E ST 521 SAMPLING METHODOLOGY SYLLABUS, FALL 2008
TIME & LOCATION: LectureTUTH1:10-2:25BC247
LabW3:30- 5:20BC115
Labs are used primarily to discuss homework and exams and to execute the class project, but may be partially used for lecture. Class meets from August 24- November 9 and is 3 credits.
Instructor: Bill Gould, Office: GU 211, Phone: 646-3986
Office Hours:TUES 2:30-3:30, WED2:00 - 3:00 OR BY APPOINTMENT
Prerequisite: EST456, EST 465, or EST 502 or EST 505 or consent of instructor.
Text (required): Thompson, S. K. (2002) Sampling. John Wiley & Sons. Second Edition
You are expected to have read the assignments before class; lectures are more valuable as a result.
______
Course Description:This course presents the basic tools and fundamental concepts of sampling. Emphasis is placed on understanding the concepts and applications to practical problems. Design-based sampling is emphasized, but model-based sampling is also introduced.
Attendance: Strongly encouraged; You are responsible for obtaining notes, assignments, etc., from other students if you miss class. The professor reserves the right to adjust reading assignments, homework, etc., based on the progress of the class.
Make Up Policy University policy states that "Students making satisfactory progress in theirclasses will be excused from classes when they are representing NewMexicoStateUniversity on a university sponsored event (e.g., ASNMSUPresident represents NMSU at legislative session, student-athletescompeting in NMSU scheduled athletic events or education field trips andconferences). Authorized absences do not relieve the student of classresponsibilities. Prior written notice of the authorized absence must beprovided to the instructor by the sponsoring department." Unless absences are excused, any late work will be severely penalized except under extenuating circumstances.
Important Dates:September 9th is the last day to drop (October 8th with a "W"). The final exam will be held Wednesday, November 15th, 3:30-5:20pm in BC 111.
Grading: There are two exams (worth 30% each), a class project (20%), and homework (20%) on which you will be graded. Tentative exam dates are Oct. 4 and November 15.
The typical grading scale is used: 100-90: A; 89-80: B; 79-70: C; 69-60: D; <60: F A grade of C or higher must be obtained to receive an ‘S’ rating for an S/U rating.
ADA Statement: If you have or think you may have a disability that interferes with your academic progress, you are encouraged to contact the Services for Students with Disabilities at 646-6840 (V) or 646-1918 (TTY) or to discuss this on a confidential basis with your instructor. Services for students with disabilities are located in Garcia Annex, Room 102. Current and appropriate documentation will be required in order to receive services. It is your responsibility to inform either your instructor or SSD representative in a timely manner ifservices/accommodations provided are not meeting your needs.
If you have a condition which may affect your ability to exit safely fromthe premises in an emergency or which may cause an emergency duringclass, you are encouraged to discuss any concerns with the instructorand/or Michael Armendariz, SSD Coordinator. Feel free to call Ms. AngelaVelasco, Interim EEO/ADA and Employee Relations Director at 646-3333 withany questions about the Americans with Disabilities Act (ADA) and/or Section 504 of the Rehabilitation Act of 1973. All medical informationwill be treated confidentially.
Academic Misconduct is not tolerated and will be subject to disciplinary action (see guidelines in student handbook). Academic misconduct includes but is not limited to cheating or knowingly assisting another student in committing an act of cheating and/or plagiarism. A grade of zero will be assigned to any persons/assignments resulting from academic misconduct.
Class Project Summary You each will be given a map of the NMSU campus on which I have delineated into a grid of 360 plots. You must select a simple random sample of 12 plots to survey from the campus map provided. The purpose of the sampling is to estimate the number of trees on campus. Much of the lab time will be spent conducting the sampling effort in your area. A written report must be submitted to me (due October 25th) that presents the objectives of the work, clearly defines the target population, the sampling frame, and the sampling units. Specific procedures to follow and questions to be answered will be clarified early on during the semester.
Course Outline*: * Instructor reserves the right to adjust material coverage and homework.
Chapter 1.Introduction Definition of sampling, population, sampling units, frame, probability-based sampling, etc. Model-based vs. Design-based sampling. Concepts of bias, precision, mean square error. HW: handout #1
Chapter 2. Simple Random Sampling
Methods for taking simple random samples, Estimation of mean, total and respective variances and estimated variances. Finite population variance, sampling with replacement, derivations. HW: handout #2 & 2.1, 2.2, 2.3, 2.4
Chapter 3. Confidence Intervals
Construction and interpretation, finite population Central Limit Theorem HW: 3.1, 3.2, 3.3
Chapter 4. Sample Size
Factors involved and calculations for population mean and total. HW: 4.1, 4.2 & handout #4
Chapter 5. Proportions and subpopulation estimation
Properties of proportions, approximate and exact confidence intervals, sample size estimation, subpopulation estimation. HW: 5.1, 5.2, 5.3
Chapter 6. Unequal probability sampling
With replacement: Hansen-Hurwitz estimator, with or without replacement: Horvitz-Thompson estimator. HW: 6.1, 6.2, 6.3 & handout #5
Chapter 11. Stratified random sampling
Sample allocation: proportional, optimal, Neyman. Post stratification. HW: Handout, 11.1, 11.2 & handout #4
Chapter 12. Cluster sampling, systematic sampling
Simple random, probability proportional to size (PPS) sampling. HW: 12.1, 12.2, 12.5
Chapter 13. Multistage sampling
Simple random at each stage, PPS sampling of primary units. HW: 13.1 & handout #5
Chapters 7 and 8 (and section 2.7). Ratio Estimation and Regression Estimation
Use of auxiliary information, ratio estimator, models in ratio estimation. HW: 7.1, 7.2, 7.3 Linear regression estimator, precision comparison, model-based intro. HW: 8.1 & handout #6
Chapter 14. Double Sampling (Time permitting)
Ratio estimation and allocation in double sampling. HW: 14.1(a)
EST 521 Learning Objectives- Introduction and Chapter 1
List the advantages of sampling compared with a census.
Distinguish between a target population and a sampled population when appropriate.
Recognize a sampling unit and propose appropriate sampling frames for selecting these units.
List the advantages of probability-based samples and classify nonprobability samples into categories.
Distinguish between sampling error and nonsampling sources of error.
Paraphrase the concepts of bias, precision, accuracy, and consistency.
Learning Objectives- Chapter 2 Simple Random Sample
Describe the 2 properties that constitute a simple random sample.
Determine the probability that a sampling unit is included in a SRS of size n.
Apply the procedure for selecting a SRS to any sampling frame.
Prove that the sample mean from a SRS is an unbiased estimator of the population mean.
Determine if the sample variance (or std. dev.) is an unbiased estimator of the finite population variance (or std. dev.).
Apply Cornfield’s method of using an indicator variable in deriving estimators of variance.
Justify the choice of sampling without or with replacement in certain scenarios.
Learning Objectives- Chapter 3 Confidence Intervals
Explain what a confidence interval represents and how it should be interpreted.
Construct confidence intervals for SRS estimators of population mean and total.
Explain what the actual coverage of a confidence interval method is.
List methods of confidence interval construction when the estimator is nonnormal.
Learning Objectives- Chapter 4 Sample Size Estimation
Explain the components underlying a margin of error (interval estimator framework).
Compute the estimated sample size for estimating a population mean (or total) within specified conditions.
Explain the difference in finite and infinite population approaches to sample size estimation
Use a relative margin of error approach to sample size estimation.
Learning Objectives- Chapter 5 Proportions
Explain the similarity in estimating a population proportion and population mean.
Judge when to use which of 3 confidence interval construction methods for proportions.
Estimate the necessary sample size for estimating a population proportion within specified conditions.
Learning Objectives- Chapter 6 Unequal Probability Sampling
Distinguish between equal probability and unequal probability samples.
Propose situations under which unequal probability sampling is reasonable.
Compute the Hansen-Hurwitz estimator of the population total and its estimated variance.
Prove the Hansen-Hurwitz estimator is unbiased. Explain the derivation of the variance estimator.
Demonstrate the use of probability proportional to size sampling and modify the Hansen-Hurwitz estimator accordingly.
Compute the Horvitz-Thompson estimator of the population total and its estimated variance.
Prove the Horvitz-Thompson estimator is unbiased. Explain the unbiasedness proof of the variance estimator.
Justify the choice of the Hansen-Hurwitz or Horvitz-Thompson estimator for a given unequal probability sample.
Learning Objectives- Chapter 11 Stratified Sampling
List the Advantages of stratified random sampling.
Apply a stratified design to a sampling problem and compute estimators of population total, mean and their associated variances.
Explain why a stratified sample is more precise than a SRS
Recognize the differences among methods of Allocation- Proportional, Optimal, Neyman- i.e., what considerations enter sample size allocation and apply them when appropriate.
Apply Poststratification to a set of data
List the advantages and disadvantages of poststratification over prestratification
Learning Objectives- Chapter 12 Systematic and Cluster Sampling
Identify the similarities and differences between the two approaches.
List the advantages and disadvantages of each approach.
Compute estimates of population total and mean and their respective estimated variances.
Explain the conditions under which gain in precision occurs
Explain the variance estimation problem with single systematic sample and present possible solutions
Recognize the use of PPS sampling within each of these approaches.
Learning Objectives- Adaptive Cluster Sampling
Describe the methodology and explain when it isadvantageous (see chapter 24 for refresher)
Learning Objectives- Chapter 13 Multistage Sampling
Distinguish between primary and secondary units
Compute estimates of the population total, primary unit average, secondary unit average and their estimated variances.
Explain the relationship to cluster sampling, and why it is useful
Estimate population size and its estimated variance using with or without replacement sampling
Recognize that each level of sampling introduces additional variance components
Learning Objectives- Chapter 7.1 and 8.1 Auxiliary Data, Ratio and Regression Estimators
Explain the benefit of using auxiliary data, ratio estimator and regression estimator
Compute estimates of the population total and mean and estimated variances using ratio or regression estimator.
Recognize why a ratio estimator is biased, and why it might be a ‘better’ estimator than an unbiased estimator.
Compare the precision of SRS, ratio and regression estimators.
Know conditions under which a ratio or regression estimator will have high precision.
Demonstrate the equality of the ratio estimator variance structure to that of a SRS.