This Dover LEdition, first published in 1960, is published through special arrangements with the Naval Weapons Center. It is an unabridged, unaltered republication of NAVORD Report 3369.
NAVAL WEAPONS CENTERChina Lake, Calif. 93555,4. The proper use of modern statistical methods provides not merely a precise summary of the conclusions that may be drawn from an experiment already performed but also, with a small amount of prior information or by making reasonable assumptions, a reliable prediction of the information that can be gained from a proposed experiment. The latter feature has to some extent already placed the planning of tests and experiments on a substantially more rational basis, and further progress can be expected in the future.
DOVER BOOKS ON INTERMEDIATE AND ADVANCED MATHEMATICSANAlthough the Manual is not intended to be a textbook or treatise on the one hand, or merely a set of tables on the other, it is rendered somewhat self-contained by the inclusion of Chapter 1, Definitions and Distributions, and the Appendix table., and charts. Attention is called particularly to Appendix Table 8, which enables confidence limits for a standard deviation to be obtained immediately, and to Charts 1I--X, which facilitate the determination of the sample sizes required for various experiments.The preparation of the Manual was first undertaken at the Naval Ordnance Test Station in 1948, under Task Assignment NOTS-36-Re3d-439-3, as authorized by Bureau of Ordnance letter NP36(Re3d)AAF:bc dated 26 October 1948, which also provided for the basic statistical study of the programs of the Station. Work on the manuscript was not completed at the termination of this task assignment, and the continuation of its preparation from 1950 to 1955 was supported by funds from explorto,-y and foundational research and general overhead. Although the book was prepared primarily for uise at the Station, it may well be of value at other Department of Defense establishments.
INTRODUCTIONThe Manual has been prepared for those engineers and scientists who wish to use statistical procedures to improve the efficiency of their experiments. Emphasis is placed on (1) consulting with a statistician while an experiment is in the planning stage, (2) specifying what a decision in a sratistical test means by stating the risks involved, and (3) attaching confidence limits to an estimate. The reader who frequently needs to use statistical methods should find a university course helpful. A group of engineers and scientists who desire specific instruction might profitably organize a seminar under the leadership of a statistician.Users of the Manual may find the following procedure useful:1. Read Chapter 1 for basic definitions. Since this is a manual rather than a textbook, the explanatory material is brief. References 1.3 and 1.5 are suggested for fuller explanations of points that seem difficult.' 2. Study the Table of Contents so that the appropriate part of the book can be found readily when a problem arises. Specific items are listed in the Index...
