ML20206Q830
| ML20206Q830 | |
| Person / Time | |
|---|---|
| Site: | Shoreham File:Long Island Lighting Company icon.png |
| Issue date: | 12/19/1986 |
| From: | Simon G LONG ISLAND LIGHTING CO. |
| To: | |
| References | |
| CON-#287-3144 86-533-01-OL, 86-533-1-OL, OL-5, NUDOCS 8704220045 | |
| Download: ML20206Q830 (100) | |
Text
3/VF gc#L TIMNSCRIPT ey J OF PRDCEEDINCE ,,: 0, OFFICE OF Si-
- 0"
UNITED STATES OF AMERICA 00CKEllitG A 'd"'HfA-BRANCH NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _x In the Matter of: :
- Docket No. 50-322-OL-5 LONG ISLAND LIGIITING COMPANY : (EP Exercise)
(Shoreham Nuclear Power Station, : (ASLBP No. 86-533-01-OL)
Unit 1) :
- - - - -x
(~\
DEPOSITION OF GARY A. SIMON Washi.ngton, D. C.
Friday, December 19, 1986 ace-FEDERAL REPORTERS, INC.
Stenotup Ik;vrtm n 444 North Capitol Street (j
Washington, D.C. 20R)1 (202) 347-3700 Nationwide Coverage 42{ggQ $N$ 2 T
970 01 01 1 CCrysimons 1 UNITED STATES OF AMERICA ,
2 NUCLEAR REGULATORY COMMISSION 3 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4 ---------------X 5 In the Matter of: :
6 LONG ISLAND LIGHTING COMPANY : Docket No. 50-322-OL-5 7 (Shoreham Nuclear Power (EP Exercise) 8 Station, Unit 1) :
9 ---------------X 10 Washington, D.C.
N 11 Friday, December 19, 1986
'Q 12 Deposition of 13 GARY A. SIMON 14 the deponent, called for examination by counsel for LILCO in 15 the offices of Hunton & Williams, 9th Floor, Conference Room 16 No. 4, 2000 Pennsylvania Avenue, N.W., Washington, D.C.
17 20036, beginning at 9:15 o' clock a.m., before Mary C.
18 Simons, a Notary Public in and for the District'of Columbia, 19 when were present on behalf of the respective parties:
20 .
21 o
RJ 22 ACE-FEDERAL REPORTERS, INC.
202447-)?00 Nationwide Coverage Iktk3MN46
970 01.01 2 COrysimons 1 APPEARANCES:
2 On Behalf of Long Island Lighting Company:
3 JESSINE A. MONAGHAN, ESO.
4 THOMAS E. KNAUER, ESO.
5 Hunton & Williams 6 2000 Pennsylvania Avenue, N.W.
7 Washington, D.C. 20036 8 On Behalf of Suffolk County:
9 MICHAEL S. MILLER, ESO.
10 Kirkpatrick & Lockhart
/~N 11 South Lobby, 9th Floor d-12 1800 M Street, N.W.
13 Washington, D.C. 20036-5891 14 On Behalf of the State of New York:
15 RICHARD J. ZAHNLEUTER, ESO.
16 Special Counsel to the Governor 17 Executive Chamber 18 Room 229 19 State Capitol 20 Albany, New York 12224 21 .
22 ACE-FEDERAL REPORTERS, INC.
m.m.m, m_ u. <_ . ..,3 -
I -----_--- - _ _ _ _ _ _ -
970'01 01 3 C;rysimons 1 On Behalf of the Nuclear Regulatory Commission:
2 ORESTE RUSS PIRFO, ESO.
3 U.S. Nuclear Regulatory Commission 4 7735 Old Georcetown Road 5 Bethesda, Maryland 20814 6 * * * ***
7 .
8 9
10
- y 11 12 13 14 15 16 17 18 19 20 21 22 ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage an33M M6
a'CR29297.0 3.1-
,aSIMONS/cg
._ _ C~O-N T E N T S 2{ WITNESS- EXAMINATION Gary A. Simon 3
by Ms.-Monaghan 4 4 by Mr.-Pirfo. 78 5.
l
'q
~6 EXHIBITS 7
i SIMON DEPOSITION NUMBER IDENTIFIED ,
Exhibit 1 12 ;
9 10
.,V_
_j 11 -j JA):
y 12 13 14 i i
15 ;
i 16 j ,
s 17 i
l 18 i !
1- i I,
' t 19 l 2 0 ,' i 21 '
i l
s 22 ' '
C i l
! i '
i j,
I
,r j~ i ACE-FEDERAL REPORTERS, INC. i
! 2CM' ?.*0)
Nat:enwide Onetage NX).3 t6-4 M6 l
970 01 01 4 PROCEEDINGS
~
marysimons 1 2 Whereupon, 3 GARY A. SIMON' 4 was called for examination by counsel for LILCO and, having 5 been first duly sworn by the Notary Public, was examined and 6 testified as follows:
7 MS. MONAGHAN: Dr. Simon, my name is Jessine 8 Monaghan and I am with the law firm of Hunton & Williams.
9 We represent Lona Island Lighting Company in this proceedina 10 for an NRC Licensing Board.
/~' 11 If at any time during the deposition today you V) 12 need to have a cuestion clarified or if-you would like to 13 - expand upon an answer that you have already given, please 14 let me know and you can do that.
15 For the benefit of the-court reporter, when you 16 respond in an affirmative fashion to a cuestion please state
-17 yes for the record so that she can record your response.
18 Nods of the head are not easy for her to record.
19 EXAMINATION 20 BY MS. MONAGHAN:
21 O Are you represented at this deposition by f)
L/ 22 counsel?
ACE-FEDERAL REPORTERS, INC.
I 202-347-3700 Nationwide Coverage 80 4 336 4 686
970 01 01 5 t :rysimons 1 A Yes.
2 O Who is that counsel?
3 A Mr. Miller.
4 O Did you bring with you any documents today?
5 A I did not.
6 0 Now did you meet with Mr. Miller or anyone else 7 in preparation for your deposition today?
8 A Yes.
9 O Who was that person?
10 A Tip Letsche.
11 O Anyone else?
12 A I met with many lawyers at Kirkpatrick, but the 13 essence of the discussion relevant to this deposition was 14 conducted with the two attorneys I mentioned.
15 O Ms. Letsche and Mr. Miller; is that correct?
16 A Yes.
17 0 When did you meet with them?
18 A I met with Tip Letsche once in November and met 19 with the lawyers again yesterday.
20 0 was anyone else present at your meetings with the 21 lawyers from Kirkpatrick and Leckhart other than Kirkpatrick 22 and Lockhart lawyers?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nat kmwide Coserage MXL336-(M6
970 01 01 6 m rysimons 1 A No.
2 O Dr. Simon, did you have other meetings in order 3 to prepare for the testimony that you are going to give in 4 this proceeding?
5 A No.
6 O So you've met only twice, once in November with 7 Ms. Letsche and yesterday to prepare for your deposition?
8 A Yes.
9 O And there have been no other meetings; is that 10 right?
11 MR. MILLER: Well, let's clarity. We are talking 12 about meetings to prepare for Dr. Simon's deposition today 13 or to prepare the testimony he may give in these 14 proceedings?
15 MS. MONAGHAN: The initial cuestion to which you 16 responded, no, no one else was present, was for preparation 17 for your deposition today.
18 Now the cuestion has changed. The cuestion now 19 is were there any other meetings which were conducted in 20 connection with preparing you to testify, not necessarily 21 limited to preparation, but at which you may have discussed
- 22 the testimony that you are coing to give in this proceeding?
ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 80fL336-6M6
- n- .,
i I
~
1 970.01'011 7
- mhrysimons 1 THE. WITNESS: It's hard to make the separation.-
2 'Much of this is simultaneous. Talking about:what we will
-l if 3 discuss at! deposition is very much related to what welwilll l 1 I 4 talk about;at testimony. So I would-have trouble making a )
i-5 separation in my mind.
6 L BY. MS. MONAGHAN: !
p
~
7- O So in connection with this case you have only.had 8- two meetings, one in November with Ms. Letsche and one
- 9. yesterday with the lawyers to prepare for your. deposition 1G today?~
11 A Yes.
b 12. O- And at those meetings there no one:present other.
-13 then Kirkpatrick and Lockhart lawyers; is that right?
14 A That is correct.
i ._
l 15 O Have you ever met with Mr. Zahnleuter before?
16 A No. I only met him yesterday.
17 0 Was he present at the meeting that-you had with 18 the Kirkpatrick and Lockhart attorneys?
19' A Well, I was there over a period of time and some
~
20 of the time was in actual meetings and with general people
,- 21. talking about nothing in particular and others were talking b
A )- s 22 about the details of my work in the case and that was just ;
i t :
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
- Q)70 01 01 8 i i q,i ~
marysimons 1- with the two attorneys that I mentioned.
2' O Mr. Miller and Ms. Letsche?
3 A Yes, correct.
4 O Would it be correct to characterize your meeting 5 of Mr. Zahnleuter yesterday as a casual meeting?
6 MR. MILLER: Well I'm not sure that Dr. Simon has 7 testified that he had a meeting with Mr. Zahnleuter.
8 BY MS. MONAGHAN:
9 O Dr. Simon, how would you characterize your 1 10 acquaintance with Mr. Zahnleuter?
11 A I was introduced to him and I said hello and d(~S 12 exchanged pleasantries and may have casually mentioned 13 aspects of the case. I would not say that any details ofe 14 the case were discussed.
15 0 All right.
16 Now at the meetings that were held with the 17 Kirkpatrick and Lockhart attorneys were you shown any 18 documents?
19 A Yes.
20 0 Do you recall what the documents were that you 21 were shown?
?O, s .L 22 A Yes.
l L
l ACE-FEDERAL REPORTERS, INC.
! 202-347-3700 Nationwide Coverage 804 336-6646
, -- ~, . - , . - . - , - .. , , - , - - . . . - . _ , , _
~
F B
970 01 01 9 mrrysimons 1 O What were they?
2 A One of the documents was the list of contentions, 3 another document was the FEMA report and at some point it 4 seemed like abstracts or selections from the evacuation 5 plan. Now you may have other identities for these 6 documents. So I hope I'm identifyino them correctly.
7 MS. MONAGHAN: I think your identification is 8 clear for my purposes.
9 Let's take a break.
10 (Brief recess taken.)
, 11 MS. MONAGHAN: Back on the record.
J J2 BY MS. MONAGHAN: (Resumino) 13 0 Now you have indicated that the documents you've 14 seen are the contentions.
15 A Yes.
16 O The FEMA report, have you seen the whole FEMA 17 report?
18 A I believe that what I had was the entire report.
19 O And you've seen abstracts from the evacuation 20 plan.
21 A Yes.
22 O I'm coina to assume that by the evacuation plan ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80} 336-6M6
970 01 01 10 m2rysimons 1 .you mean the plan and procedures. Did you see anything.that 2 had the OPIP on the-top, 0-P-I-P?
3 A ~ I do not recall exactly. The abstracted chunk 4 that I had did not even have a face page, although there 5 were fairly high page numbers on them, numbers into the 6 hundreds. I believe it was part of a lanothy document.
7 O All right. Then let's assume that you've seen 8 some parts of the plan and some parts of the procedures.
9 Do you know what the subject matter was that were 10 contained in the abstracts that you had from the plan and
(~} 11 procedures?
- _/
12 A Not in any detail, no.
13 0 Generally what it was?
14 A That particular document was one that I skimmed.
15 I would not describe it as fascinating reading. As 1 16 recall, it was a bunch of descriptions about where people 17 were supposed to be, procedures that were supposed to be 18 followed and schedules for various kinds of things.
19 O Do you remember any of the kinds of things for 20 which you saw schedules? Were there, for example, traffic 21 guides?
-g
(_.) 22 A I don' t think I could make a definitive statement ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 8043366M6
970 01 01 11 mmrysimons 1 on.any substance out of that. I mean I did look at the 2 document, I skimmed it and I observed it rather quickly.
3' This was peripheral to my main objective. So I could not 4 say that I got any deep information out of that plan.
5 0 ~ What did you consider your main objective?
6 A The details of Contention.21.
7 O And what are you going to do with the details of.
8 Contention 21?
9 A That's kind of a broad question. Well in essence 10 I'm going to support our version'of Contention 21.
'('N 11 O How do you intend to support the contention?
%f 12 A The contention makes the claim that the 13 inferences contained in the FEMA reports are inappropriate 14 and uncalled for.
15 0 What are you going to use to support the 16 conclusion that the inferences in the FEMA report are i 17 inappropriate and uncalled for?
l
! 18 A These will be statistical arguments.
19 MS. MONAGHAN: Dr. Simon, I'm going to provide 20 you with a copy of a resume has was provided to us by your i
L 21 counsel ---
r)
(_ 22 THE WITNESS: I remember it well.
i l
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 804336-6M6
970 01 01 12 m;rysimons 1 MS. MONAGHAN: ---
and ask that the court 2 reporter mark this document as Simon Exhibit No. 1.
3 (The document referred to was 4 marked Simon Deposition Exhibit 5 No. 1 for identification and 6 submitted for the record.)
7 BY MS. MONAGHAN:
8 O Dr. Simon, would you please review this copy of 9 your resume and in reviewing it would you please let me know 10 whether it accurately describes your educational background, 11 your employment background and the publications which you 12 have either reviewed or written.
13 A Okay. Are you asking me to say is this correct?
14 0 Yes, please.
15 A I believe it is. I'll ma'<e one re-reading of it.
16 (Pause as the witness reviews his resume.)
17 I believe this is correct.
18 0 Okay. So there is nothing that you would change 19 or add to that resume? Does it list all of your 20 publications?
21 A There is a book review coming out in a couple of j 22 months and I'm not even sure I remember the book now.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
970 01 01 13 murysimons 1 (Laughter.)
2 I remember the book. It was a book on 3 statistical graphics, but aside from that I wouldn't make 4 any chance.
5 O But at the present time it is correct?
6 A Yes, I would say so.
7 0 Dr. Simon, have you ever testified before in a 8 deposition?
~
9 A In a deposition, yes.
10 0 In what proceeding was that deposition given, do 11 you recall?
~
12 A That was with regard to I had worked as a 13 consultant to corporation counsel for the City of New York.
14 That's listed on page 2 of the resume. That case changed 15 names several times. When I was in it it was known as 16 Wilder versus Bernstein, and that dealt with child care in 17 the City of New York.
18 0 What was the nature of the testimony you-cave in 19 that deposition?
20 A It was purely statistical and dealt with an 21 analysis of -- well, it was the dispersion of various racial i __
22 minorities throughout the entities of the child welfare ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage M)3364646
970.01.01' 14
.mcrysimons 1 system, or child care system might be a netter name.
2 O So that what you were examining in the testimony 3 that you gave during that deposition was the frequency in 4 the population of persons affiliated with the child care 5 system in the City of New York of racial minorities; is that 6 right?
7 A I don't know if I would phrase it exactly that 8 way. The child care system in the City of New York involves 9 a number of entities, orphanage type entities, things like
. 10 group homes and religious organizations that take care of 11 -orphaned or abandoned children.
/}
12 The complicated question in that case was a 13 relicious entanglement with the State. Many of these child 14 care organizations are run or maintained by religious 15 organizations and yet they are in the position of receiving 16 public funds. Therefore it became a question worth asking, 17 do these religious entities, which are receiving public 18 funds, seem to be treating racial and ethnic minorities 19 appropriately.
20 0 Appropriately in terms of their employment of 21 those racial and ethnic minorities or the care they were 22 providing?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationside Coverage M33MM6
+
970.'01: 01 15 marysimons l' (Counsel and witness' confer.) l 2 A Could you repeat the cuestion, please?
3 0 Were you concerned with the representation of 4 racial _and ethnic minorities in employment'by those 5 -particular child care agencies?
6 A- We were concerned with the children. The child 7 care consists of these many entities, some of which are 8 perceived'as superior to others, and the question.at hand 9 was the misallocation of in particular black Protestant 10 children among these agencies.
11 O Was there a particular kind of statistical theory 12 or methodology you applied in analyzing this information?
13 A Yes.
14 O And what was that?
1,
-15 A Probably the procedure which I added to this 16 would most simply be identified as a Mantel Haenszel 17 procedure, which I will spell. M-a-n-t-e-1 18 H-a-e-n-s-z-e-l.
19 O What is the Mantel Haenszel procedure?
20 A In 25 words or less? In a case of unequal 21 dispersion of a group of interests, such as black Protestant p).
ls_ 22 children, in a large bureaucratic system, an overall i
L ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8 @ 3366646 l .. - - _ , , _ , , - . . , . , - . . _ _ . . -.,__,.-_---_.__.--__,_m.,__,-. - . - - - , , _ . - - - _
970 01 01 16 marysimons 1 analysis looking at the summary total will sometimes give 2 different picture from an analysis looking at individual 3 subcomponents.
4 This particular statistical procedure is a way of 5 taking numbers out of the subcomponents.and deciding if 6 there is an appropriate inference to be made.
7 MR. MILLER: It's clear to me.
8 (Laughter.)
9 THE WITNESS: The proper way to find out about 10 the Mantel Haenszel statistic is to find a textbook and look
/~N 11 it up in the index and read the discussion. One such book y )
12 would be Fleiss, F-1-e-i-s-s, " Statistical Methods for-Rates 13 and Proportions." That has a reasonable explanation of the 14 procedure.
15 BY MS. MONAGHAN:
16 O Do you expect to be using the Mantel Haenszel 17 procedure in analyzing the information that you are going to 18 be analyzing in connection with Contention 21?
19 A I would regard it as unlikely. I cannot exclude 20 it, but at the moment I don' t see where that's going to come 21 in.
fN
/w) 22 O Is that the only deposition that you've given ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8%33MM6
970 01:01- 17'
.marysimons'l' previously?-
2 A Yes.
! 3. O' Now have you testified in any proceeding before a1
'41 court?
5 A. No.
6 -O~ Have you testified in any proceeding before an 7 administrative agency or body?
i 8- A-- N o . --
. 9- O Have you ever offered any other testimony?
10 A No.
I(~$ 11 0 Dr. Simon, are you a member of any professional G
12 associations? I-don't note that that's on the resume.
l 13 A Yes. Did'I not' list that? .In nany cases i
L 14- memoerchip in a professional organization is eouivalent to 15 paying dues. So I'm a member of the American Statistical i 16 Association and the Biometric Society.
l 17 O Were you elected to either of those associations L
18 or are those just general membership associations, dues 19 paying?
i 20 A Those were dues paying professional l
21 associations. I was not elected to them.
22 O Are there any elected associations affiliated ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
970 01 01 18 marysimons 1 with the field of statistics?
2 A There are a number of strange entities or strange 3 arrangements. The Royal Statistical Society operates on a 4 membership by nomination from present members. There are 5 also elected positions called Fellows, F-e-1-1-o-w-s, and 6 the American Statistical Association has elected Fellows. I-7 am not one of them.
8 O What does it take to become an elected Fellow in 9 the American Statistical Association?
10 A I'm not sure of the exact requirements.
ri 11 O Dr. Simon, I think you've indicated to ne that v
12 you have listed all your publications and book reviews other 13 than the one that is currently in progress on your resume.
14 A I believe that is correct, yes.
15 0 Would it be correct to say that the discipline in 16 which you consider yourself to be an expert would be 17 statistics?
18 A Yes.
19 0 You don't consider yourself to be a sociology 20 expert, do you?
21 A I do not.
22 O So, if you will, the bounds of your expertise ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-33Mi>46 i .. .. .
970.01 01 19 marysimons.1 would be statistics?
2 A Yes. That's a fair statement'.
3 0 When did you first learn that you might be a 4 witness in this proceeding?
5 A In November.
6 O How were you notified that you would be a
[ 7 witness?
8 A I received a call from Tip Letsche.
9 O Did you receive any documents from the 10 Kirkpatrick and Lockhart lawyers in connection with that 7~'; 11 call?
L'd -
12 A No, not with the call, but we met subsequently, 13 as we mentioned before, and I don't recall at the moment 14 whether I was shown documents then or whether they were i
15 mailed later.
16 O Now, Dr. Simon, the documents that you have 17 enumerated - for me previously, the contentions, the FEMA 18 report and the abstracts from the evacuation plan and 19 procedures, did you review those in preparation for your l
l 20 deposition?
21 A Parts of them.
(~'T
(_/ 22 O What parts?
l l
ACE-FEDERAL REPORTERS, INC.
202 347 Nationwide Coverage 800 33M686 l - -
- -- - - . .. -3700-- _ , - - . _ _ _ _ . . - . - - . , _ _ . _ _ _
. _ . - ~ . , .
u 970;01'Ol' 20 marysimons 1 ..A- Well, particularly Contention No. 21, and in the.
2 process.of doing that.there were occasional cross-references-4 3 to the FEMA report.
~ 4. O . When-there was a cross-re'fere'n'ce did you look it.
5 up,in the FEMA report?
6 A Many-of them.
7 0 You did not re-review the plan and procedures in' 8 -preparation for your deposition?
9 A I did not.
10 0 Are there any other documents that yo'u reviewed
!p; . 11 in connection with preparing for your deposition?
^J 12 A No.
13 O Do you expect to testify as an expert witness in 14 this proceeding?
15 A I expect so.
16 O And the subject matter on which you expect to
- . 17 testify will be statistics; is that right?
18 A Yes.
19 O And specifically it will be statistics as related i-
> 20 to Contention 21; is that right?
I 21 A I believe that is correct.
- 22 O At this time do you know what opinions you expect 1
i ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-33MM6
_ . , , - - . . _ - . . - _ _ - , _ _ _ . _ . . _ _ . . , _ . . . . _ . . , _ _ , . . - . _ . _ _ . - - - _ . . _ - . . _ , . _ . . _ _ - , _ . . ~ . . - - _ - -
970 01 01 21 mnrysimons 1 to render in connection with your testimony on Contention 2 21?
3 A Approximately, yes.
4 0 What will those be?
5 (The witness confers with his counsel.)
6 A From what I've seen to date, my opinion is I 7 think what we noted before, namely, that the numbers and 8 information as reported in Contention 21 do not support the 9 inferences in the FEMA report.
10 0 Do you believe that you are coina to be relying 11 on any of your published works in connection with the 12 testimony that you are going to be civing in this 13 proceeding?
14 A Probably not. I can't preclude that anything I 15 know would not be brought into it, but at the moment I don't 16 see the direct application of any of these.
l 17 0 I assume that you are going to draw certainly on 18 your background knowledge of statistics as an expert in 19 order to render the testimony?
20 A Yes.
21 O My specific question to you is whether you think 22 any of the publications which you have authored will be ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coserage 8@336-6M6
970 01 01 22 m rysimons 1 directly applicable to your testimony in this case, and if 2 you would just like to take a look at the ones that are 3 listed on your resume, please do.
4 A I don' t believe that I'll be drawing from those.
5 O Now before we talked about the Mantel Haenszel 6 method and you said that it was not likely that the Mantel 7 Haenszel method would be applied to your analysis of 8 Contention 21. At this time do you know what statistical 9 method or theory you think you are going to apply to 10 Contention 21?
11 A I have an approximate idea.
12 O And what is that idea?
13 A The statistical questions that are involved here 14 are fairly elementary. The statistical questions center 15 around observed counts of successes and failures in fairly 16 small samples and questions and inferences were made about i 17 how they apply to the overall performance of an evacuation 18 plan.
19 0 Is what concerned you the small sample size that 20 was used here?
21 A That's one of the things that concerns me.
22 O What else?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cos erage 80h336-(M6
- g. ~. _ -> . .
970.;01 01~ - 23 marysimons-1. -A In addition to problems of small sample size, the 2 results that'were seen here were.not uniformly perfect. One 3 can occasionally live with a small to modest. sample. size if 4 .there is a perfect response.
5- O By a perfect response.in connection with the 6 factual issues listed in 21 what do you mean?
7- MR. MILLER: Excuse me. Are.you asking 8' specifically with respect to the subparts of Contention 21?
9 .MS. MONAGHAN: Let me rephrase the question.
10 BY MS. MONAGHAN:
("T 11. O What do you mean by the term " perfect response"?
.J 12- A 'For example, if you were to flip a coin 12 times, l'3 let's say, and observe that it comes up heads every time,-
14 you would have a reasonable suspicion that the coin in fact 15 has two heads in spite of-a small sample size.- That is,'an 16 experiment consisting of 12 runs would generally be regarded l-i 17 as a rather small experiment. But in a case like this, that 1.
l 18 is flipping a coin and getting 12 heads, you have a l:
19 reasonable suspicion that something is wrong with the coin.
20 0 If you had, for example, 66 bus drivers that were l
l 21 to go out and perform a certain task and you sampled six of 22 those bus drivers and all six of those bus drivers performed ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 804336-6M6
c 970 01 01 24 marysimons 1 perfectly, would you be able to reach any conclusions about 2 the larger population based on that sample size?
3 A You are able to reach some conclusions. Six is a 4 small sample size. For example, in the previous context of 5 flipping a coin six times and seeing six heads would not 6 necessarily overwhelmingly persuade that the coin has two 7 heads.
8 O At what point would you have the requisite level 9 of confidence using your example that the coin has two 10 heads?
11 A The problem with specifying such numbers now is
)
12 that they are context dependent, and a problem involving bus 13 drivers at this point is not ouite the same as a problem 14 involving coins.
15 In order to answer the question, alas, one has to 16 specif y things like what level of performance are we trying 17 to obtain and what statistical assurances do we want that we I
18 are within shooting distance of the target.
19 0 Now you said that your confidence level is very 20 dependent on the context in which you're looking at the 21 numbers. Other than the level of performance that you are 22 tryino to maintain and what degree of statistical assurance ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-33m
970 01 01 25 marysimons 1 you are looking'for, are there any other criteria that_you 2 use to determine what your sample size needs to be in order 3~ for you to have the requisite confidence level?
4 A We have enumerated the target level of 5 performance as one of the things you want, for example, a 90 6 percent success rate. We mention a notion of statistical 7 confidence as perhaps illustrated by a phrase, I am 99 8 percent confident that, and another entity that we might 9 want to drag into it is the difference from that target.
10 For example, if the target performance is 90 percent, we may 11 want to, for example, be 99 percent confident that the tsi}
12 result is at least 85 percent or above. I'm sorry for that 13 tangle of numbers, but that's really the way it is.
14 O That's fine.
15 I think we're probably going to get back to this i
l 16 in a little while, but did you observe the FEMA exercise?
l l
17 A No.
18 0 Have you spoken with anyone who did?
l 19 MR. MILLER: If you know the answer to that 20 ouestion, Dr. Simon.
l 21 THE WITNESS: Well, I spoke with the attorneys.
f
( 22 I have no idea if they observed it.
ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coverage 800-33MM6
2970 01 01 26 Om rysimons 1 BY MS. MONAGHAN:
2 O Have you had any discussions with anyone about 3 what they may have seen at the FEMA exercise?
4 A No.
5 0 Did you assist in any way in drafting Contention 6 21?
7 A I did not.
8 O Did you assist in drafting any of the 9 contentions?
10 A No.
11 O Now, Dr. Simon, isn't it true that if you have an 12 inadequate sample size you can't reach positive conclusions 13 about the entire population, for example ---
14 A It depends on the -- go ahead.
15 0 Let's stop the cuestion there and we'll get to 16 the example later. So ahead with your answer.
17 A What is the ouestion then? Let me make sure I 18 understand.
19 0 If you have an inadequate sample size, what you 20 consider to be an inadequate sample size, can you reach a 21 positive conclusion about the rest of the population?
j 22 A It depends on the problem. If we are debating, ACE-FEDERAL REPORTERS, INC.
202-347-37(X1 Nationwide Coserage 800-336-6N6
/
970 01 01 27 m rysimons 1 for instance, whether a coin has two heads or whether it has 2 a head and a tail and you flip it three times and get two 3 heads and one tail, we have rather substantial evidence that 4 it's not a two-headed coin. So depending on what one is N OT 5 trying to show, the sample may or may gbe ef fective. /
6 0 If you were looking at a job task, for example, 7 that has a number of different parts or tasks that one has 8 to perform, and let's take, for example, a bus driver under 9 the LERO plan and they need to do things like pick dosimetry 10 and drive to a bus station and pick up a bus a drive a bus 3
11 route and go to a transfer point, a number of tasks which 12 they need to perform, in order to ascertain whether or not 13 the population of bus drivers who are to perform those tasks 14 are able to perform those tasks, would you need a larger 15 sample size than, for example, you would need with the coin 16 toss because there are more variables involved?
17 MR. MILLER: If you understand the cuestion, you 18 can answer it. It's a rather complex ouestion.
19 THE WITNESS: I'm interpreting the question as 20 are bus drivers the same as coins. They are not in a very 21 crucial sense. Flipping a single coin many times represents
_/ 22 repeated experiments with the same material. Here we are ACE-FEDERAL REPORTERS, INC.
202-347 37(X) NationwiJe Cos erage 8(XK336-6M6
970 01 01 28 marysimons 1 talking about a set of drivers all of whom are different, 2 and the statistical problems that one gets into are a little 3 harder with drivers than with coins.
4 BY MS. MONAGHAN:
5 0 Is there some way that you can control for the 6 difference between a coin and a driver when you do a 7 statistical analysis?
8 A I don' t think I could answer it in those terms, 9 I'm sorry. I'm not trying to be difficult. It's just that 10 the coin analogy may not have been perfect for all examples, fs 11 0 Let me try it again.
g 12 When you would do a statistical analysis of bus 13 driver performance, is there some way that you can. control 14 for the fact that you're dealing with a number of different 15 persons each of whom is likely to be slightly different than 4
16 the next person?
l l
17 A There is a standard orthodox methodology for 18 doing that and that requires that you draw a simple random l
19 sample from the population of drivers. You may get good 20 ones in your sample and you may aet bad ones in your sample, 21 but we have probability calculations that can be applied to o
I- 22 all the outcomes. So at least we have under control what we ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-33MM6
T r
'L
-; . ' ' g
(
970 01 01 29 m2rysimons 1- are doing.
2 O How would you go about drawing a simple random 3 sample from a population of bus drivers?
4 MR. MILLER: I have a number of objections.. No.
5 1, it's irrelevant to what this witness is here to testify 6 about. No. 2, you've asked this witness to speculate and 7 you're asking him to draw assumptions under the LILCO plan 8 that he may not be familiar with.
9 If Dr. Simon thinks he's capable of answering 10 that cuestion, he may do so over my objections for the b'
7~ 11 record.
12 You may answer the cuestion if you can.
13 THE WITNESS: Let me describe how to draw a 14 simple random sample in general.
15 MS. MONAGHAN: All right.
16 THE WITNESS: The interconnections between what 17 I'm going to say next and what actually happened in the 18 evacuation drill, it would need to be established on a point-19 by-point basis, but again I'm dealing with this situation ex-20 post. That is, the drill has happened, it has been done, I 21 have the data as it was actually done and I wasn't asked to h
Es- 22 design it.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage En336-6646
m - - -
q 970 01 01 30 marysimons 1 BY.MS. MONAGHAN:
2 O Just for a minute before you describe on'a 3 general basis how one does a simple random sample, let me 4 just understand that the data on which you are relying is 5 what is contained in the FEMA exercise assessment; is that 6 true?
7 A My understanding is that the numbers that appear 8 in Contention 21 came out of the FEMA report and to the 9 extent that I cross-checked it, that seems to be the case.-
10 0 So for your raw data.you are relying on both
/~~N 11 what's in the contentions and the FEMA exercise report?
d 12 A Yes.
13 O Are you relying on any other documents?
14 A Not at the moment.
15 O Do you know whether other documents exist which I might assist you in your analysis?
16 17 A I really don't know that. There may be many
., 18 documents out there. I only know the ones that I've seen.
19 0 Let's assume that the players during the 20 exercise, the people who are performing the different tasks, 21 kept logs of their actions during the exercise, do you think f) x_/ 22 at this time that those logs might assist you in any way?
, ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800 336 4646
970 01'01 31
- m;rysimons 1 A You see, I'm relying on the numbers at face 2 value. At this point I'm not in the game of auditing the 3 correctness of the numbers that were recorded by FEMA or 4 then put into Contention 21. I suppose I'm capable of 5 undertaking such an audit, but I have no plans to do that 6 now.
7 O So that you are taking what is in the FEMA report 8 and the contentions at face value with respect to the 9 numbers involved?
10 A Yes.
r'] 11 O If there is a discrepancy between the contentions u
12 and the FEMA report how do you plan to deal with that?
13 A I don' t know. I'm presuming that the FEMA 14 document now has some authoritative status in this case. So 15 I really don' t know what I've cot to believe, but I don' t 16 see how I can ignore what is in the FEMA report.
17 0 okay. Let's go back to the simple random sample, 18 and you were about to tell me how one would go about drawing 19 the simple random sample on a ceneral basis.
20 A That recuires that at step one you form a roster 21 ,of all of the elements of the population, be they bus
&)
A m .J 22 drivers, the accounts for audit or be they students out of a ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6646
h I
970 01 01 32 mnrysimons 1 university. You put them down on one long list. Person No.
2 1 is designated with No. 1 and this goes up to a last person 3 who is designated capital "N".
4 Then with the aid of either a table of random o (~ d'#
5 numbers jrf a calculator or computer generated random number 6 you begin selecting random numbers of out the set 1 through 7 capital "N" until you have filled up your sample size.
8 That's the standard simple random sampling method.
9 MS. MONAGHAN: Let's go off the record a minute.
10 (Pause.)
11 MS. MONAGHAN: Back on the record.
12 EXAMINATION
-13 BY MS. MONAGHAN: (Resuming) 14 O Dr. Simon, I'm going to show you Contention 15 21(a). It's right here in this pile of papers. Due to 16 legal reasons there have been a number of draf ts of the 17 contentions, and this is the latest draft of the contentions 18 which was produced most recently by Mr. Miller's law firm 19 and it consolidates certain parts with other parts.
20 You intend to offer testimony on Contention 21; 21 is that right?
i f"
~
22 A Yes.
ACE-FEDERAL REPORTERS, INC.
202-347 37W Nationwide Coserage 800 33MM6
970 01 01 33 m;rysimons 1- 0 Now if you'll look at Contention 21, it makes the 2 statement that the sample size was too small. What do you 3 mean by the term "too small ' or what do you think that term 4 means?
5 A Well, again, with reference to some of the things 6 we mentioned before, what is too small is context dependent, 7 'and certainly in that overused example of flipping a coin 8 say 12 times and getting all heads, the small. sample size of 9 12 still permits a reasonable inference because the outcome 10 of that experiment was so overwhelming.
r~T- 11 O If the outcome of the coin flip experiment had J
12 not been so overwhelming say that you had had 8 out of 12 13 show up as heads, would you still be able to make 14 generalizations about the coin itself based on that sample 15 size?
16 A That experiment that comes out as 8 and 4 17 definitely allows you to conclude that the coin has a head 18 and has a tail. If the coin appears to really be a product 19 of the U.S. Mint, you would reasonable infer from this 8 and 20 4 breakup that the coin is probably a balanced coin.
21 If instead of a coin this were some other other
,o
(_) 22 kind of mechanical or biological entity with two possible ACE-FEDERAL REPORTERS, INC. I n ,a,. --+c-, - \
_ _ _ _ _ -_ __ _ _ _ _ _ . 1
I 1Y#
l 970 01 01 34 marysimons 1 outcomes, then an 8 versus 4 outcome would not lead you to a 2 very good estimate of the probability of the outcomes.
3 0 Is it your opinion that the samples that FEMA 4 reviewed were too small to reach the conclusions in the FEMA 5 report?
6 A Yes.
7 O Are the samples also too small to reach a 8 negative conclusion about what occurred?
9 A Well that varies by circumstance. Again, the 10 conclusion depends on the result that needs to be shown.
11 The report is filled with the phrase " reasonable assurance" 12 and no quantitative explanation of that number is ever 13 supplied.
14 O Do you plan to apply a cuantitative explanation 15 to " reasonable assurance"?
16 A I'm afraid that is on the other side of the 17 problem. That is, I have the experiment as it was actually 18 done and then could be asked to make statistical statements 19 about the overall probabilities of success. I don't know 20 what constitutes reasonable assurance in this case, but some C.tr Qky 21 of the observed outcomes cetLalu,suggest rather less than pr
() 22 100 percent performance.
ACE-FEDERAL REPORTERS, INC.
mg ~o_- - >> -
I
970 01:01 35 marysimons 1 0 Do you know what the target level of performance 2 was that FEMA expected to be achieved?
3 A I don't believe I ever saw such a statement.
4 O So absent a knowledge of the_ target level of 5 performance that was to be achieved, can you make any l 6' assessment about what the appropriate sample size would be?
7- A It would be very hard to do that. If I don't 8 know what the experiment is supposed to be showing, I-could 9 not give very cood advice about sample size.
10 0 How do you know then that the sample sizes used 11 here were too small?
/~]
s 12 A Well some of the numbers that I see as sample 13 sizes, numbers like one and two and four and eight, are 14 embarrassingly small. These are numbers that, for instance, 15 in medical literature would never be cited except in 16 articles on case studies. Nobody out there in the medical 17 world is making solid inferences based on 4 patients or 8 18 patients.
l 19 O Do you draw an analogy between what the medical 20 world is determining and what FEMA is determining here with
. 21 the FEMA exercise?
- (,/ 22 A That's sketchy. I'm throwing that in as an
(
i ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Coverage 800-336-6M6
970 01 01 36 marysimons 1 illustration of the problems of small sample sizes.
2 O How do you define the term " adequate success 3 rate"?
4 A It depends on what you're trying to do. If the 5 project is filling up with cardboard boxes with corn flakes 6 and if adequacy is defined as the correct amount of weight 7 in it, my management control system might say that 95 8 percent of the cereal boxes have to have the correct amount 9 in them at least.
10 0 Is an adequate success rate similar to a target 11 level of performance or is there a difference between the 12 two instruments?
13 A I was using this as the same idea.
14 O Now just so I understand, what criteria do you 15 use to define what constitutes an adequate success rate?
16 A It's context dependent. In the example of the 17 cereal boxes a statement like 95 percent of the boxes must 18 have at least the advertised amount of corn flakes in them 19 might be the result of a regulation or it might be the 20 result of an internal procedure. It depends on that 21 particular context. In certain situations you need very 22 high levels of success and in other situations you can ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cos erage 800-336-6M6
970 01 01 37 m rysimons-1 ' tolerate lower' levels'of. success.
2 O Now you said earlier that you didn't know what 3 the target level of performance was that FEMA has expected 4 in the FEMA exercise; is that correct?.
5 A Yes, I said that.
6 -O So, therefore, you don' t know what the adequate 7 success rate would be relative to the-tasks that were judged 8 in the FEMA' exercise; is that right?
1 9 A Would.you say that again? I'm sorry.
10 0 Sure. I'll be glad to repeat-the question.
j 11 .Can you make a determination about what the
.v )-
12 adequate success rate would be in connection with the tasks.
- 13 that were evaluated by FEMA?
14 A In a situation like'this the target success 15 probabilities of all of the players in this game is i
16 something I'm not really in a position to decide. I would 17 presume that people would want them to be high.
!- 18 0 What do you mean by high?
19 A Again, it's context dependent. The particular 20 exercise here is fairly complicated. The overall success of l 21 -
the operation depends on the successes of many of the 1
22 subparts. I would expect that a careful examination of the
- ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80433MM6
P l
l 970 01 01 38 marysimons 1 way that the parts of this evacuation link together might 2 enable one to specify what the target probabilities of 3 success are on the individual parts.
4 I'm not in a position to say what percent of the 5 bus drivers should do a good job or what percent of the e, ,
6 traffic guides should do a good job. Some where somebody 3P ss 7 ought to think about what those success probabilities should 8 be and what are adequate. There is obviously more to this 9 than what I've seen, but I have no guidance on what those 10 numbers should be.
. 11 O You said that you are not in a position to 12 determine what the success probability should be for each of 13 the different populations involved in here, and by 14 population I mean bus drivers and traffic guides; is that 15 correct?
16 A I'm willing to speculate that they should all be 17 high probabilities of success , but exactly what they should 18 be I would not be able to pin down.
19 0 Do you think that you are going to be offering an 20 opinion on that after you've had more opportunity to review 21 the documents in this case?
m 22 A It's possible. I don' t see that, but it's ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8ah3W6646
970 01 01 ,39 marysimons I certainly possible.
2 'Q- Would you look at Contention 21(a) about the 3 route-alert drivers. It'begins on page 26 of the document 4 that you have.
5 A Yes.
6 O Now this subpart of Contention 21(a) talks about 7 the route-alert drivers and the sampling that was done of 8 their performance and what was noted with respect to the 9 samples that were taken.
10 A Yes.
'~ 11 0 Are you able to tell me how many route-alert
.v}
12 drivers should have been observed for there to be a 13 statistically significant sample?
14 MR. MILLER: Objection to the relevancy of that 15 ouestion. This witness is not here to tell LILCO nor FEMA 16 nor the NRC Staff how the exercise should have been 17 conducted or structured. It also calls for this witness to 18 speculate.
19 If you can answer the question, Dr. Simon, you 20 may do so.
21 THE WITNESS: One isn't enough. One from each 13
(.) 22 area is certainly not enough.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 8@33MM6
gs. . -
l 970'01 01 40 mtrysimons 1. BY MS. MONAGHAN:
- 2. O So you're able to say that the sample size was 3 too small?
4 A Yes.
5 0 But you are unable to make a determination about-6 what the correct sample size would have been; is that true?
7 MR. MILLER: He didn't testify that he was unable-8 to do so.
9 THE WITNESS: I believe I said with a guiding 10 statement like we want to be sure that this system will be
N - 11 90 percent effective or 99 percent effective, then I could J
12 make a statement about that.
13 BY MS. MONAGHAN:
14 0 So, in other words, you would need to know what 15 the target level of performance would be or what the 16 adequate success rate would be in order for you to make a 17 determination of what would be an adequate sample size; is 18 ' that correct?
- 19 A I believe you said that correctly, yes.
20 0 Let's assume that an adequate success rate would
[ 21 be 90 percent. What formula or criteria would you apply in C'
( 22 order to determine the number of route alert drivers that l
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-336-6M6 l - - - . . , _ _ __ . . _ _ _ - . . _ _..
W 970 01 01 41 marysimons 1 would need to be sampled for you to have confidence in ---
2 MR. MILLER: I again object to the relevancy of 3 that question.
4 You may answer, Dr. Simon.
5 THE WITNESS: I'm assuming the 90 percent figure 6 is hypothetical at this point.
7 MS. MONAGHAN: That's correct.
1 8 THE WITNESS: Now the question becomes site
({
9 specific or overall. If you390 percent at each of the three d 10 locations, then we start talking about populations of size 11 20. There are exact hypergeometric calculations which you 12 can make in this instance which would guide you to the 13 sample size. Without making the calculations my quess is 14 you've got to look at 19 of 20.
15 BY MS. MONAGHAN:
16 0 And if you assumed that you wanted to look at the 17 overall performance and therefore you combined the three 18 staging areas and looked at the whole population of route-19 alert drivers, how would you go about determining the sample 20 size given a hypothetical 90 percent adequate success rate?
21 MR. MILLER: My objection is to relevancy, and 22 I'll make a continuing objection.
ACE-FEDERAL REPORTERS, INC.
> .~, ~--+<.. -
p '
e 970701 01. 42 marysimons 1- 'THE WITNESS: -The answer is the~same calculation,
~
- 2. but with other numbers. My guess is that-'if you wanted to b 3 Ltalk about a 90 percent success rate of out of 60, you are 4- ;probably going to nee'd a sample size ofiSO.-
.5 BY MS. MONAGHAN:
g 6 0 How would you'go.about doing that calculation? ,
s 7 MR. MILLER: Let me just say, so I can stop i= 8 interrupting the deposition, with respect to Contention 9 21(a), Ms. Monaghan, if you are going to ask Dr. Simon to ,
i 10 testify as to how things should have been done during.the 11 exercise in order to construct an adequate sample size, I-12 will just make a continuing objection as to this line'of i
I t
13 questions.
- j. 14 MS. MONAGHAN: What I believe I'm trying to probe h 15 here is the methodology that Dr. Simon would employ so that i
16 I can understand his expertise in the area of statistics 17 that he intends to apply to this particular contention, but
)
.i 18 your objection stands, Mr. Miller, and I understand.
19 MR. MILLER: My problem though with your 20 questions is that Dr. Simon has not been retained as an 21 expert witness by Suffolk County tell others, specifically 22 LILCO, FEMA and NRC Staff, how the exercise should have been c
4-
! ACE-FEDERAL REPORTERS, INC. !
202 347-3700 Nationwide Coserage 8 6 3366646
, - , ~ . _ . ._ _ - - , . _ _ _ _
l
- 9970101 01 43
'" Orysimons 1 conducted. He's here to testify about what was done during 2 the exercise and why, in his opinion, what was done does not 3 lead FEMA to support the conclusions that FEMA made.
4 That's why I think your questions should be
~5 focused on the contentions and not upon what Dr. Simon would 6 have done if he would have been doing it.
7 So you may answer the question.-
8 THE WITNESS: I would feel nervous about this if 9 we were covering up some great secret methodology. The 10 method of selecting sample size is not-a great secret and 11 SS I'm pLILCO people can find someone to explain it to them. '$
j x>
12 (Witness confers with counsel.)
13 Having said that, the method consists of making a 14 set of calculations that are not too hard if done by 15 computer. You imagine out of your population of capital "N" t
16 things, be they bus drivers or whatever, you imagine that l 17 there is some number of them that are adequate performers,
! 18 and this number could be anywhere from zero to capital "N".
19 You let that be a variable number, which means you put it 20 into a computer loop.
21 For every possible number of good members of the 22 population you calculate the probabilities that certain i
1 ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80tk3364M6
6Y l
970 01 01 44 mirysimons 1 numbers of them will show up in samples of certain sizes.
2 And working backward from that arithmetic you are able to 3 determine what sample size will be good enough to tell you 4 with good enough probability whether you have enough good l
5 drivers or enough good members of the population. In some 6 sense you're asking for mathematical formulas which are a 7 little difficult to convey verbally.
8 BY MS. MONAGHAN:
9 0 Is there any specific name that's applied to this 10 type of calculation that's done or a statistical theory that 11 it goes under?
12 A It's very similar to audit sampling.
13 O Looking at Contention 21(a), Dr. Simon, given the 14 sample size that was sampled there, are you able to prove a 15 negative conclusion about the rest of the population, i.e.,
16 that they would not be able to perform the task that they 17 have been assioned, the route-alert drivers?
18 A Here one should note also that the FEMA sometimes 19 gives qualified success. For example, somebody did 20 something, but did it a little slowly. I'm not sure how to 21 appraise that. Taken out of context, if you have a 22 population of size 20 and you examp?- one of them and it is ACE-FEDERAL REPORTERS, INC.
202 347-37(X) Nationaide (c tt .lge $ Nk 33MM6
970.01:01 45 marysimons 1 okay, technically it does not preclude the possibility that 2: all 20 are okay, but it would certainly be a very foolish 3 inference.
4 O And, likewise, if you had a population of 20 and 5 you sampled one and that sample was not okay, you would be 6 equally foolish in concluding or in generalizing to the 7 whole population that the whole population would be no okay; 8 is that correct?
9 A That's the parallel statement. So, indeed,.if 10 you have a group of 20 and you examine one and it is bad,
/~N 11 that does not allow you to conclude that they are all bad, x,_)
12 again a foolish inference.
13 0 You indicated that the FEMA report assigns 14 qualified success to some of the behavior that is noted in 15 the FEMA report. Do you have an idea at this time how you 16 will deal with instancos where it's a qualified success in 17 terms of your statistical analysis?
18 A Not entirely. That's a little dif ficult to sort i
19 out because a qualified success is sometimes very close to 20 failure. So I would probably need some guidance on how to i
21 deal with those.
22 0 When you have a qualified success in order to l ACE-FEDERAL REPORTERS, INC.
! 202-347-37(0 Nationnide Coserage 8 % 334 6646
970 01 01 46 marysimons 1 deal with it on a statistical basis do you have to either 2 deal with it as a success or as a failure?
3 A It again depends on the context. Sometimes a 4 qualified success isn't good enough. If I try to catch a 5 train and I'm only 37 seconds on the wrong side, it's as 6 good as a failure. If I'm preparing a batch of cookies and 7 they are 72 percent okay, well there's still something to 8 eat.
9 I don't know how to make those appraisals as in 10 the instances of players in this game showing up at assigned 11 positions late. I'm not really in a position to say what's
'~')
w_-
12 the impact of that on the evacuation.
13 O My question really coes to when you have these 14 sort of gray areas, if you will, where at this particular 15 time it's sort of qualified kind of fuzzy as to what what to 16 look at it. When you make your final determination to do a 17 statistical analysis do you have to assign to the gray areas 18 either black or white?
19 A Again, that has got to depend on the situation.
20 In the context of cookies , obviously you can have fractional 21 success and you can just average up fractional results.
22 In the case of making or not making trains, a l
ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide Coserage 80ik33MM6
970 01 01 47 mnrysimons 1 qualified success apparently is going to count the same as a 2 failure.
3 It is possible in the cases of uncertainty to 4 develop a whole set of boundaries ;&f separating success from d#
5 failure and do the analysis by every calculation. I mean 6 that is do every possible analysis. At one extreme end the 7 qualified successes will count as successes and at the other 8 extreme end the qualified successes will count as failures.
9 0 Do you know what criteria you would apply in 10 order to engage in the kind of analysis that you've just 11 described?
12 A I think I understand the question. If I have a 13 part of this in which some of the players were obviously 14 successful and some were obviously unsuccessful and a few in 15 between, I can simply repeatedly apportion various fractions 16 of the in between people to the successes and failures.
17 For instance, I can start off by putting half of 18 them in the success group and then in the next step I could 19 put three-quarters of them in the success group and a 20 Quarter in the failure and so on. It's possible to do 21 things like that.
22 Again, I would feel at least a little bit ACE-FEDERAL REPORTERS, INC.
202-347 37(t) Nationwide Coserage muk34tM6
970 01 01 ;48 marysimons 1 uncertain about.that in that, gee, I don' t really know what 2 a qualified success has as its consequence in every one of 3 these scenarios.
4 O How would you go about determining what a 5 qualified success would have as its consequence?
6 A I don't know. I think if I knew, then there 7 wouldn't be qualified successes.
8 0 Would you be able to determine that on your own 9 independently? <
10 A I doubt it. I can't imagine being able to do it
/~3 11 without some help.
u) 12 0 So you would need to consult with someone who 13 would be able to make a value judoment about the qualified 14 success of a qualified failure?
15 MR. MILLER: I assume that all these questions 16 are in the context of LILCO's emergency plan.
17 MS. MONAGHAN: Yes.
18 THE WITNESS: Oh, yes.
19 MS. MONAGHAN: I assume that that's the context 20 in which you are answering, Dr. Simon; is that correct?
1 21 THE WITNESS: Yes. I would have to get some 22 quidance from some one about how to deal with qualified i ACE-FEDERAL REPORTERS, INC.
202 347 37(M Nationwide Coserage Nn3366646
970'01_'O1: 49
-marysimons I success.
2 BY MS.~ MONAGHAN:
3 l) What about non-probability based sampling, is 4 that applicable to these kind of. situations?
5 MR. MILLER: These kind of situations meaning?
6 MS. MONAGHAN: Meaning the situations which we 7 are discussing under Contention 21(a).
8 MR. MILLER: All the subparts of Contention 21 or 9 21(a)?
- 10 MS. MONAGHAN
- 21, I'm sorry.
'/~} 11 THE WITNESS: Non-probability sampling has in v
12 general a bad taste in the mouths of statisticians. Non-
- 13 probability sampling is occasionally used as an apologetic 14 af ter-the-fact excuse for not .doing simple random sampling.
15 There do exist applications in which one would 16 want non-probability sampling. The reasons for those are 17 interesting. I do not see any evidence of those reasons l
18 coming into play in a study like this.
19 As an instance of that, in a study, let's say, 20 relating average income to educational level in a small 21 subpopulation, it's clear that on average the wealthy people
) 22 have more of the income, and one might justify a non-ACE-FEDERAL REPORTERS, INC.
202-347 37to Natiimwide Coserage IWxk33MM6
. _ . ._ . _ _ . _ _ . . . _ . . . . _ _ _ _ _ _ . . _ , _ _ . . , _ . . _ _ _ _ _ _ . . . _ , , . . _ _ _ , . . . , _ , _ _ - . _ . _ . , _ , . . . . - _ . . , _ . _ _ . _ - . ~ . . . _ .
d 970 01 01 50 marysimons 1 probability sample on the arounds that these people, the 2 wealthy people are going to provide more information on the 3 question at hand, and therefore I will interview them 4 instead of taking a simple random sample. That is one 5 instance in which you would do non-probability sampling.
6 Another instance that's a tricky one is where you 7 make the model that says the relationship of the response to 549 8 some, let'sp demographic factor, is determined by a fixed ;f#
9 model that is the same for all people and, therefore, taking 10 a collection of some people, perhaps non-randomly, still 11 gives me a random sample of the statistical noise, and 12 therefore I still have a valid estimate. That is a very 13 difficult argument to pull off in most contexts.
14 BY MS. MONAGHAN:
15 0 Where would you want to use non-probability 16 sampling?
17 A Not here.
18 MR. MILLER: When you say "here," Dr. Simon ---
19 THE WITNESS: Not in the context of LILCO's 20 evacuation plan.
21 One context in which non-probability sampling is 22 used is a case in which you want to come up with a ratio ACE-FEDERAL REPORTERS, INC.
202 347 3?m Nationwide Coserage NXb 3EN4
.970 01 01 51 mirysimons 1 estimate. The standard auditing example of this is a case 2 in which you have capital "N" files auch as insurance 3 claims. Let's say each of the capital "N" files has a 4 stated dollar amount on it. You're estimating, for example, 5 the fraud ratio in these insurance claims, and the fraud l
6 ratio is determined by looking at the total true value 7 divided by the total stated value.
8 You know in advance that the large dollar claims 9 will have a disproportionate impact on the fraud value.
10 That is, if you are investigating somebody that claims the
, 11 theft of a $13,000 automobile, that has greater potential 12 fraud than somebody who claims the loss of an S80 camera.
13 You say well, I'll probably learn more about the fraud ratio 14 by taking a non-probability sample, and in particular I'm 15 just going to dive in and get all the big claims and check 16 them out.
17 You say well, I no longer have much chance of 18 investigating the guy who has been putting in fraudulent 19 $50, S70 or $80 claims, but you say that's small potatoes in 20 this problem and I'll go for the big stuff.
21 MS. MONAGHAN: Let's take about a 10-minute x
22 break.
ACE-FEDERAL REPORTERS, INC.
nw- ~_ a c- - ,, -
970 01 01 ' 5 2 -.
COrysimons'l; (Recess taken.)
2- MS. MONAGHAN: Back on the record.
3 BY MS. MONAGHAN: .( Resuming) 4 0 Dr. Simon, if there are a number of variables 5 that could affect a result, can you establish on a 6 statistical basis a causal link with one variable?
3 7 A That is a heavy-question. I could talk-about 8 that, but.I think I would have to need a structure s- 9 including, perhaps hypothetically, what is the response and 10 what things are in the collection of possible causative 7"3 11 . variables.
[ /
12 0 Let's look at Contention 21(e) as that example.
l 13 A Okay.
14 0 I think if you look at Contention-21(e) you'll i
f i . 15 see that there is a population of 165 traffic quides who man 16 130 traffic control points.
17 A Yes.
18 0 Of that population FEMA observed 32 traffic
!' 19 guides at 27 points, and those observations showed about 20 three things.
- - 21 One is that 8 took too long to arrive at their 22 traffic control points, that 1 of 14 interviewed knew the ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationskle Coserage 8 m 3 % 6646
97'0 ' 01 Ol' - 53 marysimons 1 location of the reception center and that radio 2 communications between the Port Jefferson staging area and 3 the traffic guides was sometimes poor.
4 A- That's page 32 of your version?
5 0 Right.
6 A Yes.
7 0 Now based on that array of actions that was taken
-8 and noted in the FEMA report and the sample size, could you 9 conclude that there would be a causal link with one !
10 variable?
r~N 11 A I need a little help. What are we talking about wi 12 as the cause and what is the effect?-
13 O Well, could you establish a cause and effect here 14 as to why their performance wasn' t perfect?
15 MR. MILLER: Could Dr. Simon establish from the 16 context of Contention 21(e) a cause and effect; is that the 17 question?
18 MS. MONAGHAN: From the facts laid out in 19 Contention 21(e) could Dr. Simon establish a cause and 20 effect for the behavior exhibited by the traffic guides that 21 were sampled?
) 22 THE WITNESS: I'm having a little trouble pinning l ACE-FEDERAL REPORTERS, INC.
202-347 37W Nationside Coscrage 80lk33& N M
970 01~01 54 marysimor's _1 this down. Let me ask, for instance, with regard to item 2 Roman numeral I that says "For instance, time between.
3 deployment from the staging area and arrival at TCPs was 4 excessive." Are you asking what caused that excessive time?
5' BY MS. MONAGHAN:
6 0 What I'm asking you is can you determine what the 7 cause was for them arriving late, or could there be multiple 8 causation for that?
9 A Oh, this document seems to say nothing about why 10 they got there. There may be further details in the FEMA 11 report, but I certainly don't recall them. This just notes 12 that they were late without any stories. So from this alone 13 I certainly don't know why some were late.
14 0 Okay. And if the FEMA report didn't illuminate 15 any further stories as to why they might have been late, 16 would you be able to determine what the cause of their 17 lateness was?
18 A Not with that piece of information. Presumably 19 one could undertake a grand investigation and query these 20 people as to why they were late and see if that clears up 21 the picture any.
22 O Now, Dr. Simon, if you look at the lateness of i ACE-FEDERAL REPORTERS, INC. r 202-347 373) Nationwkle Coserage WKA33M4*M
970'01 01 55 m2rysimons 1 some of the traffic guides and the fact that'one of the 2 traf fic guides of 14 interviewed didn' t know the location of 3 the reception center and that radio communication between-4 the Port Jefferson staging area and the traffic ~ guides was 5 sometimes poor, if you look at that collection of. events,.is f 6 there any causal link that one can' draw between those events 7 and the overall performance of the traffic guides?
I 8 MR. MILLER: I hate to this this, Ms. Monaghan, 9 but other than the fact that the question is rather complex, 10 the contention states that only one of the 14 knew the
~T 11 ' location of the reception center.
O 12 MS. MONAGHAN: If I misstated it, I'm sorry. I l
13 didn't mean to do that.
14 THE WITNESS: It's easy to do.
i
, 15' BY MS. MONAGHAN:
L 16 0 Based on the text of the contention, do you
~
l 17 understand the dif ferent events that we're dealing with?
i 18 A Well, let me see if I can rephrase the question.
I 19 Are you asking am I able to make any statements in a
, 20 statistical or quantitative sense between the shortcomings, I
21 such as late arrival, failure to know the reception center 1
22 and poor radio communication, on the impact of the ACE-FEDERAL REPORTERS, INC.
202 347-37(U Nationwide rmetage *n3%fM6 l__-___,__--_-___-________--_______--____-______------_____--____.
970 01 01 56
' - tcrysimons 1 evacuation; is that correct?
2 0 No. Let me try to rephrase the ouestion again.
3 Are you able to draw any conclusions about what 4 may have caused the failures that are listed here? Could 5 you find a single cause for the problems that are listed in 6 Contention 21(e)?
7 MR. MILLER: Based upon the facts alleged in the 8 contention.
9 MS. MONAGHAN: Based upon the facts alleged in 10 the contention.
11 THE WITNESS: A single thing that caused all b' %
12 those problems?
13 MS. MONAGHAN: Yes.
14 THE WITNESS: I would say no.
15 BY MS. MONAGHAN:
16 0 Let's go back to Contention 21, subpart (a).
17 Now as I understand it, Dr. Simon, the testimony 18 which you plan to offer on Contention 21 is that the size of 19 the sample taken by FEMA was too small for FEMA to reach the 20 conclusions in the FEMA report; is that right?
l 21 A Yes.
22 0 Is there anything further that you believe you I
ACE-FEDERAL REPORTERS, INC.
202 4474 700 Nationwide Coverage Nn))M686
970 01 01 57 m;rysimons 1 will be offering testimony on in Contention 21 or have I 2 accurately described what you are going to do?
3 A I believe you have correctly described it.
4 O And that would be true for all the subparts of 5 Contention 21?
6 A In each of the parts (a) through (f) there's an 7 indication that some number of people were looked at with 8 regard to various tasks, and our contention in each of these 9 parts (a) through ( f) is that the results obtained and the 10 number of people looked at do not permit the conclusions
] 11 that appear in the FEMA report.
<J 12 O Would you.look at page 36 of the document that 13 I've placed in front of you, and I guess it's the December 14 9th version of the contentions.
15 Now for legal reasons the Board has consolidated 16 certain portions of other contentions with Contention 21.
17 Page 25 is where Contention 21 starts.
18 MR. MILLER: Maybe we'll be able to save some 19 time here, Ms. Monaghan. Let me just represent that at the 20 present time it is not the intention of the County to offer 21 Dr. Simon on any of the other contentions or contention 22 subparts subsumed within Contention 21, that is Contentions ACE-FEDERAL REPORTERS, INC.
102 347-3710 Nationwide Coverage mn3%fM6
i 970 01,01 58 marysimons 1 15(1), 16(n), 22(k) or 32.
L 2 MS. MONAGHAN: All right.
- 3. - BY MS. MONAGHAN:
4 O Looking at Contention 21(a), Dr. Simon, if you 4
5- assume that all the facts as described in Contention 21(a) 6 are true, if all the persons sampled as described in 21(a) 7 had performed perfectly, could you ganeralize about the 8 performance of the rest of the population?.
9 MR. MILLER: I have to object as asked and 10 answered, but Dr. Simon, you may answer.
c 11 THE WITNESS: Well, the number investigated, the A/
12 number examined I believe is one at each of the three f
13 locations, and if three out of three performed well, you 14 have a very shaky inference to the population of 60.
15 BY MS. MONAGHAN:
1 16 O And so as an expert you wouldn't feel confident 17 in making that inferencer is that correct?
i 18 A I can make an inference based on observing three l
19 successes out of three. The inference, however, would be so 20 wide as to be useless.
21 O Look at Contention 21(b). Again assuming that
) 22 the facts in 21(b) are correct, if all the persons sampled, ACE-FEDERAL REPORTERS, INC.
202 347 3700 Nationwide roserage 8431MM6
- - - - .- . . . - ~ . - .
e 970[01 01. 59 mrysimons 1 as described in 21(b).had performed perfectly,'could you 2 - generalize about the performance of the rest of the 3 population?
4 A Contention 21(b) deals with the drivers of whom I 4
5 believe 8 were observed over the three staging areas.. If I 6 had regarded this problem as'a sample of 8 out of 333, and 7 if hypothetically 8 out of 8 had performed perfectly, I ,
- 8- would have an inference to the entire population of 333. It i 9 would be a rather wide inference, clearly not as wide as
- _10 only the 3 out of 3 case, but I believe it would be so wide f]- 11- as to not be helpful.
- O i 12 0 . What would the inference be that you would have 13 from the population?
i 14 A Again hypothetical and again I should emphasize ,
15 I'm not here doing arithmetic, my guess on pure conjecture 4
16 if I had seen 8 out of 8 out of this collection, probably I 17 would be able to say that I've got a 95 percent confidence 1
18 statement that the probability of a successful performance
! 19 is somewhere between 30 percent and 100 percent. That's j 20 what I mean by wide inference and, indeed, the penalty for a i~ 21 small sample size.
22 0 Let's look at 21(c), the same question.
c i
ACE-FEDERAL REPORTERS, INC.
202 347 370) Nationwide Coserage 80k33M686
. . , , - , - . , - . , - ~ - , . , . - _ _ . , _ - - . . . . - - . - . _ . - . . _ . . , - . , . . . _ . - _ _ , _ _ - . _ _ _ . . - . - - , _ . _ _ _ - - . - . . . - . , . . - - - - . . . -
2970 01 01 60 Omarysimons 1 A 21(c) was ---
2 MR. MILLER: Just for my benefit, Ms. Monaghan, 3 the same question ---
4 MS. MONAGHAN: The same question and I was just 5 going to repeat it.
6 BY MS. MONAGHAN:
7 0 The same question is if all the persons sampled 8 had performed perfectly could you generalize about the 9 performance of the rest of the population?
10 A In Contention 21(c) dealing with drivers, their 11 showing up and doing various things at school buildings, and 12 I think the sample sizes were 1 and 2, something like that, 13 a sample size of 1 that does successfully simply indicates 14 that the task can be done. I really can't take the 15 infereo ' beyond that.
16 That is when a heart transplant is done and the 17 patient is alive for two hours afterwards you have simply 18 demonstrated that it can be done. You've not demonstrated 19 that it can be done with large probability nor even that 20 it's a good thing. l 21 O In rendering your opinion on Contention 21(c) are 22 you going to be looking at the number of school districts in ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coscrage MG336-6646
Y 970 01 01 61 ccrysimons 1 the emergency planning zone which are described at the 2 beginning of that contention? Will that factor into the 3 opinion' that you're going to give on 21(c)?
4 A Only marginally. Your problem here is that only 5 one -- excuse me. I think there were two buses that were 6 sent to Shoreham Wading River High School and one bus sent 7 to the Longwood School, or the Ridge School from the 8 Longwood district. Examining two out of some number doesn't 9 tell you very much. Two out of two is a useful inference 10 and even two out of three is not enough to make a definitive
N 11 statement.
e) 12 So with regard to 21(c), the samples sizes are so 13 small that the issue of whether we are talking about 2 out O(~
, 14 of 11 tea 2.out of 15 or 2 out of some other number hardly 6 15 matters.
16 O What do you view as being the population of which 17 a sample was taken in Contention 21(c)?
18 A The population is not terribly well defined in 19 this contention. That is, this contention does not 20 specifically delineate what is the unit task, whether the 21 unit consists of the school or whether the unit consists of q
X_/ 22 the driver or whether the unit consists of a bus. So the ACE-FEDERAL REPORTERS, INC.
al2-347-3700 Nationwide Coserage 800-33M646
970_01 01 62 marysimons 1 framework in this contention would have to.be clarified in
~2 order'to make other kinds of calculations.
3 O' If we assume.that the population. unit that we are 4 looking at is the schools, in your opinion is the sample 5 size inadequate to make a generalization about the entire 6- population?
7 A Well, the answer is first yes, the sample size is 8 inadequate and here crushingly so if the analysis is the 9 school. There are 33 schools, one BOCES district, and I i ,
. presume'that means one BOCES school, actually I'm not sure, 10 11 two parochial school, some nursery schools and so on. There f
j 12 are many schools here and looking at one or two is clearly t
113 not enough.
!- 14 0 If we assume that the population that is being l 15 examined in this contention is the buses, is the sample size i
l-16 inadequate to make any generalizations about the population 17 as a whole if the population is buses?
18 A well, in Contention 21(c) if the unit of analysis i
- 19. is the bus, I do not have a population size and I do not 20 know how many buses would be involved in the proposed
!' 21 evacuation. Whatever that number is, I can only note that 22 it must be many and certainly more than 30, and a sample of ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6646
. _ _ . . . _ _ ~,,_ -. ._ , . _ ... . . ._, . _ . . _ _ . . _ _ , -
970 01 01 63
}go(
marysimons 1 one or two is clearly againpenough to make a good inference. d#'
2 O Would your answer be the same if the population 3 being surveyed were the drivers of school buses?
4 A Yes. The number of drivers is approximately 5 equal to the number of buses, and I'm not saying that to be 6 humorous. I mean there's a multiple routing question. So 7 again the answer is that the small number of drivers 8 actually observed is insufficient.
9 0 If you would look at Contention 21(d) and answer 10 the question what if all the persons sampled had performed 11 perfectly, could you generalize about the performance of the 12 rest of the population?
13 A Contention 21(d) is in regard to ambulances and 14 ambulettes and I think one of each was examined. Not 15 enough. Again, insufficient for the inference.
16 0 What about 21(e)?
17 A 21(e) was the traffic guides. There were I think 18 32 of them actually. Thirty-two were actually observed.
19 This is another one of these partial success situations. I 20 thought the number was -- yes, here it is. There were 32 21 quides actually examined in the exercise. Thirty-two is not 1
c _, 22 a large sample , but at least it's not laughingly trivial.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cos erage M10-3346M6
. 970 01 01 64 m rysimons 1 Whether 32 suf fices to make an inference depends 2 on how well you want them to do. If you want them to 3 perform at a 90 percent success rate, and that is if you 4 want to be sure that they're working at a 90 percent success 5 rate, that is not a sufficient sample. If you want them to 6 be working at a 50 percent success rate, it's still a small 7 sample, but one could make some modest statements about how 8 close they are to a 50 percent success rate.
9 0 What if you were looking at say a 75 percent 10 success rate, could you draw any conclusions or 11 generalizations?
12 A I could, but the arithmetic is a little ugly for 13 me to do off the top of my head. There would be an 14 inference available. It would not be a very crisp 15 inference, and by that I mean a confidence interval might 16 have a width of 20 or 30 percentage points.
17 0 Please look at 21(f), and the question is if the 18 persons sampled had performed perfectly, could you 19 generalize about the performance of the rest of the 20 population?
21 A 21(f) was the traffic impediments. Unlike the
./ 22 previous contentions, the impediments do not represent a ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cos erage 800-336-(ww
970 01 01 65 marysimons 1 sample from a finite population. That is, in no sense could 2 there possibly exist a roster of accidents. The accidents 3 are typicnlly generated by a Poisson phenomenon, and how a many of those one would actually encounter in a real 5 evacuation is hard to know.
6 I don' t know how many traf fic impediments occur 7 in this area on a regular basis. That is, you do not need 8 an emergency at a nuclear power plant to produce traffic 9 impediments. I don't know what's a reasonable number to 10 conjecture. In this contention we say that two of them were 11 actually simulated and the responses were not quite 12 successful. That's certainly a cause for concern, but what 13 statistical issues should be addressed perhaps could be 14 formulated a little more crisply.
15 0 How would you go about determining how many 16 accidents would occur in a real evacuation?
17 MR. MILLER: How would Dr. Simon do that?
18 MS. MONAGHAN: Yes.
19 THE WITNESS: Well, I think that's outside my 20 role in this. I can make some conjectures about how that 21 might be done. I presume there exists data on the
() 22 occurrence, frequency and locations of traffic accidents and ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cm erage mB13M646
2970 01 01 66 Omarysimons 1 impediments in this area. One might make the assumption 2 that these occur at vulnerable spots in the transportation 3 network. That's where they usually occur.
4 thereafter the best thing that I could recommend 5 would be that we treat the accident probability as a 6 variable parameter and imagine simulating this thing with 7 different levels of that parameter.
8 That may be a little evasive, but indeed it is 9 hard to pin down. An evacuation that comes on a cloudless, 10 warm day is probably going to proceed more smoothly than an 11 evacuation that occurs in a 33-degree rain storm. So the 12 relevance of impediments in this evacuation is very 13 difficult to get ahold of.
14 BY MS. MONAGHAN:
15 0 Dr. Simon, if you knew that on the day that the 16 exercise was conducted that within the emergency planning 17 zone that there actually had occurred six accidents that 18 day, would the simulation of two accidents in the course of 19 the exercise give you a sufficient sample size to draw any 20 conclusions if you understood that the general population 21 that was experienced on that day was six in the real world?
) 22 A Again, this is hypothetical. You observed in ACE-FEDERAL REPORTERS, INC.
" - " - > " ~ " -
==?E_
f 2970 01 01 67 Omarysimons 1 this exerci.se two simulated impediments. It's not clear g# YN% )
2 what the quantitative.aveeestion of the individual 3 impediments involves. That is, this is not quite so simple 4 as a success or failure situation. An impediment that, 5 let's say, ties up three lanes of the Long Island Expressway 6 has got to be more tragic tha let's say an impediment on 7
- (b say a Wading River er Yaphank road or something. #
8 So what I see here is hypothetically well, we 9 checked two and there might be six. What one needs perhaps 10 is some sort of quantitative assessment of the impact of the 11 impediments that actually did occur in the simulation that 12 might occur hypothetically in a real situation.
13 I'm sorry for the vagueness of this answer. It's 14 just that the consequence of an impediment is a fairly 15 complicated notion and I don't see a great deal of 16 quantitative summary about what already happened or what was 17 covered in the FEMA report.
18 I really don' t see a useful operations analysis 19 or network analysis or any kind of operational plan putting 20 these pieces together. That's why I'm having trouble 21 formulating anything resembling a workable answer to this.
22 O What do you mean by an operations analysis?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Co rage 8tB 336 4 16
970 01 01 68 marysimons 1 A Well, that at some point somebody should have a 2 flow diagram 'i the steps involved in an evacuation and 3 consider the consequences of failures or impediments at 4 various steps of those operations.
5 0 What do you mean by a network analysis?
6 A I'm using the word casually. If other people 7 have a jargon meaning for network analysis, I'm not 8 necessarily implying that. I'm implying a diagram where 9 there are boxes with arrows leading from one box to another 10 and describing pathways or procedures or things that have to 11 be done in order for an evacuation to occur. Individual
<{)
12 boxes might represent, for instance, roadways and one would 13 have to consider the consequence of roadways being out of 14 commission.
15 0 So before you could do any kind of statistical 16 analysis about Contention 21( f) you would need to have some 17 sort of qualitative analysis done prior to that of what the 18 significance of the particular simulated impediments had 19 versus the significance of the kinds of accidents that had 20 actually occurred on that day; is that correct?
21 A Yes, and a network analysis is vastly more
() 22 complicated than simple statistical statements. An ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
970 01 01 69 marysimons 1 evacuation depends on so many subparts, and many of the -
2 si:bparts are very much unlike any others and the role of .
3 statistics gets down-played a bit.
4 You can say, oh, let's talk about the probability 5 of having an accident on the Long Island Expressway that ,
I 6 people cannot get around. Well, that's interesting and 7 amusing, but the consequences of that are terribly variable 8 and I'm not really sure how that could be anticipated in a 9 simulated experiment.
10 0 Dr. Simon, do you think that you will be offering 11 testimany on Contention 21(f)?
12 A It's possible. As I mentioned, the role of 13 impediments is very puzzling and it's a very murky 14 question. Two were tried and the outcome was not terribly 15 successful. That's certainly ominous and it's difficult to 16 put that into an overall statistical framework.
17 0 Is it possible to put it into an overall 18 statistical framework?
19 A Statisticians and other technical analysts are 20 completely fearless about modeling real world phenomena.
21 They are also completely unrestrained in attacking other
( 22 people's models of them because often these models deal with ACE-FEDERAL REPORTERS, INC.
202-347-3700 Natio . wide Coserage 800-336 f646
/
l hgj(4 70 h e
al9700101 mnrysimons 1 various conjecture versions of reality, and even >utside the '
/
2 litigious context it's hard to agree on what's ap ropriate 3 to put into these.
4 O So, in other words, if you were to construct a 5 model it would not necessarily be a model that everyone 6 would agree was the appropriate way in which to model a 7 situation? There could be variance of opinion, in other 8 words, about the validity of the model?
I 9 A I think that's a crushing understatement.
10 0 I would just like to ask you a few auestions 11 about your resume, if we could just go back to that quickly.
~}
12 In connection with your work as a consultant to 13 the Gallup Organization what did you do with them?
14 A Gee, that was 1977 to 1978. You're straining my 15 recollection. I believe that I did an analysis of fuel use 16 data. My recollection of that work was that people in the 17 Gallup Organization took a national based poll. The object 18 there was to obtain answers on a questionnaire regarding 19 fuel use habits and things like turning thermostats up and 20 down and putting in insulation and stuff like that and all l
i 21 those other hot topics in 1977. I believe that was the l
'^ , l
-) 22 substance of that. l ACE-FEDERAL REPORTERS, INC.
202-347-370t) Nationwide Ce s erage 8 5 336-(646
(-
t 970 01.01 71 marysimons 1 O Is that work in any way analogous to the kinds of 2 tasks you will be performing in connection with rendering an 3 opinion in-this case? If you recall.
4 A My guess is no because they were talking about-5 fairly large samples, several hundred respondents.
6 0 What do you do in connection with your work as a 7 consultant to the Research Testing Laboratories in Little 8 Nick, Queens?
9 A That's product testing of consumer products.
10 That stuff is more in the context of controlled experiments
('T 11 than with regard to sampling issues. For instance, the last V
12 several projects that I've done involved the comparison of 13 two products for relieving diaper rash and a comparison of 14 three dif ferent products to relieve symptoms of 15 dysmenorrhea, and in the past it has been things like 16 dandruff shampoo, underarm deodorants and hand creams and 17 all that delightful stuff which is somewhat peripheral to 18 what we're doing here.
19 0 Now when you worked as a consultant to the 20 Corporation Counsel of the City of New York, was that in 21 connection with the child care case that you described q
(_/ 22 earlier?
ACE-FEDERAL REPORTERS, INC. I 202-347-3700 Nationwide Coverage 800-336-6646
, l
[
970 01 01 72 marysimons 1 A Yes.
2 O Was there any other work that you did for the 7
3 Corporation Counsel since then?
~
4 A No. It was all in the Wilder versus Bernstein 5 case. ,
t 6 0 I also note that you are currently a consultant 7 to the Marketing Section of CITICORP. What do you do in 8 connection with that work? -
9 A The projects that I've done for them are well, ,
10 desling and analyzing results from various surveys of use of 11 credit instruments, credit card purchases.
12 The particular interest at CITICORP, well like 13 the other bank credit card companies, is finding new ways t 14 u Se-get people to make,. surf of their credit cards and finding J#l 15 new ways to target people in direct mail campaigns and so 16 on.
17 Statistical caution is needed there because the 18 credit card game has saturated -- I think it has saturated ,
19 all of America and in the process of doing new business you 20 can only do it by dipping further down in to the less 21 creditworthy segment of the population. So there are a lot 22 of interesting statistical questions that come up in that ACE-FEDERAL REPORTERS, INC.
202 147-3700 Natsonwide Coserage Mn336-6M6
l 970 01 01 73 marysimons 1 kind of work.
2 O Do you consider that work at all analogous to the 3 kinds of things you're going to be doing in this case?
4 A I would say no.
5 0 What have you done as a consultant for Consumers 6 Union?
7 A They have a number of sources of data. One of 8 their more interesting ones is the survey of their own
+
9 subscribers. These people will fill out questionnaires 10 related to various consumer products that they own, 11 including the performances of those products. The questions 12 come down to identifying good products, bad products and 13 dealing with repair rates of various kinds of appliances.
14 The statistical questions of kind of interesting 15 because the sets of products out there in people's 16 households with a particular brand name have various age 17 profiles, and since of course old machinery needs to be 18 repaired more often than new machinery, there are lot of 19 technical questions about how one should make some of these 20 comparisons. I think that is the essence of what I do for 21 them.
) 22 O And would any of the work that you do for them be ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 8(Xk336-6M6
2970 01 01 74 Omarysimons 1 analogous to what you are going to do here?
2 A Well, it's always a case of inferences from .
3 samples to population, and in those cases they are dealing 4 with substantial samples. So I don't sec it as a direct
~
5 parallel.
~
q 6 0 What about our work with Meissner, Tisch and 7 Kleinberg?
8 A Meissner, Tisch and Kleinberg represented a fuel 9 oil dealer who was sued by the City of New York for short LJ 10 delivery of fuel oil. His fuel oil company had a contract 11 with the City to delivery oil to many of the city-owned 12 buildings. The City claimed that he was under-delivering, 13 for example, saying that a delivery of 1,500 gallons 14 occurred when in fact a delivery of 1,000 gallons occurred.
15 My work in that case was related to a statistical 16 analysis of billings data, that is, numbers that the City 17 claimed to have been actually delivered versus various 18 billing data in a number of buildings serviced by this 19 vendor and also by other vendors.
20 0 Have you ever engaged in any non-probability 21 sampling analysis in connection with any of your consulting 22 work?
ACE-FEDERAL REPORTERS, INC.
- - - *~- - " -
=rE - -- -
/ .
970 01 01 75 marysimons 1 A I occasionally, I'm sure, must have come up with 2 bad samples that people are anxious to save. For instance, ,
~
3 in this Meissner, Tisch and Kleinberg case the City of New 4 York selected some buildings to audit. They did this 5 apparently by a non-probability method, which means in this ,
6 case that they just made them up. They targeted a number of .
7 buildings to investigate.
8 What they did in those cases was to take readings 9 on the fuel oil tanks using a dip stick method and then it 10 was sort of a cat and mouse game. They w ited until a day >
11 SMoS8 on which a fuel oil delivery was atu puse to occur and they gr
~
12 took dip stick readings after the fuel oil truck had left 13 and they had made an attempt then to estimate the amount of 14 short delivery of fuel oil. At the end they did not rely on 15 that data base. They relied on a data base based on billing 16 information in other buildings.
17 O Did you draw any conclusions on the basis of that 18 ton-probability sampling or did the point become moot?
19 A Well, there are at least two answers to that. If 20 you pick out an individual building, and assuming that the 21 dipping process gives correct readings, and if that 22 individual observation suggests that on a particular day 500 ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage
- Vk 336-6M6
l 970 01 01 76 marysimor gallons of fuel were paid for but'not delivered, it is i
_ certainly reasonable to conclude that a 500-gallon fraud .
~~
3 occurred on that day at that time.
4 People didn't argue about that. The question was -
5 to what extent does this generalize the entire system i
6 involving in this case thousands of buildings, and nobody 7 was willing to make that inference. We have to concede 8 that, gee, the right amount of oil probably did not get '
9 delivered to those buildings on those days. 4end that was a er Wkh V 10 general fact of the system was not something that anybody 11 was willing to get involved in.
)
12 There is one post-script to that. The fuel oil 13 dealer is bankrupt. So some of these issues became moot.
14 0 Were joined and were in fact litigated?
15 A It was a fairly complicatJd suit. They were 16 dragging RICO charges into it and everything. So the i 17 statistical work in that was only one aspect of many.
18 0 Did you offer testimony in that case at all?
19 A It never got to that.
20 O I note that one of your publications entitled 21 " Chance and Consensus in Peer Review" published in Science
( 22 was co-authored by J. Cole and S. Cole. Is that Stephen ACE-FEDERAL REPORTERS, INC.
3 202-347 3700 Nationwide Coserage &n336-6M6
970 01 01 77 marysimons 1 Cole?
2 A That's Stephen Cole whom I believe you saw .
3 yesterday, the same one.
4 O Dr. Simon, at this time do you understand that 5 you will not be offering testimony on Contention 50; is that 6 correct?
7 A That's my understanding of the way things are at 8 the moment.
l 9 0 Have you reviewed Contention 50 at all? l 10 A No.
11 O Did you come to Washington just because you were 12 going to give this deposition today?
13 A Is that a fair question?
14 MR. MILLER: Well, that's completely irrelevant, 15 but you can answer it.
16 THE WITNESS: That answer would be yes.
17 BY MS. MONAGHAN:
l 18 0 would it have been equally convenient for you to [
19 have the deposition in midtown Manhattan today as to have 20 come down here?
21 MR. MILLER: That's also completely irrelevant.
( 22 THE WITNESS: That is also irrelevant, but in ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coscrage 8tn33M646
970 01 01 78 murysimons 1 case it happens again, let me note that that would have been 2 even more convenient for me.
3 MS. MONAGHAN: Okay, fine. I have no further 4 questions.
5 MR. SAHNLEUTER: I have no questions, but I do 6 vant to state for the record that I did not meet with Dr.
7 Simon yesterday, and in fact I was at the depositions of Dr.
8 Perrow and Mr. Cole yesterday.
9 MR. PIRFO: Dr. Simon, I just have a few 10 questions.
/^'s 11 EXAMINATION Q
12 BY MR. PIRFO:
13 0 Your office I assume is at the NYU right in 14 downtown Manhattan?
15 A NYU has two offices. The Graduate School of 16 Business Administration has an office at 90 and 100 Trinity 17 Place near the World Trade Center. They also run an 18 undergraduate business program at the Washington Square 19 Campus on Fourth Street, and I spend most of my time at West 20 Fourth Street.
21 0 Do you live in Manhattan?
(-
x_/ 22 A I live in Stony Brook.
1 ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 800-33&6M6
970 01 01 79 mnrysimons 1 0 In Stony Brook. So you commute every day?
2 A Three days a week typically.
3 0 Do you live with your wife and four children in 4 Stony Brook?
5 A Yes.
6 0 And you've been there for 10 years?
7 A I moved to Stony Brook in 1975.
8 0 Did you attend the limited public appearance 9 statements that were held in any of the locations on Long 10 Island?
11 A Am I correct in assuming that these are some of 12 the hearings related to the Shoreham proceeding?
13 0 Yes.
14 A No, none of them.
15 0 Did you attend any Shoreham hearings?
16 A No.
17 0 Are you familiar with the organization of 18 Shoreham Opponents Coalition?
19 A I know there exist a number of organized Shoreham 20 opponents. I would feel very insecure about pinning down 21 the exact words that appear in any of their names.
_ 22 0 You are not a member of any of those ACE-FEDERAL REPORTERS, INC.
-- x .c _ m _ ,. - i
_ - - _ - - - - - - _ - - - _ - _ - - - - - \
970 01 01 80 marysimons 1 organizations?
2 A I am not a member of any of them.
3 0 Have you written letters to the editor of any of 4 the local newspapers up there with regard to the opening of 5 Shoreham? ,
t
~~
6 MR. MILLER: Let me just, and you can continue,
~
7 Mr. Pirfo, after I make my statement.
8 MR. PIRFO: I just have one more.
9 MR. MILLER: I've got to say these are just 10 irrelevant questions and they don' t bear at all on Dr.
11 Simon's ability to testify in this case, but Dr. Simon can 12 answer the questions.
13 MR. PIRFO: They go to his personal bias, if any 14 exists , which I don' t know, we are in discovery, and I'm 15 entitled to inquire into that when you put on an expert.
16 MR. MILLER: Go ahead, Dr. Simon.
I 17 THE WITNESS: I am not a member of any 18 organization opposed to the opening of the Shoreham power 19 plant, and I have never written a letter to the editor of 20 any publication about the Shoreham power plant. I am not 21 even necessarily an opponent of nuclear power.
) 22 BY MR. PIRFO: "
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage m)-336-6M6
/
970 01 01 81 marysimons 1 0 So is it fair to say that you have not formed an 2 opinion with regard to whether Shoreham should operate?
3 A That's a broad question because it gathers in 4 questions of the safety of the plant, the appropriateness of 5 the evacuation and one can have slightly different feelings 6 about those parts of the problem.
7 O So you have formed an opinion as to whether 8 Shoreham should operate?
9 A No.
10 0 The answer is you at this point have not formed 11 an opinion as to whether Shoreham should opergte?
'd 12 A I would say that is correctf . (D 13 0 With regard to Contention 21, you made the 14 statement that the FEMA inferences are inappropriate and 15 uncalled for. Would you just define that a little bit more 16 clearly for me, inappropriate and uncalled for and in what 17 context?
18 A There are inferences throughout the report of the 19 form that a small number of people were observed and some 20 successes and failures were noted, and then the follow-up 21 statement becomes that this provides a reasonable level of 22 assurance. I don' t take any objection to FEMA's ability to ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-33M646
2970 01 01 82 Omarysimons 1 record or note information. They seem to be reasonably 2 forthright about noting good parts and bad parts of the .
3 evacuation drill.
4 My only concern is that having s'een small samples 5 with less than total success that the follow-up statement is
~~
6 reasonable assurance.
7 O Then when you think in terms of reasonable 8 assurance, does that bring to mind as a statistical expert a 9 particular level of confidence? '
j""
6 10 A The answer to that is that it would and I would 11 have to search my soul for exactly what I think the 12 appropriate numbers are if it were close. That is, if the 13 numbers that came out of the evacuation drill were close to 14 giving a good picture of the feasibility of an evacuation 15 drill, then I would have to turn over in my soul whether 95 16 percent is a good number or whether 98 percent or 99 percent l' is a good number.
?
18 The actual results here are so far from that that 19 the precise numeric requirements don't quite come into play.
20 0 Given that the value judgment is reasonable 21 assurance, would a 75 percent level of confidence be
() 22 adequate?
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 80tk336-6646
1 l
l l
970 01 01 83 marysimons 1 A Again, to an auditor reasonable assurance would 2 have certain numeric values in a certain context, and here I ,
~
3 don't know whether we are talking about getting 75 percent 4 of the people out means reasonable assurance or whether we 5 are talking about 75 percent of the players in the ,
~~
6 evacuation drill doing the right thing or that 75 percent of
~
7 the network paths get out. So I have trouble pinning down 8 whatever was meant by that term. _ ,,
9 My previous remark, namely, that the study that O 10 they actually did is very far from anything close enough to 11 enter into debate is what bothers me. Maybe 75 percent is
)
12 plausible for some aspects of it, but I don' t think we are 13 close enough to results that would actually cause me to call 14 up the question as to whether 75 percent is the right 15 number.
16 0 I'm sure you are familiar with the term or 17 procedure of a "T" test?
18 A I have done a few thousand.
19 0 Could you explain to me the applicability, if 20 any, in the analysis of Contention 21?
21 A "T" tests are used on measured data rather than 22 counted data. In Contention 21 it appears that things were ACE-FEDERAL REPORTERS, INC.
my s m _ m c _ ,. -
~ - - - .
970 01 01 84 marysimons 1 counted up as successes of failures and therefore. I don' t 2 see the direct relevance of a "T" test.
l 3 0 Does not a "T" test indicate to the reviewer, the 4 analyzer or whatever that what is happening is not happening 5 as a result of chance as opposed to something else at work 6 in the universe or the sample?
7 A All statistical tests ask whether or not a result 8 is attributable to chance. The exact. framework in which.
9 these are done are complicated. In the case of categorical
- 10 responses, that is success and failure, people would 11 typically compare an observed fraction to a target fraction
- 12 probably using a chi-squared test. To apply a chi-squared 13 test in a situation like this I would regard as disingenuous 14 in the extreme.
15 A game that could be placed, for example, is you 16 say well, I want to show that my resulto are not i 17 significantly different from a 75-percent success rate. A 18 small sample size makes it very dif ficult to detect and it 19 makes it very difficult to reject a hypothesis that 75 20 percent is not the true rate. This is a ruse I think known 21 to most statisticians, and the possibility for doing that n
I
(_) 22 certainly exists here.
ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 800-33M646
,-~.-. . . , , . _ , , , - , - . ,, , . _ - - - - , - , _ , . - . - - . - , - - - , . - _ , - - - . - - . -- -- _ - _ _
970 01 01 85 marysimons 1 0 So you don't see that tying those together, the 2 "T" test and the chi-squared, that either one of those would 3 be applicable in the context of an analogy to Contention 21?
4 A Statistical tests with small sample sizes are not 5 recommended.
6 O But are you making an assumption there that the 7 enmple size is too small to make the test or given the small 8 sample size can you perform a "T" test or a chi-squared test 9 on it?
10 A Let me take it out of context. Suppose you 11 wanted to buy a box of Christmas tree balls for your home 12 Christmas tree and suppose they came in crates of 48 and you 13 knew that you needed 36 to decorate your tree. Since the 14 box was offered to you at a very low price you're suspicious r 15 that maybe some of the balls inside are were actually 16 broken.
17 If you performed a test which said let's pull out 18 two of the balls for a random inspection, even if the two 19 balls were broken, you might not be able to reject the 20 hypothesis that there are 36 good balls.
21 This would be an example of performing an
( 22 experiment that is too small to make the inference that you ACE-FEDERAL REPORTE" INC.
202-347-3700 Nationwide Cos erage Ol> 336-M
970 01 01 86 mnrysimons 1 need, that is if your null hypothesis, and pardon the 2 jargon, is that 36 of the balls or more are okay, an 3 experiment consisting of two draws out of that box, even if 4 the experiment is completely unfavorable, might still not be 5 able to lead you to conclude that the hypothesis is wrong.
6 Therefore, this is a case in which a sample size 7 is so small that it is not able to discriminate between the 8 statement 36 good balls and some other statement.
9 0 I attempted to set up as a given that the sample 10 was big enough, but what you're saying is that you're not 11 willing to make that assumption, that given that assumption a
12 anything that follows from that is meaningless?
13 A If we're making an hypothesis that some 14 population is 75 percent good, taking a sample of size 6 in 15 order to investigate that experiment is I am claiming a 16 disingenuous experiment in that experiments of size 6 do not 17 furnish sufficient power to detect alternatives.
18 For example, if the conjecture, and call it null 19 hypothesis, were that something is 75 percent successful and 20 you then say I'll take a sample of size six, my guess is 21 that even if your sample of size 6 turned out to be 1 good I r-l m, 22 and 5 bad, that in the formalism of hypothesis testing 1 l ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cos erage 800 336-6M6
970'01~01 1 87 marysimons 1 good and 5 bad.is not sufficient to reject the hypothesis of-2 2 75 percent good because we are playing with games and toying.
S with statistical methodology. That is, a sample with 17 4 percent good on a very small ' sample size might be 5 insufficient for-you to conclude that 75 percent is wrong.
6 I'm sorry for that --- p 7 MR. PIRFO: I'll wait until I get the transcript.
8 (Laughter.)
9 THE WITNESS: That's how it goes.
10 .MR. MILLER: Let's take a real quick break.
11 MS. MONAGHAN: Certainly. 4 d('
12 Off the record.
13 ,
(Brief recess taken.)
14 MS. MONAGHAN: Back on the record.
15 MR. MILLER: I have no questions.
16 (Whereapon, at 11:50 a.m., the deposition of Gary
- 17. Simon concluded.)
i 18 i ******
19 . j 20 21 n
V 22 ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Co erage 800-336 6646
l 970 01 01 88 m:rysimons 1 * *****
2 I have read the foregoing pages 3 [ through , inclusive, which 4 contain a correct transcript of 5 the answers made by me to the 6 questions therein recorded.
7 Signature is subject to l 8 corrections. '2 cci , L%
9 GARY A. SIMON /
10 * * ****
11 I, , Notary Public in and 12 for the '
, do hereby certify that I am 7'
13 notarizing and witnessing signature f r the Deposition of 14 GARY A. SIMON on this /3 day of ,
19 [/ .
15 Notary Public in and for the 16 My Commission expires ggs. , 8d /f/7 f 17 * ** * *
- RUTH . WDCK 18 Notary PuDikintheStateof NewYork Residing in SuMolk Ca Suffolk County Clerks #S2 45053 Gemission Enpires, Mar. 3019.
19 20 21 s
22 ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationside Coverage 80 4 336 6646
89 P
1 CERTIFICATE OF NOTARY PUBLIC 2
3 I, MARY C. SIMONS, the officer before whom the 4 feregoing deposition was taken, do hereby certify that the 5 witness whose testimony appears in the foregoing deposition
.6 was duly sworn by me; that the testimony of,said witness 7 . was taken by me in stenomask to the best of my ability and 8 thereafter reduced to word processing by me, that said
['
9 -deposition is a true record of the testimony given by said L 10 witness; that I am neither counsel for, related to, nor
.11 employed by any of the parties to the action in which this
}
12 deposition was taken; and further that I am not a relative 13 or employee of any attorney or counsel employed by the 14 parties thereto, nor financially or otherwise interested in 15 the outcome of the action.
16 17 -W/ && 2 18 Mary C. Simons l 19 Notary Public in and for the l
20 District of Columbia 21 My Commission expires 22 August 15, 1989
=3 lO
, , .- . .- . . -._ _ _ .__,,_.-.. ,._.-._-_... ._ ,-.~. _ ... ,..- __ _ . _ . _ . , _ _, ,,,__... .,,,, .. , _ _,,,_
" DEPOSITION EXHIE47 N/ W o/t /
Sary Sloon "d S O Personal Resume November 1986 Education BS Mathematics, Csenegie-Mellon University, Pittsburgh, Pennsylvania, 1966.
PhD Statistles, Stanford University, Palo Alto, California, 1972.
Positions: United States Peace Corps, worked with Department of Applied Mathematics, University of the Philippines, Los Banos, September 1%6 to August 1968.
Assistant Professor, Department of Statistics, Princeton University, Princeton, New Jersey, l September 1971 to August 1975.
Associate Professor, Department of Statistics, State University of New York, Stony Brook, New York, September 1975 to August 1981.
Associate Professor, Department of Statistics and Operations Research, Graduate School of Business f Administration, New York University, New York, New York, September 1981 to date.
l Miscellaneous Proofreader for the Wiley Series in Statistics
{ through Beatrice Shube, Wiley-Interscience, New York.
Associate Editor, Journal of the American Statistical Association, January 1979 to December 1983.
- Consultant to Gallup Organizat3on through Paul Perry, Gallup Organization, Princeton, New Jersey.
1977-1976.
1 Consultant to Committee on Science and Public Policy (COSPUP), National Academy of Sciences, 1979-1980. - #
Recipient of Chancellor's Award for Excellence an l Teaching (in the SUNY system), May 1979.
Lecturer in the Chautauqua short-course program, 1979-1980 and 1980-1981.
\> Consultant to Research Testing Laboratories, Little Neck, Queens, New York, 1981 to date.
Consultant to Corporation Counsel, City of New
] York, March 1983 to April 1984.
Consultant to marketing section, CITICORP, 575 Lenington Avenue, New York, November 1943 to date.
Consultant to Consumers Union, Mt. Vernon, New York, 1905 to date.
. Coneuttant to Meissner, Tisch, and Kleinberg, 275 Madison Avenue, New York, 1986.
f Personal Born 24 APRIL 1945, Wilkes-Barre, Pennsylvania.
Married, four children.
l BOOK REVIEWS l
Basic Concepts in Information Theory and Statistics, by A. M.
Mathat and P.M. Rathie, reviewed in Technometrics, vol 18, 1976, p 502.
R. A. Fisher, the Life of a Scient'ist, by Joan Fisher Box, reviewed in SIAM News, vol 11, #5, October 1978, pp 5-6.
s Statistics, by D. Freedman, R. Pisani, and R. Purves, reviewed in Journal of the American Statistical Association, vol 74, 1979, pp 927-928. (This review was subsequently incorporated into the publisher's advertising. )
The Analysis of Cross-Tabulated Data, by Graham J. G. Upton, reviewed in Technometrics, vol 21, 1979, pp 581-582.
j Measures of Association for Cross-Classifications, by Leo A.
I Goodman and William H. Kruskal, reviewed in Journal of the American Statistical Association, vol 75, 1980, p 1036.
Beginning Statistics with Data Analysis, by Moste11er, Rienberg, and Rourke; Modern Elementary Statistics, 6th edition, John E.
Freund, reviewed in Journal o f. the American Statistical Association, vol 80, 1985, p 240.
Statistical Sampling for Accountants, by T. M. F. Smith, and e l Practical Statistical Sampling for Auditors, by Arthur J.
Wilburn, (Joint review of two books) reviewed in Journal of the Arnerican Statistical Association, vol 80, 1985, p 1078.
.O l
l
-- - _ _ _ , _ - - , . _ . _ . . - - _ . _ _ _ _ _ . - . . - _ - - ~__ _ . . - . _ _ _ _ . . . .-- -
f, -
~
)< MALICATIONS q~ Additivity of Information in Exponential Family Probability Laws, V Journal of the herican Statistical Association, vol 64, 1973, pp 478-448.
Seinful Employment Following Head Injury, Prognostic Factors, Archives of Neurology, y 29, 1973, pp 111-116, with co-authors A.
Dresser and four others.
Alternative Analyses for the singly-Ordered Contingency Table, Journal of the American Statistical Association, vol 69, 1974, pp 971-976. .
A Nonparametric Test of Total Independence Based On Kendall's Tau, Biometrika, vol 74, 1977, pp 277-282.
l Computer Simulation Swindles, with Application to Estimates of Location and Dispersion, Applied Statistics, vol 25, 1976, pp 266-274.
Multivariate Generalizations of Kendall's Tau, with Application to Data Reduction, Journal of the American Statistical Association, vol 72, 1977, pp 367-376.
Comments on paper by Schruben and Margolin, Journal of the American Statistical Association, vol 73, 1978, pp 520-521.
Efficacies of Measures of Association for Ordinal Contingency Tables, Journal of the American Statistical Association, vol 73,
^
1978, pp 545-551.
l
' The Sign Test when Ties are Present: THe Problem of Confidence Intervals, American Statistican, vol 33, 1979, pp 140-142, with co-author John Emerson.
The Probability of Primary Tooth Exfoliation: A Reference Source for Programs that Treat Children during their Mixed Dentition, Journal of Public Health Dentistry, vol 41, #4, Fall 1981, pp 236-243, with co-authors L. Ripa and three others.
Chance and Consensus an Peer Review, Science, vol 214, 20 NOV 81, pp 881-886, with co-authors J. Cole and S. Cole.
A General Expression for the Vartance-Covartance Matrix of Estimates of Gene Frequency: the Effects of Departures from e Hardy-Weinberg Equilibrium, Annals of Human Genettes, vol 48, 1984, pp 283-286, with co-author Nancy R. Mende11.
Diagnostics for Missing Data in Least Squares Regression, Journal of the American Statistical Association, vol 81, 1986, pp 501-509, with co-author Jeffrey Simonoff.
__. _. . _ _ _ _ _ _ __._ _ _ . . _-