Precision of Simultaneous Measurement Procedures
Abstract
We consider the problem of measurement under the following conditions: The process of gathering the data is such that on any given item only one opportunity for measurement occurs, but it can be observed simultaneously by several instruments. The items to be measured are variable so that one cannot obtain replicate observations with the same instrument which would show directly,the variance of the instrument readings. Procedures, are discussed for estimating the precision of the instruments and the variability of the items being measured. An example due to Simon and Grubbs is helpful in fixing ideas. The burning times of thirty similar fuzes are determined by several different observers. We limit our discussion to the data taken by observers A and C; hence there are two determinations of the burning times of thirty different fuzes or sixty observations in all. If each of the fuzes had the same running time (which is the manufacturer's goal) and if both of the observers were absolutely accurate, then all sixty observations would be equal. However, considerable inequality in such data always occurs due to variation in the manufacturing process and inaccuracy of the observations. It then becomes desirable to use the sixty observations to answer as many questions as, possible about measurement bias and precision, mean fuze running time, and variability of burning times about their mean.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1963
- Accession Number
- ADP014614
Entities
People
- W. A. Thompson Jr.
Organizations
- National Institute of Standards and Technology