Accession Number:

AD0610159

Title:

LINEAR RELATIONSHIPS BETWEEN VARIABLES AFFECTED BY ERRORS,

Descriptive Note:

Corporate Author:

JOHNS HOPKINS UNIV BALTIMORE MD

Personal Author(s):

• Carlson,F. D.
• Sobel,E.
• Watson,G. S.

Report Date:

1964-01-01

Pagination or Media Count:

21.0

Abstract:

Although of frequent occurrence in science the problem of estimating linear relationships when all variates are subject to error, is rarely discussed in statistical courses or texts because most statisticians regard it as insoluble. This attitude arises from considering only a mathematical formulation which gives rise to an insoluble problem. It is here claimed that, in biometrical practice, additional information can be found so that routine methods can be used for the soluble problem. This is brought to the attention of biometricians and the importance of collecting the required additional data is stressed. The solution of the problem lies then in the proper design of the experiment.

Descriptors:

• *STATISTICAL ANALYSIS
• MEDICAL RESEARCH
• MATHEMATICAL ANALYSIS
• FACTOR ANALYSIS
• CORRELATION TECHNIQUES
• ERRORS
• ANALYSIS OF VARIANCE
• LINEAR SYSTEMS
• LEAST SQUARES METHOD
• MUSCLES
• HYDROLYSIS
• HEAT PRODUCTION(BIOLOGY)
• AMINO ACIDS
• PHOSPHATES
• BIOLOGY
• EXPERIMENTAL DATA

Subject Categories:

Distribution Statement:

APPROVED FOR PUBLIC RELEASE