1. Sampling 抽樣 中央大學. 資訊管理系 范錚強 mailto: ckfarn@mgt.ncu.edu.tw 2011.05 updated 11

2. 中央資管：范錚強 2 Learning Objectives Understand... The two premises on which sampling theory is based. The accuracy and precision for measuring sample validity. The five questions that must be answered to develop a sampling plan.

3. 中央資管：范錚強 3 Learning Objectives Understand... The two categories of sampling techniques and the variety of sampling techniques within each category. The various sampling techniques and when each is used.

4. 中央資管：范錚強 4 What Is a Sufficiently Large Sample? “In recent Gallup ‘Poll on polls,’... When asked about the scientific sampling foundation on which polls are based... most said that a survey of 1,500 – 2,000 respondents—a larger than average sample size for national polls—cannot represent the views of all Americans.” Frank Newport, The Gallup Poll editor in chief, The Gallup Organization

5. 中央資管：范錚強 5 The Nature of Sampling Sampling Population Element Population Census Sampling frame

6. 中央資管：范錚強 6 Why Sample? Greater accuracy Availability of elements Availability of elements Greater speed Sampling provides Sampling provides Lower cost

7. 中央資管：范錚強 7 When Is a Census Appropriate? NecessaryFeasible

8. 中央資管：范錚強 8 What Is a Valid Sample? AccuratePrecise

9. 中央資管：范錚強 9 目標母體 外部效度 External Validity 正式描述 規劃樣本 實際達成的母體 實際達成 的樣本

10. 中央資管：范錚強 10 Types of Sampling Designs Element Selection ProbabilityNonprobability UnrestrictedSimple random Convenience 便利 RestrictedComplex random Purposive 立意 SystematicJudgment ClusterQuota Stratified Snowball Double

11. 中央資管：范錚強 11 Steps in Sampling Design What is the target population? What are the parameters of interest? What is the sampling frame? What is the appropriate sampling method? What size sample is needed?

12. 中央資管：范錚強 12 When to Use Larger Sample Sizes? Desired precision Number of subgroups Number of subgroups Confidence level Population variance Small error range

13. 中央資管：范錚強 13 Simple Random Advantages Easy to implement with random dialing Disadvantages Requires list of population elements Time consuming Uses larger sample sizes Produces larger errors High cost

14. 中央資管：范錚強 14 Systematic Advantages Simple to design Easier than simple random Easy to determine sampling distribution of mean or proportion Disadvantages Periodicity within population may skew sample and results Trends in list may bias results Moderate cost

15. 中央資管：范錚強 15 Stratified Advantages Control of sample size in strata Increased statistical efficiency Provides data to represent and analyze subgroups Enables use of different methods in strata Disadvantages Increased error will result if subgroups are selected at different rates Especially expensive if strata on population must be created High cost

16. 中央資管：范錚強 16 Cluster Advantages Provides an unbiased estimate of population parameters if properly done Economically more efficient than simple random Lowest cost per sample Easy to do without list Disadvantages Often lower statistical efficiency due to subgroups being homogeneous rather than heterogeneous Moderate cost

17. 中央資管：范錚強 17 Stratified and Cluster Sampling Stratified Population divided into few subgroups Homogeneity within subgroups Heterogeneity between subgroups Choice of elements from within each subgroup Cluster Population divided into many subgroups Heterogeneity within subgroups Homogeneity between subgroups Random choice of subgroups

18. 中央資管：范錚強 18 Nonprobability Samples Cost Feasibility Time No need to generalize Limited objectives

19. 中央資管：范錚強 19 Nonprobability Sampling Methods Convenience Judgment Quota Snowball

20. 中央資管：范錚強 20 特殊的設計考量 你的對象 個人？ 企業？ 企業裡的個人 誰？ 是否能代表企業？

21. 中央資管：范錚強 21 複雜的設計 有一些構念是企業（團對、組織 … ） 有一些構念是個人 甚至是不同的個人 Matching 事後配對 發給 CEO 和 CIO ，針對回卷比對 事前配對 設計時就想好如何搭配

22. 中央資管：范錚強 22 萬一要衡量的變數是有「社會正確」 答案的 社會正確？ 你會不會照顧團體利益？ 你會不會非法 copy 軟體？ 你會不會不顧一切，堅持己見？ 得到的結果嚴重偏誤 如何能使得偏誤能降低 用兩端有張力的故事情節，挑選角色和看法

23. 中央資管：范錚強 23 Appendix 14a Determining Sample Size

24. 中央資管：范錚強 24 Random Samples

25. 中央資管：范錚強 25 Increasing Precision

26. 中央資管：范錚強 26 Confidence Levels & the Normal Curve

27. 中央資管：范錚強 27 Standard Errors Standard Error (Z score) % of AreaApproximate Degree of Confidence 1.0068.2768% 1.6590.1090% 1.9695.0095% 3.0099.7399%

28. 中央資管：范錚強 28 Central Limit Theorem

29. 中央資管：范錚強 29 Estimates of Dining Visits ConfidenceZ score% of AreaInterval Range (visits per month) 68%1.0068.279.48-10.52 90%1.6590.109.14-10.86 95%1.9695.008.98-11.02 99%3.0099.738.44-11.56

30. 中央資管：范錚強 30 Calculating Sample Size for Questions involving Means Precision Confidence level Size of interval estimate Population Dispersion Need for FPA

31. 中央資管：范錚強 31 Metro U Sample Size for Means StepsInformation Desired confidence level95% (z = 1.96) Size of the interval estimate .5 meals per month Expected range in population0 to 30 meals Sample mean10 Standard deviation4.1 Need for finite population adjustment No Standard error of the mean.5/1.96 =.255 Sample size(4.1) 2 / (.255) 2 = 259

32. 中央資管：范錚強 32 Proxies of the Population Dispersion Previous research on the topic Pilot test or pretest Rule-of-thumb calculation 1/6 of the range

33. 中央資管：范錚強 33 Metro U Sample Size for Proportions StepsInformation Desired confidence level95% (z = 1.96) Size of the interval estimate .10 (10%) Expected range in population0 to 100% Sample proportion with given attribute 30% Sample dispersionPq =.30(1-.30) =.21 Finite population adjustmentNo Standard error of the proportion.10/1.96 =.051 Sample size.21/ (.051) 2 = 81