Accession Number:

AD0653189

Title:

STUDIES IN RESEARCH METHODOLOGY. VII. THE CENTRAL LIMIT EFFECT FOR TWO DOZEN POPULATIONS AND ITS CORRELATION WITH POPULATION MOMENTS.

Descriptive Note:

Final rept., Dec 64-Oct 66,

Corporate Author:

AEROSPACE MEDICAL RESEARCH LABS WRIGHT-PATTERSON AFB OHIO

Personal Author(s):

• Bradley,James V.

Report Date:

1966-12-01

Pagination or Media Count:

88.0

Abstract:

The report shows the Central Limit Effect upon the sample mean for samples from each of 24 populations. The exact shape of each population is shown graphically and its standardized central moments are given. Along with each population is shown the nonrobustness of its sample mean at both left and right tailed nominal alphas of .50, .40, .30, .20, .10, .05, .025, .01, .005, .001, and .0005 for Ns of 2, 4, 8, 16, 32, 64, 128, 256, 512, and 1024, thus depicting the Central Limit Effect in considerable detail. It is shown that the nonrobustness of the sample mean is strongly correlated with the departure of the 3rd and 4th standardized central moments of the sampled populations from their values for a normal distribution. When the population is appreciably asymmetric, the correlation with the 3rd moment tends to be extremely strong, especially when N is moderately large. Author

Descriptors:

• (*STATISTICAL TESTS
• THEORY)
• (*POPULATION(MATHEMATICS)
• SAMPLING)
• STATISTICAL ANALYSIS
• PSYCHOLOGY
• HUMAN FACTORS ENGINEERING
• CORRELATION TECHNIQUES

Subject Categories:

• Psychology
• Statistics and Probability
• Human Factors Engineering and Man Machine Systems

Distribution Statement:

APPROVED FOR PUBLIC RELEASE