Regulatory Guide 5.36
| ML003739956 | |
| Person / Time | |
|---|---|
| Issue date: | 06/30/1974 |
| From: | Office of Nuclear Regulatory Research |
| To: | |
| References | |
| RG-5.36 | |
| Download: ML003739956 (2) | |
June 1974 U.S. ATOMIC ENERGY COMMISSION
REGULATORY GUIDE
DIRECTORATE OF REGULATORY STANDARDS
REGULATORY GUIDE 5.36 RECOMMENDED PRACTICE FOR DEALING WITH
OUTLYING OBSERVATIONS
2. On the other hand, an outlying observation may be
A. INTRODUCTION
the result of gross deviation from prescribed experi mental procedure or an error in calculating or recording Section 70.51, "Material Balance, Inventory, and the numerical value. In such cases, it may be desirable to Records Requirements," of 10 CFR Part 70, "Special institute an investigation to ascertain the reason for the Nuclear Material," requires certain licensees authorized aberrant value. The observation may even eventually be to possess special nuclear material to establish and rejected as a result of the investigation, though not maintain sufficient written material control and account necessarily so. At any rate. in subsequent data analysis.
ing procedures to enable the licensee to account for the the outlier or outliers will be recognized as probably special nuclear material in his possession. Section 70.22, being from a different process than that of the sample
"Contents of Applications," requires applicants for values.
certain AEC licenses for special nuclear material to submit to the Commission hs part of the application a The committee's purpose in developing this revised full description of such procedures, including recording, standard is to provide statistical rules that will lead the reporting, physical inventory, and measurement pro cxperilnenter almost unerringly to look for causes of cedures for special nuclear material. Essential to satis outliers when they really exist, and hence to decide factory material control and accounting procedures are whether alternative I above is the more plausible nmeasurement and statistical control. This guide identifies hypothesis to accept, as compared with alternative 2. in methods and procedures acceptable to the Regulatory order that the most appropriate action in further data staff for dealing with the problem of outlying observa analysis may be taken. The procedures covered therein tions in samples and for testing their statistical signifi cance. apply primarily to the simplest kind of experimental data, that is. replicate measurement of some property of a given material or observations in a supposedly single
B. DISCUSSION
random sample. Nevertheless, the tests suggested do cover a wide enough range of cases in practice to have Committee E-I 1 on Statistical Methods of the Ameri broad utility.
can Society for Testing and Materials (ASTM) has revised a standard that deals with the problem of Topics addressed in this standard are:
outlying observations in samples and how to test their statistical significance. An outlying observation, or "out a. Basis of statistical criteria for outliers lier." is one that appears to deviate markedly from other members of the sample in which it occurs. In this b. A single observation in a sample connection, the following two alternatives are of inter est: c. The least and the greatest observation in a sample I. An outlying observation may be merely an extreme d. The two largest or the two smallest observations as manifestation of the random variability inherent in the probable outliers data. If this is the case, the value should be retained and processed in the same manner as the other observations e. Rejection of several outliers in the sample.
the divisions Copies of published guides may be obtained by request indicating USAEC REGULATORY GUIDES desired to the US. Atomic Energy Commission. Weshington. D.C.
20646.
for Attention Direttor of Regulatory Standards Commtents and suggetions Roguiawtsiv (-fs are issued to describe and tsaice available to the public encouraged and should be sent to the Secretary itmprovemints in these guides we
"nidtht)d accetsptalis io the AEC Regulatory staff of implementing specific the Conmmsission*s getstlOns to delineate techniques *,ad by the staff in parts of of the Commrisson, US. Atomic Energy Commilssion. WashingtOn.
D.C. 2054A.
ot postulated accidents, or to provide guidance to Altenrtion: Chief. Public Proceedings Staff.
eveluai,ng specific problems applicanits Regulatory Guiode are not substitutet for regulations and compliance and solutions different from thoset out irs The guides are issued in the followring ten broad divisions:
wit*h thisrn, . not required. Methods to the guides w*l be acceptable ,f they provide a bais for the findings requisite 6. Products or license by the Covmmission. 1. Power Reactors the issualce or continuance of a permit
7. Transportation
2. Research, and Test Reectors
3. Fuels end Materiels Facilities S. Occuportional Health
4. Environffentel and Siting 9. Antitrust Revuw Pub*ishfed guides will be re'viaed periodically, as appropriate, to accommodate 10. General corsinvnts and to.reftect now information or experience.
5. Materials end Plant Protection
C. Recommended criterion using independent stan kshremt in the data, subject to the following qudli-.
dard devition Cations:
g. Recommended criteria for known standard devia I. In regard to the statistical cuitena discussed in Section tion 3.1 of the ASTM Standard, the level of sipificauice should not exceed 0.5 percent.
The revised standard was approved by ASTM Com mittee E-I I in February 1974 and will be published in 2. Statistical outlier tests should be applied to basow August 1974 with the designation E178-74. measurement control processes, but not directly to the rejection of observed MUF values.
C. REGULATORY POSITION
3. Caution should be applied to avoid unwarranted The recommended practices dealing with the problem rejection of suspect observations. Discarding an overly of outlying observations in samples and the methods for large number of false outiess could introduce bias rather testing their statistical significance contained in ASTM than eliminate it.
Standard E178-74. 1 "Recommended Practice for Dealing with Outlying Observations," are generally 4. In regard to Section 3.2, the assumption of normality acceptable and provide an adequate basis for deter. should be assessed usng Regulatory Guide 5.22. "Assess mining if an outlying observation is an aberrant value or ment of the Assumption of Normality (Emplo)ing an extreme manifestation of the random variability Individual Observed Values)," April 1974.
I'Copies may be obtained from the American Society for Testing S. All discarded outliers should be recorded as part of and Materials. 1916 Race Street. Philadelphia. Pennsylvania the records maintained for special nuclear maerial
19103. control purposes.
5.36-2