Factors Affecting Power: Size of the Difference between Population Means
(2 of 3)
In figure A, the null hypothesis
that µ
diff = 0 is true. The area shaded in black represents
Type I errors; the striped area represents
correct decisions not to reject the null hypothesis. The mean of the
sampling distribution is µ
diff which equals 0.
In figures
B-D, the null hypothesis is false. The mean of each distribution is µ
diff which
is 10 for Figure B, 20 for Figure C and 30 for Figure D. The cutoff
for significance is 19.6 and is the same for all figures. The area
to the right of 19.6

is
shaded and represents power. The area below 19.6 is shaded

and
represents Type II errors (the area below -19.6 is negligible). As µ
diff
increases, power (the proportion of the distribution larger than
19.6) increases.