ST536 Offline Quality Control: Design of Experiments for Improving Quality



Course Information

Meeting Time: MWF 10:15-11:05 Patterson 208

Instructor: Marcia Gumpertz, 203A Patterson

Office Hours: Tu,Th 2:00-3:00

Course Website: http://www.stat.ncsu.edu/~st536_info/

Textbook--- Selected chapters from Giesbrecht, F. and Gumpertz, M.
Planning, Construction, and Statistical Analysis of Comparative Experiments.

Computing Documentation

Additional Readings and References

*Indicates some reading is on electronic or regular reserve.

Grading

Grading will be based on the following 5 components, each worth 20% of the grade: (1) class participation, (2) homework, (3) projects, (4) midterm, (5) final exam.

Accommodations for Students with Disabilities

If you have trouble hearing or seeing or have any other disability, please see the instructor to discuss accommodations. Also see the website of the university office of Disability Services for Students (DSS) .

Academic Integrity

For this course you may discuss homework assignments freely with others but you are expected to write up your own work and do your own computer programming. An honor pledge will be required on exams.

The university policy on academic integrity is found in the Code of Student Conduct (http://www.fis.ncsu.edu/ncsulegal/codeof.htm).


Topics and Readings from G&G

(Revised 2/17/02)

Reading
Important concepts in experimental design. Strategy of experimentation. G&G Chapter 1
Factorial experiments where each factor has 2 levels G&G Chapter 7; Box B.1
Confounding Effects with Blocks for factors with 2 or 3 levels Chapter 9
Fractional Factorials for factors with 2 or 3 levels. Two level factorials: regular fractions, 3/4 fractions, irregular fractions. Three level factorials: Regular fractions. Box B.3; G&G Chapter 10
Plackett-Burman/Hadamard Plans. Plans for many factors, each at two levels, that require few runs. Box B.5; G&G Chapter 11
Fractional factorials and confounding for factors with 4, 5, or 6 levels. Mixed 2 and 3 level factorials Chapter 12
Plans for which run order is important, making it undesireable to randomize treatment order Chapter 13
Multi-stage experiments: split and strip-plot experiments with fractional replication Chapter 14
Orthogonal Arrays, Inner and Outer Arrays Chapters 16 and 17
Response Surface Designs Chapter 8


Homework Assignments

  1. Assignment 1. Due Jan 18.
  2. Assignment 2. Due Feb 4.
  3. Assignment 3. Due Feb 17.
  4. Project Part 1. Due Feb 27.
  5. Assignment 4. Due March 18.
  6. Project Part 2. Due April 3.
  7. Midterm is March 20, 2002. Open book, open notes.
  8. Assignment 6. Due April 22. [Paint thickness data (Wu and Hamada) for problem 3]
  9. Project Part 3. Due April 29.
  10. Assignment 7. Due May 6. [pigment data (Grize) for problem 3]

Demos

  1. Power analysis for randomized complete block design
  2. Generate a partially confounded 2x2x2 factorial in blocks of size 4
  3. Construct a normal probability plot of main effects and interactions in an unreplicated 2^4 factorial... Constructing QQ plots in SAS SAS Output.
  4. Constructing fractional factorial plans [Construct a 2^5-2 factorial using Proc Factex] [ Construct a 2^3-1 factorial using the Data step]
  5. Use PROC REG to check aliasing of a fractional factorial... and a Plackett-Burman design

Useful Tables and Websites

  1. NIST Engineering Statistics Handbook Chapter 5. Process Improvement.
  2. Russel Lenth's power and sample size computer.
  3. Finite Field Arithmetic: A Brief Tutorial by R. Beresford. Note: there is an irritating ad on this page. Close the box to get rid of it.
  4. First row of Plackett-Burman Designs.
  5. Design of experiments software
Go back to Marcia Gumpertz' home page

Go to NCSU Statistics Department home page

Last Updated 2/28/02

© Marcia L. Gumpertz 2002