Lecture 9.

Chapter 9
Sampling distributions
9.3 Introduction to sampling distribution
9.4 Sampling distribution of the sample
mean 𝐗
9.5 Sampling distribution of the sample
proportion 𝐩

9.3 Introduction to sampling

distribution
• In real life, calculating the parameters of
populations is prohibitive because populations
are very large.
• Rather than investigating the whole population, we
take a sample, calculate a statistic related to
the parameter of interest and make an
inference about the parameter of the
population (based on the statistic of the sample).
• The sampling distribution of the statistic is
the tool that tells us how close the statistic is
to the parameter.
2

1
 Recall what statistics is
Statistics
Information
Data

“Statistics is a way to get information from data”

Data: Facts, especially Information:
numerical facts, Knowledge
collected together for communicated
reference or concerning some
information. particular fact.

“Statistics is a tool for creating new

understanding from a set of numbers”
Definitions: Oxford English Dictionary

Key statistical concepts

Population Sample

Subset

Statistic
Parameter
• Populations have parameters: , 2, , p (probability)
• Samples have statistics: 𝑥, s2, s, 𝑝 (relative frequency)
• Based on values of statistics of the sample,
inferences can be made about parameters of the
population.

2
9.4 Sampling Distribution
of the Sample Mean
Example, page 346: A fair die is thrown infinitely many
times, with the random variable X = Number of spots
showing on any throw. The probability distribution of X is:

x 1 2 3 4 5 6
P(x) 1/6 1/6 1/6 1/6 1/6 1/6
and the mean and variance (of the population of
numbers of spots observed) are:

Sampling Distribution of Two Dice

A sampling distribution of the sample mean 𝐗 of size n
= 2 is created by looking at all 36 samples of two dice
and their means: 𝑥1 =1, 𝑥2 =1.5, …., 𝑥36 = 6.0.

While there are 36 possible samples of size 2, there

are only 11 values for 𝑥 , and some (e.g. 𝑥 =3.5)
occur more frequently than others (e.g. 𝑥 =1).
6

3
Sampling Distribution of Two Dice
The sampling distribution of X is shown below:
P( )
1.0 1/36
1.5 2/36
2.0 3/36
2.5 4/36
3.0 5/36
3.5 6/36
4.0 5/36
4.5 4/36
5.0 3/36
5.5 2/36
6.0 1/36

Compare…
Compare the distribution of X

1 2 3 4 5 6 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0

𝝁
The sampling
with the sampling distribution of 𝐗 distribution of the
(approximately normal distribution). statistic is the
tool that tells us
Note that: how close the
statistic is to the
parameter 𝝁
8

4
 Generalize…
We can generalize the mean and variance of the
sampling of 2 dice:

The standard
to sampling of n dice deviation of the
(roll n dice and calculate sampling
the sample mean 𝑿 ): distribution of the
sample mean is
called the
standard error.

The above formulas can be proved using the laws of

expected value and variance for:
𝑿 = (𝑿𝟏 + 𝑿𝟐 + … + 𝑿𝒏 )/n, where Xi are independent
(normally distributed) random variables with the
same mean  and variance 2. 9

n  5 (roll 5 dice)
 x  3 .5
n  10 (roll 10 dice)
 x2
 x2  .5833 ( )  x  3.5 n  25 (roll 25 dice)
5
 2  x  3 .5
 x2  .2917 ( x
)
10  x2
 x2  .1167 ( )
25
1 3.5 6

Notice that  2x is smaller

2 3.5
than 𝜎𝑥 . The larger the sample 1 6
size the smaller  2x . Therefore,
x tends to fall closer to , as the
sample size increases. 1 3.5 6 10

5
Central Limit Theorem
• If a random sample is drawn from a normal
population, then the sampling distribution of the
sample mean 𝐗 is normally distributed for all
values of n (sample size).
• If a random sample is drawn from any population,
the sampling distribution of the sample mean
𝑿 is approximately normal for a sufficiently
large sample size. The larger the sample size,
the more closely the sampling distribution of 𝐗 will
resemble a normal distribution.
• A sample size of 30 may be sufficiently large to
allow us to use the normal distribution as an
approximation for the sampling distribution of 𝐗 .

11

Sampling Distribution of the Sample Mean

1.  x   x
 x2
2.  
2
x
n
3. If X is normal, X is normal. If X is non - normal
X is approximately normally distributed for
sufficient ly large sample size.

We can standardize the sampling X

distribution of the sample mean 𝐗 as
Z 
/ n

12

6
Sampling Distribution of the Sample Mean
The summaries above assume that the population is
infinitely large. However, if the population is finite the
standard error is
 Nn
x 
n N 1
Nn
where N is the population size and N 1
is the finite population correction factor.

In practice, most applications involve populations that

qualify as large. As a consequence the finite population
correction factor is usually omitted.

13

Example 9.1 pages 355 - 357

A stock broker has observed that the annual return
in the construction industry is actually a normal
distributed random variable with a mean of 12.5%
and a standard deviation of 2.5%.
 a/ Find the probability that a randomly selected
stock in the industry will have an annual return
less than 10.825%:
X   10.825  12.5
P( X  10.825)  P(  )  P( Z  0.67)  0.2514
 2.5
 c/ Find the probability that 04 randomly
selected stocks in the industry will have the
mean annual return less than 10.825%.
X   x 10.825  12.5
P( X  10.825)  P(  )  P( Z  1.34)  0.0901
x 1.25
14

7
Example, pages 342 and 358
Salaries of a business school’s graduates:
The average weekly income of graduates one year
after graduation is $1000. Suppose the distribution
of weekly income has a standard deviation of $100.
What is the probability that 25 randomly-selected
graduates have an average weekly income of less
than $950?
Solution
Let X be the weekly income of graduates one year
after graduation. We consider sampling distribution
of sample mean 𝐗 for sample size n = 25 and have
X  950  1000
P( X  950)  P(  )  P( Z  2.5)  0.0062
x 100 25

15

Example, pages 342 and 358 (contd.)

Salaries of a business school’s graduates:
The average weekly income of graduates one year
after graduation is $1000. Suppose the distribution
of weekly income has a standard deviation of $100.
A statistical inference: If a random sample of 25
graduates actually had an average weekly income of
$950, what would you conclude about the validity of
the claim that the average weekly income is $1000?
Solution
With  = 1000, the probability of having a sample
mean of 950 is very low (0.0062). So the claim that
the average weekly income is $1000 is probably
unjustified. It would be more reasonable to assume
that the average weekly income  is smaller than
$1000, because then a sample mean of $950
becomes more probable.
16

8
Using Sampling Distributions for
Inference
– To make inferences about / related to population
parameters we use sampling distributions.
– The symmetry of the normal distribution along with the
sample distribution of the mean lead to:
X 
P (1.96  Z  1.96)  0.95, or P (1.96   1.96)  0.95
 n
- z.025 z.025
 
This can be written as P (1.96  X    1.96 )  0.95
n n
 
which becomes P (   1.96  X    1.96 )  0.95
n n
 
In general, P (   z 2  X    z 2 )  1
n n
17

Standard normal distribution Z:  = 0.05

0.025 0.025

Normal distribution of 𝑿
–1.96 0 1.96

0.025 0.025

  
  1.96   1.96
n n
18

9
 Example, pages 342 and 358 (contd.)

Substituting μ  1000, σ  100, and n  25 from the example,

100 100
P(1000  1.96  X  1000  1.96 )  0.95 or
25 25
P(960.8  X  1039.2)  0.95

Conclusion: A statistical inference can be made

 There is a 95% chance that the sample mean falls
within the interval [960.8, 1039.2] if the population
mean is 1000 and  =100.
 Since the sample mean was 950, the population mean
is probably not 1000.

19

9.5 Sampling Distribution of the

Sample Proportion
• The parameter of interest for nominal data is the
population proportion p (“theoretical relative
frequency”) of data that fall into a category, or
the proportion of time a particular outcome
(success) occurs.
• To estimate the population proportion p we use
the sample proportion p ^ (“experimental relative
frequency”).
• We prefer to use normal approximation to the
binomial distribution to make inferences about p ^.

20

10
Normal Approximation
to Binomial Distribution
Binomial distribution with n=20 and p=0.5 can be
approximated by the normal distribution with the
same mean and standard deviation.
(µ= np = 10 and  = 𝑛𝑝 1 − 𝑝 = 2.24)

21

Normal Approximation
to Binomial Distribution
• Normal approximation to the binomial distribution
works best when
– the number of experiments (sample size) is
large, and
– the probability of success p is close to 0.5.

• For the approximation to provide good results:

np  5; n(1 – p)  5.

22

11
Normal Approximation
to Binomial Distribution
To calculate
P(X=10) using
the normal
distribution, we
can find the area
under the
normal curve
between 9.5 and
10.5

P(X = 10) ≈ P(9.5 < Y < 10.5) = 0.1742

where Y is a normal random variable approximating
the binomial random variable X.
23

Explanation
 = np = 20(0.5) = 10; 2 = np(1 – p) = 20(0.5)(1 – 0.5) = 5;  = 2.24

The exact probability is P(X = 10) = 0.176.

P(9.5<YNormal<10.5)
the approximation

9.5 10 10.5

9.5  10 10.5  10
P(XBinomial = 10) ~= P(9.5<Y<10.5)  P( Z )  .1742
2.24 2.24
24

12
More Normal Approximation Exercises

P(X  8)  P(Y < 8.5)

8
P(X 14)  P(13.5 < Y < n + 0.5) 8.5
 P(Y > 13.5)

For large n the effect of the

continuity correction factor 0.5 is 13.5 14
very small and will be omitted.
25

Approximate Sampling Distribution of a

Sample Proportion 𝒑
𝑋
• We have: the sample proportion 𝑝 = 𝑛 where X is the
number of successes in n trials and is binomially distributed,
with mean np and variance np(1-p).
• If both np  5 and n(1–p)  5, then 𝑋 can be approximated by
Y, normally distributed with the same mean and is
approximately standard normally distributed. Hence
X  np np̂  np p̂  p
Z  or Z
np(1  p) np(1  p) p(1  p)
n
can be considered as having standard normal distribution.
(𝑝 can be considered as normally distributed with mean p and
standard deviation 𝑝(1 − 𝑝)/𝑛 .
26

13
Example 9.2 revised, page 363
The Laurier company’s brand has a market share of
45.6%. In a survey, 300 consumers were asked which
brand they prefer. What is the probability that more than
50% of the respondents say they prefer the Laurier
brand?
Solution: The number of respondents who prefer Laurier
is binomial with n = 300 and p = 0.456. Also, np =
300(0.456) = 136.8 > 5, and n(1 – p) = 300(1 – 0.456)
= 163.2 > 5. Therefore, 𝑝 is normal with mean p =
0.456 and standard error
p (1  p ) 0.456(1  0.4561)
 pˆ    0.0288
n 300
Hence
 pˆ  p 0.50  0.456 
P ( pˆ  0.50)  P   P ( Z  1.53)  0.0618
 p (1  p ) n 0.0288 

27

Summary: page 366

Home assignment:

- Section 9.4 Exercises pages 359 - 361: 9.6, 9.9,

9.15, 9.20

- Section 9.5 Exercises page 364: 9.24, 9.25, 9.28,

9.30

28

14