110年第1學期-1583 抽樣調查 課程資訊
If the class size is too large, or impossible to alternate attendance, the class may be held online with the consent of the teacher and students, provided that the quality of online instruction and related teaching and learning records are maintained. If the epidemic alert is lowered, classes should resume physical classroom.
本課程名額為 70人，已有50 人選讀，尚餘名額20人。
The course aims to equip the students with sampling techniques, which include the method of sampling (one-stage and two-stage), the corresponding estimator, the variance (and the estimated variance) of the estimators. It is expected that the students have sufficient knowledge to design survey and analyze survey data.
This course will cover the following topics:
1. simple random sampling (with/without replacement)
2. stratified random sampling (optimum allocation)
3. poststratified estimator 4. ratio estimator
5. cluster sampling (one-stage and two-stage, systematic sampling)
6. sampling with unequal probability (probability proportional to size (PPS) with/without replacement)
7. Horvitz–Thompson estimator.
Sampling is widely used in the modern world. The statistical offices of many nations have sample surveys conducted on topics of interest such as unemployment, size of labor force etc.. Furthermore, sample surveys are often conducted by company on topics of the behavior of consumers.
Sampling design determines the precision of the estimates. Thus, the way a sample is drawn is as important as the mathematical form of the estimator. Sample design consists of both a sample selection plan and an estimation procedure. In this course, we shall ﬁrst deﬁne population, sampling units (primary, secondary, etc.); then introduce many sampling schemes, such as simple random sampling (with or without replacement), stratiﬁed sampling (optimum allocation of sampling units to various strata), multistage sampling (e.g. two-stage stratiﬁed cluster sampling), systematic sampling, double sampling, and sampling with unequal probabilities (with or without replacement). We shall also introduce ratio estimator, regression estimator, and poststratiﬁcation estimators.
Finally, we shall derive the variances of the proposed estimators and the estimators of their variances. For estimating the variance of the estimators, the topics of Jackknife method and bootstrap method are also included in this course.
1. Sampling Methods for Applied Research by Peter Tryfors, John Wiley and Sons, Inc.
2. SAS 統計軟體與資料分析 by 沈葆聖