STAT 325 Experimental Design (3)
Experimental design and analysis for a variety of problems: completely
randomized, randomized complete block, Latin square, completely randomized with
factorial treatments, unbalanced and/or incomplete, random effects, mixed
effects, nested; multiple comparisons; introduction to SAS®.
Prerequisite: STAT 292. (Offered odd years, spring semester.)
| Required Course Materials: |
R. Ott and M. Longnecker, An Introduction to Statistical Methods and
Data Analysis, 5th edition, ISBN 0-534-25122-6, Duxbury, 2001.
Prerequisite course material: J. Devore, Probability and Statistics for Engineering and the Sciences, 7th edition, ISBN 0-495-38217-5, Brooks/Cole, 2008.
L. Marlin Eby, Ph.D., Associate Professor of Mathematics and Statistics
- Students majoring in mathematics or minoring in statistics.
- Students wanting a more rigorous experimental design course.
- This course can be used to meet the elective requirement for mathematics
majors.
- To intuitively understand each concept.
- To understand, when possible and appropriate, the rigor of a mathematical
proof.
- To integrate topics by identifying commonalties.
- To understand the limitations of each analysis through consideration of
assumptions.
- To express general concepts in terms of the application.
- To communicate results, clearly and completely, in a manner appropriate to
nonquantitative audiences.
- To use the computer for analysis only after understanding how to perform
the analysis manually.
- Understanding of basic one-way and two-way ANOVA's and a multiple
comparison procedure.
- Understanding of multiple linear regression including inferential
analyses.
- Familiarity with one-sample and two-sample nonparametric hypothesis
testing.
- Ability to use the computer to perform statistical analyses.
- Introduction to SAS®: overview.
- Completely Randomized Design: one-way ANOVA, properties and assumptions,
statistical model, regression approach, design specifications, evaluating
assumptions, power calculations and sample size determination, and
nonparametric method.
- Multiple Comparisons: procedures and properties.
- Randomized Complete Block Design: partial two-way ANOVA, properties and
assumptions, statistical model, regression approach, design specifications
and evaluation, and nonparametric method.
- Latin Square Design: partial three-way ANOVA, properties and assumptions,
statistical model, regression approach, design specifications and
evaluation.
- Random Assignment of Treatments to Experimental Units: assignment of
treatments with CRD, RBD, and LSD.
- Completely Randomized Designs with Factorial Treatments: complete two-way
ANOVA, properties and assumptions, statistical model, regression approach,
design specifications, and extension to k factors.
- Analysis of Variance for Unbalanced and/or Incomplete Designs: matrix
approach to multiple linear regression, application to ANOVA models–estimation,
ANOVA, and multiple comparisons.
- Random-Effects and Mixed-Effects Designs: ANOVA, properties and
assumptions, multiple comparisons, statistical model, and design
specifications.
- Nested Designs: ANOVA, properties and assumptions, multiple comparisons,
statistical model, and design specifications.
- Computer labs with MINITAB®, a highly respected and internationally-known
interactive system and SAS®, the premier system of its kind in
the world–integrating statistical analysis, data management, and report
writing.
- Murray Library.
- Outside speakers.
- Discussion of career and graduate school opportunities in Statistics.
- Discussion of recommended supporting courses.
- Breadth of exposure to topics is desirable, but it should never be
realized at the expense of understanding.
- Primarily lecture with an open format allowing for substantial
instructor-student interaction.
- Multiple handouts designed to facilitate the organization and synthesis of
topics.
- Two-part homework projects: Part 1 manual solution and Part 2 SAS®
solution to Part 1 plus an expanded solution facilitated by usage of the
computer.
Revised: January 2008