Sampling Theory

Determining the distribution of Sample

statistics
 Sampling Theory
sampling distributions
Note:It is important to recognize the dissimilarity
(variability) we should expect to see in various
samples from the same population.

• It is important that we model this and use it

to assess accuracy of decisions made from
samples.
• A sample is a subset of the population.
• In many instances it is too costly to collect
data from the entire population.
Statistics and Parameters

A statistic is a numerical value computed from a

sample. Its value may differ for different samples.
e.g. sample mean x , sample standard deviation s, and
sample proportion p̂.

A parameter is a numerical value associated with a

population. Considered fixed and unchanging. e.g.
population mean m, population standard deviation s,
and population proportion p.
Observations on a measurement X
x1, x2, x3, … , xn
taken on individuals (cases) selected at random from a
population are random variables prior to their
observation.
The observations are numerical quantities whose
values are determined by the outcome of a random
experiment (the choosing of a random sample from
the population).
The probability distribution of the observations x1, x2,
x3, … , xn
is sometimes called the population.
This distribution is the smooth histogram of the the
variable X for the entire population
0.07

0.06

0.05

0.04

0.03

0.02

0.01

0
0 10 20 30 40 50 60
the population is unobserved (unless all observations
in the population have been observed)

0.07

0.06

0.05

0.04

0.03

0.02

0.01

0
0 10 20 30 40 50 60
A histogram computed from the observations
x1, x2, x3, … , xn
Gives an estimate of the population.
0.07

0.06

0.05

0.04

0.03

0.02

0.01

0
0 10 20 30 40 50 60
A statistic computed from the observations
x1, x2, x3, … , xn
is also a random variable prior to observation of the
sample.
A statistic is also a numerical quantity whose value is
determined by the outcome of a random experiment
(the choosing of a random sample from the
population).
The probability distribution of statistic computed
from the observations
x1, x2, x3, … , xn
is sometimes called its sampling distribution.
This distribution describes the random behaviour of
the statistic
0.09

0.08

0.07

0.06

0.05

0.04

0.03

0.02

0.01

0
0 10 20 30 40 50 60
It is important to determine the sampling distribution
of a statistic.
It will describe its sampling behaviour.
The sampling distribution will be used the assess the
accuracy of the statistic when used for the purpose of
estimation.
Sampling theory is the area of Mathematical Statistics
that is interested in determining the sampling
distribution of various statistics
Many statistics have a normal distribution.
This quite often is true if the population is Normal
It is also sometimes true if the sample size is
reasonably large. (reason – the Central limit theorem,
to be mentioned later)
Combining Random Variables
 Combining Random Variables

Quite often we have two or more random variables

X, Y, Z etc

We combine these random variables using a

mathematical expression.

Important question
What is the distribution of the new random variable?
Example 1: Suppose that one performs two
independent tasks (A and B):
X = time to perform task A (normal with mean 25
minutes and standard deviation of 3 minutes.)
Y = time to perform task B (normal with mean 15
minutes and std dev 2 minutes.)
Let T = X + Y = total time to perform the two tasks

What is the distribution of T?

What is the probability that the two tasks take more
than 45 minutes to perform?
Example 2:

Suppose that a student will take three tests in the next

three days

1. Mathematics (X is the score he will receive on this

test.)
2. English Literature (Y is the score he will receive on
this test.)
3. Social Studies (Z is the score he will receive on this
test.)
Assume that

1. X (Mathematics) has a Normal distribution with

mean m = 90 and standard deviation s = 3.
2. Y (English Literature) has a Normal distribution
with mean m = 60 and standard deviation s = 10.
3. Z (Social Studies) has a Normal distribution with
mean m = 70 and standard deviation s = 7.
 Graphs
0.14 X (Mathematics)
m = 90, s = 3.
0.12

0.1

0.08
Z (Social Studies)
0.06
m = 70 , s = 7.

0.04
Y (English Literature)
m = 60, s = 10.
0.02

0
0 20 40 60 80 100
Suppose that after the tests have been written an overall
score, S, will be computed as follows:

S (Overall score) = 0.50 X (Mathematics) + 0.30 Y

(English Literature) + 0.20 Z (Social Studies) +
10 (Bonus marks)

What is the distribution of the overall score, S?

Sums, Differences, Linear Combinations of R.V.’s

A linear combination of random variables, X, Y, . . . is

a combination of the form:
L = aX + bY + … + c (a constant)
where a, b, etc. are numbers – positive or negative.
Most common:
Sum = X + Y Difference = X – Y
Others
Averages = 1/3 X + 1/3 Y + 1/3 Z
Weighted averages = 0.40 X + 0.25 Y + 0.35 Z
Sums, Differences, Linear Combinations of R.V.’s

A linear combination of random variables, X, Y, . . . is

a combination of the form:
L = aX + bY + … + c (a constant)
where a, b, etc. are numbers – positive or negative.
Most common:
Sum = X + Y Difference = X – Y
Others
Averages = 1/3 X + 1/3 Y + 1/3 Z
Weighted averages = 0.40 X + 0.25 Y + 0.35 Z
Means of Linear Combinations

If L = aX + bY + … + c

The mean of L is:

Mean(L) = a Mean(X) + b Mean(Y) + … + c
mL = a mX + b mY + … + c
Most common:
Mean( X + Y) = Mean(X) + Mean(Y)
Mean(X – Y) = Mean(X) – Mean(Y)
Variances of Linear Combinations

If X, Y, . . . are independent random variables and

L = aX + bY + … + c then
Variance(L) = a2 Variance(X) + b2 Variance(Y) + …
 L2  a 2 X2  b 2 Y2 
Most common:
Variance( X + Y) = Variance(X) + Variance(Y)
Variance(X – Y) = Variance(X) + Variance(Y)

The constant c has no effect on the variance

Example: Suppose that one performs two
independent tasks (A and B):
X = time to perform task A (normal with mean 25
minutes and standard deviation of 3 minutes.)
Y = time to perform task B (normal with mean 15
minutes and std dev 2 minutes.)
X and Y independent so T = X + Y = total time is normal
with mean   25  15  40
standard deviation   32  2 2  3.6
What is the probability that the two tasks take more than 45
minutes to perform?
 45  40 
P T  45  P Z    P Z  1.39  .0823
 3.6 
Example 2:
A student will take three tests in the next three days

1. X (Mathematics) has a Normal distribution with

mean m = 90 and standard deviation s = 3.
2. Y (English Literature) has a Normal distribution
with mean m = 60 and standard deviation s = 10.
3. Z (Social Studies) has a Normal distribution with
mean m = 70 and standard deviation s = 7.
Overall score, S = 0.50 X (Mathematics) + 0.30 Y
(English Literature) + 0.20 Z (Social Studies) +
10 (Bonus marks)
 Graphs
0.14 X (Mathematics)
m = 90, s = 3.
0.12

0.1

0.08
Z (Social Studies)
0.06
m = 70 , s = 7.

0.04
Y (English Literature)
m = 60, s = 10.
0.02

0
0 20 40 60 80 100
Determine the distribution of
S = 0.50 X + 0.30 Y + 0.20 Z + 10
S has a normal distribution with
Mean mS = 0.50 mX + 0.30 mY + 0.20 mZ + 10
= 0.50(90) + 0.30(60) + 0.20(70) + 10
= 45 + 18 + 14 +10 = 87

 0.5    0.3    0.2  

2 2 2
s  2
X
2
Y
2
Z

 0.5 3   0.3 10   0.2  7

2 2 2
 2 2 2

 2.25  9  1.96  13.21  3.635

 Graph
0.12

0.1
distribution of
0.08
S = 0.50 X + 0.30 Y + 0.20 Z + 10

0.06

0.04

0.02

0
0 20 40 60 80 100
 Sampling Theory

Determining the distribution of Sample

statistics
Combining Random Variables
Sums, Differences, Linear Combinations of R.V.’s

A linear combination of random variables, X, Y, . . . is

a combination of the form:
L = aX + bY + … + c (a constant)
where a, b, etc. are numbers – positive or negative.
Most common:
Sum = X + Y Difference = X – Y
Others
Averages = 1/3 X + 1/3 Y + 1/3 Z
Weighted averages = 0.40 X + 0.25 Y + 0.35 Z
Means of Linear Combinations

If L = aX + bY + … + c

The mean of L is:

Mean(L) = a Mean(X) + b Mean(Y) + … + c
mL = a mX + b mY + … + c
Most common:
Mean( X + Y) = Mean(X) + Mean(Y)
Mean(X – Y) = Mean(X) – Mean(Y)
Variances of Linear Combinations

If X, Y, . . . are independent random variables and

L = aX + bY + … + c then
Variance(L) = a2 Variance(X) + b2 Variance(Y) + …
 L2  a 2 X2  b 2 Y2 
Most common:
Variance( X + Y) = Variance(X) + Variance(Y)
Variance(X – Y) = Variance(X) + Variance(Y)

The constant c has no effect on the variance

Normality of Linear Combinations
If X, Y, . . . are independent Normal random variables
and
L = aX + bY + … + c
then L is Normal with
mean 2

 L  a X  b Y    c
and standard deviation

 L  a   b  
2 2
X
2 2
X
In particular:
X + Y is normal with mean  X  Y
standard deviation  X2   Y2
X – Y is normal with mean  X  Y
standard deviation  X2   Y2
The distribution of the sample
mean
 The distribution of averages (the mean)

• Let x1, x2, … , xn denote n independent random

variables each coming from the same Normal
distribution with mean m and standard deviation s.
• Let n

x i
1 1 1
x i 1
   x1    x2      xn
n n n n
What is the distribution of x ?
 The distribution of averages (the mean)
Because the mean is a “linear combination”
1 1 1
 x     x1     x2       xn
n n n
1 1 1 1
             n    
n n n n
and
2 2 2
1 2 1 2 1 2
     x1     x2       xn
2
x
n n n
2 2 2
1 2 1 2 1 2 2 2
              n 2 
n n n n n
Thus if x1, x2, … , xn denote n independent random
variables each coming from the same Normal
distribution with mean m and standard deviation s.
Then n

x i
1 1 1
x i 1
   x1    x2      xn
n n n n
has Normal distribution with
mean  x   and
 2
variance  x2 
n

standard deviation  x 
n
 Graphs
0.08
The probability
distribution of
0.06 the mean

0.04 
n
The probability
distribution of
0.02 individual
observations


0
150 170 190 210  230 250 270 290 310
Summary
• The distribution of the sample mean x is Normal.
• The distribution of the sample mean has exactly the
same mean as the population (m).
• The distribution of the sample mean has a smaller
standard deviation then the population.

compared to  
n
• Averaging tends to decrease variability
• An Excel file illustrating the distribution of the
sample mean x
Example
• Suppose we are measuring the cholesterol level of
men age 60-65
• This measurement has a Normal distribution with
mean m = 220 and standard deviation s = 17.
• A sample of n = 10 males age 60-65 are selected and
the cholesterol level is measured for those 10 males.
• x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, are those 10
measurements
Find the probability distribution of x ?
Compute the probability that x is between 215 and 225
Solution

Find the probability distribution of x

Normal with  x    220

 17
and  x    5.376
n 10
P  215  x  225
 215  220 x  220 225  220 
 P   
 5.376 5.376 5.376 
 P  0.930  z  0.930  0.648
 The Central Limit Theorem

The Central Limit Theorem (C.L.T.) states that if n

is sufficiently large, the sample means of random
samples from any population with mean m and finite
standard deviation s are approximately normally
distributed with mean m and standard deviation  .
n
Technical Note:
The mean and standard deviation given in the CLT
hold for any sample size; it is only the “approximately
normal” shape that requires n to be sufficiently large.
 Graphical Illustration of the Central Limit Theorem

Original Population Distribution of x:

n=2

10 20 30 x 10 20 30 x

Distribution of x: Distribution of x:
n = 10 n = 30

10 x 10 20 x
 Implications of the Central Limit Theorem

• The Conclusion that the sampling distribution of the

sample mean is Normal, will to true if the sample size
is large (>30). (even though the population may be non-
normal).
• When the population can be assumed to be normal, the
sampling distribution of the sample mean is Normal, will
to true for any sample size.
• Knowing the sampling distribution of the sample mean
allows to answer probability questions related to the
sample mean.
 Example
Example: Consider a normal population with m = 50 and s = 15.
Suppose a sample of size 9 is selected at random. Find:

1) P ( 45 £ x £ 60)
2) P ( x £ 47.5)

Solutions: Since the original population is normal, the distribution of the

sample mean is also (exactly) normal

1) m x = m = 50

2) s x = s n = 15 9 = 15 3 = 5
 Example

45 50 60 x
- 1.00 0 2.00 z

x- æ45 - 50 60 - 50ö

z= ; P (45 £ x £ 60) = Pç £z £ ÷
s n è 5 5 ø
= P(-1.00 £ z £ 2.00)
= 0.8413 - 0.0228 = 0.8185
 Example

0.3085

47.5 50 x
-0.50 0 z

x- æ x - 50 47.5 - 50ö

z= ; P ( x £ 47.5) = Pç £ ÷
s n è 5 5 ø
= P( z £ -.5)
= 0.5000 - 01915
. = 0.3085
 Example
Example: A recent report stated that the average cost of a hotel
room in Toronto is $109/day. Suppose this figure is correct
and that the standard deviation is known to be $20.

1) Find the probability that a sample of 50 hotel rooms selected at

random would show a mean cost of $105 or less per day.
2) Suppose the actual sample mean cost for the sample of 50 hotel
rooms is $120/day. Is there any evidence to refute the claim of
$109 presented in the report?
Solution:
• The shape of the original distribution is unknown, but the sample
size, n, is large. The CLT applies.
• The distribution of x is approximately normal
m x = m = 109 s x = s n = 20 50 » 2.83
 Example
1)

0.0793

105 109 x
 141
. 0 z

x- æ 105 - 109ö

z= ; P( x £ 105) = Pçz £ ÷
s n è 2.83 ø
= P ( z £ -141
. )
= 0.0793
2) • To investigate the claim, we need to examine how likely an
observation is the sample mean of $120

• Consider how far out in the tail of the distribution of the sample
mean is $120
x- æ -
z= ; P ( x ³ 120) = Pçz ³ 120 109ö ÷
s n è 2.83 ø
= P ( z ³ 389
. )
= 1.0000 - 0.9999 = 0.0001

• Since the probability is so small, this suggests the observation of

$120 is very rare (if the mean cost is really $109)
• There is evidence (the sample) to suggest the claim of  = $109 is
likely wrong
 Summary
• The mean of the sampling distribution of x is equal to
the mean of the original population:  x  
• The standard deviation of the sampling distribution of x
(also called the standard error of the mean) is equal to the
standard deviation of the original population divided by
the square root of the sample size: 
x 
n
• The distribution of x is (exactly) normal when the
original population is normal
• The CLT says: the distribution of x is approximately
normal regardless of the shape of the original
distribution, when the sample size is large enough!
Sampling Distribution of a
Sample Proportion
 Sampling Distribution for Sample Proportions
Let p = population proportion of interest
or binomial probability of success.
Let X no. of succeses
pˆ  
n no. of bimomial trials
= sample proportion or proportion of
successes.
Then the sampling distributi on of p̂
is approximately a normal distribution with
p (1  p )
mean  pˆ  p  pˆ 
n
 Logic
Recall X = the number of successes in n trials has a
Binomial distribution with parameters n and p (the
probability of success).
Also X has approximately a Normal distribution
with
mean m = np and
standard deviation   npq  np (1  p )
X 1
Then the sampling distribution of pˆ   X
n n
is a normal distribution with
1 1
mean  pˆ        np  p
n n
1 1 p (1  p )
and  pˆ        np (1  p ) 
n n n
 Sampling distribution of p̂

30
p  1  p
 pˆ 
25 n
20

15  p̂  p
c
10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Example Sample Proportion Favoring a
Candidate
Suppose 20% all voters favor Candidate A.
Pollsters take a sample of n = 600 voters. Then
the sample proportion who favor A will have
approximately a normal distribution with
mean  pˆ  p  0.20

p (1  p ) 0.20(0.80)
 pˆ    0.01633
n 600
 Sampling distribution of p̂
30

25

20

15
c
10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Using the Sampling distribution:
Suppose 20% all voters favor Candidate A. Pollsters
take a sample of n = 600 voters.

Determine the probability that the sample proportion

will be between 0.18 and 0.22

i.e. the probability, P  0.18  pˆ  0.22

 Solution:
Recall  pˆ  p  0.20

p (1  p ) 0.20(0.80)
 pˆ    0.01633
n 600

 0.18  0.20 pˆ  0.20 0.22  0.20 

P  0.18  pˆ  0.22  P   
 00.1633
.01633 0.1633
0.01633 0.01633 
0.1633

 P  1.225  z  1.225  0.8897  0.1103  0.7794

Sampling Theory - Summary
The distribution of sample statistics
 Distribution for Sample Mean

If data is collected from a Normal distribution

with mean m and standard deviation s then:
the sampling distribution of x
is a normal distribution with

mean  x   and standard deviation  x 
n
 The Central Limit Thereom

If data is collected from a distribution (possibly non

Normal) with mean m and standard deviation s then:
the sampling distribution of x
is a approximately normal (for n > 30) with

mean  x   and standard deviation  x 
n
 Distribution for Sample Proportions
Let p = population proportion of interest
or binomial probability of success.
Let X no. of succeses
pˆ  
n no. of bimomial trials
= sample proportion or proportion of
successes.
Then the sampling distributi on of p̂
is approximately a normal distribution with
p (1  p )
mean  pˆ  p  pˆ 
n
Sampling distribution of a
differences
Sampling distribution of a difference in two
Sample means
Recall

If X, Yare independent normal random variables, then :

X – Y is normal with

mean  X  Y
standard deviation  X2   Y2
 Comparing Means

Situation
• We have two normal populations (1 and 2)
• Let m1 and s1 denote the mean and standard deviation of
population 1.
• Let m2 and s2 denote the mean and standard deviation of
population 2.
• Let x1, x2, x3 , … , xn denote a sample from a normal
population 1.
• Let y1, y2, y3 , … , ym denote a sample from a normal
population 2.
• Objective is to compare the two population means
We know that:
1
x is Normal with mean  x  1 and  x 
n
and
2
y is Normal with mean  y  2 and  y 
m

Thus D  x  y is Normal with mean

 x  y   x   y  1 -  2

 2
 2
 x  y =  x2   y2  1
 2
n m
Example
Consider measuring Heart rate two minutes after a twenty
minute exercise program.

There are two groups of individuals

1. Those who performed exercise program A (considered to be
heavy).
2. Those who performed exercise program B (considered to be
light).

The average Heart rate for those who performed exercise

program A was m1 = 110 with standard deviation, s1 = 7.3, while
the average Heart rate for those who performed exercise
program B was m2 = 95 with standard deviation, s2 = 4.5.
 0.1 Heart rate for
0.09 program B
0.08 Heart rate for
0.07 program A
0.06

0.05

0.04

0.03

0.02

0.01

0
80 90 100 110 120 130
-0.01
Situation
• Suppose we observe the heart rate of n = 15 subjects on
program A.
• Let x1, x2, x3 , … , x15 denote these observations.
• We also observe the heart rate of m = 20 subjects on
program B.
• Let y1, y2, y3 , … , y20 denote these observations.
• What is the probability that the sample mean heart rate for
Program A is at least 8 units higher than the sample mean
heart rate for Program B?
We know that:
7.3
x is Normal with mean  x  110 and  x 
15
and
4.5
y is Normal with mean  y  95 and  y 
20

and D  x  y is Normal with mean

 x  y   x   y  110 - 95  15

 2
 2
7.32
4.5 2
 x  y =  x2   y2  1
 2    2.1366
n m 15 20
 distn of
0.45
sample mean
0.4
for program B distn of
0.35 sample mean
0.3 program A
0.25

0.2

0.15

0.1

0.05

0
80 90 100 110 120 130
-0.05
 0.2

0.18 distn of difference

0.16 in sample means, D
0.14

0.12

0.1

0.08

0.06

0.04

0.02

0
0 5 10 15 20 25 30
• What is the probability that the sample mean heart rate for
Program A is at least 8 units higher than the sample mean
heart rate for Program B?

Solution
want P  x  y  8  P  x  y  8  P  D  8
 D  15 8  15 
 P    P  z  3.28
 2.1366 2.1366 

 1  0.0005  0.9995
Sampling distribution of a difference in two
Sample proportions
 Comparing Proportions
Situation
• Suppose we have two Success-Failure experiments
• Let p1 = the probability of success for experiment 1.
• Let p2 = the probability of success for experiment 2.
• Suppose that experiment 1 is repeated n1 times and
experiment 2 is repeated n2
• Let x1 = the no. of successes in the n1 repititions of
experiment 1, x2 = the no. of successes in the n2 repititions
of experiment 2.
x1 x2
pˆ1 = and pˆ 2 =
n1 n2
x1 x2
What is the distribution of D  pˆ1  pˆ 2 =  ?
n1 n2
We know that:
x1
pˆ1 = is Normal with mean  pˆ1  p1
n1
p1  1- p1 
and  pˆ1 
n1
x2
Also pˆ 2 = is Normal with mean  pˆ 2  p2
n2
p2  1- p2 
and  pˆ 2 
n2
Thus D  pˆ1  pˆ 2 is Normal with mean
 pˆ1  pˆ 2   pˆ1   pˆ 2  p1 - p2
p1  1  p1  p2  1  p2 
 pˆ1  pˆ 2 =   
2
pˆ1
2
pˆ 2  
n1 n2
Example
The Globe and Mail carried out a survey to investigate
the “State of the Baby Boomers”. (June 2006)

Two populations in the study

1. Baby Boomers (age 40 – 59) (n1 = 664)
2. Generation X (age 30 – 39) (n2 = 342)
One of questions
“Are you close to your parents? – Yes or No”
Suppose that the proportions in the two populations were:
• Baby Boomers – 40% yes (p1 = 0.40)
• Generation X – 20% yes (p2 = 0.20)

What is the probability that this would be observed in

the samples to a certain degree?

^ – p^ ≥ 0.15]?
What is P[p1 2
Solution:
x1
pˆ1 = is Normal with mean  pˆ1  p1  0.40
n1
p1  1- p1 
and  pˆ1 
n1
0.40  1- 0.40 
  0.019012
664
x2
Also pˆ 2 = is Normal with mean  pˆ 2  p2  0.20
n2
p2  1- p2 
and  pˆ 2 
n2

0.20  1- 0.20 
  0.02163
342
 distn of sample
25 proportion for Gen X

20
distn of sample proportion
for Baby Boomers
15

10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Now D  pˆ1  pˆ 2 is Normal with mean

 D   pˆ1  pˆ 2   pˆ1   pˆ 2  p1 - p2  0.4  0.2  0.2

p1  1  p1  p2  1  p2 
 D   pˆ1  pˆ 2 =   
2
pˆ1
2
pˆ 2  
n1 n2

0.4  1  0.4  0.2  1  0.2 

 
664 342
 0.028797
 Distribution of D  pˆ1  pˆ 2
16

14

12

10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Now

P  pˆ1  pˆ 2  0.15  P  D  0.15

 D   D 0.15  0.2 
 P  
 D 0.028797 
 P  z  1.74  1  0.0409  .9591
Sampling distributions

Summary
 Distribution for Sample Mean

If data is collected from a Normal distribution

with mean m and standard deviation s then:
the sampling distribution of x
is a normal distribution with

mean  x   and standard deviation  x 
n
 The Central Limit Thereom

If data is collected from a distribution (possibly non

Normal) with mean m and standard deviation s then:
the sampling distribution of x
is a approximately normal (for n > 30) with

mean  x   and standard deviation  x 
n
 Distribution for Sample Proportions
Let p = population proportion of interest
or binomial probability of success.
Let X no. of succeses
pˆ  
n no. of bimomial trials
= sample proportion or proportion of
successes.
Then the sampling distributi on of p̂
is approximately a normal distribution with
p (1  p )
mean  pˆ  p  pˆ 
n
Distribution of a difference in two sample Means
D  x  y is Normal with mean
 x  y   x   y  1 -  2

 2
 2
 x  y =  x2   y2  1  2
n m
Distribution of a difference in two sample proportions

D  pˆ1  pˆ 2 is Normal with mean

 pˆ1  pˆ 2   pˆ1   pˆ 2  p1 - p2

p1  1  p1  p2  1  p2 
 pˆ1  pˆ 2 =   
2
pˆ1
2
pˆ 2  
n1 n2
The Chi-square (c 2) distribution
 The Chi-squared distribution
with
n degrees of freedom

Comment: If z1, z2, ..., zn are independent

random variables each having a standard
normal distribution then
U= z1
2
 z 2
2    z 2


has a chi-squared distribution with n

degrees of freedom.
 The Chi-squared distribution
with
0.18 n degrees of freedom

0.12

0.06

0
0 10 20

n - degrees of freedom
0.5
2 d.f.
0.4

0.3
3 d.f.

0.2
4 d.f.

0.1

2 4 6 8 10 12 14
Statistics that have the Chi-squared
distribution:
x  Eij 
2
c r c r
     rij2
2 ij
1.
j 1 i 1 Eij j 1 i 1

The statistic used to detect independence

between two categorical variables
d.f. = (r – 1)(c – 1)
Let x1, x2, … , xn denote a sample from the
normal distribution with mean m and
standard deviation s, then
r

 x  x 
2
i
2. U i 1
2
(n  1) s 2

2

has a chi-square distribution with d.f. = n – 1.

Example
Suppose that x1, x2, … , x10 is a sample of
size n = 10 from the normal distribution with
mean m =100 and standard deviation s =15.
Suppose that r

 x  x 
2
i
s i 1
n 1
is the sample standard deviation.
Find P  10  s  20 .
 Note r

 x  x 
2
i
(n  1) s 2
(9) s 2
U i 1
 
 2
2 (15) 2

has a chi-square distribution with

d.f. = n – 1 = 9
P  10  s  20  P 100  s  400 
 2

 9  100  9 s 2 9  400  
 P 2
 2  2 
 15 15 15 
 P  4  U  16
chi-square distribution with d.f. = n – 1 = 9

P  4  U  16

We do not have tables to compute this area

The excel function
CHIDIST(x,df) computes P  x  U 

P  4  U  16

P x  U 

x
P  4  U  16  CHIDIST(4,9)-CHIDIST(16,9)
= 0.91141 - 0.06688 = 0.84453
Statistical Inference