ML20214G599
| ML20214G599 | |
| Person / Time | |
|---|---|
| Site: | Shoreham File:Long Island Lighting Company icon.png |
| Issue date: | 05/18/1987 |
| From: | Mileti D LONG ISLAND LIGHTING CO. |
| To: | |
| Shared Package | |
| ML20214G546 | List: |
| References | |
| OL-5, NUDOCS 8705270117 | |
| Download: ML20214G599 (7) | |
Text
__
~
9 s
LILCO, M3y 18, 1987
_O i
9 00LMETED i
UNITED STATES OF AMERICA US M NUCLEAR REGULATORY COMMISSION
87 liAY 21 P6 :25 Before the Atomic Safety and Licensino Board 0FFIC4: Or N
. Us 00ChLIIE A W /fCf.
BRANCH In the Matter of
)
)
LONG ISLAND LIGHTING COMPANY
)
Docket No. 50-322-OL-5
)
(EP Exercise)
(Shoreham Nuclear Power Station, )
Unit 1)
)
SUPPLEMENTAL TESTIMONY OF DENNIS S. MILETI ON CONTENTION EX 21 1.
Q.
Dr. Mileti, what, in your view, is the relevant unit of analysis for judging the performance of LERO?.
A.
The LERO emergency response organization has multiple goals and objectives which must be met *were it to re-spond to an actual emergency (these goals and objectives are numerous and listed at length in many places, for example, NUREG-0654).
Some of these goals and objec-tives are scheduled to be met by a relatively few people (for example, event classification); others are to be serviced by relatively many people (for example, traffic control).
Additionally, goals and objectives in organi-zations can often be met in multiple ways and through I
redundant (for example, independent and identical) or diverse (for example, functionally equivalent although not identical) backup mechanisms, or both.
These orga-nizational mechanisms provide flexibility that is de-signed to assure that organizational goals and 8705270117 870518 PDR ADOCK 05000322 PDR T
l l
m 3.
objectives are~ met even if some individuals fail at achieving their assignments.
The relevant unit of anal-ysis for emergency response, therefore, must be the or-ganization since the organization could achieve its goals and objectives despite some failures by individu-als.
Additionally, it is also possible for individuals to be successful at their unique jobs, without the orga-nization's achieving its general goals or objectives.
The evaluation of any emergency response organization, including LERO, must therefore be performed at the orga-nizational level since it is only at this level (and not the individual level) that a determination can be made regarding the effectiveness of meeting emergency goals and objectives.
Consider, for example, the objective of detecting road impediments in an evacuation.
Redundant and di-verse mechanisms exist in LERO for detection to occur.
These include detection by route spotters, bus drivers, traffic guides, the public, the news media and others.
If a bus driver did not report an impediment but a route spotter did then the objective would be met.
In this example, there would be individual " failure," but an or-ganizational " success."
None of this suggests that a plan is the only basis for evaluating emergency response effectiveness.
Exer-cises are useful in judging if normative systems (the
.--.--.--..,,....n.
w...
7 M
pl's.da)'can be translated in adequate performance systems (behavior).
However, what is relevant to consider in
' judging exercises is whether or not organizational goals 1
and objectives can be met.
Sampling techniques to judge organizational effectiveness are best cast in terms of making judgments toward that end, and should not neces-sarily be sampling techniques applicable to other forms of social science research, for example, statistically valid samples for citizen surveys or polls.
2.
Q.
On pages 26 and 27 of their testimony, Drs. Simon and Cole argue representative samples can be drawn for small populations.
Do you agree?
A.
Standard convention clearly defines a minimum standard for concluding that one has a statistical basis for gen-eralizing from a sample to a population.
This minimum standard, as I am familiar with its use in sociological practice, is the 0.05 level of significance.
In order to achieve this minimum standard, one must honor the basic assumption that the mean of the sample is rea-sonably equal to the mean of the sampling distribution.
The logic of this assumption is explained as follows.
The basis for generalizing from a sample to a popu-l lation is based on the concept of the sampling distribu-tion.
A sampling distribution is a theoretical proba-bility distribution of sample statistics.
A sampling distribution is generated by taking all possible random samples (each with at least one unique element) with a 7
l
L.
u i fixed size.from a population.
A mean could be computed
~
.for each of-these unique samples.
The distribution of the means of all of these samples would constitute a sampling. distribution of means.
A sampling distribution is typically a univariate distribution ~(based only one one factor) and may there-fore be described in terms of its central tendency and variability, among other factors.
The central tendency for a sampling distribution is customarily referred to in terms of the expected value of a statistic.
The expected value is the average value a statistic assumes for its sampling distribution, that is, the arithmetic mean of the sampling distribution of the statistic.
This is the value that would be obtained if all of the statistics in the sampling distribution were added and then divided by their number.
If the av-erage or expected value of a statistic is equivalent to the parameter.(population value) which it estimates, then the statistic is said to be an unbiased estimator of the parameter.
A basic assumption necessary to gen-l eralize from a sample.to a population is that the mean of the sample obtained for research is equal to the mean of the sampling distribution.
Another sampling distribution characteristic to j
consider is the extent to which the sample statistics l
vary around their parameter.
Standard error is a i
,. summarizing measurement that illustrates the variability of a sample statistic around the parameter.
Standard errors are dependent, among other factors, on the size of the sample from which the statistic is computed.
The law of large numbers suggests that as sample size in-creases, the standard error decreases.
The larger the sample, the more closely a statistic will approximate Its parameter.
S,nall samples f rom small populations are destined to have high standard error and will fall below minimum acceptable levels needed to conclude that a statisti-cally representative sample exists.
Even larger samples 5
from small populations can fall prey to the dramatic change that could occur in a sample mean by the mere ex-clusion or inclusion of one more unit in the sample.
- =
From a statistical viewpoint, larger samples from small populations are "better" than small samples, but that does not mean that one can conclude that the sample is representative.
Additionally, sample data can only be l
used to generalize to mean population data; a sample of 19 out of 20 cannot be used to generalize to the indi-vidual excluded from the population since that particu-lar individual may have a score above or below the sam-ple mean.
i I
\\
l
^ -- ---
- i 3-
'p
[
LILCO, M:y 18,1987 000 XE itT usmc CERTIFICATE OF SERVICE
'87 MAY 21 P6 :25 In the Matter of 0FFicE <ihG itivicr.
- v.. r LONG ISLAND LIGHTING COMPANY 00CKL1 (Shoreham Nuclear Power Station, Unit 1)
BRANCH Docket No. 50-322-OL-5 I hereby certify that copies of LILCO'S MOTION TO STRIKE PORTIONS OF SUFFOLK COUNTY'S TESTIMONY ON CONTENTION EX 21, OR IN THE ALTERNA-TIVE, TO FILE SUPPLEMENTAL TESTIMONY, and SUPPLEMENTAL TESTIMONY OF DENNIS S. MILETI ON CONTENTION EX 21 were served this date upon the following by hand as indicated by an asterisk (*), by Federal Express as indicated by two asterisks
("), or by first-class mall, postage prepaid.
John H. Frye, III, Chairman
- Atomic Safety and Licensing Atomic Safety and Licensing Board Panel Board U.S. Nuclear Regulatory Commission U.S. Nuclear Regulatory Commission Washington, D.C. 20555 East-West Towers 4350 East-West Hwy.
Oreste Russ Pirfo, Esq.
- Bethesda, MD 20814 Edwin J. Reis, Esq.
U.S. Nuclear Regulatory Commission Dr. Oscar H. Paris
- 7735 Old Georgetown Road Atomic Safety and Licensing (to mailroom)
Board Bethesda, MD 20814 U.S. Nuclear Regulatory Commission East-West Towers Herbert H. Brown, Esq.
- 4350 East-West Hwy.
Lawrence Coe Lanpher, Esq.
Bethesda, MD 20814 Karla J. Letsche, Esq.
Kirkpatrick & Lockhart Mr. Frederick J. Shon
- South Lobby - 9th Floor Atomic Safety and Licensirg 1800 M Street, N.W.
Board Washington, D.C. 20036-5891 U.S. Nuclear Regulatory Commission East-West Towers, Rm. 430 Fabian G. Palomino, Esq.
- 4350 East-West Hwy.
Richard J. Zahnleuter, Esq.
Bethesda, MD 20814 Special Counsel to the Governor Executive Chamber 4
Secretary of the Commission Room 229 Attention Docketing and Service State Capitol Section Albany, New York 12224 U.S. Nuclear Regulatory Commission 1717 H Street, N.W.
Mary Gundrum, Esq.
Washington, D.C. 20555 Assistant Attorney General 120 Broadway Atomic Safety and Licensing Third Floor, Room 3-116 Appeal Board Panel New York, New York 10271 U.S. Nuclear Regulatory Commission Washington, D.C. 20555 l
-,-,_,----,-nc------
w..-.
,-,.,,r-
,----,-,.-a,
-n 7,,
,,4
t i
\\ Spence W. Perry, Esq.
- Ms. Nora Bredes William R. Cumming, Esq.
Executive Coordinator Federal Emergency Management Shoreham Opponents' Coalition Agency 195 East Main Street 500 C Street, S.W., Room 840 Smithtown, New York 11787 Washington, D.C. 20472 Gerald C. Crotty, Esq.
Mr. Jay Dunkleberger Counsel to the Governor New York State Energy Office Executive Chamber Agency Building 2 State Capitol Empire State Plaza Albany, New York 12224 Albany, New York 12223 Martin Bradley Ashare, Esq.
Stephen B. Latham, Esq. **
Eugene R. Kelly, Esq.
Twomey, Latham & Shea Suffolk County Attorney 33 West Second Street H. Lee Dennison Building P.O. Box 298 Veterans Memorial Highway Riverhead, New York 11901 Hauppauge, New York 11787 Mr. Philip McIntire Dr. Monroe Schneider Federal Emergency Management North Shore Committee Agency P.O. Box 231 26 Federal Plaza Wading River, NY 11792 New York, New York 10278 Jonathan D. Feinberg, Esq.
New York State Department of Public Service, Staff Counsel Three Rockefeller Plaza Albany, New York 12223 R
b~
Esk B. Zeup (/
.Hunton & Williams 707 East Main Street P.O. Box 1535 Richmond, Virginia 23212 DATED: May 18,1987