The FACTEX Procedure


Example 7.1 Completely Randomized Design

Note: See A Completely Randomized Design in the SAS/QC Sample Library.

An experimenter wants to study the effect of cutting speed (Speed) on the surface finish of a component. He considers testing the components at five levels of cutting speed (100, 125, 150, 175, and 200) and decides to test five components at each level.

The design used is a single-factor completely randomized design with five levels and 25 runs. The following statements generate the required design:

proc factex;
   factors Speed / nlev=5;
   size design=25;
   output out=SurfaceExperiment randomize(713)
      Speed nvals=(100 125 150 175 200);
run;
proc print data=SurfaceExperiment;
run;

The RANDOMIZE option in the OUTPUT statement randomizes the run order; the random seed, 713 here, is optional. The design saved in the data set SurfaceExperiment is displayed in Output 7.1.1.

Output 7.1.1: A Completely Randomized Design

Obs Speed
1 200
2 175
3 200
4 125
5 100
6 150
7 175
8 125
9 100
10 100
11 100
12 200
13 125
14 125
15 150
16 175
17 175
18 150
19 175
20 150
21 200
22 125
23 200
24 150
25 100



If you are working through this example on your computer, you might find a different run order in your output. This is due to the difference in the seed value of the random number generator. You can specify a seed value with the RANDOMIZE option. For syntax, see Randomize Design.