STAT 368: Introduction to Design and Analysis of Experiments
office hours MW 13:00-15:00 in
CAB 441, or by appointment.
- TIME AND VENUE
- CALENDAR DESCRIPTION AND PREREQUISITES
Basic principles of experimental design, completely randomized
design-one way ANOVA and ANCOVA, randomized block design, Latin square
design, Multiple comparisons. Nested designs. Factorial experiments.
Prerequisites: STAT 266 and a course in Linear Algebra; MATH 225
- COURSE OBJECTIVE
To give students solid foundations in understanding techniques of
design and analysis of experiments with response involving
uncertainty, with strong emphasis on hands-on computational
Statistics for Experimenters: Design, Innovation, and Discovery,
2nd Edition by G. E. P. Box, J. S. Hunter, W. G. Hunter.
Wiley, 2005. (Required).
Introduction to R, by W. N. Venables, D. M. Smith and the R
Development Core Team. (Recommended. Available freely from the
Internet, as well as other R documentation and texts.)
Design and Analysis of Experiments by D. C. Montgomery. Wiley,
2008 (7th edition). (Supplementary. A more mathematical
textbook, available in many editions.)
- ASSIGNMENTS AND DATES
Homework 35% (five assignments, equally weighted: due Jan 21,
Feb 4, Mar 6, Mar 20, Apr 3)
Midterm exam 15% (Monday February 25, in class)
Project 20% (report due April 12)
Final exam 30% tentatively
like Tuesday, April 23, 9:00-12:00.
There are no classes February 18-22 (reading week), March 29 (Good
Friday), April 1 (Easter Monday).
- IMPLEMENTATION OF THE GRADING SYSTEM
The raw cumulative score will be computed at the end of the term, with
the weights as indicated above, and used to determine the letter grade
using the combination of absolute and relative standing as follows:
everybody with 91% or more percent will be guaranteed to receive A;
everyone with 76% or more B; everybody with 51% or more C. The
standards for + and - modifications will be set similarly. All the
cutpoints may be somewhat lowered (but not raised), should the
situation require that.
Samples of past or representative evaluative course material will be
made available through the course web page.
- COMMENTS ON THE WORK IN THE COURSE
Attendance: although no marks are given for class
participation, students are expected to attend classes and actively
participate in discussions.
Homework: the weight of the assignments signals that it may take a
couple of days to complete; plan accordingly. The papers do not have
to be typed, but must be readable and document all the steps necessary
to obtain the solution.
Exams: closed book, closed notes. The policy regarding use of
simple calculators will be announced if necessary.
Project: will consist in designing, carrying out, and analyzing
an experiment, and submitting a printed report documenting all its
aspects: the purpose of the experiment, the design used, the analysis
of the experimental output and the conclusions. The evaluation of the
project will be based on originality, relevance of the subject to the
course, quality and sophistication of the statistical analysis, and
clarity and professional appearance of the report; the latter is
normally supposed not to extend 10 printed pages (and not to contain
any direct computer output except for the relevant graphical
displays). The work on the projects will be done in groups of up to
three students; the composition of the groups has to be finalized
before the Reading Week. Rather than a one-time assignment, the work
on the projects should be seen as a continuing process; I will be
happy to discuss plans and look at the partial results at any time.
All computations in the course (textbook, lectures, exams) will be
done in the statistical language R, an open source software available
at CRAN and its
mirrors. R may be
downloaded and installed
for free, on a variety of platforms; it is also available in CAB 341
and 345 labs. The use of R is aimed at gaining a working understanding
of the course concepts, and allows for possible bypassing of cumbersome
mathematical details without compromising the quality of the learning
The computations may require installing certain add-on packages,
either from CRAN, or from my private repository; the way how to do it
is described in Lecture 2. The package BHH22,
developed specially for and during the course, should be reinstalled
frequently, so that students use always the latest version.
General paradigm of deductive-inductive iterative reasoning. The
components of experiment: response, factors, replication.
Comparing two entities: randomization and blocking. Review of
significance tests. Permutation tests, their approximations by random
sampling. Tests using normal and t distributions. Discrete and count
Estimation by least squares, decomposition of the sum of squares,
ANOVA tables. Randomized experiments with a simple factor. Graphical
analysis of variance. Inferences: tests, multiple comparisons,
Randomized block designs, Latin and Graeco-Latin squares. Incomplete
Factorial designs. Analysis of variance. Two factors, Tukey test of
additivity. General factorial design. Factorial designs with two
levels. Blocking and confounding.
Fractional factorial designs, nested and split-plot designs,
experiments with random effects.
Also, some prerequisite concepts will be reviewed, if necessary: these
include notions of random sampling, frequency distribution,
probability density, expectation, covariance, and correlation, and
various probability distributions.
- STAT 501
Students taking this course as STAT 501 will be graded separately.
Otherwise, the same rules apply for them, except for the project,
which they will do individually and besides the report prepare also a
presentation, which will count for 5% (out of the total 20%) of the
mark for the project.
Lecture transparencies (will be available after classes):
Midterm review sheet |
Midterm warm-up |
368 at U of A eClass (for grades) |
Normal probability paper
- FINE PRINT (OBLIGATORY ANNOUNCEMENTS)
There are neither deferred term exams nor extended deadlines for the
homework. A student who cannot complete the assignment or misses a
term examination because of an incapacitating illness, severe domestic
affliction or other compelling reasons can apply for a deferral of the
missed component weight to the final exam.
A student who cannot write the final examination because of an
incapacitating illness or is suffering from severe domestic affliction
or other compelling reasons can apply for a deferred final
examination. Such an application must be made to the student's Faculty
office within 48 hours of the missed examination and must be supported
by a Statutory Declaration (in lieu of a medical statement form) or
other appropriate documentation. (University of Alberta Calendar
23.5.6.) If granted, the deferred final exam will be held from
9:00 to 12:00 on Saturday, May 4, 2013 (meeting outside
of CAB 357).
The University of Alberta is committed to the highest standards of
academic integrity and honesty. Students are expected to be familiar
with these standards regarding academic honesty and to uphold the
policies of the University in this respect. Students are particularly
urged to familiarize themselves with the provisions of the
Code of Student
Behaviour and avoid any behaviour which could potentially result
in suspicions of cheating, plagiarism, misrepresentation of facts
and/or participation in an offence. Academic dishonesty is a serious
offence and can result in suspension or expulsion from the University.
All forms of dishonesty are unacceptable at the University. Any
offence will be reported to the Senior Associate Dean of Science who
will determine the disciplinary action to be taken. Cheating,
plagiarism and misrepresentation of facts are serious offenses. Anyone
who engages in these practices will receive at minimum a grade of zero
for the exam or paper in question and no opportunity will be given to
replace the grade or redistribute the weights. As well, in the Faculty
of Science the sanction for cheating on any examination will include a
disciplinary failing grade (no exceptions) and senior students should
expect a period of suspension or expulsion for the University of
Cell phones are to be turned off during lectures, labs and seminars;
they are not to be brought to exams.