;
Notes:
;
1. Data is input after the cards
statement. Enter power alpha ;
diff sd and n separated by spaces
either POWER or n should ;
be set to −9 (value to be
estimated). ;
;
2. Sd is the pooled standard deviation
derived from both ;
samples.
;
;
3. Diff is the difference in sample
means to be detected. ;
;
4. This is a two-sided test, for a one-
sided test change ;
alpha/2 to alpha in the second line
of code after the data. ;
** 5. If sample sizes are
unequal the harmonic mean, n, of
n1 **;
** and n 2 should be used.
This is evaluated
as **;
** 2 (n 1 n 2 ) / (n 1 +n 2 ).
Alternatively, two
power **;
** estimates can be
determined, one for each
n. **;
********************************
********************************
*****;
data a;
input power alpha diff sd n;
cards;
0.910 0.05 1.5 1.2 − 9
0.80 0.05 5 10 − 9
;
data a; set a;
pza=abs(probit(alpha/2)) ; /*
pza=significance level as a z
Appendix 356