Estimation and Hypothesis Testing:

Proportion

Lecture Handouts
St. Luke’s College of Nursing
SY. 2017 - 2018
 Qualitative Data (Review)
Not numerical in nature

Numbers are used only to label

e.g. Assigning 1 for male, and 2 for female

Cannot apply basic mathematical operations to these

variables (i.e. cannot compute the mean, it’s for
quantitative)

Instead, we count the frequency and measure them

using proportion per category, therefore it is
used for qualitative variable
 PARAMETERS
Binomial Population
One in which the elements belong to either one of
two mutually exclusive and collectively
exhaustive categories

elements can belong to one and only one

category
e.g. Variable = Nursing Licensure Exam
Category can only be either: Pass or Fail

Variable = Dominant Hand

Category can only be either: Right or Left
 PARAMETERS
Binomial Population
One in which the elements belong to either one of
two mutually exclusive and collectively exhaustive
categories

elements are distributed to the category to which

they belong
e.g. Population = NLE Takers
*all NLE Takers will have a result, no taker will have no
grading mark in the exam

Population = SLCN Students

*all SLCN Students have a dominant hand, no student
does not have a dominant hand.
 BinomialPARAMETERS
Population: Example
Variable : Blood Pressure Status
Categories : Hypertensive (SBP > 120)
Normotensive (SBP < 120)
Blood pressure Category
150/90 Hypertensive
115/70 Normotensive
120/80 Hypertensive
110/65 Normotensive
125/75 Hypertensive
 Binomial Population: Example
Subdividing the population into those who ‘possess’ the
characteristic or attribute of interest and those who ‘do
not possess’ the characteristic of interest

The Proportion of those who possess the characteristic of

interest in the population is denoted by Ρ
Blood pressure Category
Interest
150/90 Hypertensive
Hypertensive
115/70 Normotensive
120/80 Hypertensive
Frequency
3 out of 5
110/65 Normotensive
125/75 Hypertensive Proportion P
0.6 or 60%
 Binomial Population: Example
Subdividing the population into those who ‘possess’ the
characteristic or attribute of interest and those who ‘do
not possess’ the characteristic of interest

The Proportion of those who do not possess the

characteristic of interest is denoted by Q (or 1 – P)
Blood pressure Category
Opposite of Interest
150/90 Hypertensive
Normotensive
115/70 Normotensive
120/80 Hypertensive
Frequency
2 out of 5
110/65 Normotensive
125/75 Hypertensive Proportion Q
0.4 or 40%
 Binomial Population: Example
For the variable of Hypertensive Status, the population
can only be exhausted to be either Hypertensive or
Normotensive, categorized exclusively either:

Ρ = 0.6 or Q = 0.4
3/5 2/5
Blood pressure Category
150/90 Hypertensive
115/70 Normotensive
120/80 Hypertensive
110/65 Normotensive
125/75 Hypertensive
 Standard Normal Distribution
Z-score
Represents the number of standard deviations a
score away from the proportion

Where the Proportion =0

& S.D. =1

10 15 20 25 30 35 40
 Standard Normal Distribution
Z-score Formula
𝑝−𝑃
𝑃𝑄
𝑛

10 15 20 25 30 35 40
 Interval Estimate
Need to transform the sampling
distribution of ρ to standard normal using
general formula for z:
*taking the 95% of sample proportions, ρ, are within
+ 1.96 standard deviations from Ρ
ie.,
𝑝𝑞 𝑝𝑞
𝑝 − 1.96 𝑝 + 1.96
𝑛 C.I. 95% 𝑛

Lower Limit of the Upper Limit of the

Interval Interval
 Interval Estimate
*taking the 90% of sample proportions, ρ, are within
+ 1.64 standard deviations from Ρ
ie.,
𝑝𝑞 𝑝𝑞
𝑝 − 1.64 𝑝 + 1.64
𝑛 C.I. 90% 𝑛

Lower Limit of the Upper Limit of the

Interval Interval
*taking the 99% of sample proportions, ρ, are within
+ 2.58 standard deviations from Ρ
ie., 𝑝𝑞 𝑝𝑞
𝑝 − 2.58 𝑝 + 2.58
𝑛 C.I. 99% 𝑛
Interval Estimate: Problem Solving
Exercise:
In a farmland of a nearby province, a
screening for Filariasis was done to rice-field
workers . From all the 90 samples screened,
54 showed positive.

What is the variable of interest? Filariasis Status

What is the value of p? 54 / 90 = 0.6 or 60%

What is the Parameter being estimated? Rice-field

workers
Interval Estimate: Problem Solving
What is the interval estimate of the parameter using 99%
confidence interval?

Given: n = 90
p = 60% q = 40% CI = 99% ; z = 2.58

Formula:
𝑝𝑞 𝑝𝑞
𝐿𝐿 = 𝑝 − 2.58 𝑈𝐿 = 𝑝 + 2.58
𝑛 𝑛
 Interval Estimate: Problem Solving
Solution:
60𝑥40 60𝑥40
𝐿𝐿 = 60 − 2.58 𝑈𝐿 = 60 + 2.58
90 90

= 60 − 2.58 26.67 = 60 + 2.58 26.67

= 60 − 13.32 = 60 + 13.32

𝐿𝐿 = 46.68 % 𝑈𝐿 = 73.32 %
Interval Estimate: Problem Solving

Interpretation:
“It is with 99% confidence that the
proportion of Filariasis among Rice-field
workers is between 46.68 and 73.32
percent.”
 Hypothesis Testing
Test Statement of Symbol
Non- Difference Ρ1 ≠ Ρ2
Directional
or Two- “Not equal to”
Tailed “Significantly Different”
Directional Direction Ρ1 > Ρ2 or
or One- Ρ1 < Ρ2
Tailed “Higher than”, or
“Lower than”
“Greater than”, or
“Lower than”
“Increased”, or
“Decrease”
 Hypothesis Testing
Test Example Symbol
Non- The proportion of students who Ρ1 ≠ Ρ2
Directional passed the board exam last year is
not equal to that of the students
or Two- this year
Tailed
Directional The prevalence of malnutrition is Ρ1 > Ρ2 or
or One- higher in Sitio Catmon compared to
Sitio Wawa Ρ1 < Ρ2
Tailed
The incidence of medical error in
Hospital de Luna is lower than the
set average standard.
 Hypothesis Testing: Steps
1. State the null and alternative Hypothesis

2. State the level of significance

3. Analyze: p-value (SPSS)

Confidence Interval (manual computation)

4. Make a statistical Decision

5. Interpret
 Hypothesis Testing: Proportion
Case 1:
This year’s goal of the Association of Chief Nursing
Officer (CNO) of the Philippines is to keep the
medication errors of all I.C.U. nurses below 5%.

They sampled 120 I.C.U. nurses around the

Philippines and found that only 5 nurses or 4.16%
committed atleast one medication error in that
year.

Did the association statistically meet the goal by

having a significantly lower percentage compared
to the set limit?
 Hypothesis Testing: Proportion

What is the variable of interest? Medical error

What is the value of p? 4.16

What is the Parameter being estimated?

ICU Nurses in PI
 Hypothesis Testing: Proportion
Step 1
H0: P=5%
“the proportion of I.C.U. nurses that committed
medication error is significantly the same with the set
limit”

Ha: P<5%
“the proportion of I.C.U. nurses that committed
medication error is significantly below the set limit”

Step 2
Level of Significance: α = 0.05
 Hypothesis Testing: Proportion
Step 3 : p-value
 Hypothesis Testing: Proportion
Step 3 : p-value
 Hypothesis Testing: Proportion
Step 3 : p-value

Hypothesized
Value
 Hypothesis Testing: Proportion
Step 3 : p-value
 Hypothesis Testing: Proportion
Step 4

Probability Value (p-value) Decision

Level of
α = 0.05 Do not reject H₀
Significance
p-value 0.442
The p-value 0.442 is greater than α = 0.05
 Hypothesis Testing: Proportion
Step 5

Interpretation:
“There is no sufficient evidence to conclude
that the proportion of I.C.U. nurses in the
Philippines who committed medication error
is significantly below the set limit.”
 Hypothesis Testing: Proportion
Case 2:
In the beginning of this year 2016 (January),
Region IV-A has the highest proportion of
Dengue in the Philippines with 18.9%
according to the Department of Health.

By the middle of the year (July), the country

has 59,585 dengue cases, and the proportion
of dengue cases in Region IV-A is at 12.7%. Is
there a significant decrease in Dengue
proportion in the said region?
 Hypothesis Testing: Proportion

What is the variable of interest? Dengue

What is the value of p? 12.7%

What is the Parameter being estimated?

Region W-A
 Hypothesis Testing: Proportion
Step 1
H0: P = 18.9 %
“the proportion of dengue in Region IV-A last July 2016
has no significant difference with that of January 2016”

Ha: P < 18.9 %

“the proportion of dengue in Region IV-A last July 2016 is
significantly lower compared from January 2016”

Step 2
Level of Significance: α = 0.05
 Hypothesis Testing: Proportion
Step 3 : Confidence Interval

Given: n = 59,585
p = 12.7% q = 87.3% CI = 95% ; z = 1.96

Formula:
𝑝𝑞 𝑝𝑞
𝐿𝐿 = 𝑝 − 1.96 𝑈𝐿 = 𝑝 + 1.96
𝑛 𝑛
 Hypothesis Testing: Proportion
Solution:
12.7𝑥87.3 12.7𝑥87.3
𝐿𝐿 = 12.7 − 1.96 𝑈𝐿 = 12.7 + 1.96
59,585 59,585

= 12.7 − 1.96 0.0186 = 12.7 + 1.96 0.0186

= 12.7 − 0.27 = 12.7 + 0.27

𝐿𝐿 = 12.43 % 𝑈𝐿 = 12.97 %
 Hypothesis Testing: Proportion
Step 4

Confidence Interval (C.I.) Decision

Hypothesized
value 18.9 % Do not
reject H₀
95% C.I. 12.43 to 12.97 %
The Hypothesized value 18.9% is not within the
confidence interval 12.43 to 12.97%
 Hypothesis Testing: Proportion
Step 5

Interpretation:
“This 2016, the mid-year (July) proportion of
dengue in Region IV-A of 12.7% is significantly
lower compared to its proportion last January
which is 18.9%.”
 Other References

Lecture notes on Statistical Inference: Sampling

Distribution of Proportions and Estimation of the
Population Proportion for Biostatistics 201 by K.L.
Cochon, Department of Epidemiology and Biostatistics,
UP Manila College of Public Health, 2013.