Random Variables

Lessandro Estelito O. GARCIANO

Professor
Department of Civil Engineering
De La Salle University
E-mail: [email protected]
 Random Variable
A random variable is defined as a function that maps events onto
intervals on the axis of real numbers.

Schematic representation of a random variable as a function

 Basic Functions
o The probability mass function (PMF): 𝑝𝑥 (𝑥) is the probability that a
discrete random variable 𝑋 is equal to a specific value 𝑥 where 𝑥 is
a real number.

𝑝𝑥 (𝑥) ≤ 𝑃(𝑋 = 𝑥)

𝑝𝑥 1 = 𝑃 𝑋 = 1 = 0.05
𝑝𝑥 2 = 𝑃 𝑋 = 2 = 0.20
𝑝𝑥 3 = 𝑃 𝑋 = 3 = 0.65
𝑝𝑥 4 = 𝑃 𝑋 = 4 = 0.10

A probability mass function (PMF)

Note: the random variable (with an uncertain value) is denoted by a capital letter, whereas a specific value or realization of the variable is denoted
by a lower case letter.
 Basic Functions
o The cumulative distribution function (CDF): 𝐹𝑥 (𝑥) is the total sum
(or integral) of all probability functions (continuous and discrete)
corresponding to a value less than or equal to 𝑥.
𝑝𝑥 1 = 𝑃 𝑋 = 1 = 0.05
𝑝𝑥 2 = 𝑃 𝑋 = 2 = 0.20 𝐹𝑥 𝑥 = 𝑃(𝑋 ≤ 𝑥)
𝑝𝑥 3 = 𝑃 𝑋 = 3 = 0.65
𝑝𝑥 4 = 𝑃 𝑋 = 4 = 0.10

A probability mass function (PMF) A CDF for a discrete random variable

 Basic Functions
o For continuous random variables, the probability density function
(PDF) 𝑓𝑥 (𝑥) is defined as the first derivative of the CDF.
𝑥
𝑑
𝑓𝑥 𝑥 = 𝐹𝑥 𝑥 𝐹𝑥 𝑥 = න 𝑓𝑥 𝜉 𝑑𝜉
𝑑𝑥 −∞

A probability density function (PDF) A CDF for a discrete random variable

 Properties of Probability Functions
Several important properties of the CDF and enumerated below. Any function that satisfies
these six conditions can be considered a CDF.
1. The definition of a CDF is the same for both discrete and continuous RVs.
2. The CDF is a positive, non-decreasing function whose value is between 0 and 1.

0 ≤ 𝐹𝑥 𝑥 ≤ 1

3. If 𝑥1 < 𝑥2 , then 𝐹𝑥 𝑥1 ≤ 𝐹𝑥 𝑥2 ,
4. then 𝐹𝑥 −∞ = 0.
5. then 𝐹𝑥 +∞ = 1.
6. For continuous random variables,

𝑃 𝑎 ≤ 𝑥 ≤ 𝑏 ≤ 𝐹𝑥 𝑏 − 𝐹𝑥 𝑎 = න 𝑓𝑥 𝜉 𝑑𝜉
𝑎

Graphical representation of Fx(b) – Fx(a)

 Parameters of a Random Variable
Basic Parameters
Consider the random variable 𝑋. The mean value of 𝑋 is denoted by 𝜇𝑥 .

For a continuous random variable, the mean value is defined as

+∞

𝜇𝑥 = න 𝑥 𝑓𝑥 𝑥 𝑑𝑥
−∞

For a discrete random variable, the mean value is defined as

𝜇𝑥 = ෍ 𝑥𝑖 𝑝𝑥 𝑥𝑖
𝑎𝑙𝑙 𝑥𝑖
 Parameters of a Random Variable
Basic Parameters
The expected value of 𝑋 is commonly denoted by E 𝑋 and is equal to the mean
value of the variable as previously defined.
𝐸 𝑋 = 𝜇𝑥
The nth moment of 𝑋 is defined for continuous random variables as
+∞

𝐸 𝑋 𝑛 = න 𝑥 𝑛 𝑓𝑥 𝑥 𝑑𝑥
−∞

and discrete random variables, the nth moment is defined as

𝐸 𝑋 𝑛 = ෍ 𝑥𝑖𝑛 𝑝𝑥 𝑥𝑖
𝑎𝑙𝑙 𝑥𝑖
 Parameters of a Random Variable
Basic Parameters
The variance of 𝑋 is commonly denoted as by 𝜎𝑥2 , is defined as the expected value of
𝑋 − 𝜇𝑥 2 and is equal to.
+∞

𝜎𝑥2 = න 𝑥 − 𝜇𝑥 2 𝑓𝑥 𝑥 𝑑𝑥
−∞
for a continuous RV and

𝜎𝑥2 = ෍ 𝑥 − 𝜇𝑥 2𝑝
𝑥 𝑥𝑖
𝑎𝑙𝑙 𝑥𝑖

for a discrete RV.

 Parameters of a Random Variable
Basic Parameters
An important relationship exists among the mean, variance and second moment of a
RV 𝑋.
𝜎𝑥2 = 𝐸 𝑋 2 − 𝜇𝑥
The standard deviation of 𝑋 is defined as the positive square root of the variance.

𝜎𝑥2 = ෍ 𝑥 − 𝜇𝑥 2𝑝
𝑥 𝑥𝑖
𝑎𝑙𝑙 𝑥𝑖

The non-dimensional coefficient of variation, 𝑉𝑥 or COV, is defined as the standard

deviation divided by the mean.
𝜎𝑥
𝑉𝑥 =
𝜇𝑥
 Parameters of a Random Variable
Sample Parameters
If a set of n observations 𝑥1 , 𝑥2 , ⋯ , 𝑥𝑛 , is obtained for a particular random variable
𝑋, then the true mean 𝜇𝑥 can be approximated by the sample mean 𝑥ҧ and the true
standard deviation can be approximated by the sample standard deviation 𝑠𝑥 .
The sample mean is calculated as
𝑛
1
𝑥ҧ = ෍ 𝑥𝑖
𝑛
𝑖=1

The sample standard deviation is calculated as

σ𝑛𝑖=1 𝑥𝑖 − 𝑥ҧ 2 σ𝑛𝑖=1 𝑥𝑖2 − 𝑛 𝑥ҧ 2

𝑠𝑥 = =
𝑛−1 𝑛−1
 Parameters of a Random Variable
Standard form
Let 𝑋 be a random variable. The standard form of 𝑋, denoted by 𝑍, is defined as.
𝑋 − 𝜇𝑥
𝑍=
𝜎𝑥
The mean of the standard form of a random variable is 0 and its variance is 1.
 Common Random Variables
o Uniform
o Normal (Gaussian)
o Lognormal
o Gamma
o Extreme Value (Types I, II and III)
o Poisson
 Uniform Random Variable
For a uniform (or rectangular) random
variable, the pdf has a constant value for
all possible values of the random variable
on the interval 𝑎, 𝑏 .
1
𝑃𝐷𝐹 = 𝑓(𝑥) = ቐ 𝑏 − 𝑎 , 𝑎≤𝑥≤𝑏
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

The CDF for this random variable is given

by
𝑥−𝑎
𝐹(𝑥) =
𝑏−𝑎 PDF and CDF of a uniform random variable
 Normal Random Variable
o Gaussian or the Normal distribution N(m, s)
1  1  x − m 2 
f (x) = exp  −    - x
s 2  2  s  

m and s are the parameters of the distribution.

 Normal Random Variable
o Standard Normal distribution N(0,1)

1 2
1 − ( x)
f (x) = exp 2 - x
2

o Due to its wide usage, a special notation F(z) is commonly

used to designate the CDF of the standard normal variates.

Φ(𝑧) = 𝐹𝑧 (𝑧)
 Standard Normal Random Variable

o CDF of N(0,1)
o The value of the standard normal variate at a cumulative
probability p would be denoted as.

s p = F −1 ( p)

o The value for Φ𝑍 can be

obtained by using:

Φ𝑍 = 1 − Φ −𝑍
 Standard Normal Random Variable
o Let X be a random variable. The standard form of X, denoted by Z, is
defined as
𝑋 − 𝜇𝑋
𝑍=
𝜎𝑋

o By the definition of CDF, we can write:

𝑥−𝜇𝑥
𝐹𝑋 𝑋 = 𝜙 = 𝐹𝑍 𝑧
𝜎𝑋

o A relationship can be derived relating the PDF of any normal

random variable to the PDF of the standard normal variable
𝑑 𝑑 𝑥 − 𝜇𝑥 1 𝑥 − 𝜇𝑥
𝑓𝑋 𝑥 = 𝐹𝑋 𝑥 = 𝜙 = 𝜙
𝑑𝑥 𝑑𝑥 𝜎𝑋 𝜎𝑋 𝜎𝑋
 Standard Normal Random Variable

o The distribution functions for a normal random variable

have some important properties which are summarized as
follows
1. The PDF 𝑓𝑋 𝑥 is symmetrical about the mean 𝜇𝑋
𝑓𝑋 𝜇𝑋 + 𝑥 = 𝑓𝑋 𝜇𝑋 − 𝑥
2. Because the normal distribution is symmetrical about its
mean value, the sum of 𝐹𝑋 𝜇𝑋 + 𝑥 and 𝐹𝑋 𝜇𝑋 − 𝑥 is equal
to 1.
𝐹𝑋 𝜇𝑋 + 𝑥 + 𝐹𝑋 𝜇𝑋 − 𝑥 = 1
Appendix B
Appendix B
Appendix B
 Example Problem

If Z is a standard normal random variable, and z = -

2.16, what are the PDF and CDF values?
If z = +1.51, what is 𝜙 1.51 .
Given 𝜙𝑍 = 0.80 × 10−4 , what is the corresponding
value of Z ?
If Z is a standard normal random variable, and z = -2.16,
(a) what are the PDF and CDF values?
(b) If z = +1.51, what is 𝜙 1.51
(c) Given 𝜙𝑍 = 0.80 × 10−4 , what is the corresponding value of Z?

a) From Appendix B, Φ 𝑧 = −2.16 = 0.0154.

1 −1 𝑧 2
𝜙 𝑧 = 𝑒 2 = 𝜙 𝑧 = −2.16 = 0.0387
2𝜋

b) From Appendix B, Φ 𝑧 = −1.51 = 0.0655

𝜙 𝑧 = 1 − 𝜙 −𝑧 = 1 − 0.0655 = 0.9345

c) From Appendix B, we note that this value of Φ 𝑧 does not

correspond to a specific tabulated value of z. Therefore, we must use
interpolation. For z = -3.77, Φ = 0.816 × 10−4 and for z = -3.78,
Φ = 0.784 × 10−4 . Interpolating between these values, the desired
value of z is about -3.775.
 Example Problem
Assume X is a normal random variable with 𝜇𝑋 = 1500 and
𝜎𝑋 = 200. The PDF for the variable is shown. Calculate the
following:
(a) Fx (1300);
(b) Fx (1900);
(c) Fx (1700);
(d) fx(x) for x = 1300;
(e) fx(1500)
Assume X is a normal random variable with 𝜇𝑋 = 1500 and 𝜎𝑋 = 200. The PDF for the
variable is shown. Calculate the following: (a) Fx(1300); (b) Fx(1900); (c) Fx(1700); (d) fx(x)
for x = 1300; (e) fx(1500)

c) 𝐹𝑥 1500 + 200 = 1 − 𝐹𝑥 1500 − 200

𝐹𝑥 1500 + 200 = 1 − 0.159
𝐹𝑥 1500 + 200 = 0.841

1 𝑥−𝜇𝑋 1 1300−1500 1
d) 𝑓𝑥 𝑥 = 𝜙 = 𝜙 = 𝜙 −1
𝑥−𝜇𝑥 𝜎𝑋 𝜎𝑋 200 200 200
a) 𝐹𝑥 𝑥 = 𝜙 1 1 −1 𝑧 2 1
𝜎𝑥
𝑓𝑥 𝑥 = 𝜙 −1 = 𝑒 2 = 0.242 = 0.00121
1300 − 1500 200 2𝜋 200
𝐹𝑥 1300 = 𝜙 = 𝜙(−1)
200
From Appendix B, 𝜙 −1 = 0.159

1 𝑥−𝜇𝑋 1 1500−1500 1
e) 𝑓𝑥 𝑥 = 𝜙 = 𝜙 = 𝜙 0
𝜎𝑋 𝜎𝑋 200 200 200
𝑥−𝜇𝑥
b) 𝐹𝑥 𝑥 = 𝜙 1 1 −1 𝑧 2 1
𝜎𝑥 𝑓𝑥 𝑥 = 𝜙 0 = 𝑒 2 = 0.399 = 0.00199
1900 − 1500 200 2𝜋 200
𝐹𝑥 1900 = 𝜙 = 𝜙(+2)
200
From Appendix B, 𝜙 2 = 1 − 𝜙 −2
𝜙 2 = 1 − 0.0228 = 0.977
 Lognormal Random Variable

o Lognormal distribution

1  1  ln x − l  
2

f (x) = exp  −   
(z x ) 2  2  z  

x0

l and z are the parameters of

the distribution.
 Example Problem

Let X be a lognormal random variable with a mean

value of 250 and a standard deviation of 30. Calculate
Fx(200) and fx(200).
Let X be a lognormal random variable with a mean value of 250 and a
standard deviation of 30. Calculate Fx(200) and fx(200)

𝜎𝑥 30 ln 𝑥 − 𝜇ln 𝑥
𝑉𝑥 = = = 0.12 𝐹𝑥 (200) = 𝜙
𝜇𝑥 250 𝜎ln 𝑥
ln 200 − 5.51
2 2 𝐹𝑥 200 = 𝜙 = 𝜙(−1.77)
𝜎ln 𝑥 = ln 𝑉𝑥 + 1 = 0.0143 0.1196
𝐹𝑥 200 = 0.0384
𝜎ln 𝑥 = 0.1196
1 ln 𝑥 − 𝜇ln 𝑥
1 2 𝑓𝑥 (200) = 𝜙
𝜇ln 𝑥 = ln 𝜇𝑥 − 𝜎ln 𝑥 𝑥𝜎ln 𝑥 𝜎ln 𝑥
2 1
𝜇ln 𝑥 = ln 250 − 0.5 0.0143 = 5.51 𝑓𝑥 200 = 𝜙 −1.77
200 0.1196
0.0833
𝑓𝑥 200 = = 0.00348
23.92
 Gamma Distribution

o Gamma distribution

𝜆 𝜆𝑥 𝑘−1 𝑒 −𝜆𝑥
𝑓𝑋 𝑥 = 𝑓𝑜𝑟 𝑥 > 0
Γ 𝑘

l And k are the distribution

parameters

o The mean and variance can be

calculated as follows
PDFs of gamma random variables
𝑘 𝑘
𝑢𝑥 = 𝜎𝑥2 = 2
𝜆 𝜆
 Extreme Type I
(Gumbel Distribution, Fisher-Tippett I)

o Extreme value distribution are useful to characterize the

probabilistic nature of the extreme values of some
phenomenon overtime.
𝐹𝑋 𝑥 = exp − exp −𝛼 𝑥 − 𝑢 𝑓𝑜𝑟 − ∞ < 𝑥 < ∞

o The mean and standard deviation

can be calculated as follows
0.577 1.282
𝑢𝑥 ≈ 𝑢 + 𝜎𝑥 ≈
𝛼 𝛼

PDF of an extreme type I random variable

 Extreme Type II

o Extreme type II gives the best approximation of the

distribution of the maximum seismic load applied to a
structure.
𝑢 𝑘
−
𝐹𝑋 𝑥 = 𝑒 𝑥 𝑓𝑜𝑟 0 < 𝑥 < ∞

o The mean and variance can be

calculated as follows
1
𝑢𝑥 = 𝑢Γ 1 − 𝑓𝑜𝑟 𝑘 > 1
𝑘

2 1
𝜎𝑥2 = 𝑢2 Γ 1 − − Γ2 1 − 𝑓𝑜𝑟 𝑘 > 2
𝑘 𝑘

PDF of an extreme type II random variable

 Extreme Type III
Weibull Distribution

o Extreme type III is defined by three parameters. There is

different function for the largest and smallest value
For largest value, the CDF is: For smallest value, the CDF is:
𝑤−𝑥 𝑘 𝑥−𝜀
− 𝑤−𝑢 −
𝐹𝑋 𝑥 = 𝑒 𝑓𝑜𝑟 𝑥 ≤ 𝑤 𝐹𝑋 𝑥 = 1 − 𝑒 𝑘 𝑢−𝜀 𝑓𝑜𝑟 𝑥 ≥ 𝜀

o The mean and variance can be calculated as follows

1 1
𝑢𝑥 = 𝑤 − 𝑤 − 𝑢 Γ 1 + 𝑢𝑥 = 𝜀 + 𝑢 + 𝜀 Γ 1 +
𝑘 𝑘

2 1 2 1
𝜎𝑥2 = 𝑤 − 𝑢 2
Γ 1+ − Γ2 1 + 𝜎𝑥2 = 𝑢 − 𝜀 2
Γ 1+ − Γ2 1 +
𝑘 𝑘 𝑘 𝑘
 Probability Distribution Models
o Generalized Extreme Value (GEV)
distribution
1
−
𝜉 𝜉
𝐺 (𝑥 ) = 𝑒𝑥𝑝 1 − (𝑥 − 𝜇)
𝜎

o m , s (>0) and x parameters are

the location, scale and shape
parameters, respectively.
 Poisson Distribution

o Poisson Distribution is a discrete probability distribution

that can be used to calculate the PMF for the number of
occurrences of a particular event in a time or space interval
(0, t). 𝑣𝑡 𝑛
𝑃 𝑁 = 𝑛 𝑖𝑛 𝑡𝑖𝑚𝑒 𝑡 = 𝑒 −𝑣𝑡 𝑛 = 1,2, … , ∞
𝑛!
o The mean and standard deviation can be calculated as follows
𝑢𝑁 = 𝑣𝑡
𝜎𝑁 = 𝑣𝑡

o The return period is simply the reciprocal of the mean

occurrence rate v: 1
𝜏=
𝑣
 Example Problem
Suppose that the average occurrence rate of earthquakes (with
magnitude between 5 and 8) in a region surrounding Los
Angeles, California, has been determined to be 2.14
earthquakes/year. Determine
a) The return period for earthquakes in this magnitude range
b) The probability of exactly three earthquakes (magnitude between
5 and 8) in the next year
c) The annual probability of an earthquake with magnitude between
5 and 8
 Suppose that the average occurrence rate of earthquakes (with magnitude between 5 and 8) in a region
surrounding Los Angeles, California, has been determined to be 2.14 earthquakes/year. Determine
a) The return period for earthquakes in this magnitude range
b) The probability of exactly three earthquakes (magnitude between 5 and 8) in the next year
c) The annual probability of an earthquake with magnitude between 5 and 8

a) c)
1 1
𝜏= = = 0.47 𝑦𝑒𝑎𝑟𝑠 𝑃 𝑎𝑡 𝑙𝑒𝑎𝑠𝑡 𝑜𝑛𝑒 𝑒𝑎𝑟𝑡ℎ𝑞𝑢𝑎𝑘𝑒 = 1 − 𝑃 𝑛𝑜 𝑒𝑎𝑟𝑡ℎ𝑞𝑢𝑎𝑘𝑒𝑠
𝑣 2.14
or

b)
𝑃 𝑁 ≥1 = 1−𝑃 𝑁 =0
3
2.14 1
𝑃 𝑁 = 3 𝑖𝑛 1 𝑦𝑒𝑎𝑟 = 𝑒− 2.14 1
3!
therefore,
0
𝑃 𝑁 = 3 𝑖𝑛 1 𝑦𝑒𝑎𝑟 = 0.192 2.14 1
𝑃 𝑁 ≥1 = 1− 𝑒− 2.14 1
0!
− 2.14 1
𝑃 𝑁 ≥1 = 1−𝑒
𝑃 𝑁 ≥ 1 = 0.88
 Other Probability Distribution Models
o Bernoulli Sequence and the Binomial Distribution
o Geometric Distribution
o Negative Binomial Distribution
o Poisson Process and the Poisson Distribution
o Exponential Distribution
o Geometric Distribution
o Gamma Distribution
o Hypergeometric Distribution
o Beta Distribution
 Common Distributions and Their Parameters
Distribution PDF or PMF Parameters Relations to moments
𝑎+𝑏
𝜇𝑥 =
1 2
Uniform 𝑓(𝑥 ) = ; 𝑎≤𝑥≤𝑏 a,b
𝑏−𝑎 2
1
𝜎𝑥 = (𝑏 − 𝑎 )2
12

1 1 𝑥−𝜇 2 𝜇𝑥 = 𝜇
( )
𝑓 𝑥 = 𝑒𝑥𝑝 − ;
Gaussian 2𝜋𝜎 2 𝜎 m ,s
(Normal) 𝜎𝑥2 = 𝜎 2
−∞ ≤ 𝑥 ≤ +∞

1
1 1 𝑙𝑛𝑥 − 𝜆 2 𝜇𝑥 = 𝑒𝑥𝑝 𝜆 + 𝜁 2
2
Lognormal 𝑓(𝑥 ) = 𝑒𝑥𝑝 −
2 𝜁
; l,z
𝜁𝑥 2𝜋 2
𝜎𝑥2 = 𝜇𝑥2 𝑒 𝜁 − 1 39/71
 Common Distributions and Their Parameters
Distribution PDF or PMF Parameters Relations to moments

𝑛 𝜇𝑥 = 𝑛𝑝
Binomial 𝑓(𝑥 ) = (1 − 𝑝)𝑥−1 ; p
𝑘
𝜎𝑥2 = 𝑛𝑝(1 − 𝑝)
𝑥 = 0,1,2, ⋯ 𝑛

0.5772
Gumbel 𝑓(𝑥 ) = 𝑒𝑥𝑝 −𝑒 −(𝑥−𝑢 ) 𝜇𝑥 = 𝑢 +
𝛼
(Type I u,a
2
largest) 𝜋
𝜎𝑥2 = 2
6𝛼
1
𝜇𝑥 = 𝜀 + (𝑥 − 𝜀 )Γ 1 +
Weibull 𝑘 𝑥−𝜀 𝑘−1 𝑥−𝜀 𝑘 𝑘
−
(Type III ( )
𝑓 𝑥 = 𝑒 𝑤 −𝜀 ;
k,w 2
𝑤−𝜀 𝑤−𝜀 2 2
𝜎𝑥 = (𝑤 − 𝜀 ) Γ 1 +
smallest) 𝑘
𝑥≥𝜀
1
−Γ 2 1 +
𝑘
40/71
 Probability Papers
o Consider a normal random variable 𝑥 with mean value 𝜇𝑥 and
standard deviation 𝜎𝑥 .

o Imagine a transformation in which the S-shaped CDF is

straightened as shown.

o The transformation is such that each point on the original CDF can
only move vertically up or down.

o This transformation is accomplished by altering the scale of the

vertical axis as shown. Take note that the values on the left vertical
axis are not evenly spaced.
S-Shaped CDF for a normal random variable

o If the coordinate pairs for a normal random variable 𝑋 are plotted

on normal probability paper, then the graph will be a straight line.
 Probability Papers
o Probability papers are graph papers that are constructed for plotting
observed data and their corresponding cumulative probabilities (or
frequencies). Probability papers are constructed such that a given
probability paper is associated with a particular probability distribution

o Different probability papers correspond to different probability

distributions

o All CDFs are non-decreasing functions of the random variable and thus has
an “S-shape”. The basic idea is to re-define the vertical scale so that the
CDF will plot as a straight line.

o Conversely, if a set of data plotted on normal probability paper plots as a

straight line, then it is reasonable to model the data using a normal CDF.

Normal Probability Paper

o The slope and y-intercept of the graph can be used to determine the mean
and standard deviation of the distribution.
Normal Probability Paper
Normal Probability Paper

Interpretation of a straight-line plot on normal probability paper

in terms of the mean and standard deviation of the normal
random variable.
 Procedure
1. Arrange the data values 𝑥 in 4. For each 𝑝𝑖 , determine 𝑧𝑖 =
increasing order. Once ordered, the Φ−1 𝑝𝑖 .
first (lowest) value of 𝑥 will be
denoted as 𝑥1 , the next value as 𝑥2
and so on up to the largest value 𝑥𝑁 . 5. Plot the coordinates 𝑥𝑖 , 𝑧𝑖 on
Do not discard repeated values. standard linear graph paper by hand
or using a computer.

2. Associate with each 𝑥𝑖 a cumulative

probability 𝑝𝑖 equal to 6. If the plot appears to follow a straight
line, then it is reasonable to conclude
𝑖 that the data can be modeled using a
𝑝𝑖 =
𝑁+1 normal distribution.
3. If commercial normal probability If the data do not appear to follow a
paper is being used, then plot and go straight line, then the normal
to step 6. Otherwise go to Step 4. distribution is probably not
appropriate.
 Example Problem

Consider the following set of data points:

{x} = {6.5, 5.3, 5.5, 5.9, 6.5, 6.8, 7.2, 5.9, 6.4} (N = 9).
Plot the data on normal probability paper.
Consider the following set of data points:
{x} = {6.5, 5.3, 5.5, 5.9, 6.5, 6.8, 7.2, 5.9, 6.4}(N= 9).
Plot the data on normal probability paper.