6.
3 Confidence Intervals for the Population Proportion
Recall the probability of success in a single trial of a binomial experiment is p.
This probability is a population proportion. In this section, you will learn how to estimate a population proportion p using a
confidence interval.
As with confidence intervals for m, you will start with a point estimate.
The point estimate for p, the population proportion of successes, is given by the proportion of successes in a sample and
is denoted by
x
pn = Sample proportion
n
where x is the number of successes in the sample and n is the sample size.
The point estimate for the population proportion of failures is qn = 1 - pn .
The symbols pn and qn are read as “p hat” and “q hat.”
EXAMPLE
Finding a Point Estimate for p
In a survey of 1000 U.S. teens, 372 said that they own smartphones. Find a point estimate for the population
proportion of U.S. teens who own smartphones. (Adapted from Pew Research Center)
Solution
Using n = 1000 and x = 372,
x
pn = Formula for sample proportion
n
372
= Substitute 372 for x and 1000 for n.
1000
= 0.372 Divide.
= 37.2%. Write as a percent.
So, the point estimate for the population proportion of U.S. teens who own smartphones is 37.2%.
Check Point
In a survey of 2462 U.S. teachers, 123 said that “all or almost all” of the information they find using
search engines online is accurate or trustworthy. (Pew Research Center)
a. Identify x and n.
b. Use x and n to find pn .
6.3 Confidence Intervals for the Population Proportion
2
Constructing a confidence interval for a population proportion p is similar to constructing a confidence interval for a
population mean. You start with a point estimate and calculate a margin of error.
A c@confidence interval for a population proportion p is
pn - E 6 p 6 pn + E
where
pn qn
E = zc . Margin of error for p
Bn
The probability that the confidence interval contains p is c, assuming that the estimation process is repeated a large
number of times.
GUIDELINES
Constructing a Confidence Interval for a Population Proportion
IN WORDS IN SYMBOLS
1. Identify the sample statistics n and x.
x
2. Find the point estimate pn . pn =
n
3. Verify that the sampling distribution npn Ú 5, nqn Ú 5
of pn can be approximated by a
normal distribution.
4. Find the critical value zc that
corresponds to the given level
of confidence c.
pn qn Level of
5. Find the margin of error E. E = zc
Bn Confidence zc
6. Find the left and right endpoints Left endpoint: pn - E 90% 1.645
and form the confidence interval. Right endpoint: pn + E 95% 1.96
Interval: pn - E 6 p 6 pn + E 99% 2.575
ESTIMATING p
0. Verify that np̂ 15 and n(1 p̂) 15 so the normal distribution can be used
1. STAT
2. Left arrow key once to TESTS menu
3. ALPHA [A] to get 1-PropZInt
4. Enter values of x (the number of “successes”), n (the number of trials), and c (confidence level)
5. Select Calculate
6. Report CI as lower bound p upper bound and interpret
3
6.3 Confidence Intervals for the Population Proportion
EXAMPLE
Constructing a Confidence Interval for p
Use the data in Example in page 1 to construct a 95% confidence interval for the population proportion
of U.S. teens who own smartphones.
Solution
pn = 0.372. So, the point estimate for the population
proportion of failures is
qn = 1 - 0.372 = 0.628.
Using n = 1000, you can verify that the sampling distribution of pn can be
approximated by a normal distribution.
npn = 11000210.3722 = 372 7 5
and
nqn = 11000210.6282 = 628 7 5
Using zc = 1.96, the margin of error is
pn qn 10.372210.6282
E = zc = 1.96 ≈ 0.030.
Bn B 1000
Next, find the left and right endpoints and form the 95% confidence interval.
Left Endpoint Right Endpoint
pn - E ≈ 0.372 - 0.030 pn + E ≈ 0.372 + 0.030
= 0.342 = 0.402
0.342 6 p 6 0.402
0.342 0.402
0.372
x
0.33 0.34 0.35 0.36 0.37 0.38 0.39 0.40 0.41
Interpretation With 95% confidence, you can say that the population proportion of U.S. teens who own
smartphones is between 34.2% and 40.2%.
6.3 Confidence Intervals for the Population Proportion
4
EXAMPLE
Constructing a Confidence Interval for p
The figure at the right is from a survey of 498 U.S. adults.
Construct a 99% confidence interval for the population Who are the more dangerous drivers?
proportion of U.S. adults who think that teenagers are the
more dangerous drivers. (Source: The Gallup Poll)
71%
Teenagers
People over
65
25%
Solution
From the figure, pn = 0.71. So, 4% No opinion
qn = 1 - 0.71 = 0.29.
Using these values and the values n = 498 and zc = 2.575, the margin of
error is
pn qn
E = zc
Bn
10.71210.292
≈ 2.575
B 498
≈ 0.052.
Next, find the left and right endpoints and form the 99% confidence interval.
Left Endpoint Right Endpoint
pn - E ≈ 0.71 - 0.052 pn + E ≈ 0.71 + 0.052
= 0.658 = 0.762
0.658 6 p 6 0.762
0.658 0.762
0.71
x
0.64 0.66 0.68 0.70 0.72 0.74 0.76 0.78
Interpretation With 99% confidence, you can say that the population proportion of U.S. adults who think that
teenagers are the more dangerous drivers is between 65.8% and 76.2%.
6.3 Confidence Intervals for the Population Proportion 5
F I N D I N G A M I N I M U M S A M P L E S I Z E TO E S T I M AT E p
Given a c@confidence level and a margin of error E, the minimum sample size n needed to estimate the
population proportion p is
zc 2
n = pn qn a b .
E
This formula assumes that you have preliminary estimates of pn and qn. If not, use pn = 0.5 and qn = 0.5.
EXAMPLE
Determining a Minimum Sample Size
You are running a political campaign and wish to estimate, with 95% confidence, the population proportion of registered
voters who will vote for your candidate. Your estimate must be accurate within 3% of the population proportion.
Find the minimum sample size needed when (1) no preliminary estimate is available and (2) a preliminary estimate
gives pn = 0.31. Compare your results.
Solution
1. Because you do not have a preliminary estimate of pn , use pn = 0.5 and
qn = 0.5. Using zc = 1.96 and E = 0.03, you can solve for n.
zc 2 1.96 2
n = pn qn a b = 10.5210.52 a b ≈ 1067.11
E 0.03
Because n is a decimal, round up to the nearest whole number, 1068.
2. You have a preliminary estimate of pn = 0.31. So, qn = 0.69. Using zc = 1.96
and E = 0.03, you can solve for n.
zc 2 1.96 2
n = pn qn a b = 10.31210.692 a b ≈ 913.02
E 0.03
Because n is a decimal, round up to the nearest whole number, 914.
Interpretation With no preliminary estimate, the minimum sample size should be at least 1068 registered
voters. With a preliminary estimate of
np = 0.31, the sample size should be at least 914 registered voters. So, you will 24 have had an HIV test. (Source:
CDC/NCHS, National Health Interview Survey)