This preview shows pages 1–3. Sign up to view the full content.
EXST 7003 Lab 09
Fall 2011
EXST 7003, Lab 09:
Decision Processing and OneWay ANOVA in SAS
Week of:
November 7
st
to November 11
th
DUE:
November 15
th
(section 1) or November 18
th
(section 2 or 3)
The Program
When you want to compare the means of more than two populations, the two sample ttest is no
longer an efficient procedure. While you could do multiple ttests for each possible pair of
populations, this would inflate your error rate, as well as being a tedious method. The analysis of
variance (ANOVA) is an appropriate solution to this problem.
There are many ways to approach the analysis of variance in SAS. We will focus only two of
them: PROC ANOVA, PROC GLM and PROC MIXED. These procedures are quite similar and
are used extensively in other courses in statistics.
dm
'log;clear;output;clear'
;
OPTIONS
nodate
nocenter
pageno=
1
ls
=
78
ps
=
55
;
Title1
'EXST 7003 Lab 08'
;
Title2
'Your name'
;
DATA
one;
infile
'C:\Users\shuyu\Desktop\shoedata.txt'
M
=
','
;
DL
LENGTH
region $
25
product $
15
subsidiary $
15
;
INPUT
region $ product $ subsidiary $ stores sales inventory
returns;
IF
region =
'United States'
;
IF
subsidiary =
'Chicago'
THEN
subnum =
1
;
IF
subsidiary =
'Los Angeles'
THEN
subnum =
2
;
IF
subsidiary =
'Minneapolis'
THEN
subnum =
3
;
IF
subsidiary =
'New York'
THEN
subnum =
4
;
IF
subsidiary =
'Seattle'
THEN
subnum =
5
;
IF
subsidiary =
'Baton Rouge'
THEN
subnum =
6
;
DROP
inventory returns;
RUN
;
PROC
PRINT
data
=one;
Title3
'Printout of Pacific areas sales data'
;
RUN
;
PROC
ANOVA
data
=one;
class
subnum;
model
sales=subnum;
means
subnum/
lsd
clm
;
run
;
PROC
GLM
data
=one;
Title3
'Proc GLM ANOVA analysis of U.S. sales data'
;
Class
subnum;
Model
sales = subnum/
ss3
;
Means
subnum /
HOVTEST
=Levene
Tukey
CLDIFF
;
Output
out
=glmout p=pred r=resid;
1
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentEXST 7003 Lab 09
Fall 2011
RUN
;
PROC
UNIVARIATE
data
=glmout
NORMAL
PLOT
;
Title3
'Univariate analysis of resids'
;
VAR
resid;
RUN
;
PROC
BOXPLOT
data
=one;
TITLE3
‘Box Plot of Sales  by Subsidiary’;
PLOT
sales*subnum /
CBOXES
= dagr
CBOXFILL
= ywh;
RUN
;
The data
The data set to be utilized for this week’s consists of sales, inventory, and returns data for
a large, multinational department store (e.g. Macy’s).
The data set has been stored as an
individual file, separate and external to the SAS program, much like the file that was utilized in
Lab 5.
This file was originally stored as a Microsoft Excel spreadsheet file, but was converted
into a comma separated values (CSV) file which will be read in by the SAS software package
using a suitable statement structure.
For this lab, we are interested in testing to see if varied
stores in the United State have different Sales data; hence we use the subsetting IF statement to
omit records from other regions and countries.
Decision Processing
This is the end of the preview. Sign up
to
access the rest of the document.
 Fall '08
 Moser,E

Click to edit the document details