|Details of the FACTEX Procedure
- SIZE size-specification ;
You use the SIZE statement to specify the size of the design, which
is the number of runs in the design.
statement is required for designs of less than a
full replicate (for example, fractional factorial designs).
By default, the design consists of one full
replication of all possible combinations of the factors.
The two simplest explicit size-specifications that you can use are
- DESIGN=n, which specifies the number of runs (n)
in the design
- FRACTION=h, which specifies 1/h
of one full replicate
Use only one of these two options.
In all, there are six mutually exclusive size-specifications that
you can use, as described by the following list:
specifies the actual number of runs in the design. The number
of runs must be a power of the number of levels q for the
factors in the design. (See the
If the last FACTORS statement does not contain the
NLEV= option, then q=2 by default, and as a result, n must
be a power of 2. For an example, see
specifies the fraction of one full replication of all
possible combinations of the factors. For instance, FRACTION=2 specifies a
half-fraction, and FRACTION=4 specifies a quarter-fraction, and so on. In
general, FRACTION=h specifies a design with 1/h of
the runs in a full replicate. If the design has f factors,
each with q levels, then the size of the design is
qf/h. If you use FRACTION=h,
h must be a power of q.
See Example 15.4.
specifies the number of run-indexing factors in the design.
The design contains one run for each possible combination of the
levels of the run-indexing factors.
Run-indexing factors are the first m factors
for a design in qm runs. All possible
combinations of the levels of the run-indexing
factors occur in the design.
As a result, if each factor has q levels, the number of
runs in the design is qm.
For details on run-indexing factors, see "Types of Factors"
and "Structure of General Factorial Designs"
constructs a design with the minimum number of runs
(no larger than one full replicate) given all of the
other characteristics of the design.
In other words, the design
size is optimized. You cannot
specify this option if you specify any of the block size optimization
features in the BLOCKS statement
Equivalence of Specifications
The three explicit size-specifications are related to
each other, as demonstrated by the following example.
Suppose you want to construct a design for 11 two-level factors in
128 runs. Since 128=211/16=27, three equivalent size specifications for
this design are
Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.