Power of tests of normality for detecting contaminated normal samples

Conference · Wed Dec 31 23:00:00 EST 1980

OSTI ID:5578689

Thode, Jr, H C; Smith, L A; Finch, S J

Seventeen tests of normality or goodness of fit were evaluated for power at detecting a contaminated normal sample. This study used 1000 replications each of samples of size 12, 17, 25, 33, 50, and 100 from six different contaminated normal distributions. The kurtosis test was the most powerful over all sample sizes and contaminations. The Hogg and weighted Kolmogorov-Smirnov tests were second. The Kolmogorov-Smirnov, chi-squared, Anderson-Darling, and Cramer-von-Mises tests had very low power at detecting contaminated normal random variables. Tables of the power of the tests and the power curves of certain tests are given.

• Research Organization: Brookhaven National Lab., Upton, NY (USA)
• DOE Contract Number: AC02-76CH00016
• OSTI ID: 5578689
• Report Number(s): BNL-30744; CONF-810842-4; ON: DE82010899
• Country of Publication: United States
• Language: English

Related Subjects

510000 -- Environment
Terrestrial-- (-1989)
54 ENVIRONMENTAL SCIENCES
560300* -- Chemicals Metabolism & Toxicology
63 RADIATION, THERMAL, AND OTHER ENVIRON. POLLUTANT EFFECTS ON LIVING ORGS. AND BIOL. MAT.
99 GENERAL AND MISCELLANEOUS
990200 -- Mathematics & Computers
BIOLOGICAL MATERIALS
COMPARATIVE EVALUATIONS
CONTAMINATION
DATA
INFORMATION
MATERIALS
MATHEMATICAL MODELS
NUMERICAL DATA
SAMPLING
SIMULATION
STATISTICAL DATA
STATISTICAL MODELS