Lecture Eight

Condence Interval for Population

Proportion

Prof. O.E. Olubusoye

Department of Statistics,

University of Ibadan,

Nigeria.

2020/21 Academic Session

1/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Introduction

Introduction

In our last lecture, we discussed how to construct condence intervals for

the population mean. You have learnt how to identify acceptance and
rejection region of a standard normal distribution. You have also learnt
how to obtain various critical values from statistical tables using
R-programming language. In this lecture, you will learn how to construct
condence interval and limits for proportion when sample size is large.

2/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Objectives

At the end of this lecture, you should be able:

2 To understand problems relating to condence interval for proportion.
2 To understand how to compute condence interval and limit for
proportion.
2 To understand how to compute the condence limits for sums and
dierences of two proportions.

3/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Margin of Error for Proportion

Margin of Error for sample proportion

When reporting a statistical survey, it is important to include the margin of errors

(MOE).
The number of standard errors that we need to add or sustract to get the MOE
depends on how condent we want to be in our result. This condent, is the
condence level (as discussed in the previous lecture).
We discussed that, a typical condence level is 95%. Thus, the basic idea is to
add or substract about 2 standard errors (1.96, to be exact, i.e Zα/2 = Z0.05/2 )
in order to get the margin of error. As a result of this, one would be able to
account for about 95% of all possible results that may have occurred with
repeated sampling.
Consequently, to be 99% condent implies that, we add or sustract 2.58 ( i.e
Zα/2 = Z0.01/2 ) standard errors.
If we use a larger condence percentage, then our margin of error will be larger
too.

4/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Margin of Error for Proportion Cont'd

The general formula for the margin of error for a sample proportion is
given as:

p pq
±Z α2 n

Two conditions must be met in order to use Z-value in the formula:

1. You need to be sure that np is at least 10
2. You need to be sure that nq is at least 10.
If this condition fail, Z-value can not be use. The reason is that, the
sample must be large enough. We have discussed lage sample cases in
the previous lecture.
For instance, at 95% condence, based on the current level of
performance, we are sure that 95% of STA 121 students will passed
±4.3% margin of error.

5/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Condence Interval and limits for Proportion

Condence Interval and limits for Proportion

Let p denote the proportion of a population that has a certain attribute
pn the proportion of the sample of size `n' drawn from the population
that possess the attribute.
The condence interval and limits for the proportion p are dened
respectively as:

1. pn ± Z α2 pqn , thus pn − Z 2
p p pq p pq
n ≤ p ≤ pn + Z 2
α α
n
2. pn ± Zα n , thus pn − Zα n ≤ p ≤ pn + Zα n
p pq p pq p pq

3. pn − Zα pq n , thus, p ≥ pn − Zα n , for 1- tailed (lower)

p p pq

4. pn + Zα n , thus, p ≤ pn + Zα n , for 1- tailed (upper)

p pq p pq

6/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Condence Interval for a Population Proportion

Given a sample of size n and there arer items of a certain characteristic.

Then, sample proportion of items with that characteristic is p = nr
The distribution of P is approximately
P(1 − P)
 
N, Z ∼ N(0, 1)
n
Let the true (population) proportion be
Q

The estimate of the true (Population) proportion is

p(1 − p) p(1 − p)
r r
p − 1.96 ≤ p + 1.96 [95% C .I ]
Q
≤
n n
p(1 − p) p(1 − p)
r r
p − 2.58 ≤ p + 2.58 [99% C .I ]
Q
≤
n n
p(1 − p) p(1 − p)
r r
p − 1.645 ≤ p + 1.645 [90% C .I ]
Q
≤
n n
Y
⇒ p ± Z1−α S.E (p)

where Z( ) is standardized value from the standard normal table and α is the
signicance level. ⇒ If [α = 5%, 1 − α = 95%]
7/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Examples

Example:
In a community with 300 households, 123 ha colour TV. Obtain a 95%
condence interval for the (true population) proportion of households
with colour TV in the whole town.
for Practicals:

Example: The weight of 10 tubers of yam from a farm are

9.8, 9.9, 10.3, 10.1, 10.4, 10.6, 10.2, 9.7, 10.5 and 9.8.
Find the (i) 95% and (ii) 99% condence interval for the mean of all yam
in the farm.

8/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Solution Cont'd
123
n = 300, p = 300 = 0.41, Z0.95 = 1.96

0.41(1 − 0.41)
r
S.E(p) =
300
0.41(0.59)
r
=
300
= 0.0284
Y
= p ± Z0.95 S.E(p)
= 0.41 ± 1.96(0.0284)
= 0.41 ± 0.05566
= (0.3543, 0.4657)
If 1.96 is appropriated to 2, we have
= 0.41 ± 2.00(0.0284)
Y

= 0.41 ± 0.0568
= (0.3532, 0.4668)
9/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Solution Cont'd

Example I
Given a population of size N = 4 as (3, 5, 8, 2)
The population mean is:
N
P
Xi
i=1
µ=
N
3+5+8+2
=
4
= 4.5
And the population variance is:
(3 − 4.4)2 + (5 − 4.5)2 + (8 − 4.5)2 + (2 − 4.5)2
σX2 =
4
21
=
4
= 5.25
10/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Class Work

Class work
1. Suppose that UI admission oce latest statistics shows that out of
every 3500 students that apply for undergraduate admissions, only
1500 are qualied. At 95% and 99% condence level,
a. what is the percentage of admission intake based on the
statistics?
b. what is the margin of errors of the qualied candidates?
c. construct the condence limit and interval of the qualied
candidates.
2. Suppose 80% out of every `n' that enrol in a military course
graduated. Given that the margin error is ±1.43% at 95%
condence level. What is
a. the value of `n'?
b. construct the condence limit and interval for the given
information.

11/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Exercises

Exercises
Exercise 1: If 450 out of every 500 iron rods produced by a Steel
Manufacturing Company are non-defective. Construct
(a) the percentage of defective iron rods
(b) the margin of errors of non-defective l and obtain
(c) condence limits and intervals for the proportion of non-defective iron rods
at 99% and 95% condence leve.
Exercise 2: The recent US presidential pre-election poll shows that, 2000 out of
every 5000 American voters will vote the Republican candidate in the
forthcoming election at 99% condence level. Construct,
(i) the percentage of voters not in favour of Republican candidate
(ii) the margin of error of voters in support of Republican candidate
(iii) the margin of error of voters that will oppose the Republican candidate
(iv) the condence limit and interval of voters for and against the Republican
candidate.

12/13
Prof. O.E. Olubusoye C.I%
 Lecture Eight

Q&A

Q&A

13/13
Prof. O.E. Olubusoye C.I%