/* 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;