APM 625
Sampling Techniques
Fall, 2013
Description: APM 625 focuses on sampling methods for environmental science applications. Many of the
basic concepts and techniques are shared by other traditional application areas of sampling methods (e.g.,
agriculture, business, and health), but primarily biological and environmental ("natural resource")
applications are discussed. General techniques and concepts of probability sampling are emphasized. This
course focuses on problems in which the primary objective is description: what are the characteristics (e.g.,
means, totals, proportions) of the collection of objects (the population) you are studying?
Lecture:
Tuesday, Thursday 11:00-12:20, Bray 313
Instructor:
Dr. Steve Stehman
322 Bray
470-6692
svstehma@syr.edu
Office Hours: Tuesday and Thursday 1:00-2:00
By appointment (scheduling by email works best)
Prerequisite:
Introductory statistics (sample mean and variance, confidence intervals)
Text:
No required text, purchase lecture notes on ESF campus (basement Bray Hall)
Optional Texts (on reserve at Moon Library):
Sampling Techniques, W. G. Cochran
The Basic Ideas of Scientific Sampling, A. Stuart
Sampling, S. K. Thompson
A Sampler of Inventory Topics, K. Iles
Elementary Survey Sampling, Scheaffer, Mendenhall, and Ott
Sampling Techniques for Forest Resource Inventory, Shiver and Borders
Sampling Strategies for Natural Resources and the Environment, Gregoire and Valentine
Sampling: Design and Analysis, Lohr
Objectives - At the end of the course, you should be able to:
1. Explain the basic terminology and definitions of sampling.
2. Explain what constitutes a "sampling problem" and decide when the techniques of the course can be
applied to a specific situation.
3. Describe different sampling designs and estimators (i.e., demonstrate awareness of basic techniques and
options that make up the sampling "toolbox").
4. Correctly implement the sampling techniques contained in the toolbox described in Objective 3 (i.e.,
compute estimates, standard errors).
5. Define statistical criteria for evaluating and comparing sampling strategies.
6. Explain the theoretical basis of the criteria defined in Objective 5.
7. Using the criteria for evaluating sampling strategies (Objective 5), list advantages and disadvantages of
different sampling strategies.
8. Identify characteristics of a population that are favorable to particular strategies.
1
Evaluation:
Homework
Exam 1
Exam 2
Final
Total Possible
125
275
275
325
1000
Exams: Exams may include in-class and take-home components, the former to evaluate understanding of
basic skills and concepts, and the latter to evaluate ability to apply these skills and concepts. For all exams,
you may bring your class notes and homework exercises. You do not need to memorize formulas, but you
do need to know when and how to use them.
Exam Dates:
Exam 1
Exam 2
Final
Thursday, October 3
Thursday, November 14
Thursday, December 12, 3-5 pm (313 Bray)
The final exam is scheduled by the College and not subject to change. Please be aware that Exam 1 and
Exam 2 dates may change by one week depending on our progress through the course material. Exam 1 is
the easiest of the three exams, so please do not be surprised or discouraged if exam scores decrease with
each exam – the degree of difficulty increases as we progress through the semester.
Course Grades: Your final grade will be determined by the Total Points earned on the items listed in the
Evaluation section.
A
AB+
B
935-1000
895-934
875-894
800-874
BC+
C
F
740-799
700-739
600-699
<600
Homework: Homework assignments will be collected approximately weekly. Most homework problems
are designed to reinforce use of formulas, basic concepts, and notation. You may work together on
homework. You will be awarded full credit for homework if you have put forth reasonable effort, and you
hand the assignment in on time.
The above schedule and procedures in this course are subject to change in the event of extenuating
circumstances.
2
Outline of Topics (may be presented in an order different from that listed):
I. Fundamentals
- What is a sampling problem?
- Developing structures for sampling problems: universe, population, frame, sample, sample space
- Describing the population and sample: parameter, statistic, variable, estimator
- Sampling strategy: combination of a sampling design (protocol for selecting the data) and an estimator
(formula for estimating a parameter of interest -- analysis of the sample data)
- Basic designs: simple random, systematic, cluster, stratified
II. Inference in Sampling: techniques for making statements about the population based on the sample
- Statistical inference in sampling: sample space and random variable representations
- Design-based inference
- Properties of estimators: expected value, bias, variance, mean square error
- Choosing a sampling strategy
III. Estimation
- Basic estimation: means, totals, variances, estimated variances
- Inclusion probabilities: definition of a probability sample, importance to biological sampling
- Horvitz-Thompson estimation: generalization, use for non-linear parameters (e.g. ratios, correlation)
- Miscellaneous estimation: population proportions, subpopulation parameters, distribution functions
IV. Sampling strategies when a list frame is not available
- Line-intercept sampling
- Point-grid sampling
- Fixed-area plot sampling (quadrats, circular plots, strip transects)
- Variable radius plot sampling (also called prism, point, angle-gauge, and Bitterlich sampling)
V. Using Auxiliary Information, Part 1: Information known for the universe
- Estimation: difference, ratio, and regression estimators; poststratification
- Design: variable probability sampling
- Variable probability independent (VPI) sampling
- Probability proportional to prediction (3-P) sampling
VI. Using Auxiliary Information, Part 2: Information known only for the sample
- Two-stage sampling (of clusters)
- Two-phase sampling
VII. Special Topics (time permitting)
- Simulation (Monte Carlo) methods used to evaluate sampling strategies
- Adaptive sampling - sampling rare, but clustered items (rare plants, marine mammals)
- Analytic surveys - conventional statistical analyses as applied to complex surveys
3
APM 625 – Extended Syllabus
Learning Resources (and how to use them)
1. Lecture: The approach is a traditional (some might call it “old school”) approach of “cover the
material” using the lecture notes provided to you. Most of the factual material and details (e.g.
definitions, formulas, description of examples) are provided in the lecture notes so you do not have
to spend a lot of effort copying basic information in lecture. However, the lecture notes have space
for you to write additional information as topics are discussed in class. In general, I assume you are
encountering the information for the first time when I present course material in lecture (you of
course may read the notes prior to lecture but I do not assume that you have). In lecture, my intent
is to have you see and hear the course material presented in an order and within a context that I
think makes the most sense for learning. Lecture introduces basic facts and methods, and on
occasion, I will try to provide the “big picture” and context for the topic being presented. However,
you should also invest some time trying to develop your own view of the “big picture” that makes
the most sense to you. In general, lecture will not be an “active learning” environment in that you
will not be learning sampling from a personal voyage of “self discovery” approach. The active
learning component will take place outside of class when you review notes and work homework
exercises.
2. Homework: 10-12 assignments, approximately one per week. Homework provides practice
using formulas and interpreting results of various analyses. Homework is where you “learn by
doing”. I will provide solutions to the homework problems prior to the due date. You should
check your answers with the solutions, and if you have any questions, write those questions on your
homework and I will respond. I have moved to this system because it is a way to provide
immediate feedback on your work, and you can assess if your numerical answers are correct as
easily as I can if you have the solutions. This approach does place a high level of responsibility on
you to put in a good faith effort trying to work problems before looking at the solutions.
3. Exams: Two exams and a final (comprehensive) will be given. Exams are open book and notes
so it is not necessary to memorize formulas or definitions. However, I do expect that you have
spent considerable time working with the formulas and notation so that you have very rapid recall,
and that you have organized the information so that you can find it quickly (either in your lecture
notes or on summary/synthesis notes you have prepared). The exams provide motivation for you to
review and synthesize course content, and they also provide feedback on where additional work is
necessary. Exams will be cumulative to some extent because ideas presented early in the course
will continue to be used as we proceed through more advanced topics.
4. Practice tests (usually 2 practice tests for each exam). These practice tests are exams from
previous years covering the same content. I consider these an important learning resource that you
should use after you have thoroughly prepared for an exam. That is, the practice tests should be
taken as if they were a real exam after you have completed most of your study and review. Practice
exam questions are usually more challenging than homework questions and often extend the basic
ideas of the course. Practice tests are old exams so they give you a good idea of how I expect you
to be able to apply what you are learning.
4
5. Optional review session before each exam. These sessions allow you to ask questions about
anything (related to the course!), but typically questions focus on practice tests with a few questions
related to clarification of lecture notes. I do not formally prepare a review lecture and simply
respond to your questions. However, often I use your questions as a jumping off point to review
key ideas or topics I suspect might need clarification or synthesis to help your review.
6. Office hours: 2 hours per week, also available for appointments and email correspondence.
Most office hour and email correspondence are about homework problems.
7. “Group Learning” (Working with Other Students). Collaboration with other students is
allowed and encouraged on homework. I expect some students will informally and voluntarily
meet to study or review together.
Expected or typical student “learning” process
1. Attend lecture (not necessary to have read material beforehand); write additional notes on lecture
handouts as you find necessary.
2. Review notes before next lecture (this is very important!); re-work any numerical examples from
lecture, make sure definitions and notation are familiar; make up your own “glossary” of terms and
notation lists to help your short-term recall.
3. Read your “text” (the chapters in the back of the Course Notes your purchased) if there is a
section relevant to the lecture notes; work through any examples in the text.
4. Work homework problems, check solutions for any difficulties, correct mistakes, and if there are
questions, write down where you are “stuck” or “I don’t even know where to begin” on the
homework to obtain help.
5. 7-10 days before exam, start reviewing past material, focusing on details but also thinking some
about the “big picture” (i.e. what ideas go together, why are certain methods advantageous, what
general theory exists to guide what we do in practice).
6. After completing your review, work the practice tests (as if they were a real exam) and check
solutions; revisit any topics that you do not understand well, ask questions in office hours or via
email; also review homework problems and solutions for the portion of the course covered by the
exam.
7. Attend review session to ask questions and to hear questions other students may have.
8. After the exam has been returned, look over any problem areas, and ask questions if you still are
confused about some of the problems and solutions.
5