Nonparametric Methods in Statistics with SAS Applications

70 TESTS FOR CATEGORICAL DATA

run;

The second way to create a data set is by specifying cell counts. This is how it is done:

data psoriasis2;

input group $ 1-7 relief $ count;

cards;

Placebo yes 2

Placebo no 8

Drug A yes 8

Drug A no 2

Drug B yes 6

Drug B no 4

;

proc freq data=psoriasis2 order=data;

weight count;

tables relief*group/fisher;

run;

The P-value in this example is

Fisher’s Exact Test

Pr <= P 0.0366

Based on SAS output, the P-value is 0.0366. It is between 0.01 and 0.05, thus, we can

reject the null hypothesis at the 5% significance level and conclude that proportions

of patients who experienced relief differ across the groups. At the 1%, however, we

conclude that the proportions are the same. 

Exercises for Chapter 3

Exercise 3.1 A student in marketing ...