You are on page 1of 25

Lecture 7.

SAMPLING
▪ Sampling
▪ Sampling distribution
▪ Point Estimate
▪ Acceptance Interval

▪ [1] Chapter 7. pp. 298 - 335

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 1


Lecture 7. SAMPLING
▪ Case study
• Engineers designs
process control
system for an
industrial production
line.
• Further reading: Central Limit Theorem. Follow CLT, a
production process is “in control” if the sample
mean with 𝑛 > 30 has normally distributed.
𝜎
• 95,5% of 𝑥ҧ lies in the interval: 𝜇 ± 2 (2 sigma)
𝑛
𝜎
• 99,74% of 𝑥ҧ lies in the interval: 𝜇 ±3 (3 sigma)
𝑛
PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 2
Lecture 7. SAMPLING
▪ Case study
• Six employees,
whose year of
experience are:
2, 4, 6, 6, 7, 8
• Pop. Mean: 𝜇 = 5.5

• Two of these
employees are to be
selected randomly

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 3


Inferential Statistics
▪ Inferential Statistics: Deduce information of Population
from Sample data.
▪ Population size: 𝑵, 𝑿 = (𝒙𝟏, 𝒙𝟐, … , 𝒙𝑵 )
▪ Parameters:
Population Mean 𝝁
Population Proportion 𝒑
Population Variance 𝝈𝟐

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 4


Example
▪ Population: 4 8 5 7 6 7 6 2 5 8
▪ Parameter: 𝜇 = 5.8 𝜎 2 = 3.16
4
▪ Proportion of “odd value” 𝑝 = = 0.4
10

▪ Sample 1: 4 8 5
1
𝑥ҧ1 = 5.67, 𝑠12 = 4.33, 𝑝1ҧ = = 0.33
3
▪ Sample 2: 8 6 7 8
1
𝑥ҧ2 = 7.25, 𝑠22 = 0.917, 𝑝2ҧ = = 0.25
4

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 5


7.1. Random Sample
Population data: from census
▪ Exactly
▪ Maybe impossible
▪ Difficult to gather
▪ Costly, much time
Sample data: from surveys
▪ Possilbe to gather
▪ Easier than census
▪ Less cost and time

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 6


Random Sample
▪ A Simple Random Sample of size 𝑛 from a finite
population of size 𝑁 is a sample selected
such that each possible sample of size 𝑛 has the same
probability of being selected.
▪ A Random Sample of size 𝑛 from an infinite
population is a sample selected such that the
following conditions are satisfied.
• Each element is selected from the same population
• Each element is selected independently.
Consequences:
• 𝐸 𝑥1 = ⋯ = 𝐸 𝑥𝑛 = 𝜇
• 𝑉 𝑥1 = ⋯ = 𝑉 𝑥𝑛 = 𝜎 2
PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 7
Sampling Method
▪ [1] p.331
▪ Stratified sampling

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 8


Sampling Method
▪ [1] p.332
▪ Cluster sampling

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 9


Sampling Method
▪ [1] p.332
▪ Systematic sampling

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 10


Sampling Method
▪ [1] p.332
▪ Convenience sampling

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 11


Sampling Method
▪ [1] p.333
▪ Judgment/Purposive sampling

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 12


7.2. Sampling Distribution
▪ A sampling distribution is a distribution of all possible
values for a statistic of a given sample selected from a
population
• Distribution of Sample Mean
• Distribution of Sample Proportion
• Distribution of Sample Variance

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 13


Distribution of Sample Mean
▪ Sample Mean (Random Sample)
∑𝑥𝑖
𝑥ҧ =
𝑛
▪ For a random sample, 𝑥ҧ is treated as a random variable
▪ 𝐸 𝑥ҧ = 𝜇𝑥ҧ = 𝜇
𝜎2
▪ 𝑉 𝑥ҧ = 𝜎𝑥2ҧ =
𝑛
𝜎
▪ 𝜎𝑥ҧ = = S.E(𝑥)ҧ (S.E stands for “Standard Error”)
𝑛
▪ Sample mean has same expectation with 𝑋 , but
smaller variance (depends on sample size 𝑛).

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 14


Correction Factor
▪ In case of large sample or finite population, and 𝑛 is
relatively large in comparison to 𝑁
𝑁−𝑛
▪ The finite population correction factor is
𝑁−1
𝑁−𝑛 𝜎2
• 𝑉 𝑥ҧ = ×
𝑁−1 𝑛
𝑁−𝑛 𝜎
• 𝜎𝑥ҧ = ×
𝑁−1 𝑛

▪ In lecture: sample is not large

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 15


Normal Distribution
▪ If Population has normally distributed: 𝑋~𝑁(𝜇, 𝜎 2 ) or
hasn’t normal distributed but 𝑛 > 30 then:
2
▪ 𝑥~𝑁(𝜇
ҧ 𝑥ҧ 𝑥ҧ )
, 𝜎
𝜎2
▪ 𝑥~𝑁
ҧ 𝜇,
𝑛

Ex. Population 𝑋~𝑁(20,42 )


▪ Sample 𝑛 = 16
▪ 𝑥~𝑁ҧ 20, 12

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 16


Example
Example 7.1. Workers’ salary is Normal distributed with
Mean of 300 $ and Standard deviation of 20 $.
a) What is the probability that salary of a worker chosen
randomly exceeds 305?
b) Random choose 10 workers, what is the probability
that sample mean exceeds 305?
c) What is the probability that sample mean of 100
workers exceeds 305?
d) With the probability of 0.67, what is the maximum of
sample mean of 10 workers?
e) With the probability of 0.67, what is the maximum of
sample mean of 100 workers?
PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 17
Distribution of Sample Proportion
▪ Population Proportion = Probability = 𝑝
▪ Sample proportion = 𝑝ҧ
𝑝(1−𝑝)
▪ 𝐸 𝑝ҧ = 𝑝 ; 𝑉 𝑝ҧ =
𝑛
▪ With 𝑛 ≥ 100
𝑝 1−𝑝
𝑝~𝑁
ҧ 𝑝,
𝑛

Example 7.2. Probability that candidate pass the exam is


0.4. Find the probability that proportion of pass in 200
candidates is greater than 45%
PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 18
7.3. Acceptance Interval
▪ Assume Population is known
▪ Parameter 𝜇, 𝑝, 𝜎 2 are known
▪ Deduce for statistics in sample
▪ With probability of 95%, 90%, or (1 − 𝛼)

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 19


Sample mean
▪ With probability of 95%
𝑃 −1.96 < 𝑍 < 1.96 = 0.95
𝑥ҧ − 𝜇
𝑃 −1.96 < 𝑛 < 1.96 = 0.95
𝜎
▪ Acceptance interval 95% of sample mean
𝜎 𝜎
𝜇 − 1.96 < 𝑥ҧ < 𝜇 + 1.96
𝑛 𝑛
𝜎
▪ Or: 𝜇 ± 1.96
𝑛

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 20


Sample mean
▪ In general, probability of 1 − 𝛼
▪ The acceptance interval for sample mean:
𝝈 𝝈
𝝁−𝒛 𝜶 ഥ<𝝁+𝒛
<𝒙 𝜶
𝟐 𝒏 𝟐 𝒏
𝝈
or 𝝁 ± 𝒛𝜶
𝟐 𝒏

Example 7.3. Worker income ($) has normally distributed


with mean of 300 and variance of 400. What is the
interval that average income of 25 workers falls into, with
probability of 95%, 90%, 80% ?

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 21


Sample Proportion
▪ If population parameter 𝑝 is known
▪ 𝑝 is population proportion or probability
▪ Acceptance interval of sample proportion 𝑝ҧ
𝒑(𝟏 − 𝒑) 𝒑(𝟏 − 𝒑)
𝒑 − 𝒛𝜶 ഥ < 𝒑 + 𝒛𝜶
<𝒑
𝟐 𝒏 𝟐 𝒏
▪ Or
𝒑(𝟏 − 𝒑)
𝒑 ± 𝒛𝜶
𝟐 𝒏

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 22


Sample Proportion
Example 7.4. Probability that a visitor buying at least one
item in the shopping mall is 0.3.
(a) What is the probability “there are 65 or more of 200
visitors buy at least one item”.
(b) At probability level of 95%, in 200 visitors, what are
acceptance interval of relative frequency of number of
buyers, and acceptance interval of number of buyers?

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 23


Key Concepts
▪ Random Sample
▪ Sampling Distribution
▪ Acceptance Interval

Exercise
[1] Chapter 7
▪ (309) 13, 16,
▪ (320) 21, 27, 29
▪ (326) 37, 38

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 24


CENTRAL LIMIT THEOREM

PROBABILITY & STATISTICS – Nguyen Hai Duong – NEU – www.mfe.edu.vn/nguyenhaiduong 25

You might also like