Republic of the Philippines

EASTERN VISAYAS STATE UNIVERSITY

Tacloban City
College of Arts and Sciences

CHAPTER VI

Estimation of a Population
Proportion
In partial fulfillment of the subject Stat 112
Introduction to Sampling Design

Submitted by:
Ivy Ruth L. Oracion
BS Stat 3A

Submitted to:
Mrs. Wilma A. Perante
Instructor

FEBRUARY 2016
Systematic Sampling

Objective

o To estimate the population proportion using 1 – in – k systematic

sample.
o To present a bound on the error of estimation of the population
proportion.

Formula and its components

To determine the proportion, we use a 1 – in – k systematic sample. The

estimator of the population proportion 𝑝 obtained from systematic sampling is
denoted by 𝑝̂ . As in the simple random sampling, the properties of 𝑝̂ parallel those of
the sample mean 𝑦̅ if the response measurements are defined as follows. Let 𝑦𝑖 = 0
if the ith element sampled does not possess the specified characteristics and 𝑦𝑖 = 1 if
it does. The estimator 𝑝̂ is then the average of 0 and 1 values from the sample.

o Estimator of the population 𝑝:

∑𝒏𝒊=𝟏 𝒚𝒊
̂𝒔𝒚 = 𝒚
𝒑 ̅𝒔𝒚 =
𝒏
where,
𝑝̂ = The sample proportion
n = number of observation in a sample
𝑦𝑖 = Elements that possess specified characteristic

o Estimated variance of 𝑝̂𝑠𝑦 :

𝒏 𝒑𝒔𝒚 𝒒𝒔𝒚
̂ (𝒑
𝑽 ̂𝒔𝒚 ) = (𝟏 − )
𝑵 𝒏−𝟏
Where,
𝑝̂ = Sample proportion
𝑞̂ = ( 1 − 𝑝̂ ) assuming a randomly ordered population
𝑛 = Number of observation in a sample
N = total number of observation in a population

𝑛
The (1 − 𝑁) can be ignored if the population size N is unknown but
can be assumed large relative to n. Again, note that the estimated variance of
𝜌̂ is identical to the estimated variance of 𝑝̂ using simple random sampling.
This result does not imply that the corresponding population variance are
equal, however, if N is large and if the observations within a systematic
sample are unrelated (i.e., 𝑝 = 0), the two population variances will be equal.
o Bound on the error of estimation:

𝑛 𝑝̂𝑠𝑦 𝑞̂𝑠𝑦
2√𝑉̂ (𝑝̂𝑠𝑦 ) = 2√(1 − 𝑁) 𝑛−1
It is simply, 2 times the square root of the estimated variance
Example 1:
A 1 – in – 6 systematic sample is obtained from a voter registration list to estimate
the proportion of voters in favor of the proposed bond issue. Several different random
starting points are used to ensure that the results of the sample are not affected by
periodic variation in the population. The coded results of this preelection survey are
as shown in the accompanying table. Estimate 𝑝, the proportion of the 5775
registered voters in favor of the proposed bond issue (N = 5775). Place a bound on
the error of estimation.

Voter Response

4 1
10 0
16 1
. .
. .
. .
5760 0
5766 0
5772 0

∑962
𝑖=1 𝑦𝑖 = 625

Solution:
the sample proportion is given by

∑962
𝑖=1 𝑦𝑖 652
𝑝̂𝑠𝑦 = = = 0.678
962 962

the bound on the error of estimation is

𝑛 𝑝̂𝑠𝑦 𝑞̂𝑠𝑦
2√𝑉̂ (𝑝̂𝑠𝑦 ) = 2√(1 − 𝑁) 𝑛−1

962 (0.678)(1 − 0.678)

= 2√(1 − ) ≈ 0.028
5775 962 − 1

Thus, we estimate that 67.8% of the registered voters favor the proposed bond issue.
We are relatively confident that the error of estimation is less than 2.8%.
Example 2:
The management of a particular company is interested in estimating the proportion of
employees favoring a new investment policy. A 1 – in – 10 systematic sample is
obtained from employees leaving the building at the end of a particular workday. Use
the data in the accompanying table to estimate 𝑝, the proportion in favor of the new
policy and place a bound on the error of estimation. Assume N = 2000

Employee Sampled Response, 𝒚𝒊

3 1
13 0
23 1
. .
. .
. .
1993 1

∑200
𝑖=1 𝑦𝑖 = 132

Solution:
the sample proportion is given by

∑200
𝑖=1 𝑦𝑖 132
𝑝̂𝑠𝑦 = = = 0.66
200 200

the bound on the error of estimation is

𝑛 𝑝̂𝑠𝑦 𝑞̂𝑠𝑦
2√𝑉̂ (𝑝̂𝑠𝑦 ) = 2√(1 − 𝑁) 𝑛−1

200 (0.66)(1 − 0.66)

= 2√(1 − ) ≈ 0.064
2000 200 − 1

Thus, the estimated proportion of employees favoring the new investment policy is
66% with less than 6.4% error of estimation.
Example 3:
The highway patrol of a particular state is interested about the proportion of motorist
who carry their licenses. A 1 – in – 7 systematic sample is obtained from a
checkpoint on a major highway. Use the data in the accompanying table to estimate
𝑝, the proportion of drivers carrying their licenses. Place a bound on the error of
estimation. Assume that N = 2800 cars pass the checkpoint during the sampling
method.

Car Response

4 1
11 1
18 0
. .
. .
. .
2797 1

∑400
𝑖=1 𝑦𝑖 = 324

Solution:
The sample proportion is given by:
324
𝑝̂𝑠𝑦 = = 0.81
400
The bound on the error of estimation is
𝑛 (0.81)(1 − 0.81)
2√𝑉̂ (𝑝̂𝑠𝑦 ) = 2√(1 − ) ≈ 0.36
𝑁 400 − 1
Thus, the estimated proportion of motorist carrying their licenses is 81% and the error
of estimation is less than 3.6%.
 Reference
 Scheaffer, R., Mendenhall, W., Ott, L. R., & Gerow. K. (2012, 2006, 1996).
Elementary Survey Sampling (7thed.). Boston: Books/Cole, Cengage Learning