Statistical Analysis for Education and Psychology Researchers

and Castallan, (1988). Table 5 in Appendix A4 provides small sample critical values for
selected significance levels.

Computer Analysis

The Wilcoxon signed ranks test is automatically performed by the procedure PROC
UNIVARIATE in SAS when the response variable analyzed is a difference score. SAS
code for data entry and analysis, using the same data as in the worked example, is shown
in Figure 7.5.

data a;

input best poor @@;

diff=best-poor;

cards;

4 2 3 1 5 3 2 2 3 1 5 1 1 1 4 3 3 4 4 2

;

proc univariate data=a;

var diff;

title ‘Difference between best and poor

relationships two-tailed test’;

run;

Figure 7.5: SAS code for computation

of the Wilcoxon Signed Rank statistic

Output from this SAS programme is shown in Figure 7.6:
Difference between best and poor relationships two-tailed test Univariate
Procedure
Variable=DIFF
Moments
N 10 Sum Wgts 10
Mean 1.4 Sum 14
Std Dev 1.429841 Variance 2.044444
Skewness −0.03421 Kurtosis 0.215703
USS 38 CSS 18.4
CV 102.1315 Std Mean 0.452155
T:Mean=0 3.096281 Pr>|T| 0.0128
Num^=0 8 Num>0
M(Sign) 3 Pr>=|M| 0.0703
Sgn Rank 16.5 Pr>=|S| 0.0234

Figure 7.6: Output from PROC

Univariate for paired difference best-

poor relationship

Statistical analysis for education and psychology researchers 230