ML20106A336
| ML20106A336 | |
| Person / Time | |
|---|---|
| Site: | Limerick  |
| Issue date: | 05/08/1984 |
| From: | Iversen G SWARTHMORE COLLEGE, SWARTHMORE, PA |
| To: | Romano F ROMANO, F. |
| References | |
| I-AWPP-3, OL, OL-I-AWPP-3, NUDOCS 8408170164 | |
| Download: ML20106A336 (5) | |
Text
._
(
go -a sa va an yp 4
\\
...ers o e co<<ee Sgf SWARTHMOR E. PENNSY LV AM1 A 9 966 9 ese:=vmanv or warnemaries April 26,1984 18881
- 7-7 Mr. Frank Romano 61* Forest Avenua
~
Ambler, PA 19002
Dear Mr. Romano:
As I told you on the phone yesterday, I am a statistician by traning and profession.
I did undergraduate work in mathematics, sociclogy and statistics.
Af ter receiving masters degrecs in mathematics and sociology, I went on to get my Ph.D. in statistics from Harvard University in 1969.
I came to Swarthmore College as an associate professor of statistics and was promoted to professor of statistics seven years ago. Since 1972, I have been a statistical consultant at Swarthmore College's Center for Social and Policy Studies. I have been accepted as an expert witness in statistics in federal district court in Phila-delphia, and I have testified for the defense in two cases, Presseisen
- v. Swarthmore College and Kunda v. Muhlenberg College.
In addition, I have worked on several other Title VII cases. I am a member of the American Statistical Association.
s Sincerely,
'N
%%%4 4'\\lk Cudmund R. Iversen, Ph.D.
Professor of Statistics CRI/g W.s,2t1128 EHir.ATU*lY 19594 HM N >.d_
l)
$$/N3M To f.'11 Identified..
- e..si 8408170164 840500
~
PDR ADOCM 05000352 Date:
g>
- T PDR
/
g,g V
(.g1 4 h
_____*"I__~.-___**._________***~.*_ _ _ - - -
- O
'* * ^
n.
AIR and WATER
.u #
Pollution Patro!
BROAD AXE, PA.
May 1, 1984
/
j U.S. Nuclear Regulatory Commission
/
Washington, D.C.
20555 BEFORE THE ATOMIC SAFETY AND LICENSING BOARD In The Matter of PHILADELPHIA FLECTRIC COMPANY Docket Nos. 50-352 and 50-353 (Limerick Generating Station Units 1 and 2)
TESTIMONY OF GUDMUND R. IVERSEN CONCERNING AUDITING AND SAMPLING AS IT RELATES TO QUALITY ASSURANCE RE WELDING AT THE LIMERICK GENERATING STATION Question il Would you state your name and position, please?
Answer il My name is Gudmund R.
Iversen, Ph.D.,
Professor of Statistics, Swarthmore College.
Question #2 Have your prepared a statement of your prof-essional qualifications?
Answer
- 2 Yes.
A statement of my qualifications (letter of April 26, 1984 to Mr. Frank Rom-ano), is attached to this testimony.
Question #3 What is the purpose of your testimony?
Answer
- 3 The purpose of my testimony is to state the requirements of statistical methods, pro-cedures, etc., as it relates to highest probability of validity of conclusions made from sampling a population, whatever that -
population.
guestion #4 Will you discuss the terms generally used I
in statistics sdch as sample, sampling, random sampling, sample size, population, probability, deviation, standard deviation.
l confidence levels and other important ter-rirology used in statistical analysis, pro-cedures and results?
l
-- --- ----- o--
- 1 r-m-
-c a1
y C. O AIR and WATER Pollution Patrol
,/
BROAD AXE, PA.
/
(2),
1 Testimony of Gudmund R.
Iverson continued:
When examining a set of objects (population), one can examine all of them or a properly selected statistical 1
Answer
- 4 sample.and use probability theory to generalize from sample to population.
The main sampling methods are simple random, stratified and cluster sampling.
The riample size is determined before the sample is drawn in such a way that the sampling distribution has a small enough standard deviation (standard error) to permit the chosen confidence level to produce confidence intervals that are short enough for the problem under stt Question
- 5 Will you discuss the relevance and importance of sample selection to Quality Control and Quality Assurance?
Answer 45 In order to generalize from sample to population the sample must be chosen in accordance with proper I
statistical procedures, and the sample must be large enough to give the desired precision for the analysis.
Question
- 6 Does the nature of an activity, for example, welding at a nuclear reactor, require a great-er degree of confidence in sampling than weld-ing in an office building?
Antwer 96 The importance of the subject matter determines the confidence level for estimating properties of the population, and nuclear reactors are often thought to be more important than office buildings.
Question
- 7 As it related to P.E. audit report on page 244 of 5/12/81 can you show any weak points.from a statistician's view point?
~
Answer 17 Nothing is said about how the hangers were selected.
If this is not a proper statistical sample, then the results apply only to the selected hangers and it is not possible to say anything about the population o hangers Question 18 As it relates to P.E.
audit report Page 308 of 6/2 to 6/15/82, can you show any weak points from a statistician's standpoint?
+
w
.q
+
e s.==eme==-
e
'n_
- W*
I
,.u
/
.-i.
o.
~f
/c. 1J AIR and WATER W
Pollution Patrol
/
/
BROAD AXE, PA.
(3)
Testimony of Gudmund R. Iversen continued:
f i
1 Answer 98 Same as answer 7.
Question
- 9 Same for Audit report No. P316 of 6/30-7/7/82.
Answer
- 9 The audit does not state how large the population of small pipe hangers is, how large the sample of inspected hangers is and how the sample was selected.
Question fl0 Same re No. P 331 of 8/26 to 9/3/82.
Answer
- 10 Same as answer 7.
1 l
Question
- 11 Same re page354 of 11/8/82.
Answer
- 11 A 10% sample rate, assuming a properly drawn samole, seems low.
With that sample size, the orobability of zero defective welds is larger than 0.05 as long as the percentage defective welds in the population lies between 0 and 11.
Question
- 12 Same re page 374 of 2/8 -3/4/83.
Answer
$12 Same as answer 7.
6 Question
- 13 same re page 374 of 3/10/83.
Answer
- 13 Same as answer 7.
9 e
[.
/ --
/(
AIR and WATER
^
/
Pollution Patrol
/
BROAD AXE, PA.
/
(4)
Testimony of Gudmund R. Iversen continued:
Question
- 14 "Same re page 285 of 2/19/82 Answer
- 14 Same as answer 7.
Question
- 15 Same as page 289 of 3/22/82.
Answer
- 15 Same as answer 7.
Question
- 16 Same re page 307 of 6/1/82.
Answer
- 16 Same as answer 7.
Question
- 17 A group of other Quality Assurance infractions have been sent to the ASLB by Air & Water Polu-tion Patrol, will you, as an expert statistician comment on them from a statistical viewpoint?
Answer
- 17 Yes I will.
e a
G s'
G e
O
, --