You are on page 1of 8

1

Statistics and Probability


Module 6: Estimating Population Proportions

Objectives: At the end of the unit, I can

1. identify the point estimator for the population proportion;


2. compute the point estimate of the population proportion;
3. identify the appropriate form of the confidence interval estimator for the population
proportion based on the central limit theorem;
4. compute the confidence interval estimate of the population proportion;
5. solve problems involving confidence interval estimation of the population proportion;
and
6. draw conclusion about the population proportion based on its confidence interval
estimate.

Introduction:

Decision makers, and even students like you, often need to estimate a population proportion, that
is, an amount that is part of the entire population. This is because, sometimes, decisions are based on the
number of favorable observations in relation to the total number of observations. On the part of
businessmen, majority of their decisions, say market share, is based on the percentage of people
consuming their goods. For students and teachers, decision is made after considering a significant sample
size, e.g., the number of students most likely to participate in an event. Here, many factors will be studied
before finalizing their decisions on whether or not to carry out a plan.
In this module, you will be computing accurate estimates of the population proportion – both point
estimates and interval estimates.

The Sample Proportion as a Point Estimator

Suppose you want to find out the percentage of brown-eyed Filipinos as of 2015.
Since there are about a hundred million Filipinos as of 2015, it would be impossible to
survey or to test every Filipino in order to determine the desires percentage. To simplify
the process, you might just consider taking a random sample of 100 Filipinos, for
example, and determine the number of people who are indeed brown-eyed. Then, using
the definition of percentage,
28
= 0.28 or 28%
100

Thus, you would say that 28% of Filipinos are brown-eyed. Of course, you know
that this may not be true, as you would find out later in this module. This value is called
a sample proportion, denoted as 𝑝̂ , read as “p-hat”. Therefore, s sample proportion
can be determined using the formula:
𝑥
𝑝̂ =
𝑛
where x = number of favorable observations and n = sample size.

Example 1: In a sample of 50 individuals taken from a certain population, 10 are left-handed. What is the
proportion of left-handed individuals on the sample?
2

Solution:
Given: number of favorable observation (x) = 10
Sample size = 50
Use the formula:
𝑥
𝑝̂ =
𝑛
10
𝑝̂ = = 0.2 Substitute the given in the formula.
50

Hence, the proportion of left-handed individuals in the sample is 0.2

Distribution of Sample Proportions

Going back to the brown-eyed population problem on the previous page, you
know that the value of the sample proportion varies depending on the actual samples
chosen. Thus, the sample proportion can be considered a random variable with a
corresponding sampling distribution. By convention, this distribution is approximated to
be normal when both the following conditions are satisfied:
np > 10 and np(1 – p) ≥ 10.
The population proportion, denoted as p, may be estimated using the sample
proportion. Like in sampling distributions of the sample mean, the mean of the sample
proportion is approximately equal to
√𝑝(1 – 𝑝) .
𝑛

Characteristics of the Sampling Distribution of Sample Proportion


1. The sampling distribution of sample proportion is approximately normal provided
That:
a. np > 10 and
b. np(1 – p) ≥ 10.
2. In sampling distributions of the sample proportion,
𝜇𝑝̂ = p ,
Where 𝜇𝑝̂ is the mean of the sample proportion and
p is the population proportion.

𝑝(1 – 𝑝)
𝜎𝑝̂ = √ ,
𝑛

Where 𝜎𝑝̂ is the standard deviation or standard error of the sample proportion,
p is the population proportion, and n is the number of sample.

When the population proportion is not given, you can estimate it using the sample proportion,
i.e., p ≈ 𝑝̂ .

Study the example below:

Example 2: In a sample of 50 individuals taken from a certain population, 10 are left-handed. Describe the
distribution of the sample proportion.
3

Solution: Given: n = 50 and p ≈ 𝑝̂ = 0.2


Prove that: np(1 – p) ≥ 10
np = 50(0.2) = 10 Substitute the given in the formula.
np = 10
np(1 – p) =
10(0.8) = 8
Prove that: np(1 – p) ≥ 10
10 ( 1-0.2) ≥ 10 Substitute the value of np=10; p=0.2 in the formula.
10(0.8) ≥ 10
8 ≥ 10 Is false.
Hence, the sampling distribution is not normally distributed.
To solve for the standard error of the sample proportion, use the formula,

𝑝(1 – 𝑝)
𝜎𝑝̂ = √ Substitute the given, p=0.2, n=50.
𝑛

0.2(1 – 0.2)
𝜎𝑝̂ = √ ≈ 0.06
50

The description of the distribution is summarized below:


a. Since the np(1 – p) ≥ 10, then the sampling distribution is not normally distributed.
b. The mean of the distribution of the sample proportion is 0.2
c. The standard error of the proportion is equal to 0.06.

Probabilities Involving Distribution of Sample Proportions

To find the probabilities involving distributions of sample proportions, which are


approximately normal, apply the z-score conversion formula that is similar to the
distribution of the sample mean and then use the normal distribution table to
determine the desired probability.

𝑝̂−𝑝 𝑝̂−𝑝
z= =
𝜎𝑝̂ √𝑝(1 – 𝑝)
𝑛

Study the example below.

Example 3: Suppose the proportion of all high school students who joined the math club in the past
year is 0.35. A researcher selects a sample of 100 students at random representative of the
whole population. What is the probability that the proportion of students from this sample
who have joined the club in the past year is greater than 0.43?

Solution:
Given: p=0.35; n=100; sample proportion (𝑝̂ )=0.43 .

It is given that p = 0.35 and n = 100 and you are asked to find P(𝑝̂ > 0.43). To do this, you need
to compute first the standard error of proportion(𝜎𝑝̂ ) in order to convert the given sample
proportion into z-score. Use the formula for the standard error,

𝑝(1 – 𝑝)
𝜎𝑝̂ = √
𝑛
4

𝑝(1 – 𝑝)
𝜎𝑝̂ = √
𝑛

0.35(0.65)
= √ ≈ 0.05 Substitute p=0.35, n=100.
100

𝜎𝑝̂ ≈ 0.05 The standard error of the proportion.

Then, convert greater than 0.43 to z-score. We use the formula,


𝑝̂−𝑝
z= Substitute the given in this formula.
𝜎𝑝̂

0.43−0.35
P(𝑝̂ > 0.43) = P(𝑍 > ) = 1.6
0.05

Locate 1.6 in the z-table under 1.6 column 0.00. The area is 0.4452

0.4452

Z > 1.6 = 0.0548 (area)

Since half of the curve is 0.5, using the z-distribution table, P(𝑍 > 1.6) = 0.5 – 0.4452 = 0.0548 or
5.48%.

Example 4: The proportion of all left-handed Filipinos is 0.15. A random sample with size 200 is chosen.
What is the probability that the proportion of left-handed individuals in the sample is less than
0.12?

Solution:
Given: n = 200, population proportion (p) = 0.15;

First, solve for the standard error: Use the formula below,
𝑝(1 – 𝑝)
𝜎𝑝̂ = √
𝑛

0.15(1−0.15)
𝜎𝑝̂ = √
200
0.15(0.85)
= √
200
0.1275
= √
200
= √0.0006375
𝜎𝑝̂ ≈ 0.03 Round up to the nearest hundredths.

Convert the given sample proportion (𝑝̂ = 0.12 ) to z-score. Use the formula:

𝑝̂−𝑝
z=
𝜎𝑝̂

0.12−0.15 −0.03
z= = Substitute the given to the formula.
0.03 0.03

z = -1
5

Find the probability that the proportion is less than (<) 0.12. Determine the z= 1 in the z-table. The
negative in 1 is the direction in the curve as shown below. Locate the value of z=1 column 0.00 whose area
is 0.3413.

Z< -1

P(𝑝̂ < 0.12) = P(𝑍 < −1)

P(𝑝̂ < 0.12) = P(𝑍 < −1) = 0.5 – 0.3413 = 0.1587 or 15.87%.

Hence, the probability is 15.87%

Assessment 6.1:

Estimating the Population Proportion

Methods similar to estimation population mean with or without the population variance
can be used to estimate the population proportion. As you learned previously, the Central
Limit Theorem for sample proportions led to the following formula:
𝑝̂−𝑝
z=
√𝑝(1 – 𝑝)
𝑛

Since the population proportion is not given, then you can use the point estimate 𝑝̂ to
find the standard deviation of the sample proportions. The formula then becomes:

𝑝̂−𝑝
z=
̂ (1 – 𝑝
̂)
√𝑝
𝑛

Solving for p results in the following equation:

𝑝̂(1 – 𝑝̂) 𝑝̂(1 – 𝑝̂)


p = 𝑝̂ – 𝑧𝑎 √ < p < 𝑝̂ + 𝑧𝑎 √
2 𝑛 2 𝑛

where 𝑝̂ = sample proportion, n = sample size, and p = population proportion.


𝑝̂(1 – 𝑝̂)
Note that in this formula, 𝑝̂ is the point estimate and 𝑧𝑎 √ is the maximum error of
2 𝑛
estimation.

Study the examples below:


6

Example 5: A survey from 87 randomly selected private schools with online enrolment procedure revealed
that 39% of their students use online enrolment to remedy the time of waiting in school during
enrolment. Use this information at 95% confidence level to estimate the true population
proportion of students using online enrolment procedures.

Solution: Given: n = 87; 𝑝̂ = 0.39; C=95% or 0.95


𝐶 0.95
= = 0.475
2 2

0.475 0.475
0.025 0.025

Solve for 𝛼: Use the formula: 1-C


𝛼=1-C
= 1-0.95 = 0.05 Solve for 𝛼. Substitute C in the equation.
𝛼 = 0.05
𝛼 0.05 𝛼
Solve = Solve for . Substitute the value of 𝛼.
2 2 2
𝛼
= 0.025
2

Observe the curve above on how you arrive to 0.025. Do not forget that the whole curve has an
area of 1. Locate the area =0.475 in the z-table. The value of z is 1.9 column 0.06.
Hence, 𝑧𝑎 = 𝑧0.025 = 1.96 .
2

Solve for the population proportion. Use the formula below:

𝑝̂(1 – 𝑝̂) 𝑝̂(1 – 𝑝̂)


𝑝̂ – 𝑧𝑎 √ < p < 𝑝̂ + 𝑧𝑎 √
2 𝑛 2 𝑛

0.39(1−0.39) 0.39(1− 0.39)


0.39 – 1.96 √ < p < 0.39 + 1.96 √ Substitute the given in the formula.
87 87

0.39(0.61) 0.39(0.61)
0.39 – 1.96 √ < p < 0.39 + 1.96 √ Simplify.
87 87

0.2379 0.2379
0.39 - 1.96 √ < p < 0.39 + 1.96 √ Simplify what is inside the radical symbol.
87 87

0.39 - 1.96 (0.05229) < p < 0.39 + 1.96 (0.05229) Round down to five decimal places.
0.39 - 0.10249 < p < 0.39 + 0.10249 Round up to five decimal places.
0.28751 < p < 0.49249 Round off to the nearest hundredths.
0.29 < p < 0.49

This interval suggests that the population of students using online enrolment procedures is
somewhere between 0.29 and 0.49 based on the point estimate of 0.39 with a maximum error of 1.10 and
at least 95% level of confidence.

Example 6: In a certain school for special children, a sample of 169 students was chosen and among these,
26 were diagnosed with mild autism.
a. Compute the point estimate of the population proportion.
b. Compute a 90% confidence interval that estimates the proportion of students with mild
autism.
7

Solution:
Given: a. n = 169 and x = 26. Thus, the point estimate of the population proportion is:
26 2
𝑝̂ = =
169 13

b. The confidence level is 90% or C= 90% or 0.90


𝐶 0.90
For = = 0.450
2 2

𝐶
This means that the area of in the curve is 0.450. Locate this area in the z-table. 0.450 is
2
in between the area of 0.4505 and 0.4515 with their corresponding z-value of 1.65 and 1.66, respectively as
shown below.
1.65+1.66 3.31
1.65 ??? 1.66 = = 1.655
2 2
0.4515 0.450 0.4515

What we did before was we get the average instead of interpolation.

This implies that 𝛼 = 0.10 and 𝑧𝑎 = 𝑧0.05 = 1.655. Thus,


2

𝑝̂(1 – 𝑝̂) 𝑝̂(1 – 𝑝̂)


𝑝̂ – 𝑧 √ < p < 𝑝̂ + 𝑧 √
𝑛 𝑛

2 2 2 2
2 (13)(1− 13 ) 2 (13)(1− 13 )
– 1.655 √ < p < ( ) + 1.655 √
13 169 13 169

2 13 2 2 13 2
2 (13)(13−13) 2 (13)(13−13) 13
– 1.655 √ < p < ( ) + 1.655 √ Recall: 1=
13 169 13 169 13

2 11 2 11
2 (13)(13) 2 (13)(13)
– 1.655 √ < p < ( ) + 1.655 √
13 169 13 169

22 22
2 (169) 2 (169)
– 1.655 √ < p < ( ) + 1.655 √
13 169 13 169
2 2
– 1.655 (0.02775) < p < ( ) + 1.655 (0.02775) Round down to 5 decimal places.
13 13

(0.15385- 0.04593) < p < 0.15385 + 0.04593

0.10792 < p < 0.19978

0.10 < p < 0.20 Round up to the nearest hundredths.

The proportion of students in the school with mild autism is between 0.11 and 0.20.

Assessment 6.2:

An accurate estimation of the population proportion can help guide decisions and further
actions to be undertaken by some respective entities. So, with our choices and
decisions in life for whatever we sow, we will surely reap.
8

Congratulations! You are now ready to proceed to our next module…

References:
Canlapan, R. and Campena, F. 2016. Diwa Senior High School Series: Statistics and Probability.
Makati City, Philippines. University Press First Asia.

Creative commons.org
Online Math learning
Statistics for Managers Using Microsoft excel 2004 Prentice-Hall.
https://bolt.mph.ufl.edu/

You might also like