STATS 345.3
"DESIGN and ANALYSIS
of EXPERIMENTS"

FALL 2004

LECTURES: MWF 10:30-11:20 AM
LAB: TBA


Latest Modification: June 28, 2004

A free download of "Acrobat Reader" is available here.

Instructor:

C. D. O'Shaughnessy

239 McLean Hall

Phone:  966-6103

Email:  oshaughn@snoopy.usask.ca

 

 

Text:

DESIGN OF EXPERIMENTS:  Statistical Principles of Research Design and Analysis

2nd Edition (2000), by Robert O Kuehl  (Duxbury)

 

Primary Reference:

DESIGN AND ANALYSIS OF EXPERIMENTS, 5th Edition (2001)

by Douglas C Montgomery  (Wiley)

 

 

Prerequisite:   STATS 242.3 or 245.3

(or a comparable introductory course in statistical methodology such as

ECON 204.6 or PSY 233.3/234.3 or STATS 244.3+SOC 240.3 or COMM 104.3/207.3 or GE 210.3 or PL SC 314.3).

 

Course Outline:  As many of the following topics as possible will be discussed.

The order of presentation may differ from the order given below.

The presentation will stress modeling and applications.

Chapter 0.      Statistical Review and

             Introduction to Design of Experiments

  

a.       One Mean and the Difference between Two Means Problems

          i.  Estimation and Confidence Intervals

         ii..  Hypothesis Testing

   b.       Principles and language of experimental design

 

Chapter 1.      One-Way Analysis of Variance (ANOVA):

               the Completely Randomized design.

 

   a.       Comparing the Means of several populations

   b.      Linear Models

   c.       Estimating Model Parameters

   d.      Regression Approach

 

Chapter 2.      Treatment Comparisons and Diagnostics

 

  a.       Contrasts (and "preplanned" comparisons)

   b.      Multiple Comparisons (and "data snooping")

   c.       Residuals and Diagnostics

 

Chapter 3.      Two- and Three-Way ANOVA

  

 a.       Two-factor experiments:  Main effects; Interactions

   b.      Fixed and Random Effects

   c.       Crossed versus Nested Designs

   d.      Expected Mean Squares

   e.       Three and more factor experiments

   f.        Unequal replications (unbalanced data)

 

Chapter 4.      Block Designs

 

   a.       Randomized Complete Block Designs

   b.      Latin Square and Related Designs

   c.       Incomplete Block Designs

 

Chapter 5.      Selected Topics from the following list

  

 a.       Factorial Designs

          i.    2n designs

         ii.    Blocking and Confounding

        iii.    Fractional Factorials

        iv.    3n designs

   b.      Analysis of Covariance (ANCOVA or ANOCOV)

   c.       Split-Plot Designs

   d.      Repeated Measures Designs

   e.       Cross-Over Designs

   f.        Incomplete Block Designs

   g.       Regression Models

   h.       Response Surface Methodology

   i.         Non-Parametric Methods in ANOVA

   j.        Multivariate ANOVA