/* SAS Example 6: Chapter 11.2 Question 11.12 */
/* That the starting salaries of new accounting graduates would
differ according to geographic regions of the United States seems
logical. A random selection of accounting firms is taken from
three geographic regions, and each is asked to state the starting
salary for a new accounting graduate who is going to work in
auditing. Use a 1% level of significance */
OPTIONS PS=52;
OPTIONS LS=78;
OPTIONS NODATE;
OPTIONS PAGENO=1;
DATA SALARY;
INPUT REGION $ SALARY;
CARDS;
South 28000
South 29000
South 27500
South 28500
South 29000
Northeast 38500
Northeast 37000
Northeast 36500
Northeast 35500
Northeast 37000
West 33000
West 31000
West 32500
West 34000
West 33500
;
RUN;
PROC PRINT DATA=SALARY;
TITLE 'Regional Starting Salary Data';
RUN;
/* The PROC GLM statement is what we will use for our one-way ANOVA.
We could have used PROC ANOVA to get the same results, however,
PROC ANOVA will give incorrect results if we use it with
unbalanced designs */
PROC GLM;
TITLE 'One-way ANOVA for Regional Starting Salary Data';
/* Our class statement is where we define the classification or
treatment variable whose levels we want to look for a difference
between */
CLASS REGION;
MODEL SALARY = REGION;
/* The means statement will show us the mean of our response variable
at all levels of the explanatory variables that we list. In the
case below, we are asking for the mean salary for each region. */
MEANS REGION /TUKEY ;
RUN;