Nonparametric Methods in Statistics with SAS Applications

16 HYPOTHESES TESTING FOR TWO SAMPLES

data learning program;

input program $ GPA @@;

datalines;

yes 3.98 no 3.42 yes 3.45 no 2.56 yes 3.66

no 2.00 yes 3.78 no 3.19 yes 3.90 no 3.00

yes 4.00 no 3.56 yes 3.78 no 3.56 yes 3.12

no 4.00 yes 3.45 no 2.78 yes 3.97 no 3.44

;

proc npar1way data=learning program wilcoxon;

class program;

var GPA;

exact;

run;

The output is

Wilcoxon Two-Sample Test

Statistic (S) 136.5000

Exact Test

One-Sided Pr >= S 0.0078

The output test statistic is the same as the one computed by hand. The P-value is

below 0.01, which validates our conclusion. 

Example 1.8 The performance of a new wound healing bioimplant is studied retro-

spectively. In ...