SURVMETH 612
Methods of Survey Sampling
Summer 2006
| CLASS MEETINGS: | 9:30-12:00, Monday-Friday, July 17 – August 11, 2006 Beijing University |
| LABORATORY MEETINGS: | 9:30-12:00, Monday-Friday, July 17 – August 11, 2006 Beijing University |
| INSTRUCTORS: | James M. Lepkowski, University of Michigan Email: jimlep@umich.edu |
| OFFICE HOURS: | TBD |
Methods of Survey Sampling/Applied Sampling is an applied statistical methods course. It differs from most statistics courses because it is concerned almost exclusively with the design of data collection rather than with analysis of collected data. The focus of the course will be on problems of sampling human populations. While the principles of sample selection can be applied to many other types of problems, sample survey practices are widely used in the study of human populations. Sampling human populations poses unique and challenging problems that illustrate the application of the methods.
The course is presented at a moderately advanced statistical level. We will develop few of the mathematical aspects of sampling theory. A sound background in applied statistics is necessary, since a few mathematical derivations will be presented. And though the emphasis will not be on mathematical derivations, a thorough understanding of mathematical formulations will be needed.
The aims of the course are to teach basic ideas of sampling from an applied perspective and to provide experience with real problems. The course will cover the main techniques used in sampling practice: simple random sampling, stratification, systematic selection, cluster sampling, multistage sampling, and probability proportional to size sampling. These methods will be examined further in the context of two particular types of sample designs, area sampling and telephone sampling. The course will also cover topics such as sampling frames, cost models, sampling error estimation techniques, non-sampling errors, and compensating for missing data.
Class sessions are daily, include lecture and discussion of homework and examinations, and provide opportunity for questions. Students will have lecture notes available. Lecture notes, homework problems, homework solutions, and other course materials will possibly available on a course web site, accessible only to course students.
Homework and Examinations
The syllabus presents approximate dates of lecture topics and reading assignments, and the due dates for homework assignments. Students are expected to have completed the assigned reading prior to the lecture.
Homework and examinations are important learning devices. It is important to keep up with homework assignments. Each assignment is to be handed in when due at the start of the class session.
The homework assignments will be graded on a simple three category scale: Ö+, Ö, and Ö-. These marks will be converted to scores of 100, 90, and 80, and if the assignment is not turned in, a zero will be assigned. Late submission will incur a penalty of 20 points.
The mid-term examination is in-class, scheduled for Friday, July 28, 9:30-12. The final examination is also in-class and scheduled for Friday, August 11, 9:30-12. The examinations will be cumulative, covering all material taught in the course prior to the examination. Examinations are open-book and open-notes. Cheating will not be tolerated. Each exam will be scored on a 100 point scale.
Study groups are useful for preparing answers to homework, and are encouraged. However, group answers are not acceptable; each student must submit their own work. If at any point during the course students desire additional review or discussion, they may suggest that the instructor schedule special sessions for interested students.
Final grades will be a weighted composite of homework (approximately 45%) and examination scores (approximately 55%).
Textbooks
The principal text for the course will be Survey Sampling by Leslie Kish (John Wiley and Sons, Inc., New York, 1965). There is a translation of the text in Chinese that is available in Beijing. Students may find that the following texts serve as useful supplemental reading to several lecture topics: Introduction to Survey Sampling by Graham Kalton (Sage Publications, Beverly Hills, 1983), Sample Survey Methods and Theory, Volume 1, by Morris Hansen, et al. (New York: John Wiley and Sons, Inc., 1953), and Sampling Techniques, 3rd edition, by William G. Cochran (New York: John Wiley and Sons, Inc., 1977). There are also assigned readings of two papers. These will be distributed in class in the latter part of the course.
[1] Rust, K. "Variance estimation for complex estimators in sample surveys," Journal of Official Statistics, 1(4): 381-397.
[2] Kalton, G. and D. Kasprzyk. "The treatment of missing survey data," Survey Methodology, 12:1-16.
Course Syllabus
| Date | Topic | Readingsa | HW | |
| July | 17 | Course introduction. Principles in sample selection. | Kish 1.0-1.7 | -- |
| 18 | Simple random sampling. | Kish 2.1-2.5 | 1 | |
| 19 | Frame problems. Weights and weighted estimators. | Kish 2.6-2.7 | 2 | |
| 20 | Cluster sampling. | Kish 5.1-5.2 | 3 | |
| 21 | Two stage sampling. Intra-cluster homogeneity. | Kish 5.3-5.4 | 4 | |
| 24 | Stratified sampling. Sample allocation to strata. | Kish 3.1-3.3 | 5 | |
| 25 | Stratification topics. | Kish 3.4-3.5 | 6 | |
| 26 | Systematic sampling. | Kish 4.1-4.2 | 7 | |
| 27 | Unequal sized cluster sampling. | Kish 6.1-6.2 | 8 | |
| 28 | Midterm examination. | -- | -- | |
| 31 | Stratified cluster sampling. Complex sampling. | Kish 6.3-6.5 | -- | |
| Aug | 1 | Probability proportionate to size selection. | Kish 7.1-7.2 | 9 |
| 2 | Probability proportionate to size selection. | Kish 7.3-7.5 | 10 | |
| 3 | Area sampling. | Kish 9 | 11 | |
| 4 | Area sampling. | Kish 10 | 12 | |
| 7 | Variance estimation: balanced and jackknife repeated replication | Kish 14.1-14.3 | 13 | |
| 8 | Variance estimation: Generalized variances. Software. | [1] | 14 | |
| 9 | Total survey error. Observation error. Response error models. | Kish 13 | 15 | |
| 10 | Non-observation error. Compensating for missing data | [2] | 16 | |
| 11 | Final examination. | -- | -- |
* Readings are from the textbooks by Kish, or from specified papers