*Sample Design in Business Research*(Wiley, 1960)

Sample design in business research / by W. Edwards Deming

Deming, W. Edwards (William Edwards), 1900-

New York : Wiley, 1960

517 p. : ill. ; 24 cm

Sample Design in Business Research (Wiley Classics Library) (Paperback) by W. Edwards Deming (Author) "Need for understanding of responsibilities. Good statistical work is the product of several kinds of knowledge working skilfully together..." (more)

126

根據 WSJ

"美國一項最新調查顯示﹐數學家﹑精算師﹑

此書第一部分有八十幾頁在談 STANDARDS OF STATISTICAL PRACTICEPreface 中說

"...appreciate sampling as a trustworthy instrument, and as an indispensable tool of scientific inference in management and in research.

My recommendation, for use of the book as a textbook, is to begin with Part II and to study it concurrently with Part I, with daily lessons in Part III. The order of the chapters is purposive, not random, to enable nonmathematical readers to begin with Chapter 1 and to continuw far enough to appreciate sampling as the scientific method of inference before mathematics looks too heavy, the premise being that one must learn to hum or strum a tune and must like music before he can derive much good from study of harmony and counterpoint. ..." (pp. v-vi)

這本書十六年之後，1976年，台大商學系才買，第4刷。平均每年約一人借閱。

我2015.7.16 借出

《統計推論的界限》：

第三種為實務統計學家，許多年前偉大的老赫胥黎（Thomas Henry Huxley, 1825-95 ) 這樣說：「所謂實務者，不過是重複祖先的錯誤而已。」 (“The practical man is the man who practices the errors of his forefathers.” )此種實務人士對任何行業而言，都會有危害。

《企業研究的樣本設計》的前言和第15章 (

**Chapter 15 Stratified Sampling)章頭都引用。**

**Chapter 1 Responsibilities in Planning a Survey**

“The statistician who supposes that his main contribution to the planning of an experiment will involve statistical theory, finds repeatedly that he makes his most valuable contribution simply by persuading the investigator to explain why he wishes to do the experiment, by persuading him to justify the experimental treatments, and to explain why it is that the experiment, when completed will assist him in his research.” Gertrude M. Cox

“For that theory (mathematical theory of statistics) is solely concerned with working out the properties of the theoretical models, whereas what matters – and what in one sense is most difficult – is to decide what theoretical model best corresponds to the real world-situation to which statistical methods must be applied. There is great danger that mathematical pupils will imagine that a knowledge of mathematical statistics alone makes a statistician.” D. G. Champernowne

**Chapter 2 Some Remarks on the Theory of Sampling**

“It is noteworthy that the etymological root of the word theatre is the same as that of the word theory, namely a view. A theory offers us a better view.” Raymond J. Seeger

**Chapter 3 The Frame and the Elements of a Sampling Plan**

“I’ve got a little list – I’ve got a little list…I’ve got him on the list…They’d none of ‘em be missed.” Gilbert and Sullivan

**Chapter 4 Operational Definitions of Expected Value and of Standard Error**

“Of that there is no manner of doubt, No probable, possible, shadow of doubt, No possible doubt whatever.” Gilbert and Sullivan

**Chapter 5 Uncertainties not Attributable to Sampling**

“The fact that a general impression is more or less universal can not in itself be a guarantee of its validity.” P. C. Mahalanobis

**Chapter 6 Some Simple Replicated Designs**

“How I envy the clarity of vision that comes to the travelling salesman in a railway buffet-car at the third highball! How simple the great problems become!” Clarence B. Randall

**Chapter 7 A Survey of Business Establishments with Correction for Nonresponse**

“O time, thou must untangle this, not I.” Viola to Malvolio in Shakespeare’s Twelfth Night

**Chapter 8 Examples in Sampling Accounts**

“You see, but you do not observe. The distinction is clear.” Sherlock Holmes to Watson

**Chapter 9 Evaluation of Inventory of Materials**

“Sir, - In your issue of December 31 you quote Mr. B.S. Morris as saying that many people are disturbed because about half the children in the country are below average in reading ability. This is only one of many similarly disturbing facts. About half the church steeples in the country are below average height; about half our coal scuttles below average capacity; and about half our babies below average weight. The only remedy would seem to be to repeal the law of averages.” Letters to the editor

**Chapter 10 Exercise in a Replicated Survey of a Small Urban Area**

“Alas, my lord, your wisdom is consumed in confidence.” Calpurnia to Julius Caesar, Shakespeare

**Chapter 11 General Theory and Procedure for Replicated Sampling of a Large Area**

“No skill in pedagogy, and no luster of personality, can atone for teaching errors instead of trugh. Errors are very likely to be taught by those who do no research, and then, the more skillful the pedagogic indoctrination, the greater the harm.” Harold Hotelling

**Chapter 12 Field Procedure for the Creation of Segments and for the Selection of People within Frames**

“There is no knowledge of external reality without the anticipation of future experience. … what the concept denotes has always some temporal spread and must be identified by some orderly sequence in experience. … There is no knowledge without interpretation. … Thus, if there is any knowledge at all, some knowledge must be a priori.” C. I. Lewis

**Chapter 13 A Statistical Aid to Supervision**

“For whoso despiseth wisdom and nurture, he is miserable, and his hope is in vain, his labours unfruitful, and his works unprofitable.” Wisdom of Solomon

**Chapter 14 Sampling New Material**

“6. Some man holdeth his tongue because he hath not to answer: and some keepeth silence knowing his time. 7. A wise man will hold his tongue till he see opportunity: but a babler and a fool will regard no time. 8. He that useth many words shall be abhorred.” Ecclesiasticus

**Chapter 15 Stratified Sampling**

“The practical man is the man who practices the errors of his forefathers.” Thomas Huxley

**Chapter 16 Evaluation of Expected Value and of Bias in Sampling Procedures**

“And I have heard of thee, that thou canst make interpretations, and dissolve doubts: now if thou canst read the writing, and make known to me the interpretation thereof, thou shalt be clothed with scarlet, and have a chain of gold about thy neck, and shalt be the third ruler in the kingdom.” Daniel 5:16

**Chapter 17 Theory of Variances**

“How poor are they that have not patience!” Iago to Roderigo in Shakespeare’s Othello

**Chapter 18 The Poisson Series and the Square-Root Transformation**

“The greatest achievements are those not accomplished.” Ralph Roeder

**Chapter 19 Optimum Number of Segments per Block**

“Ye suffer fools gladly, seeing that ye yourselves are wise.” Corinthians 11:19

**Chapter 20 Theory for the Formation of Strata**

“I held my tongue and spake nothing: I kept silence, Yea, even from good words; but it was pain and grief to me.” Psalm 39

**Chapter 21 Choice of Zoning Interval and Number of Subsamples**

“The advantage of obtaining an explicit mathematical statement of the optimum is that it aids in thinking about good sample design, since it points out which factors are effective in determining the amount of information obtained per dollar.” Hansen, Hurwitz, and Madow

我們現在拜全文搜索之賜 ，可以做許多沒放在索引的字彙調查 。譬如說 quality

**1-10 of 20 pages with references to**

**quality:**

on Page 3: | |

"... CHAPTER 1 Responsibilities in Planning a SurveyThe statistician who supposes that his main contribution to the planning of an experiment will involve statistical theory, finds repeatedly that he makes his most valuable contribution simply by persuading the investigator to explain why he wishes to do the experiment, by persuading him to justify the experimental treatments , and to explain why it is that the experiment, when completed, will assist him in his research.-Gertrude M. Cox, Lecture in Washington, Department of Agriculture, I I January 1951.For that theory (mathematical theory of statistics) is solely concerned with working out the properties of the theoretical models, whereas what matters-and what in one sense is most difficult-is to decide what theoretical model best correspond to the real world- situation to which statistical methods must be applied. There is great danger that mathematical pupils will imagine that a knowledge of mathematical statistics alone makes a statistician. -- -D. G. Champernowne, in a discussion on the teaching of mathematical statistics, Journal q f the Royal Statistical Society, volume 118, 1955: page 203."..." |

on Page 23: ... CHAPTER 2 Some Remarks on the Theory of Sampling

It is noteworthy that the etymological root of the word theatre is the same as that of the word theory, namely a view. A theory offers us a better view. --Raymond J. Seeger, ... 1946

p.23

on Page 61: | |

"... Mahalanobis, The National Sample Survey, General Report No. 1, Ministry of Finance, Delhi, December 1952. Reasons for studying all the sources ..." |

on Page 87: "... CHAPTER 6 Some Simple Replicated Designs

How I envy the clarity of vision that comes to the travelling

**salesman**in a railway buffet-car at the third highball! How simple the great problems become! --Clarence B.

**Randall**, as reported in Business Week (New York), 12 June 1954: page 192.

A. INTRODUCTION TO REPLICATED SAMPLING Purpose of this .."

. | on Page 38: |

"... -Gilbert and Sullivan, The Mikado. The five elements of a sampling plan. Once there is a decision to go ahead on a survey, and the ..." |

**Mikado**.

on Page 50: "Operational Definitions of Expected Value and of Standard Error

Of that there is no manner of doubt,

No probable, possible, shadow of doubt,

No possible doubt

**whatever**.

..."GILBERT and Sullivan,The Gondoliers.

on Page 61: "... The fact that a general impression is more or less universal can not in itself be a guarantee of its

**validity**.-P. C. ..."

on Page 102: | |

"... CHAPTER 7 A Survey of Business Establishments with Correction for Nonresponse O time, thou must untangle this, not I.- Viola to Malvolio, in Shakespeare's Twelfth Night, Act 11, Scene ii. ..." |

on Page 110: "Examples in Sampling Accounts

A. QUICK STUDY OF ACCOUNTS IN LEDGERS*

You see, but you do not observe. The distinction is clear.-

**Sherlock**Holmes to Watson, in A. Conan Doyle's A Scandal in Bohemia.

CHAPTER 9 Evaluation of Inventory of Materials*

Sir,-In your issue of December 31 you quote Mr. B. S.

**Morris**as saying that many people are disturbed because about half the children in the country are below average in reading

of many similarly disturbing facts. About half the church steeples in the country are below average height; about half our coal

**scuttles**below average capacity, and about half our babies below average weight. The only remedy would seem to be to

**repeal**the law of averages. Yours faithfully, etc.-Letter to the editor of the Times, London, January 1954.

The nature of the problem

..."

on Page 165: | |

"... CHAPTER 10 Exercise in a Replicated Survey of a Small Urban Area Alas, my lord, your wisdom is consumed in confidence.- Calpurnia to Julius Caesar, Shakespeare's Julius Caesar, Act II, Scene ii. ..." |

on Page 186: | |

"... CHAPTER 11 General Theory and Procedure for Replicated Sampling of a Large Area* No skill in pedagogy, and no lustre of personality, can atone for teaching errors instead of truth. ... Errors are very likely to be taught by those who do no research, and then, the more skillful the pedagogic indoctrination, the greater the harm.-Harold Hotelling, "On the teaching of statistics, " a paper read at a meeting of the Institute of Mathematical Statistics at Dartmouth College, September 1945: Annals of Mathematical Statistics, vol. xix, 1948: pp. 95-115; page 106 in particular. |

on Page 243: "... CHAPTER 13 A Statistical Aid to Supervision

For whoso despiseth

**wisdom**and nurture, he is miserable, and his hope is in vain, his labours unfruitful, and his works unprofitable. -The

**Wisdom**of Solomon, iii: 11

Chapter 12

Field Procedure for the Creation of Segments and for the Selection of People within Families

There is no knowledge of external reality without the anticipation of future experience. ... what the concept denotes has always some temporal spread and must be identified by some orderly sequence in experience. ... There is no knowledge without interpretation. ... Thus, if ther is knowledge at all, some knowledge must be a prior. -- C. I. Lewis, Mind and World-Order (Charles Scribner's Sons, 1929), p.157 (Author's note: knowledge a priori means theoretical description--the statistician's probability model.)

p.226

on Page 255: "...

CHAPTER 14 Sampling New Material

6. Some man

**holdeth**his tongue because he hath not to answer: and some keepeth silence knowing his time.

7. A wise man will hold his tongue till he see opportunity: but a

**babbler**and a fool will regard no time.

8. He that useth many words shall be

**abhorred**.-Ecclesiasticus 20.

Range and shape of the distribution of the frame. ..."

on Page 276: | |

"... CHAPTER 15 Stratified Sampling The practical man is the man who practices the errors of his forefathers.-Thomas Huxley. A. PURPOSE AND GENERAL REMARKS The purpose of stratification. ..." |

on Page 361: "...

Evaluation of Expected Value and of Bias in Sampling Procedures

And I have heard of thee, that thou canst make interpretations, and

**dissolve**doubts: now if thou canst read the writing, and make known to me the interpretation thereof, thou shallbe clothed with scarlet, and have a chain of gold about thy neck, and shalt be the third ruler in the kingdom.-

**Daniel**5:16.

Purpose and content of Part III. ..."

on Page 380: | |

"... CHAPTER 17 Theory of Variances How poor are they that have not patience!-Iago to Roderigo in Shakespeare's Othello, Act 111, Scene iii. A. ..." |

on Page 455: | |

"... CHAPTER 18 The Poisson Series and the Square-Root Transformation The greatest achievements are those not accomplished.-Ralph Roeder, Juarez and His Mexico (Viking, 1947); page 179. ..." |

1. | on Page 468: |

"... CHAPTER 19 Optimum Number of Segments per Block Ye suffer fools gladly, seeing that ye yourselves are wise.-ii Corinthians 1 1 : 19.Statement of the problem. ..." |

on Page 4: "The test of

**talent**in management is ability to perceive and to formulate questions that if answered would help to find out what is wrong, or would improve the business, or would discover what we need to know: and in addition,

**wisdom**enough to use the results of a statistical survey or experiment that is designed to throw light on the problem. This view is contrary to the evaluation that some people make themselves when they claim ability to make the right decision on the basis of experience.

on Page 58: "... ... Efficiency,

**validity**, and usefulness.

A sample-design is efficient for a certain estimate if it produces at low cost a small standard error ..."

1. | on Page 56: |

"... bigger sample, with a smaller standard error, but whose definitions, * I am indebted to my colleague Professor P. C. Mahalanobis, F.R.S., for pointing out this fact to me. ..." | |

2. | on Page 59: |

"... sampling units in the frame, to produce almost any desired result? (Yes.) * Private communication to the author, about 1937. Mahalanobis was taking the same steps in Bengal about the same time. ..." | |

3. | on Page 61: |

"... Mahalanobis, The National Sample Survey, General Report No. 1, Ministry of Finance, Delhi, December 1952. Reasons for studying all the sources ..." | |

4. | on Page 71: |

"... , count, inspection, evaluation, calibration.) Meanwhile, Mahalanobis in India had for years been measuring the errors of sampling, and calibrating the work of observers, by means of ..." | |

5. | on Page 116: |

"... as the average in Class A. Let A, B, * Called to my attention in 1946 by Professor P. C. Mahalanobis, F.R.S. ..." | |

6. | on Page 186: |

"... Replication originated with Mahalanobis in 1936, and it is a pleasure to express my appreciation for the privilege of studying his methods in India, ..." | |

7. | on Page 251: |

"... 1. P. C. Mahalanobis, "Recent experiments in statistical sampling in the Indian Statistical institute," J. Roy. Statist. Soc., Vol. 109, 1946: pp. 325-370. 2. ..." | |

8. | on Page 303: |

"... C. Mahalanobis, "On large-scale sample-surveys," Phil. Trans. Roy. Soc., vol. 23IB, 1944: pp. 329-451. See also Hansen, Hurwitz, and Madow, Sample Survey ..." | |

9. | from Index: |

"... 309, 424, 425, 431, 445, 449, 450, 473, 481, 483, 485 Lillian MADOW 125 Elbert T. MAGRUDER 331 P. C. MAHALANOBIS 56, 59, 61, 77, 116, 186, 251, 303, 449 Management, report to, 12, 14, 16-22, 109, 117, 162; limits of ..." | |

10. | from Front Matter: |

"... estimation of the standard error and of any mathematical bias that exists in the estimator used. The method is essentially Mahalanobis's interpenetrating samples, which he introduced in 1936 in his surveys of Bengal. Acting ..." |

blemish, jack pots,

**least**

**squares**. It originated with Gauss about 1813, who derived it for use in astronomy, where originated much modern statistical theory. A ..."

. | on Page 41: |

"... of concentration of one or more chemicals, specified rates of flow to test, specified temperatures, and other experimental conditions. In analytic studies, which constitute most studies of the consumer and of test of product and of tests of process and methods, ..." | |

5. | on Page 356: |

"... STRATIFICATION FOR ANALYSIS Enumerative and analytic uses of data.* Thus far in this chapter the problem has been to allocate the sample to get the best ...This problem occurs frequently in marketing research. For example, one might need to estimate the number of people that use soluble coffee, and information as well on the reasons why they do, and on the difference between the average per capita consumption in 2 areas. The estimate of the number of users is an enumerative problem, and this is what we ... The study of the reasons that people give, and the estimate of the per capita differences, is analytic ,ands requires another kind of sample design.In agriculture, in industrial, and in science, one meets the analytic problem even oftener than he meets the enumerative one. Wherever there is a study of the causes of variation in dimensions, hardness, sales, production rates, error rates, the problem is analytic. We shall learn now something about the best allocation of ..." |

## 沒有留言:

張貼留言