Dunn Procedure:
Suitable for pair-wise as well as complex comparisons
(or contrasts).
Does not require
equal sample sizes.
Controls the type I error
rate at the familywise (or experimentwise)
level
It is based on the
Bonferroni Inequality theorem and student
t-distribution
Declares a contrast to be
significant if it exceeds the MSD:
MSD= 
where 
critical value is obtained from Table E.14
(p.829)
SAS
commands for carrying out the Dunn
Procedure:
PROC GLM;
CLASS
treat;
MODEL
score=treat;
MEANS treat/BON
ALPHA=.10;
Dunn-
Procedure:
An improvement over
the Dunn procedure
Does not require equal
sample sizes
Controls the type I error
rate at the familywise level
It is based on the
Multiplicative inequality and the student
t-distribution
Declares a contrast to be
significant if it exceeds the MSD:
MSD= 
where
critical value is obtained from Table E.15
(p.830-832)
SAS
commands for carrying out the
Dunn-Procedure:
PROC GLM;
CLASS
treat;
MODEL
score=treat;
MEANS treat/SIDAK
ALPHA=.10;
(1) Review Section 4.4
in Kirk
(2) Do questions 17(a),
17(b), 18(a), and 18(b) of Chapter 4 in
Kirk.
|
Please note the
corrections needed for questions 17(b) and
18(b).
|
(3) Apply Dunn and
Dunn-
procedures to questions 2,3,4,5,7, and 8 from Chapter
5 in Kirk. For Dunn and Dunn-
procedures, assume C=number of all pairwise
comparisons.
(4) Preview Sections 4.3 and
4.4 in Kirk.
