Professional Documents
Culture Documents
Course number: STAT 3850/5850. New course, beginning January, 2008. (Graduate students
can take an enhanced version of the course with an additional project, with credit for STAT
5850, according to University of Lethbridge Faculty of Graduate Studies regulations)
Textbook: Oehlert, G.W. 2000. A first course in design and analysis of experiments.
W.H. Freeman and Co., New York; Software: JMP IN
- and selected readings from literature and case studies; provided in class, or on-line
Outline: The purpose of this course is to learn, apply and discuss methods of experimental
design and analysis, for the purposes of identifying probable causality, quantifying variability,
distinguishing differences, detecting changes, estimating precision, and establishing confidence.
Replication, randomization, restrictions on randomization, sampling, and model development
will be emphasized, in presentations of both theory and practical applications, with computation.
Worked examples will apply simple but powerful software, mainly JMP, SAS and SPSS.
Lecture topics will generally follow the text (a more detailed list of lectures and lab topics, by
date, will be supplied in the first lecture). The core topics to be covered in the course are as
follows:
Hypothesis testing, estimation and scientific knowledge
Components of effective experiments
Randomization and design
Completely Randomized Designs
Regression and Analysis of Variance; contrasts and multiple comparisons
Generalized Type I error rates
Power and sample size
Factorial treatment structure, factorial models
Random effects and fixed effects
Nesting, crossing, mixed effects, and expected mean squares
Complete Block Designs, and Incomplete Block Designs
Factorials in Incomplete Blocks
Confounding
Split Plot Designs, and related designs
Design and analysis with covariates
Contrasts and means comparisons
Other topics: Multivariate data and analysis
Logistic regression and related methods
1
Additional general topics will include experimental, inference, data types, hypothesis
testing, parameter estimation, and a short history of statistical analysis, notation,
conventions, schools and people. Most methods will concern single dependent variables,
but some multivariate methods will be discuss and compared.
Lecture format: Lectures will consist of presentations, primarily by the instructor, with notes
available on a website that will be available as the class progresses. Methods, examples, and
case studies will be presented. A class website will be used to distribute notes and examples.
Lab format: Each lab will consist of a case study and problem to be analyzed and discussed. A
handout or presentation on the required methods of calculation, tests, inference or interpretation
will be provided. A complete list of lab topics will be handed out during the course. Some labs
will be primarily technical exercises, and others will involve academic work such as searching
for scientific data and literature, and preparing an interpretation and short written summary.
Each lab will result in a short report. Lab reports are due one week following the lab session,
and will be returned the following week. The instructor will attend the labs, describe each case
study and method of analysis, provide additional support and answer questions.
Readings will be assigned from the text, and selected articles (including case studies and
science publications).
Marking: Examinations will concern material from the text, lectures, lab instruction and
assigned readings. Especially important material will be identified in class and in review
sessions and handouts. A review before each exam will clarify what will be on the final exam.