Statistical Analysis for Education and Psychology Researchers

poolvar=

((n1−1)*v1+ (n2−1)*v2) /

(ndiff−2);

sediff=sqrt (poolvar*(1/n1 +

l/n2));

t=round (tinv (1−alpha/2,

ndiff−2),.001);

lc=round(ydiff−t *sediff,

.001);

uc=round(ydiff+t*sediff,

.001);

proc print data=b split='*'

noobs;

where t ne .;

var n1 n2 y1 y2 v1 v2 alpha t

1c uc;

title1 'Confidence interval

for difference between means';

label alpha='Alpha'

t = 'Critical *t-value'

y1 = 'mean*(gp 1) '

y2 = 'mean*(gp 2)'

n1 = 'sample*size*(gp

1)'

n2 = 'sample*size*(gp

2)'

v1 = 'variance*(gp 1)'

v2 = 'variance*(gp 2)'

lc = 'Lower Confidence

Limit'

uc = 'Upper Confidence

Limit';

run;

Figure 17: Program for

confidence intervals for

difference between two

independent sample

means

Appendix A4 Statistical Tables

Table 1: The normal distribution (Z

deviates)

Probability values in the body of the table correspond to a given Z
score. This Z score (deviate) represents the proportion of the total area

Appendix 393