Anda di halaman 1dari 226

Confidence Intervals

Inferential Statistics
• The process of using information
obtained using a sample and
generalizing and using this
information to draw conclusions
about or to make generalizations
about a population
Recall
• Parameters – calculations made from
population data
ƒ For example, the population mean, μ,
and the population standard deviation,
σ
• Statistics – calculations made from
sample data
ƒ For example the sample mean, x , and
the sample standard deviation, s
Estimation
• Using sample data to estimate a
value for an unknown parameter
such as the population mean, μ, or
the population standard deviation, σ
Central Limit Theorem
• Given a random variable X with
population mean μ and (population)
standard deviation σ, for a random
sample of size n taken from this
population, the sampling distribution
of x becomes approximately normal
as the sample size n increases. The
mean of the distribution is μ⎯x = μ
and the standard deviation is σx= σ .
n
Point Estimate
• A point estimate for a parameter
ƒ The value of a statistic that
estimates the value of the
parameter
• Example: sample mean, x, is a
point estimate for the population
mean,μ
Other Point Estimates for the
Population Mean
• In addition to using the sample
mean, x , as a point estimate for
the population mean, μ, we could
use
ƒ Sample median
ƒ Mode
Other Point Estimates for the
Population Mean
• In addition to using the sample
mean, x , as a point estimate for
the population mean, μ, we could
use
ƒ Sample median
ƒ Mode
• Any measure of center
Other Point Estimates for the
Population Mean
• In addition to using the sample
mean, x, as a point estimate for
the population mean, μ, we could use
ƒ Sample median
ƒ Mode
• Q: Which point estimate should be
used to estimate the population
mean?
Use Sample Mean as Point
Estimate for Population Mean
• Unbiased estimator: A statistic is
an unbiased estimator if its
expected value is the same as the
value of the parameter.
Use Sample Mean as Point
Estimate for Population Mean
• Unbiased estimator: A statistic is
an unbiased estimator if its
expected value is the same as the
value of the parameter. Recall - a
statistic is unbiased if it does not
systematically overestimate or
underestimate the value of the
parameter that it estimates
Use Sample Mean as Point
Estimate for Population Mean
• Unbiased estimator: A statistic is an
unbiased estimator if its expected value
is the same as the value of the
parameter. Recall – by Central Limit
Theorem, if we took many samples
of size n and calculated the mean
for each sample then the mean of
the sample means would be μ
Use Sample Mean as Point
Estimate for Population Mean
• Consistent: the sample mean
provides consistent estimates of the
population mean since as the sample
size increases, the value of the
sample mean gets closer to the
value of the population mean
Use Sample Mean as Point
Estimate for Population Mean
• Efficiency: the majority of sample
means will be “close” to the value
of the population mean
Best Point Estimate
• The sample mean, x, is the best
point estimate for the population
mean, μ
Caution
• In using point estimates, we never
know
ƒ If the statistic is correct
• We can determine if the procedure
used to produce the statistic is
ƒ Unbiased
ƒ Consistent
ƒ Efficient
Confidence Interval
• A confidence interval estimate of a
parameter consists of an interval of
numbers together with a measure
of the likelihood that the interval
contains the unknown parameter
Confidence Interval
• A confidence interval estimate of a
parameter consists of an interval of
numbers together with a measure
of the likelihood, the Level of
Confidence, that the interval
contains the unknown parameter
Confidence Interval
• A confidence interval estimate of a
parameter consists of a pair of
numbers with an associated
probability, the Level of
Confidence, that the unknown
parameter will lie between the two
numbers
Level of Confidence
• The level of confidence in a
confidence interval is the
percentage of intervals that will
contain μ if a large number of
repeated samples are taken.
Level of Confidence
• The level of confidence in a
confidence interval is the
percentage of intervals that will
contain μ if a large number of
repeated samples are taken.
• Associated formula: (1 – α) 100%

Values of α in (1 - α)•100%
Level of Confidence Value of α
(1 – α)•100%
90% 0.1
95% 0.05
96% 0.04
99% 0.01
Values of α in (1 - α)•100%
Level of Confidence Value of α Area in each “Tail”,
(1 – α)•100% α/2
90% 0.1 0.05
95% 0.05 0.025
96% 0.04 0.02
99% 0.01 0.005
Values of α in (1 - α)•100%
Level of Value of α Area in each Critical Value,
Confidence “Tail”, Zα/2
(1 – α)•100% α/2
90% 0.1 0.05 1.645
95% 0.05 0.025 1.96
96% 0.04 0.02 2.05*
99% 0.01 0.005 2.575

*Note: Value from table of Standard Normal Probabilities.


Values of α in (1 - α)•100%
Level of Value of α Area in each Critical Value,
Confidence “Tail”, Zα/2
(1 – α)•100% α/2
90% 0.1 0.05 1.645
95% 0.05 0.025 1.96
96% 0.04 0.02 2.053748911
99% 0.01 0.005 2.575

*Note: Value obtained using TI-84 Plus Silver Edition.


Values of α in (1 - α)•100%
Level of Value of α Area in each Critical Value,
Confidence “Tail”, Zα/2
(1 – α)•100% α/2
90% 0.1 0.05 1.645
95% 0.05 0.025 1.96
96% 0.04 0.02 2.054
99% 0.01 0.005 2.575

*Note: Approximate value expressed to three decimal places


Values of α in (1 - α)•100%
Level of Value of α Area in each Critical Value,
Confidence “Tail”, Zα/2
(1 – α)•100% α/2
90% 0.1 0.05 1.645
95% 0.05 0.025 1.96
96% 0.04 0.02 2.054
99% 0.01 0.005 2.576*

*Using the z* row of the t-Distribution table available at


http://www.framingham.edu/~smabrouk/online_stats/handout
s/pdf/t-Distribution_Table.pdf
Values of α in (1 - α)•100%
Level of Value of α Area in each Critical Value,
Confidence “Tail”, Zα/2
(1 – α)•100% α/2
90% 0.1 0.05 1.645
95% 0.05 0.025 1.96
96% 0.04 0.02 2.054
99% 0.01 0.005 2.576*

*Using the z* row of the t-Distribution table available at


http://www.framingham.edu/~smabrouk/online_stats/handout
s/pdf/t-Distribution_Table.pdf
Values of α in (1 - α)•100%
Level of Value of α Area in each Critical Value,
Confidence “Tail”, Zα/2
(1 – α)•100% α/2
90% 0.1 0.05 1.645
95% 0.05 0.025 1.96
96% 0.04 0.02 2.054
99% 0.01 0.005 2.576
Interpretation of
Confidence Interval
• A (1 - α)•100% confidence interval
tells us that if we were to obtain
many simple random samples of size
n from a population for which the
mean μ is unknown then
approximately (1 - α)•100% of the
intervals would contain the value of
the population mean μ.
Central Limit Theorem
• Given a random variable X with
population mean μ and (population)
standard deviation σ, for a random
sample of size n taken from this
population, the sampling distribution
of x becomes approximately normal
as the sample size n increases. The
mean of the distribution is μx =μ
and the standard deviation is σx = σ .
n
By the Central Limit Theorem

• We know that
ƒ The distribution for the sample
means is approximately normal
ƒ The standard deviation for the
sample means is σx = σ
n for known
population standard deviation σ
Constructing a (1 - α)•100%
Confidence Interval about μ with Known
Population Standard Deviation σ
Constructing a (1 - α)•100%
Confidence Interval about μ with Known
Population Standard Deviation σ
• Suppose a simple random sample size n is
taken from a population with unknown
mean μ and known standard deviation σ.
• A (1 - α)•100% confidence interval for
μ is determined by
ƒ Lower bound: x −zα/2 σn
ƒ Upper bound: x + zα/2 σn
for zα/2 the critical z-value
• Time in Bed: How much do Americans
sleep each night? Based on a random
sample of 1120 Americans 15 years of
age or older, the mean amount of sleep
per night is 8.17 hours according to
the American Time Use Survey
conducted by the Bureau of Labor
Statistics. Assuming the population
standard deviation for the amount of
sleep per night is 1.2 hours, construct
a 95% confidence interval for the mean
amount of sleep per night for
Americans 15 years of age or older.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Note: It is not realistic to know the population


standard deviation especially when we are trying to
determine the confidence interval for the population
mean.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• We know the values for the sample size, n = 1120,


the sample mean, x = 8.17 , and the population
standard deviation, σ = 1.2.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• For a 95% confidence interval, α = 0.05


and α/2 = 0.025.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• For a 95% confidence interval, α = 0.05


and α/2 = 0.025.
• Then, zα/2 = 1.96 is the critical value.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Using these values (n = 1120, x = 8.17 , σ = 1.2,


and zα/2 = 1.96), we calculate the lower bound and
the upper bound for the confidence interval.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Lower bound: x-zα/2 σ =8.17-1.96 1.2


n 1120
≈ 8.099720558
≈ 8.10
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Lower bound: x-zα/2 σ =8.17-1.96 1.2


n 1120
≈ 8.099720558
≈ 8.10
Notice that we do NOT approximate any part of this
calculation.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Lower bound: x-zα/2 σ =8.17-1.96 1.2


n 1120
≈ 8.099720558
≈ 8.10
We only approximate the final value …
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Lower bound: x-zα/2 σ =8.17-1.96 1.2


n 1120
≈ 8.099720558
≈ 8.10
We only approximate the final value, never any of
the intermediate parts.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Upper bound: x + zα/2 σ =8.17 +1.96 1.2


n 1120
≈ 8.240279442
≈ 8.24
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Upper bound: x + zα/2 σ =8.17 +1.96 1.2


n 1120
≈ 8.240279442
≈ 8.24
Notice that we do NOT approximate any part of this
calculation.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Upper bound: x + zα/2 σ =8.17 +1.96 1.2


n 1120
≈ 8.240279442
≈ 8.24
We only approximate the final value …
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Upper bound: x + zα/2 σ =8.17 +1.96 1.2


n 1120
≈ 8.240279442
≈ 8.24
We only approximate the final value, never any of
the intermediate parts.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Confidence Interval:
ƒ 8.10 ≤ μ ≤ 8.24 (Inequality Notation)
ƒ μ ∈ [8.10, 8.24] (Interval Notation)
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• We are 95% confident that the population mean, μ,


is such that 8.10 ≤ μ ≤ 8.24.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• We are 95% confident that the population mean, μ,


is in the interval [8.10, 8.24].
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• We are 95% confident that the population mean, μ,


is between 8.10 hours and 8.24 hours, inclusive.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• We are 95% confident that the population mean, μ,


is between 8.10 hours and 8.24 hours, inclusive.

This would be a meaningful statement for the


confidence interval.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• We are 95% confident that the population mean, μ,


is between 8.10 hours and 8.24 hours, inclusive.

• What does this mean?


• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Our statement tells us that if we were to select


many different samples of size 1120 and construct
the corresponding confidence intervals, 95% of these
confidence intervals would actually contain the value
of the population mean.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• Note: The level of confidence, in this case 95%,


refers to the success rate of the process being used
to estimate the mean.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• CAUTION: The level of confidence, in this case


95%, is NOT a probability!!! It would be
INCORRECT to say that 95% of sample means would
be between 8.10 hours and 8.24 hours, inclusive.
• Time in Bed: How much do Americans sleep each
night? Based on a random sample of 1120 Americans
15 years of age or older, the mean amount of sleep
per night is 8.17 hours according to the American
Time Use Survey conducted by the Bureau of Labor
Statistics. Assuming the population standard
deviation for the amount of sleep per night is 1.2
hours, construct a 95% confidence interval for the
mean amount of sleep per night for Americans 15
years of age or older.

• CAUTION: The level of confidence, in this case


95%, is NOT a probability!!! It would be WRONG to
say that the probability that the population mean is
between 8.10 hours and 8.24 hours, inclusive, is
0.95.
Margin for Error
• The margin for error in a
confidence interval for which σ is
known is determined by
E =zα/2 σ
n

for sample size n.


Margin for Error
• The margin for error in a
confidence interval for which σ is
known is determined by
E =zα/2 σ
n

for sample size n.

• Note: The distribution must be


approximately normal or the sample
size must be at least 30.
Margin for Error
• If we know the margin for error,
E =zα/2 σ ,
n

that we are willing to tolerate then


we can determine the sample size
necessary to achieve the desired
margin for error.
Margin for Error
• If we know the margin for error,
E =zα/2 σ ,
n

that we are willing to tolerate then


we can determine the sample size
necessary to achieve the desired
margin for error.

• HOW???
Margin for Error
• If we know the margin for error
that we are willing to tolerate then
we can determine the sample size
necessary to achieve the desired
margin for error.

• Solve E =zα/2 σ for n.


n

E = zα / 2 ⋅ σ
n
Multiply both sides
E n = zα / 2 ⋅σ by the square root
of n and divide both
sides by E.
zα / 2 ⋅σ
n=
E
2
z

⋅σ ⎟ ⎞
Square both sides and,
n= ⎜ α/2
⎜ ⎟


E ⎟

to simplify the
expression, square
each value on the
2 right-hand side, if
z

⎜⎜ ⎟⎟

⋅σ2 desired. Either form
n= ⎝ α/2⎠ can be used.
E2
Determining the Sample Size n
• The sample size n required to estimate
the population mean μ with a level of
confidence of (1 - α)•100% with a
specified margin of error E is
determined by
2
n= z •σ
⎛ ⎞
⎜ ⎟
⎜ α/2 ⎟

⎜⎜

E ⎟
⎟⎟

where n is rounded up to the nearest


whole number.
• In “real life” will we know the
population standard deviation???
• In “real life” will we know the
population standard deviation???

Most likely not!


• How Much TV Do Teenagers Watch? A
Gallup poll conducted January 17 –
February 6, 2005, asked 1028
teenagers 13 years of age to 17 years
of age , “Typically, how many hours per
week do you spend watching TV?” The
teenagers who participated in the survey
watched, on average, 13.0 hours of
television per week with a standard
deviation of 2.3 hours per week.
Construct a 95% confidence interval for
the average number of hours that
teenagers watch TV each week.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We would like to construct a 95% confidence interval.


• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• For a 95% confidence interval, α = 0.05


and α/2 = 0.025.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• It is important to note that the mean and standard


deviation given in the scenario are the
ƒ sample mean and
ƒ the sample standard deviation.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• That is, since the


ƒ sample mean and
ƒ the sample standard deviation
are provided, the population standard deviation, σ, is
unknown.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We know the values for the sample size, n = 1028,


the sample mean, x = 13.0 , and the sample standard
deviation, s = 2.3.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We know the values for the sample size, n = 1028,


the sample mean, x = 13.0 , and the sample standard
deviation, s = 2.3.
• Notice that the sample size, 1028, is at least 30.
ƒ That is, n ≥ 30.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We cannot use the standard normal distribution to


determine the critical value for the confidence
interval since we do not know the population standard
deviation.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Instead, we use the t-distribution to determine the


critical value for the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• That is, we use the t-distribution to determine the


critical value in order to calculate the lower bound
and the upper bound for the confidence interval.
What is the t-Distribution?
Properties of t-Distribution
• The t-Distribution changes based on
the number of degrees of freedom
Properties of t-Distribution
• The t-Distribution is centered at 0
and is symmetric about 0.
Properties of t-Distribution
• The area under the curve is 1.
Properties of t-Distribution
• The area under the curve is to the
left of 0 is ½.
Properties of t-Distribution
• The area under the curve is to the
right of 0 is ½.
Properties of t-Distribution
• The curve never touches the horizontal
axis.
Properties of t-Distribution
• The t-distribution is similar to the
standard normal distribution.
Properties of t-Distribution
• The area in the “tails” of the t-
Distribution is a little greater than the
area in the tails of the standard
normal distribution.
Properties of t-Distribution
• As the sample size n increases, the
density curve of t gets closer to the
standard normal density curve. (Law of
Large Numbers)
Properties of t-Distribution
• As the sample size n increases, the
density curve of t gets closer to the
standard normal density curve. (Law of
Large Numbers)

Check this out …


Properties of t-Distribution
• As the sample size n increases, the
density curve of t gets closer to the
standard normal density curve. (Law of
Large Numbers)

Check this out using the Probability Density


Function Comparison Tool …
Properties of t-Distribution
• As the sample size n increases, the
density curve of t gets closer to the
standard normal density curve. (Law of
Large Numbers)

Check this out using the Probability Density


Function Comparison Tool available at
http://www.framingham.edu/~smabrouk/Interactive/.
Constructing a (1 - α)•100%
Confidence Interval
about μ and σ unknown
Constructing a (1 - α)•100%
Confidence Interval
about μ and σ unknown
• Suppose a simple random sample size n is
taken from a population with unknown
mean μ and an unknown standard deviation
σ.
• A (1 - α)•100% confidence interval for
μ is determined by
x − t s
ƒ Lower bound: α/2
n
x + t s
ƒ Upper bound: α/2
n
for tα/2 computed with n-1 degrees of
freedom
Constructing a (1 - α)•100%
Confidence Interval
about μ and σ unknown
• Suppose a simple random sample size n is
taken from a population with unknown
mean μ and an unknown standard deviation
σ.
• A (1 - α)•100% confidence interval for
μ is determined by
x − t s
ƒ Lower bound: α/2
n
x + t s
ƒ Upper bound: α/2
n
for critical value tα/2 computed with n-1
degrees of freedom
Margin for Error
• The margin for error, E, in a
confidence interval for which sample
standard deviation is known is
determined by
E = tα/2 s
n

for sample size n.


Margin for Error
• If we know the margin for error,
E, that we are willing to tolerate
then we can determine the sample
size, n, necessary to achieve the
desired margin for error.
Using Margin for Error to
Determine Sample Size
• In order to determine the sample
size, we must know
ƒ the margin for error, E,
ƒ the level of confidence (from
which we determine the critical
value, tα/2), and
ƒ an approximate value for sample
standard deviation, s.
Using Margin for Error to
Determine Sample Size
• If we know margin for error E,
critical value tα/2 , and an
approximate value for sample
standard deviation s then we can
solve
E = tα/2 s
n

for sample size, n.


Multiply
E = tα/2 s both sides
n

by n .
nE = tα/2s
Divide both
n = tα/2 s

sides by
E E.
2
n= t


⎜ α/2 •
s ⎞
⎟ Square both
E




⎠ sides.
Using Margin for Error to
Determine Sample Size
• If we know the margin for error,
E, the level of confidence (from
which we determine the critical
value, tα/2), and an approximate value
for sample standard deviation, s,
2
then the sample size is
n = tα/2

⎜ s⎟

E
⎜ ⎟
⎜ ⎟
⎝ ⎠
rounded up to the nearest whole
number.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Since n = 1028, there are 1027 degrees of freedom.


Since our table does not list 1027 degrees of
freedom, we will use the next closest, 1000 degrees
of freedom …
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Since n = 1028, there are 1027 degrees of freedom.


Since our table does not list 1027 degrees of
freedom, we will use the next closest, 1000 degrees
of freedom, without going over n – 1 = 1027.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• That is, since our t-distribution table,


http://www.framingham.edu/~smabrouk/online_stats/h
andouts/pdf/t-Distribution_Table.pdf, does not
include a row for 1027, we must use the next closest
row, 1000 …
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• That is, since our t-distribution table,


http://www.framingham.edu/~smabrouk/online_stats/h
andouts/pdf/t-Distribution_Table.pdf, does not
include a row for 1027, we must use the next closest
row, 1000, without going over 1027.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• If we had a table containing information for 1027


degrees of freedom, we could use that table.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• If we had a table containing information for 1027


degrees of freedom, we could use that table.
ƒ There are such tables available online as well as
electronic tables and software that can provide
this information as well.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• It is important to note that we cannot use the z*


row: the z* row corresponds to an infinite number of
degrees of freedom, and the number of degrees of
freedom in this case is 1027 and not infinite.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• It is important to note that we cannot use the z*


row: the z* row corresponds to an infinite number of
degrees of freedom.
ƒ The z* row is used when critical values are taken
from the standard normal table.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• The critical value corresponding to α = 0.025 and


1000 degrees of freedom is tα/2 = 1.962.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We use our values n = 1028, x = 13.0 , s = 2.3, and


tα/2 = 1.962 to determine the lower bound and the
upper bound for the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Lower bound: x − tα/2 s =13.0-1.962 2.3


n 1028
≈12.85925587
≈12.9
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Lower bound: x − tα/2 s =13.0-1.962 2.3


n 1028
≈12.85925587
≈12.9
Notice that we do NOT approximate any part of this
calculation.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Lower bound: x − tα/2 s =13.0-1.962 2.3


n 1028
≈12.85925587
≈12.9
We only approximate the final value …
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Lower bound: x − tα/2 s =13.0-1.962 2.3


n 1028
≈12.85925587
≈12.9
We only approximate the final value, never any of
the intermediate parts.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Upper bound: x + tα/2 s =13.0 +1.962 2.3


n 1028
≈13.14074413
≈13.1
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Upper bound: x + tα/2 s =13.0 +1.962 2.3


n 1028
≈13.14074413
≈13.1
Notice that we do NOT approximate any part of this
calculation.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Upper bound: x + tα/2 s =13.0 +1.962 2.3


n 1028
≈13.14074413
≈13.1
We only approximate the final value …
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Upper bound: x + tα/2 s =13.0 +1.962 2.3


n 1028
≈13.14074413
≈13.1
We only approximate the final value, never any of
the intermediate parts.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Confidence Interval:
ƒ 12.9 ≤ μ ≤ 13.1 (Inequality Notation)
ƒ μ ∈ [12.9, 13.1] (Interval Notation)
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the population mean, μ, is


such that 12.9 ≤ μ ≤ 13.1.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the population mean, μ, is


in the interval [12.9, 13.1] .
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the population mean is


between 12.9 hours per week and 13.1 hours per
week, inclusive.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the population mean is


between 12.9 hours per week and 13.1 hours per
week, inclusive.
• What does this mean in context of this scenario?
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the population mean is


between 12.9 hours per week and 13.1 hours per
week, inclusive.
• That is, what is the interpretation of the confidence
interval?
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the population mean is


between 12.9 hours per week and 13.1 hours per
week, inclusive.
• That is, what is the interpretation of the confidence
interval, in the context of this scenario?
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that, on average, teenagers


13 years of age to 17 years of age spend between
12.9 hours per week and 13.1 hours per week,
inclusive, watching TV.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that teenagers 13 years of


age to 17 years of age spend, on average, between
12.9 hours and 13.1 hours, inclusive, watching TV
each week.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the average amount of


time that teenagers 13 years of age to 17 years of
age spend watching TV each week is between 12.9
hours and 13.1 hours, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the mean amount of time


that teenagers 13 years of age to 17 years of age
spend watching TV each week is between 12.9 hours
and 13.1 hours, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the average number of


hours that teenagers 13 years of age to 17 years of
age spend watching TV each week is between 12.9
hours and 13.1 hours, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the mean number of


hours that teenagers 13 years of age to 17 years of
age spend watching TV each week is between 12.9
hours and 13.1 hours, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the average number of


hours that teenagers 13 years of age to 17 years of
age watch TV each week is between 12.9 hours and
13.1 hours, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the mean number of


hours that teenagers 13 years of age to 17 years of
age watch TV each week is between 12.9 hours and
13.1 hours, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the average number of


hours that teenagers 13 years of age to 17 years of
age watch TV is between 12.9 hours per week and
13.1 hours per week, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• We are 95% confident that the mean number of


hours that teenagers 13 years of age to 17 years of
age watch TV is between 12.9 hours per week and
13.1 hours per week, inclusive.
This is an interpretation of the confidence interval.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Notice that none of these interpretations of the


confidence interval mentions any of the sample
information. This confidence interval mentions the
mean …
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Notice that none of these interpretations of the


confidence interval mentions any of the sample
information. The interpretation of the confidence
interval mentions the mean and
ƒ this mean is the population mean.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• So, if we were to select many different samples of


size 1028 and construct the corresponding confidence
intervals, 95% of these confidence intervals would
actually contain the population mean.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Again, the 95% is the success rate of the process


being used to estimate the population mean.
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• Note: The 95% is NOT a probability.


• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• CAUTION: Saying that there is a 95% chance that


the population mean is in the confidence interval is
INCORRECT!!!
• How Much TV Do Teenagers Watch? A Gallup poll
conducted January 17 – February 6, 2005, asked
1028 teenagers 13 years of age to 17 years of age ,
“Typically, how many hours per week do you spend
watching TV?” The teenagers who participated in the
survey watched, on average, 13.0 hours of television
per week with a standard deviation of 2.3 hours per
week. Construct a 95% confidence interval for the
average number of hours that teenagers watch TV
each week.

• CAUTION: It is WRONG to say that 95% of sample


means are in the confidence interval.
Choosing between the Normal
Distribution and the t-Distribution
for Confidence Intervals
σ known and population
Use Normal Distribution normally distributed
(z) OR
σ known and n ≥ 30
σ not known and
population normally
Use t-Distribution distributed
(t) OR
σ not known and n ≥ 30
Round-off Rule for Confidence
Intervals used to Estimate μ
• When using original sample data to
construct a confidence interval, round
the limits of the confidence interval to
one more decimal place than the original
data.
• When using given values of ⎯x, n, and s
with the sample data unknown, round
the limits of the confidence interval to
the same number of decimal places as
the sample mean.
Constructing a (1 - α)•100% Confidence
Interval for the Population Proportion p
Constructing a (1 - α)•100% Confidence
Interval for the Population Proportion p
• Suppose a simple random sample size n is
taken from a population.
A (1 - α)•100% confidence interval for p is
determined by

ƒLower bound: p̂ − zα/2


(
pˆ 1 − pˆ )
n

ƒUpper bound: p̂ + zα/2


(
pˆ 1 − pˆ )
n

for which …
Constructing a (1 - α)•100% Confidence
Interval for the Population Proportion p
• Suppose a simple random sample size n is
taken from a population.
A (1 - α)•100% confidence interval for p is
determined by
p̂ − zα/2
(
pˆ 1 − pˆ )
ƒLower bound:
n

p̂ + zα/2
(
pˆ 1 − pˆ )
ƒUpper bound:
n
(
ˆ 1 − pˆ ≥ 10
for which both n ≤ 0.05N and np )
must be satisfied.
Constructing a (1 - α)•100% Confidence
Interval for the Population Proportion p
• For a simple random sample size n,

ƒ Lower bound is p̂ − zα/2


(
pˆ 1 − pˆ )
n

ƒ Upper bound is p̂ + zα/2


(
pˆ 1 − pˆ )
n
ˆ 1 − pˆ ≥ 10
for which both n ≤ 0.05N and np ( )
must be satisfied.

The margin for error is E = zα /2


(
pˆ 1 − pˆ ).
n
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.*
(a) Construct and interpret the 99%
confidence interval for the proportion of U.S.
households having a telephone.
(b) Based on this confidence interval, should
those conducting surveys by telephones be
concerned?

*based on data from the U.S. Census Bureau


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒThe households were randomly selected.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent,
~ that is, n ≤ 0.05N.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent,
~ that is, n ≤ 0.05N.

~ By examining this inequality, n ≤ 0.05N,


we are determining if the sample size,
n, is no more than 5% of the population
size, N.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent.
~ The functional form of this inequality,

n ≤ 0.05N, is n/0.05 ≤ N since we do


not know the exact population size, N,
but we do know the sample size, n.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent.
~ Since 4276 households were surveyed,

n = 4276.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent.
~ Since 4276 households were surveyed,

n = 4276.
n 4276
=
0.05 0.05
= 85,520
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent.
~ Since there are far more than 85,520

U.S. households, the sample size is no


more than 5% of the population size and

• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that these households
are independent.
~ Since there are far more than 85,520

U.S. households, the sample size is no


more than 5% of the population size and,
thus, the households surveyed are
independent.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒNext, we must make sure that distribution
for the sample proportion is approximately
normal.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒNext, we must make sure that distribution
for the sample proportion is approximately
normal, that is, np( )
ˆ 1 − pˆ ≥ 10 .
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.
n = 4276
4019
p̂ =
4276
4019
( )
1 − pˆ = 1 −
4276
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.
n = 4276
4019
p̂ =
4276
257
( )
1 − pˆ =
4276
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.

⎛ 4019 ⎞ ⎛ 257 ⎞
( )
npˆ 1 − pˆ = 4276 ⎜ ⎟ ⎜ ⎟
⎝ 4276 ⎠ ⎝ 4276 ⎠
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.

⎛ 4019 ⎞ ⎛ 257 ⎞
( )
npˆ 1 − pˆ = 4276 ⎜ ⎟ ⎜ ⎟
⎝ 4276 ⎠ ⎝ 4276 ⎠
⎛ 257 ⎞
= 4019 ⎜ ⎟
⎝ 4276 ⎠
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.

⎛ 4019 ⎞ ⎛ 257 ⎞
( )
npˆ 1 − pˆ = 4276 ⎜
4276 ⎟ ⎜ 4276 ⎟
⎝ ⎠⎝ ⎠
≈ 241.553554724
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.

⎛ 4019 ⎞ ⎛ 257 ⎞
( )
npˆ 1 − pˆ = 4276 ⎜
4276 ⎟ ⎜ 4276 ⎟
⎝ ⎠⎝ ⎠
≈ 241.6
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.

⎛ 4019 ⎞ ⎛ 257 ⎞
( )
npˆ 1 − pˆ = 4276 ⎜ ⎟ ⎜ ⎟
⎝ 4276 ⎠ ⎝ 4276 ⎠
≥ 10
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Before we can construct the 99% confidence
interval we have some checks to perform.
ƒWe must make sure that distribution for the
sample proportion is approximately normal.

Since we have established that


( )
npˆ 1 − pˆ ≥ 10 , we know that the
distribution for the sample proportion is
approximately normal.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Having established the necessary conditions,
ƒthe surveyed households are independent
ƒthe distribution for the sample proportion is
approximately normal
we can construct the 99% confidence interval.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Having established the necessary conditions,
ƒthe surveyed households are independent
ƒthe distribution for the sample proportion is
approximately normal
we can construct the 99% confidence interval.

For a 99% confidence interval, …


Values of α in (1 - α)•100%
Level of Value of α Area in each Critical Value,
Confidence “Tail”, Zα/2
(1 – α)•100% α/2
90% 0.1 0.05 1.645
95% 0.05 0.025 1.96
96% 0.04 0.02 2.054
99% 0.01 0.005 2.576
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Having established the necessary conditions,
ƒthe surveyed households are independent
ƒthe distribution for the sample proportion is
approximately normal
we can construct the 99% confidence interval.

For a 99% confidence interval, the critical


value is zα/2 = 2.576.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is p̂ − zα/2


(
pˆ 1 − pˆ ).
n
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is p̂ − zα/2


(
pˆ 1 − pˆ )…
n
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is


⎛ 4019 ⎞ ⎛ 4019 ⎞
⎜ ⎟ ⎜ 1− ⎟
pˆ (1 − pˆ ) 4019 ⎝ 4276 ⎠ ⎝ 4276 ⎠
p̂ − z = − 2.576
.
n 4276 4276
α/2
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is


⎛ 4019 ⎞ ⎛ 257 ⎞
pˆ (1 − pˆ ) 4019
⎜ 4276 ⎟ ⎜ 4276 ⎟
p̂ − z = − 2.576
⎝ ⎠ ⎝ ⎠
α/2
n 4276 4276 .
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is

p̂ − z
(
pˆ 1 − pˆ ) =
4019
− 2.576
(
4019 257 ).
3
n 4276 4276
α/2
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is

p̂ − z
(
pˆ 1 − pˆ ) =
4019
− 2.576
(
4019 257 ).
3
n 4276 4276
α/2

Notice that we do NOT approximate any part


of this calculation.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is

p̂ − z
(
pˆ 1 − pˆ ) =
4019
− 2.576
(
4019 257 ).
3
n 4276 4276
α/2

We only approximate the final value …


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is

p̂ − z
(
pˆ 1 − pˆ ) =
4019
− 2.576
(
4019 257 ).
3
n 4276 4276
α/2

We only approximate the final value, never


any of the intermediate parts.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is

p̂ − z
(
pˆ 1 − pˆ ) ≈ 0.930534103214 .
n
α/2

We only approximate the final value, never


any of the intermediate parts.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The lower bound is

p̂ − z
(
pˆ 1 − pˆ ) ≈ 0.93 .
n
α/2

We only approximate the final value, never


any of the intermediate parts.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is p̂ + zα /2


(
pˆ 1 − pˆ ) .
n
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is p̂ + zα /2


(
pˆ 1 − pˆ ) …
n
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is


⎛ 4019 ⎞ ⎛ 4019 ⎞
⎜ ⎟ ⎜ 1− ⎟
pˆ (1 − pˆ ) 4019 ⎝ 4276 ⎠ ⎝ 4276 ⎠
p̂ + z = + 2.576 .
n 4276 4276
α/2
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is


⎛ 4019 ⎞ ⎛ 257 ⎞
pˆ (1 − pˆ ) 4019
⎜ 4276 ⎟ ⎜ 4276 ⎟
p̂ + z = + 2.576 ⎝ ⎠ ⎝ ⎠
α/2
n 4276 4276 .
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is

p̂ + z
(
pˆ 1 − pˆ ) =
4019
+ 2.576
(
4019 257 ).
3
n 4276 4276
α/2
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is

p̂ + z
(
pˆ 1 − pˆ ) =
4019
+ 2.576
(
4019 257 ).
3
n 4276 4276
α/2

Notice that we do NOT approximate any part


of this calculation.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is

p̂ + z
(
pˆ 1 − pˆ ) =
4019
+ 2.576
(
4019 257 ).
3
n 4276 4276
α/2

We only approximate the final value …


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is

p̂ + z
(
pˆ 1 − pˆ ) =
4019
+ 2.576
(
4019 257 ).
3
n 4276 4276
α/2

We only approximate the final value, never


any of the intermediate parts.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is

p̂ + z
(
pˆ 1 − pˆ ) ≈ 0.949260096974 .
n
α/2

We only approximate the final value, never


any of the intermediate parts.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

The upper bound is

p̂ + z
(
pˆ 1 − pˆ ) ≈ 0.95 .
n
α/2

We only approximate the final value, never


any of the intermediate parts.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

Since the lower bound is 0.93 and the upper


bound is 0.95, the confidence interval is …
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

Since the lower bound is 0.93 and the upper


bound is 0.95, the confidence interval is …
ƒ 0.93 ≤ p ≤ 0.95 (Inequality Notation)
ƒ p ∈ [0.93, 0.95] (Interval Notation)
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that the population


proportion, p, is such that 0.93 ≤ p ≤ 0.95.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that the population


proportion, p, is such that p ∈ [0.93, 0.95].
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that the population


proportion, p, is in the interval [0.93, 0.95].
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• What does confidence interval actually


mean????
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• What does confidence interval actually mean in


the context of this scenario????
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• What does this actually mean????


ƒThat is, what is the interpretation of the
confidence interval?
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that the proportion of


U.S. households that have a telephone is
between 93% and 95%, inclusive.

This is an interpretation of the confidence interval.


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that the proportion of


U.S. households having a telephone is between
93% and 95%, inclusive.

This is an interpretation of the confidence interval.


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that the percentage of


U.S. households that have a telephone is
between 93% and 95%, inclusive.

This is an interpretation of the confidence interval.


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that the percentage of


U.S. households having a telephone is between
93% and 95%, inclusive.

This is an interpretation of the confidence interval.


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• We are 99% confident that between 93% and


95%, inclusive, of U.S. households that have
a telephone.

This is an interpretation of the confidence interval.


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• Notice that none of our interpretations


includes any of the sample information. The
interpretation of the confidence interval
mentions the proportion or the percentage and
ƒthis proportion or percentage is the
population proportion.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• So, if we were to select many different


samples of size 4276 and construct the
corresponding confidence intervals, 99% of
these confidence intervals would actually
contain the population proportion.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• Again, 99% is the success rate for the


process being used to estimate the population
proportion.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• Note: The 99% is NOT a probability.


• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• CAUTION: Saying that there is a 99%


chance that the population proportion is in the
confidence interval is INCORRECT!!!
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
99% confidence interval for the proportion of
U.S. households having a telephone

• CAUTION: It is WRONG to say that 99% of


sample proportions are in the confidence
interval.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Based on this confidence interval, should
those conducting surveys by telephones be
concerned?
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Based on this confidence interval, should
those conducting surveys by telephones be
concerned?

• Since, according to the 99% confidence


interval, between 93% and 95%, inclusive, of
U.S. households that have a telephone, those
conducting surveys by telephone can reach at
least 93% of U.S. households.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Based on this confidence interval, should
those conducting surveys by telephones be
concerned?

• However, this also means that there may be


as much as 7% of U.S. households that cannot
be reached by telephone.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Based on this confidence interval, should
those conducting surveys by telephones be
concerned?

• So, while the percentage of U.S. households


that have a telephone has increased from 35%
in 1920 to at least 93% now, not all U.S.
households have a telephone.
• In 1920, only 35% of U.S. households had
telephones. A recent survey of 4276
randomly selected households showed that
4019 households had telephones.
Based on this confidence interval, should
those conducting surveys by telephones be
concerned?

• So, while the percentage of U.S. households


that have a telephone has increased from 35%
in 1920 to at least 93% now, not all U.S.
households have a telephone. Therefore, as of
yet, not all U.S. households can be reached
by telephone.
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
ƒ For example,
~ population mean

~ population proportion
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
• You MUST be careful not to approximate any
values used in calculating
ƒ the lower bound or
ƒ the upper bound
for the confidence interval.
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
• You MUST be careful not to approximate any
values used in calculating
ƒ the lower bound or
ƒ the upper bound
for the confidence interval.
• You can ONLY approximate the final value for
the lower and upper bound for the confidence
interval.
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
• You MUST be careful not to approximate any
values used in calculating
ƒ the lower bound or
ƒ the upper bound
for the confidence interval.
• You CANNOT calculate the lower and upper
bound piecemeal, approximate any values used
in the calculations, and then combine the
parts.
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
• You MUST be careful not to approximate any
values used in calculating
ƒ the lower bound or
ƒ the upper bound
for the confidence interval.
• You CANNOT calculate the lower and upper
bound piecemeal, approximating any values.
Doing so, produces errors and combining such
values makes the lower and upper bounds
useless and meaningless.
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
• You MUST be careful not to approximate any
values used in calculating
ƒ the lower bound or
ƒ the upper bound
for the confidence interval.
• You CANNOT calculate the lower and upper
bound piecemeal, approximating any values.
• You are interested in obtaining the most
accurate values possible.
• Notice that confidence intervals, which are
determined based on sample information,
provide a great deal of information about
population parameters.
• You MUST be careful not to approximate any
values used in calculating
ƒ the lower bound or
ƒ the upper bound
for the confidence interval.
• You CANNOT calculate the lower and upper
bound piecemeal, approximating any values.
• Without precision and accuracy, statistics and
parameters are meaningless.