The rationale and computational
formulae for SPF pr.q design
SAS analysis of data collected from an SPF pr.q
Design
The complete specification approach with
every effect and every error term included in the MODEL
statement.
The leave-one-out approach with the
second error term left out from the MODEL statement.
The interpretation of results obtained
from SAS outputs based on a SPF pr.q design.
(1) Review Section 12.10 in
Kirk.
(2) Reanalyze data from p. 457, Table
10.6-1 according to a SPF 2.33 design. To start on this problem,
you need to first delete the 5th subject from the data set. Then
divide the remaining 4 subjects evenly into two levels of a
between-subject variable. Finally regard both factors A and B as
two within-subject variables. Thus, you have a balanced SPF p.qr
design, ready to be analyzed by SAS.
(3) Reanalyze Problem 14 of Chapter 10 as
a SPF 2.32 design. In order to have a balanced design, I ask you
to once again delete the 5th subject from the entire data set.
Treat the remaining 4 subjects as blocks of two levels of a
between-subject variable. Treat both factors A and B as two
within-subject variables. In this way, you will be able to handle
the data set as one based on a SPF 2.32 design.
(4) Preview Sections 15.1-15.3 in
Kirk.
