leading to a 95% confidence interval of
0 : 51 Gð 1 : 96 Þð 0 : 05 Þ¼ð 0 : 41 ; 0 : 61 Þ
(b) In 1980, the estimated rate was
p 2 ¼
43
100
¼ 0 : 43
with its standard error
SEðp 2 Þ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð 0 : 43 Þð 1 0 : 43 Þ
100
r
¼ 0 : 05
leading to a 95% confidence interval of
0 : 43 Gð 1 : 96 Þð 0 : 05 Þ¼ð 0 : 33 ; 0 : 53 Þ
It can be seen that the two confidence intervals, one for 1965 and one for
1980, are both quite long and overlapsed, even though the estimated rates show
a decrease of 8% in smoking rate, because the sample sizes are rather small.
Example 4.10 A study was conducted to look at the e¤ects of oral contra-
ceptives (OC) on heart disease in women 40–44 years of age. It is found that
amongn 1 ¼5000 current OC users, 13 develop a myocardial infarction (MI)
over a three-year period, while among n 2 ¼ 10 ;000 non-OC users, seven
develop an MI over a three-year period. Application of the method described
above yields the following 95% confidence intervals for the MI rates:
(a) For OC users, the estimated rate was
p 1 ¼
13
5000
¼ 0 : 0026
with its standard error
SEðp 2 Þ¼
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð 0 : 0026 Þð 1 0 : 0026 Þ
5000
r
¼ 0 : 0007
ESTIMATION OF PROPORTIONS 163