just one question.

19             I -- I found some strange cause of deaths; for example,

20     number 53, toxic effect of alcohol.  And number 49, poisoning by

21     diuretics and other unspecified drugs, medicaments, and biological

22     substances.

23             How are they related to violent war-related death?

24             THE WITNESS:  Well, in this table, the causes are presented as

25     reported in the DEM-2 databases.  For these two cases, we also have an


Page 28217

 1     indication that these causes were -- these two cases were related to war.

 2     They were war -- indicated as war victims.  Well, how they could be

 3     related to war victims?  Obviously poisoning is not the same as shooting,

 4     shelling, or sniping.  Obviously it is a completely different category of

 5     cause of death.  This category, this cause of death, I can imagine would

 6     be easily possible in a territory under siege and wouldn't be listed

 7     under causes like shelling or sniping or other shooting but would be

 8     listed in a separate category which we also had in Sarajevo -- in the two

 9     previous Sarajevo reports as indirect, for instance, cases of war deaths.

10             JUDGE KWON:  Thank you.

11             Yes, Ms. West.

12             MS. WEST:

13        Q.   So if we go back to the handout, page 13 of the handout, and go

14     on the table from the left side, where we see numbers in green, to the

15     table on the right side where we don't see numbers in the green unknown

16     section, can you just tell us in a very brief --

17        A.   Yes.

18        Q.   -- general way how you redistributed those numbers --

19        A.   Yes.

20        Q.   -- and what the basis for that was?

21        A.   The green areas contain unknowns on both civilian/military status

22     and on cause of death.  About the unknowns, we assumed that they

23     contained both categories: war-related victims, and natural deaths.  So

24     what we did, we statistically distributed the unknowns proportionally to

25     the observed distribution highlighted in the -- I don't know, yellowish


Page 28218

 1     colour, something like that.  I don't know exactly how to call this other

 2     colour in the table.

 3        Q.   So this was a distribution by proportion?

 4        A.   Yes.  This is the allocation of unknowns among the two types of

 5     well-defined categories using the proportionality assumption.

 6        Q.   Okay.  Turn the page in the handout.  We're now going to move to

 7     wounded.  Tell us about the source of the information for this table.

 8        A.   This table is based on hospital records.  It includes all the

 9     records reported in the three hospitals.  These are cases of hospitalised

10     wounded persons only.  The total of cases for the entire siege period is

11     3.036.  Of those, we identified 405 listed in the last row of the table

12     as the cases of hospitalised wounded persons related to the

13     Dragomir Milosevic case.  It is, again, by selecting according to the

14     time criterion.  Area of all cases is relevant.

15        Q.   All right.  So this number is based on actual observed numbers,

16     correct, this is a minimum number?

17        A.   Yes.  It is a very, very low absolute minimum number.

18        Q.   All right.  I'd like to turn -- to skip a page in the handout and

19     go directly to page 16 and talk about timing in the Milosevic case.

20             Did you do the same analysis here that you had done in Karadzic

21     and Galic?

22        A.   Yes.  Timing was analysed exactly the same way as it was in the

23     first and second report.

24        Q.   All right.  And in those reports we saw different types of

25     tables, and here again is a third type of table.  And this regards


Page 28219

 1     numbers of deaths per day.

 2             Please tell us about this.

 3        A.   Well, this table gives a summary of what we observed on the

 4     so-called low days and high days, just a summary, not listed here day by

 5     day, but just summary of numbers of killed civilians and soldiers on low

 6     and high days.  Low days are those when one to two deaths per day were

 7     observed.  High days are days with three or more deaths per day were

 8     observed.

 9             If we move to the second column of the table, then we see the

10     number of days.  How many were there in the siege?  There were 224 days

11     that can be considered low, and 107 that can be considered high.  Out of

12     a total of 487.

13             487, for transcript.

14             On the low days, the number of killed civilians was, in total,

15     160; and soldiers, 102; 46 were unknowns.  The number of civilians, as we

16     see, is considerably higher than the number of killed soldiers.

17             On the high days, the number of killed civilians is 289; and of

18     soldiers, 317.  The number of killed civilians is lower - not

19     considerably, though - than the number of killed soldiers.

20        Q.   So, Doctor, what conclusion did you come to based on those

21     numbers?

22        A.   The conclusion is exactly the same and [sic] in the two previous

23     reports, but I want to note that conclusion is drawn on many more

24     analysis than what we just seen in this table.  This table is a small

25     selection of all ways of analysis that is completed and presented in the


Page 28220

 1     report.

 2             So the conclusion is that the mechanisms killing civilians and

 3     soldiers are different, and significantly so, and civilians must have

 4     been targeted specifically because their numbers are so high on many,

 5     many days in every episode of the siege.

 6             MS. WEST:  May we have 65 ter 12135, page 5, please.

 7        Q.   If we can go to the bottom of the page.  This is your Milosevic

 8     report.  In the last paragraph, you noted, in talking about this timing,

 9     that:

10             "The above findings suggest that the civilian population of

11     Sarajevo had systematically been the target for military attacks

12     disregarding the actual presence of heavy combat activity [sic].  This

13     also indicates that a targeted terror campaign against civilians likely

14     continued for most of the period from August of 1994 to

15     November of 1995."

16             Dr. Tabeau, can you explain to us how your analysis supports a

17     targeted terror campaign in Sarajevo?

18        A.   It is -- the basis for this conclusion is the significant

19     difference between the timing of killings of civilians and soldiers and

20     of wounding of civilians and soldiers.  On many days in every episode,

21     numbers of killed civilians were higher than the number of killed

22     soldiers.  If the mechanism would have been the same and it would be

23     combat, killing both, this observation would not hold.  So the mechanism

24     killing civilians must have been related to targeting them, disregarding

25     of combat activities going on in the siege.


Page 28221

 1             MS. WEST:  Mr. President, I'm going to move on to the Srebrenica

 2     reports.  I wonder if you would like me to do that or is this the time to

 3     stop?

 4             JUDGE KWON:  I wanted to suggest to you to stop for today here.

 5             In addition, there are a couple of matters I'd like to deal with.

 6     And I was also informed that Mr. Karadzic has something to raise.

 7             Shall we excuse the witness?  Or ...

 8             THE ACCUSED: [No interpretation]

 9             JUDGE KWON:  I'm not sure we are getting the interpretation.

10             THE ACCUSED: [Interpretation] As for what the witness would like

11     to raise, I would like the witness to be present.  I can repeat it.

12             So regarding the issue that I would like to raise, I would like

13     Dr. Tabeau to be present in the courtroom.

14             JUDGE KWON:  Yes.  Please continue, Mr. Karadzic.

15             THE ACCUSED: [Interpretation] I would like to inform you that

16     Dr. Tabeau was kind and agreed to meet the Defence, and we agreed that

17     all scientific findings, hers, as in any scientific study, can be tested

18     and repeated if they are checked using -- applying the same methods and

19     using the same materials, so I would appreciate it if Dr. Tabeau could

20     provide us during the day today or tomorrow all the materials that she

21     used.  Especially the census for 1996 that Dr. Tabeau rejected.  Then

22     1997 and 1998 census materials.

23             I would also like to have the ethnic composition of the

24     authorities, the organ in power in -- the government organs in 1996,

25     1997, and 1998, as elected after the elections.


Page 28222

 1             I would also like - and I'm not sure if I have that - a census

 2     from 1991 with all the relevant names, the address, father's name, and

 3     the JMBG number.

 4             I would also like Dr. Tabeau to kindly forward to us the

 5     techniques and the tools and procedures that she used and that she meant

 6     a moment ago that she -- that she mentioned a moment ago that she applied

 7     in eliminating duplicates.  She probably did mention them, but we do not

 8     have them in the reports.

 9             We would also appreciate getting information of displaced persons

10     and refugees for the entire country and especially according to their

11     ethnic structure.  Also, information on the dead and missing for the

12     entire territory of the country, and especially in reference to what

13     Dr. Tabeau said on page 89, to obtain the analysis that were used to get

14     the information that were entered in the last table that were used and

15     that were not included in the report.

16             I would appreciate receiving this so that I can prepare for my

17     cross-examination over the weekend.

18             JUDGE KWON:  Before I hear from you, Dr. Tabeau, I will ask you,

19     Mr. Karadzic, why could you not have requested such request on your own?

20     Why only now?

21             THE ACCUSED: [Interpretation] Your Excellency, I did not know how

22     these results would be presented and what the conclusions would be that

23     she would mention in the examination-in-chief.

24             So it is obvious that I will have to repeat all these processes

25     and -- because I would like to test them and see what the results would


Page 28223

 1     be that we would obtain.  Also, in view of the general context, it turns

 2     out now that it is necessary to request this information because I've

 3     just learned that the 1996 census was rejected, although it is

 4     immediately following the war, as opposed to 1997 and 1998.  And also I

 5     realise that there was information that needs to be tested and checked by

 6     the Defence.

 7             JUDGE KWON:  Well, you must have received all of her reports a

 8     long time ago.  But, very well.

 9             Do you have any observation, Ms. West?

10             MS. WEST:  My observation is just that, Your Honour.  It is quite

11     late in the game to be asking for this material.  It's certainly material

12     that the defendant himself should have used to come to his own

13     conclusions.  The main basis of all of these reports are the 1991 census

14     and the voters' registration from 1997 and 1998.  Those he had received.

15             I do note that he just said something about a 1996 census.  It

16     must be a mistranslation.  It must be something other than a census,

17     because there wasn't one.  But nonetheless, anything that happened in

18     1996 doesn't now make his request in 2012 more relevant today than it

19     could have been two years ago.

20             In addition to the matching criteria that he asked for, that is

21     in the report.  There's nothing new that he is asking for that he has not

22     had this whole entire time.

23             And finally, the Karadzic report on municipalities which was

24     provided in, I believe, 2009 is the exact same methodology and sources

25     that are in the Galic report.  There is nothing new being done here.  The


Page 28224

 1     most recent report that he was given - we'll get to this - is the

 2     Srebrenica report, and that was recently, because Dr. Tabeau did it off

 3     the recent ICMP database information.  But again, that's just additional

 4     numbers.  That has nothing to do with methodology or sources.

 5             JUDGE KWON:  In the previous session you said you would come back

 6     to us as to the flagged material, 1991 census, as regards counting of

 7     displaced persons.

 8             MS. WEST:  We're trying to do that right now.

 9             JUDGE KWON:  Yes.  Oh.

10             THE ACCUSED: [Interpretation] Just a correction.  I did not say

11     the census from 1996 but, rather, the voters' lists of 1996.

12             JUDGE KWON: [Overlapping speakers] ... I know I was overlapping.

13             That's a separate matter.

14                           [Trial Chamber confers]

15             JUDGE KWON:  Dr. Tabeau, do you have any observation with respect

16     to Mr. Karadzic's request?

17             THE WITNESS:  Yes.  If I may, first of all, I am aware of the

18     practice of the OTP from many other cases when it comes to disclosure of

19     sources used for demographic and other reports.  All sources, I'm sure, I

20     used in my reports have been disclosed by the OTP, and I can imagine long

21     time ago.  So the population census of 1991; the voters' registers,

22     1997/1998; the military lists.  Further, Sarajevo household survey,

23     Bakije funeral home, DEM-2 databases.  I'm sure it all has been

24     disclosed.  It is impossible that it is otherwise.  So that is one point.

25             It is the practice, in scientific research, to disclose the


Page 28225

 1     sources.  And when the methods are described, it is possible for anybody

 2     else to redo the analysis using the same data.  So I think in order to

 3     challenge what I have done, Mr. Karadzic has all he needs for this.  And

 4     it is very naive to believe that this work can be redone over the

 5     weekend.

 6             This is my conclusion.  We are speaking of millions and millions

 7     of records that have to -- have been analysed.  Overdoing this in one

 8     weekend, in three days, even though it is a long weekend, I don't think

 9     is feasible.

10                           [Trial Chamber confers]

11             THE ACCUSED: [Interpretation] May I respond?

12             JUDGE KWON:  Yes.  Yes, briefly, Mr. Karadzic.

13             THE ACCUSED: [Interpretation] Well, simply, we will take a random

14     sample and test the data.  On the other hand, I'm not certain that I

15     received all the information for the entire country, and that would be

16     absolutely necessary because this was a closed system, it had very hard

17     borders, and it was difficult to leave the country.

18             So the fluctuation of the population within the country was huge,

19     and I would need to have all this information related to Muslims for the

20     entire country and not just for the nine or 27 municipalities.

21             JUDGE KWON:  Having heard the parties and Dr. Tabeau, the Chamber

22     would leave it at that and, if necessary, that may be a subject of

23     separate litigation.

24             I'll note the time, but given that we started a bit later today,

25     I think I have time to issue a couple of oral rulings.


Page 28226

 1             First the Chamber will issue an oral decision on the accused's

 2     motion for public admission of exhibits under seal filed confidentially

 3     on 10th of April, 2012.

 4             In the motion, the accused requests that the Chamber reassign a

 5     new exhibit number to all exhibits admitted under seal in this case

 6     except for witness statements and plea agreements and admit them publicly

 7     in order for the parties and the Chamber to refer to them in public

 8     filings and in the final judgement without them being associated with a

 9     particular witness.

10             The Prosecution filed its response on the 19th of April, 2012,

11     requesting that the Chamber dismiss the accused's motion on the basis

12     that the accused fails to identify any specific confidential exhibit

13     which should be made public and to substantiate his request for

14     reconsideration of all the Chamber's decisions to admit exhibits

15     confidentially, and, finally, because his suggested approach is

16     unworkable and ineffective.

17             The Chamber first recalls that in balancing the importance of

18     conducting a public trial while protecting the identity of witnesses and

19     victims, the Chamber has exercised utmost caution in determining whether

20     to place exhibits under seal throughout the trial.  The Chamber further

21     recalls countless examples where it has inquired proprio motu about the

22     necessity of placing an exhibit under seal, and, whenever possible, the

23     Chamber has ordered the parties to produce public redacted versions of

24     exhibits placed under seal.

25             Moreover, the Chamber recalls that in the motion, the accused has


Page 28227

 1     not provided any specific examples of exhibits admitted under seal which

 2     it argues should be made public.  In this respect, the Chamber considers

 3     the accused's request to be vague and unsubstantiated, bordering on

 4     frivolous, and the Chamber will not engage in a review of each of its

 5     decisions to admit exhibits under seal in this case.

 6             The accused's motion is therefore denied.

 7             Next, the Chamber will issue an oral ruling on the accused's

 8     request filed yesterday to admit into evidence the public and under seal

 9     versions of KDZ612's supplemental statement now that the required

10     attestation under Rule 92 bis (B) has been provided.

11             Having reviewed the under seal version of the supplemental

12     statement which has been uploaded into e-court as 1D5557, the Chamber is

13     satisfied that the certification procedure fulfils the formal

14     requirements of Rule 92 bis (B).  The Chamber therefore requests the

15     Registry to record the confidential version of KDZ612's supplemental

16     statement is admitted into evidence and to assign an exhibit number to

17     it.

18             In relation to the public redacted version of KDZ612's

19     supplemental statement which has been uploaded into e-court as 1D5558,

20     the Chamber notes that the Rule 92 bis attestation has not been attached

21     to it.  The Chamber therefore requests the accused to attach a redacted

22     version of the 92 bis attestation to the document currently uploaded as

23     1D5558.  The Chamber further instructs the Registry to assign an exhibit

24     number to this document and to record that it is admitted into evidence.

25             Doctor, we'll resume next week on --


Page 28228

 1             THE ACCUSED:  May I just --

 2             JUDGE KWON:  -- on Tuesday at 11.00.

 3             THE ACCUSED:  May I just, in transcript, small intervention in

 4     transcript.

 5             I said:  Not only for the Muslims and not only for 27

 6     municipalities.  In transcript, it is concerning only Muslims.

 7             JUDGE KWON:  Thank you.  That's noted.

 8             Hearing is adjourned.

 9                           [The witness stands down]

10                           --- Whereupon the hearing adjourned at 3.06 p.m.,

11                           to be reconvened on Tuesday, the 1st day of

12                           May, 2012, at 11.00 a.m.

13

14

15

16

17

18

19

20

21

22

23

24

25