EXST 4012
Introduction To Sampling Techniques

Course Description: EXST 4012 - Introduction To Sampling Techniques (3 Credit hours). Simple random, stratified random, cluster, systematic, multistage, multiphase, and unequal probability sampling procedures and applications; ratio and regression estimation; non-response and non-sampling errors.

Course Prerequisites: Introduction To Statistical Analysis (EXST 2201) or equivalent.

Course Objectives: General objectives include the following:

• Familiarize the student with sampling terminology and the general objectives of sampling studies
• Familiarize the student with several of the "well known" statistical sampling methods and conditions and situations where each is appropriate
• Provide practice in applying the sampling methods discussed in class and the calculations of desired estimates and their measures of precision
• Provide a foundation for further study in the area of statistical sampling
• Stimulate each student's general interest in and an appreciation for statistics and its application

Instructor: Dr. Barry Moser, 149A Ag. Administration Bldg, 578-8376, bmoser@lsu.edu

Office Hours: Tu, Th 10:45a-11:30a or by appointment.

Text: Lohr, S.L. 1999. Sampling: Design and Analysis. Duxbury Press, Pacific Grove, CA, 494pp.

Supplementary Text: Scheaffer, R.L., W. Mendenhall, and L. Ott. 1996. Elementary Survey Sampling, fifth edition. PWS-Kent Publishing Co., Boston, 464pp.

Assignments and Grading: Grades will be based upon a midterm exam (100 pts), homework problem sets (100 pts total), a sampling project (100 pts), and a final exam (100 pts) with the total score (400 pts max) divided by 4 and rounded to the nearest integer to produce the final score. Missed exams will be scored as 0 points unless authorization to make the exam up is first obtained from Dr. Moser prior to the exam, or due to emergencies recognized by the University (Provost or Dean of Students) as such. Homework turned in late will be discounted by 10 points for each day that it is overdue. Sampling projects will not be accepted late. All work MUST BE COMPLETED by the end of the final exam for the course. No additional projects or works will be permitted for additional course credit.

Grading Scale: Letter grade assignments will be made according to the following scale based upon the final course score:

Final Exam: Monday, December 9, 2002

Content Overview

1. Sample surveys overview: Purposes of statistical sampling, basic concepts, sampling and statistical terminology and notation
2. Mathematical preliminaries: Expected values, variances, covariances
3. Simple random sampling (SRS): Selecting the sample, estimation of the mean, total, and proportion and other population parameters, variances of the estimators, sample size estimation, confidence intervals on the parameters
4. SRS: Estimation of parameters for subdomains
5. Unequal probability sampling: Horvitz-Thompson estimator, Hansen-Hurwitz estimator
6. Ratio estimation: Using auxiliary information, estimators, estimation of the variance of a ratio estimator
7. Difference and regression estimation: Estimators, comparisons with ratio estimation
8. Double sampling: Uses of multiphase sampling, Applications with ratio estimation
9. Stratified random sampling (StRS): Selecting the sample, estimation of the parameters, sample size determination and allocation, optimal strategies, post-stratification of a simple random sample
10. Double sampling: double sampling applied to stratified random sampling
11. Combining ratio or regression estimation with stratified random sampling: Estimation approaches
12. Relative efficiency of estimators: Comparing sampling and estimation strategies
13. Systematic sampling: Selecting the sample, parameter estimation, measures of precision for estimates, repeated systematic sampling
14. Single stage cluster sampling: Selecting the sample, parameter estimation and sample sizes, clusters of equal size, clusters of unequal size (ratio estimation and pps sampling), relationship to systematic sampling
15. Two stage cluster sampling: Selecting the sample, parameter estimation and sample sizes
16. Adaptive random sampling: Defining a network, selecting the sample, estimation

Topics as time permits

17. Non-response and non-sampling errors: Sample imputation approaches, methods to reduce non-response
18. Sampling models: Randomized response, weighted distributions and size bias, capture-recapture, line-intercept, line-transect sampling
19. Variance estimation: Methods for complex surveys

The course will focus on material covered in chapters 1-6 and chapter 12 of Lohr 1999 with material added or supplemented in class. This text covers the basics of statistical sampling, but also has material related to model-based sampling, sampling weights, and variance estimation which will be introduced in the class as time permits. To help with your understanding of the material you should work problems listed at the end of each chapter. It is not uncommon for these or similar problems to be included on exams. The material in each chapter will be supplemented in lecture with additional information, estimation methods, and sampling plans, and may be varied depending upon the makeup and interest of the class.  

Objectives and Purpose: The sampling class project provides an opportunity for the student to demonstrate that he/she has knowledge and understanding of the sampling principles discussed in class, and that he/she can design, implement, analyze, and report on a "real" sampling problem. Further, the project provides additional practice and review of the sampling methods discussed in class.

Project Instructions: Select a project for which sampling can be used to estimate an appropriate population parameter or parameters. You will have to design the appropriate sampling plan and then collect and analyze the data. Note that the analysis of previously collected data is not acceptable although some data sets might be acceptable as "populations." You will need permission of the instructor before you sample from an existing data base. One purpose of the project is to actively engage the student in the process of statistical sampling. Write a report that clearly describes the sampling objectives, the sampling design, the data collection procedures, estimators, data analysis, results, and a discussion of the results and of the sampling method that was employed. This discussion should include the appropriateness of the sampling plan for the sampling problem and suggestions about alternative sampling plans, if any, and why they might be preferred. The report should be typed and include title and author information, introduction, methods, results, and discussion sections, and should include an appendix of the data, and an appendix of the calculations and computer programs that you used. The report is to be limited to no more than 5 standard single sided pages, excluding the title page, appendices, tables, and figures. You will be asked to make an oral presentation of your sampling project on or after the due date. This will give all participants the opportunity to learn from your project.

Due Date: The project is due at the beginning of class on Thursday, December 5, 2002, no exceptions.

Grading: Several factors will be considered in evaluating the projects and are listed below. In general, the projects will be graded on the appropriateness of the sampling design to the problem chosen, the correct implementation of the sampling plan, correctness of the data analysis and estimation, clear and concise presentation of the methods, results, and discussion, and the overall impact of the project as relates to the objectives and purposes described above for the project. To encourage students to use sampling plans and estimation methods other than simple random sampling, a score multiplier is applied to each project (see below). Specifically, the project scoring protocol will be:

1. States background and objectives of the study

0 pt: Fails to state background and objectives
1 pt: States objective but fails to set it in a context
2 pt: Objective is ambiguous or vague
3 pt: Clearly states objective within a context
 
2. Identifies population, sampling frame or frames, sampling unit or units, and other pertinent information

0 pt: Fails to identify pertinent sampling information
1 pt: States sampling information but fails to set it in a context
2 pt: Sampling information is not clearly stated or is ambiguous or vague
3 pt: Clearly identifies the sampling information within context
 
3. Selects sampling plan or design appropriate to problem

0 pt: Sampling design is clearly inappropriate for the problem
1 pt: Sampling design could be used but is clearly not a reasonable choice
2 pt: Sampling design is good but certain features of the problem have been ignored
3 pt: Sampling design is very appropriate to the problem
 
4. Correctly implements the sampling plan or design

0 pt: Sampling plan is incorrectly implemented, not implemented at all, or inappropriate
1 pt: Sampling plan is implemented but with major errors
2 pt: Sampling plan is implemented but with minor errors
3 pt: Sampling plan is properly implemented
 
5. Describes the statistical analysis and estimation methods

0 pt: Methods are not described or considered
1 pt: Methods of analysis or estimation are not appropriate for 
        the sampling plan
2 pt: Methods are appropriate but not clearly described
3 pt: Methods are appropriate and clearly described
 
6. Correctly performs data analysis and estimation

0 pt: No analysis is performed or is clearly in error
1 pt: Analysis is performed but with some major errors
2 pt: Analysis is performed but with some minor errors
3 pt: Analysis is complete and correct
 
7. Provides clear and concise presentation of results

0 pt: No presentation of results is given or results do not match 
        with the data analysis
1 pt: Results are poorly presented; some results do not match the 
       data analysis
2 pt: Results are presented and consistent, but appropriate tables 
        and figures are missing
3 pt: Results are clearly presented and consistent with the data analysis 
        and appropriate tables and figures are used
 
8. Discusses results, makes appropriate inferences, evaluates sampling plan

0 pt: Discussion and/or inferences are absent
1 pt: Inferences are not in context and are poorly discussed
2 pt: Discussion and inferences along with evaluation of sampling plan 
        are vague or incomplete
3 pt: Inferences are made in context and fully discussed along with an 
        evaluation of the sampling plan
 
9. Provides complete and organized report

0 pt: Important sections are absent or neglected
1 pt: Report is complete but poorly organized
2 pt: Report is complete but with some lack of organization
3 pt: Report is complete and well organized
 
10. Provides well written and presented report

0 pt: Report is difficult to read; frequent spelling and grammatical errors
1 pt: Report is poorly written with several spelling and grammatical errors
2 pt: Report is well written but occasional spelling or grammatical errors
3 pt: Report is well written with almost no spelling or grammatical errors  

To encourage the use of some designs and estimation methods over others, a multiplication factor will be applied to the sum of the above point values to determine the project score on a 50 point scale. The multiplication factors are:

If your methodology is not listed above, please check with the course instructor.

Examples: Below are examples of projects that have been done in the past. They are provided solely to provide an idea of what might be done. Remember that the quality of the project depends upon a variety of factors.

1. Estimate the proportion of references found in papers in the Journal of the American Statistical Association that are themselves papers in the Journal of the American Statistical Association
2. Estimate the number of desks in LSU classrooms that are in need of repair
3. Estimate the average monthly housing expenditure for graduate students majoring in Experimental Statistics
4. Determine the proportion of students taking an introductory communications course that plan to take an additional course in communications
5. Estimate the average amount of time that graduate students spend studying per week
6. Estimate the number of children living in a certain subdivision
7. Estimate the number of new and used car dealers in Baton Rouge
8. Estimate property values in a specific region of Baton Rouge
9. Estimate the mean DBH and basal area of pines on a four acre tract of land
10. Estimate the proportion of LSU faculty with a PhD degree

Student Name ______________________________________________________________________.

Points provided in the blanks below correspond with the project scoring protocol identified for projects in the course syllabus.

1. ________ States background and objectives of the study
2. ________ Identifies population, sampling frame or frames, sampling unit or units,
                    and other pertinent information
3. ________ Selects sampling plan or design appropriate to problem
4. ________ Correctly implements the sampling plan or design
5. ________ Describes the statistical analysis and estimation methods
6. ________ Correctly performs data analysis and estimation
7. ________ Provides clear and concise presentation of results
8. ________ Discusses results, makes appropriate inferences, evaluates sampling plan
9. ________ Provides complete report
10. ________ Provides well written report

Last Modified: August 26, 2002