Sampling Distribution of The Sample Proportion

Sampling distribution of the sample
proportion
In the same way that we were able to find a sampling distribution for the
sample mean (SDSM), we can find a sampling distribution for the sample
proportion (SDSP).
In other words, if we’re dealing with a population proportion p, instead of a

population mean μ, then we’ll be trying to create a sampling distribution of
the sample proportion, as opposed to a sampling distribution of the
sample mean.
Sampling distribution of the sample proportion

Often we’ll want to calculate a population proportion p, which is the
number of subjects in our population that meet a certain condition.
For example, maybe we want to know how many students in our school
have brown hair. If there are 5,000 students who attend our school, it might
not be possible to survey everybody. So instead we could take a random
sample of 100 students and see how many of them have brown hair. This is
the sample proportion, since it’s the proportion of students in the sample
with brown hair, which is given by
x
p̂ =
n
where p̂ is the sample proportion, x is the number of students in the

sample with brown hair (the number of “successes”), and n is the sample
size (we surveyed 100 students for our sample).
235
Just like for the SDSM, the sampling distribution of the sample proportion
(SDSP) is created when we take every possible sample from our
population, calculate the sample proportion for each sample, and then plot
all of those sample proportions into a probability distribution.
In other words, the SDSP is the probability distribution of all possible

sample proportions p.̂
Central Limit Theorem

Remember that, whenever we’re dealing with a proportion, the distribution
is a binomial distribution, since every outcome is either a “success” or a
“failure.” So in the case of a population proportion, the original population
will be modeled by a binomial distribution, not a normal distribution.
But even though the population follows a binomial distribution, we can still
use the Central Limit Theorem to create a sampling distribution of the
sample proportion. Just like the SDSM, the CLT tells us that the SDSP is
only guaranteed to be normally distributed when we use a sample size of
at least n = 30.
Mean, variance, and standard deviation

The mean of the sampling distribution of the sample proportion μp̂ will be
equal to the population proportion p.
μp̂ = p
236
The standard deviation of the sampling distribution of the sample
proportion σp,̂ also called the standard error of the proportion, will be
p(1 − p)
SE = σp̂ =
n
where p is the population proportion and n is the sample size. We use this
formula for standard error of the proportion if our population is infinite, or
if the population is finite but large in comparison to our sample size (if
sample size is no more than 5 % of the population, n /N ≤ 0.05).
Of course, based on this formula for standard error, we can say that the
variance of the sampling distribution of the sample proportion is
p(1 − p)
σp2̂ =
n
If the population proportion is unknown, we estimate the population

proportion p using the next best thing, the sample proportion p,̂ and the
formulas for standard error and variance become
̂ − p)̂
p(1
SE = σp̂ =
n
̂ − p)̂
p(1
σp2̂ =
n
Finite population correction factor
237
In the case where the population is finite and n /N > 0.05, we have to apply
the finite population correction factor, and in that case the correct formula
for the standard error of the proportion is then
p(1 − p) N−n
SE = σp̂ =
n N−1
where p is the population proportion, n is the sample size, and N is the size
of the population. If we’re applying the FPC, then the formula for variance
of the SDSP is
( )(N − 1)
p(1 − p) N−n
σp2̂ =
n
And again, if the population proportion is unknown, we approximate it with

the sample proportion, and our formulas for standard error and variance
with the FPC are
̂ − p)̂
p(1 N−n
SE = σp̂ =
n N−1
( )(N − 1)
̂ − p)̂
p(1 N−n
σp2̂ =
n
Let’s do an example so that we can see how these formulas work.
Example
A group of 4 people have the following hair color: brown, brown, brown,
blonde. Find all possible random samples of size 2 if we’re sampling with
replacement. If we define brown hair as a “success,” then find the sample
238
proportion for every sample. Determine the probability distribution of the
sample proportion, the mean of the SDSP p,̂ and the standard error σp.̂
Let’s first determine the total number of possible samples, using N n, given
N = 4 and n = 2.
N n = 42 = 16
The complete sample space, and the proportion for each sample, is
239
Sample Sample proportion
brown, brown 1
brown, brown 1
brown, brown 1
brown, blonde 1/2
brown, brown 1
brown, brown 1
brown, brown 1
brown, blonde 1/2
brown, brown 1
brown, brown 1
brown, brown 1
brown, blonde 1/2
blonde, brown 1/2
blonde, brown 1/2
blonde, brown 1/2
blonde, blonde 0
Build a table for the probability distribution of the sample proportion.

Because there are 16 total samples, the probability of each sample
proportion will be given by the number of times that sample proportion
occurs, divided by the total number of possible samples, so “count/16.”
240
Sample proportion P(pi)
0 1/16
1/2 6/16
1 9/16
Now we can calculate the mean of the sampling distribution of the sample
proportion, μp,̂ where pî is a given sample proportion, P( pî ) is the
probability of that particular sample proportion occurring, and N is the
number of samples.
N
∑
μp̂ = pî P( pî )
i=1
( 16 ) 2 ( 16 ) ( 16 )
1 1 6 9
μp̂ = 0 + +1
3 9
μp̂ = +
16 16
12
μp̂ =
16
3
μp̂ =
4
Because we’re sampling with replacement, we would expect this mean of

the SDSP to be equivalent to the population proportion, μp̂ = p, and we can
see that it is if we calculate the population proportion.
3 people with brown hair 3

p= =
4 people in the population 4
241
Both proportions are μp̂ = p = 3/4. The variance of the SDSP would be
N
σp2̂ = ( pî − p)2 P( pî )
∑
i=1
( 4 ) ( 16 ) ( 2 4 ) ( 16 ) ( 4 ) ( 16 )
2 2 2
2 3 1 1 3 6 3 9
σ p̂ = 0 − + − + 1−
( 4 ) ( 16 ) ( 4 ) ( 16 ) ( 4 ) ( 16 )
2 2 2
3 1 1 6 1 9
σp2̂ = − + − +
16 ( 16 ) 16 ( 16 ) 16 ( 16 )
9 1 1 6 1 9
σp2̂ = + +
9 6 9
σp2̂ = + +
256 256 256
24
σp2̂ =
256
3
σp2̂ =
32
and then the standard error would be
3
σp̂ =
32
3
σp̂ =
4 2
6
σp̂ =
8
242
σp̂ ≈ 0.31
243

Sampling Distribution of The Sample Proportion

Uploaded by

Copyright:

Available Formats

Sampling Distribution of The Sample Proportion

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sampling Distribution of The Sample Proportion

Uploaded by

Copyright:

Available Formats

Sampling distribution of the sample

In other words, if we’re dealing with a population proportion p, instead of a

Sampling distribution of the sample proportion

where p̂ is the sample proportion, x is the number of students in the

In other words, the SDSP is the probability distribution of all possible

Central Limit Theorem

Mean, variance, and standard deviation

If the population proportion is unknown, we estimate the population

Finite population correction factor

And again, if the population proportion is unknown, we approximate it with

Let’s do an example so that we can see how these formulas work.

brown, blonde 1/2

brown, blonde 1/2

brown, blonde 1/2

blonde, brown 1/2

blonde, brown 1/2

blonde, brown 1/2

Build a table for the probability distribution of the sample proportion.

Because we’re sampling with replacement, we would expect this mean of

3 people with brown hair 3

and then the standard error would be

You might also like