Course Information: BTRY 604 & ILRST 414: http://courses.cit.cornell.edu/btry604/
The pre-requisites for Statistical Methods IV aka Applied Design (BTRY 604 & ILST 414) are Biological Statistics II (BTRY 302, NTRES 413), Statistical Methods II (BTRY 602) or equivalent. This course focuses on principles and methods of experiments design and analysis. Unrestricted randomization is discussed for a single-factor or several factors completely randomized designs with fixed-effects, random-effects and mixed-effects factors. Restricted randomization is introduced through a single-block factor, two blocking factors with both resolvable and non-resolvable row-column designs. Full and fractional two-level or three-level factors designs along with response surface designs are covered. Restrictions on randomization with different sizes of experimental units are emphasized for split plots, split blocks designs and variations of these two types of designs. Multiple comparisons of means and testing significance of contrasts will be discussed in the context of designed experiments. Both design of experiments and data analysis are performed using standard statistical software packages including SAS and/or Gendex as needed. 

 

Instructor:  Freedom King, fk35@cornell.edu, 1192 Comstock Hall, 255-0297
Office Hours: after class or by appointment.

Lectures:  lecture attendance is mandatory.
Monday and Wednesday 0840 AM – 0955 AM, B60 WN

Computer Labs: computer lab attendance is mandatory.
1          Tuesday         0255 PM – 0410 PM  B60 WN
  
Materials:

  • Required #1: Fundamental Concepts in the Design of Experiments, 5th Edition by Hicks R. C., and Turner V. K., 1999, Oxford University Press, Inc., ISBN: 978-0-19-512273-2.
  • Required #2: Variations on Split Plot and Split Block Experiment Designs, by Federer W. T.,  and King F., 2007, Wiley & Sons, ISBN: 978-0-470-08149-5
  • Suggested reference #3: Design and Analysis of Experiments, by Angela Dean and Daniel Voss, 1999, ISBN: 0-387-98561-1
  • Suggested reference #4: Models for discrete data, Revised Edition, Daniel Zelterman, 2006, ISBN: 0-19-856701-4.
  • Suggested reference #5: Design of Experiments: Statistical Principles of Research Design and Analysis, 2nd Edition by Kuehl O. R., 2000, Brooks/Cole, A Division of Thompson Learning.
  • Suggested reference #6: SAS for Mixed Models, 2nd Edition, by Little R. C., Milliken G. A., Stroup W. W., Wolfinger R. D. and Schabeneberger O., 2006, SAS Institute Inc., Cary, NC, USA.
  • Suggested reference #7: Generalized Linear Mixed Models with Applications in Engineering and the Sciences, by Myers R. H., Montgomery C. D., and Vining G. G., John Wiley & Sons, Inc.
  • Suggested reference #8: Learning SAS in the Computer Lab, Second Ed., by Elliot J. R., Duxbury Press, 2000, ISBN: 0-534-35925-6.
  • A non-programmable calculator capable of square roots, logarithms, and exponentials.

Statistical Software:

  • Design of Experiments: Gendex, SAS 9.1
  • Data analysis: SAS

Tentative Course Outline:


Topics

Material

Source

1

Research design principles

H&T: Chap. 1

2

Single factor designs: fixed-effects, random factors

H&T: Chap. 2-3

3

Two-factor studies: fixed-effects, random effects, mixed effects

H&T: Chap. 5, 6

4

More than two factor studies: fixed-effects, random effects, mixed effects

H&T: Chap. 5, 6

6

Nested and partially nested designs: random effects, mixed effects

H&T: Chap. 7,8

7

Randomized complete block designs: one factor, two or more factors

H&T: Chap. 4, 8

8

Latin square designs and lattice designs

H&T: Chap. 8

9

Incomplete block designs, row-column designs

H&T: Chap. 2-3

10

Two-level and three-level factorial experiments

H&T: Chap. 9, 10

11

Factorial experiments with confounding in blocks

H&T: Chap. 12

12

Fractional factorial replications

H&T: Chap. 13

13

Response surface designs

 

14

Split plot and Split block designs

H&T: Chap. 11 F&K: Chap. 1-5

15

Variations of split plot, split block designs

F&K: Chap. 1-5

16

Cross-over designs

 

17

Repeated measures designs

 

Legend: H&T = Hicks and Turner's 1999 textbook, F&K: Federer and King's 2007 textbook.
The topics listed above will be covered, based on a book in progress and the two required textbooks . Emphasis will be put on producing designs and data analysis (both normal and non-normal responses will be discussed). The topics listed above will be covered as time permits and the outline is subject to change.
Examinations:                     
Three equal weight exams (Prelim I, Prelim II, and Final) will account for 60% of the total course grade. Exams will be closed notes, closed book unless otherwise indicated by the course instructor. Students are allowed to write formulas on three 8.5 by 11 inches regular paper size (both sides) for each exam. Statistical tables will be provided as needed. Students should bring a non-programmable hand calculator to the exam. Graded hard copies of preliminary exams will be returned during lab times. Graded hard copies of the final exam will be available for pick up at our main office (CK 1998). Information in this regard will be posted on the course web page when available. Dates and venues for the exams are as follows:

Prelim 1: To be announced after consulting with students to avoid conflicts.
Prelim 2: To be announced after consulting with students to avoid conflicts.
Final exam: Thursday, May 15/2008, from 9 AM - 11 AM, in 201 Warren Hall.

Final Course Project:
The final course project will account for 20% of the total course grade. Key topics for the final course project are: cross-over designs, repeated measures, variations on split-plot and split-block experiment designs. Response variable should be either binomial or Poisson and the design should include both fixed and random factors. Data from published research papers are acceptable. Additional details on the final course project will be provided by mid March 2008. 

Homework Assignment and Computer Labs:
Homework assignments (10%) and computer labs (10%) will account for 20% of the total course grade. Homework assignments will be posted on the course web page on Fridays in the afternoon. There will be no homework assignment the week of exams. Each homework will go out on Friday and will be due on Friday the following week, unless otherwise indicated. Turn in a hard copy of your homework assignment, to Beatrix Johnson, on Fridays (if applicable) no later than 2:30 PM in 1198 Comstock Hall. Please, do not email your homework assignment file to the course instructor unless you have been formally instructed to do so. Make sure to type your homework solutions in Microsoft Word and paste the appropriate portions of your computer output whenever applicable. No late homework assignments will be accepted without prior permission from the course instructor. Permission for a homework extension should be made by email to the course instructor at least four days before it is due unless there are last minute extenuating circumstances. A homework extension will only be granted if there are extenuating circumstances (illness, accident, death in the family and so forth). Thus, failure to adequately manage time is not a sufficient reason for being granted an extension. Graded hard copies of homework assignments will be returned during lab times (exceptions may apply).
Computer labs will be graded at the end of each computer lab session, unless otherwise indicated by the course instructor. Remember that lab attendance is mandatory. For each computer lab, students will sign-up their names, as evidence of their lab participation.

Course Grading:
Course grades will be based on homework assignments/Labs (20%), 2 evening prelims (20 % each), final exam (20%), and a final course project (20%). We EMPHASIZE that you are not in competition with the other students; your grade is based solely on your performance. The breakdown of the total score is as follows:
Two prelims at 200 points possible each: 400
Homework assignments/labs (of equal weight): 200
Final exam, 200 points possible: 200
Final course project, 200 points possible: 200
Maximum total score: 1000

Incompletes will only be given under sufficiently dire circumstances, such as an auto accident or a death in the family. Failure to manage your time is not a sufficient reason for an incomplete. The final course grade will be computed as shown below.

Letter Grade Total Score
A+  970- 1000
A              930- 969
A-             900- 929 
B+             870- 899
B               830- 869
B-             800- 829
C+           770-799
C             730-769
C-            700-729
D+           670-699
D             630-669
D-           600-629
F                 0 –599

    

Re-grade Policy
Sometimes we make mistakes when grading homework or exams. If you think you have received an unfair grade we would be happy to regrade it. We would like to distinguish two cases. If a simple error has been made (like the points added incorrectly) then just contact the course instructor to have it corrected and properly recorded. On the other hand, if you feel a question has not been properly graded, write a short supporting explanation, staple it to your exam hard copy and give it to the course instructor for consideration. Please hand in any material to be re-graded within one week of receiving it, or it won’t be accepted.

Policy On Working Together                                                                              
We encourage you to form study groups to discuss the ideas from this course. You will learn a lot more by being able to talk things over with other people. However, all work that you hand in must be your own. That means that all homework must be completely your own work, though you may freely discuss homework problems with your classmates.              


Academic Integrity
Each student in this course is expected to abide by the Cornell University Code of Academic Integrity.  Students must in no way misrepresent their work, fraudulently or unfairly advance their academic status, or be a party to another student’s failure to maintain academic integrity.  The code may be found at: http://cuinfo.cornell.edu/Academic/AIC.html
                                    
Accommodations for Students with Disabilities
In compliance with the Cornell University policy and equal access laws, I am available to discuss appropriate academic accommodations that may be required for student with disabilities. Requests for academic accommodations are to be made during the first three weeks of the semester, except for unusual circumstances, so arrangements can be made. Students are encouraged to register with Student Disability Services to verify their eligibility for appropriate accommodations.