STA 8200 — Design of Experiments for Research Workers
Lab 3 – Due Friday, October 1
Example:
As part of the production of a liquid medication, the producer, a large pharmaceutical
company, mixes batches of medication in large vats. To determine the degree to which
their product is consistent from batch to batch, the company takes a random sample of
ﬁve batches from a month’s production. From each batch four samples were obtained and
for each sample, an important characteristic of the medication was measured, yielding
the data below.
Batch
1
2
3
4
5
3.2 2.6 3.4 4.2 1.8
3.8 2.9 3.9 4.4 2.3
3.5 2.8 3.3 4.3 1.9
3.0 2.0 3.1 4.2 2.1
The ﬁle “drug.sas” contains a SAS program for analyzing these data with a oneway
random eﬀects ANOVA model. Copy this program from the public data directory to
your home directory and run the program in SAS.
•
drug.sas illustrates the use of both PROC GLM and PROC MIXED to implement the
oneway random eﬀects ANOVA model. First, we use PROC GLM:
•
To test
H
0
:
σ
2
a
= 0, the hypothesis that there is no batchtobatch variability we use
F
=
MS
A
/MS
E
= 2
.
987
/
0
.
0918 = 32
.
53 (p.2 of the output). Comparing this statistic
to its distribution,
F
(
a

1
,N

a
) =
F
(4
,
15) gives a
p

value of less than 0.0001, so we
reject the null hypothesis and conclude that there is signiﬁcant batchtobatch variability.
•
To estimate the variance components: We estimate the withinbatch variance (the error
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '08
 Staff
 Normal Distribution, PROC GLM, random effects ANOVA, batchtobatch variability, eﬀects ANOVA model

Click to edit the document details