In: Statistics and Probability
An NHANES report gives data for 654 women aged 20 to 29 years. The mean BMI of these 654 women was 26.8. On the basis of the sample, estimate the mean BMI in the population of all 20.6 million women in the age group. Let us assume that the sample is from a normal population with standard deviation of 7.5.
a) Construct a 95% confidence interval for mean BMI of all women
b) Construct both a 90% and 99% confidence interval for the mean BMI in this population?
c) How does increasing the confidence level change the margin of error of a confidence interval when the
sample size and population standard deviation remain the same?
d) Suppose that the survey above had a sample size of just 100 women. What is the 95% confidence interval
for this situation? Is it different from the initial case?
e) Construct a 95% confidence interval for samples with 400 and 1600 women. How are these confidence
intervals different?
x̅ = 26.8, σ = 7.5, n = 654
a) 95% Confidence interval :
At α = 0.05 two tailed critical value, z_c = ABS(NORM.S.INV(0.05/2)) = 1.96
Margin of error , E = z-crit*σ/√n = 1.96*7.5/√654 = 0.5748
95% Confidence interval :
Lower Bound = x̅ - E = 26.8 - 0.5748 = 26.2252
Upper Bound = x̅ + E = 26.8 + 0.5748 = 27.3748
b)
90% Confidence interval :
At α = 0.1 two tailed critical value, z_c = ABS(NORM.S.INV(0.1/2)) = 1.645
Margin of error , E = z-crit*σ/√n = 1.6449*7.5/√654 = 0.4824
Lower Bound = x̅ - E = 26.8 - 0.4824 = 26.3176
Upper Bound = x̅ + E = 26.8 + 0.4824 = 27.2824
99% Confidence interval :
At α = 0.01 two tailed critical value, z_c = ABS(NORM.S.INV(0.01/2)) = 2.576
Margin of error , E = z-crit*σ/√n = 2.5758*7.5/√654 = 0.7554
Lower Bound = x̅ - E = 26.8 - 0.7554 = 26.0446
Upper Bound = x̅ + E = 26.8 + 0.7554 = 27.5554
c)
The margin of error increases as the confidence level increases when the sample size and population standard deviation remain the same
d)
95% Confidence interval :
At α = 0.05 two tailed critical value, z_c = ABS(NORM.S.INV(0.05/2)) = 1.96
Margin of error , E = z-crit*σ/√n = 1.96*7.5/√100 = 1.47
Lower Bound = x̅ - E = 26.8 - 1.47 = 25.33
Upper Bound = x̅ + E = 26.8 + 1.47 = 28.27
Yes it is different from the initial case.
e)
For n = 400
At α = 0.05 two tailed critical value, z_c = ABS(NORM.S.INV(0.05/2)) = 1.96
Margin of error , E = z-crit*σ/√n = 1.96*7.5/√400 = 0.735
95% Confidence interval :
Lower Bound = x̅ - E = 26.8 - 0.735 = 26.065
Upper Bound = x̅ + E = 26.8 + 0.735 = 27.535
For n = 1600:
At α = 0.05 two tailed critical value, z_c = ABS(NORM.S.INV(0.05/2)) = 1.96
Margin of error , E = z-crit*σ/√n = 1.96*7.5/√1600 = 0.3675
95% Confidence interval :
Lower Bound = x̅ - E = 26.8 - 0.3675 = 26.4325
Upper Bound = x̅ + E = 26.8 + 0.3675 = 27.1675
As the sample size increases the length of confidence interval decreases.