What is a Random Variable?

A variable is any characteristic, observed or measured. A variable

can be either random or constant in the population of interest.

Note this differs from common English usage where the word
variable implies something that varies from individual to individual.

For a defined population, every random variable has an associated

distribution that defines the probability of occurrence of each
possible value of that variable (if there are a finitely countable
number of unique values) or all possible sets of possible values (if
the variable is defined on the real line).
Distributions of Random Variables
RVDist-5
RVDist-6
 Probability Distribution
A probability distribution (function) is a list of the probabilities of
the values (simple outcomes) of a random variable.
Table: Number of heads in two tosses of a coin

y P(y)
outcome probability
0 1/4
1 2/4
2 1/4

For some experiments, the probability of a simple 0 p( y ) 1

outcome can be easily calculated using a specific
probability function. If y is a simple outcome and
p(y) is its probability.  p( y ) 1
all y
RVDist-9
RVDist-10
 Discrete Distributions

Relative frequency distributions for “counting” experiments.

• Binomial Distribution
• Poisson Distribution
 Binomial Distribution
• The experiment consists of n identical trials (simple experiments).
• Each trial results in one of two outcomes (success or failure)
• The probability of success on a single trial is equal to  and 
remains the same from trial to trial.
• The trials are independent, that is, the outcome of one trial does not
influence the outcome of any other trial.
• The random variable y is the number of successes observed during
n trials.
n!
P( y )   y (1   ) n  y
y!(n  y )!
n!=1x2x3x…x n
 n Mean
Standard deviation
  n (1   )
RVDist-13
RVDist-14
RVDist-15
 Poisson Distribution
A random variable is said to have a Poisson Distribution with rate
parameter , if its probability function is given by:

e=2.718…
y
P( y )  e  , for y 0,1,2,...
y!

  ,   2
Mean and variance for a Poisson

RVDist-19
 Probability Density Function
A function which integrates to 1 over its range and from which
event probabilities can be determined.

Area under curve f(x)

sums to one.

Random variable range

A theoretical shape - if we were able to sample the

whole (infinite) population of possible values, this is
what the associated histogram would look like.

A mathematical abstraction
 Normal Probability Density Function

 Recall: continuous
Figure: Age distribution
random variables are of a pediatric population
described with with overlying Normal
probability density pdf
function (pdfs)
curves
 Normal pdfs are
recognized by their
typical bell-shape

23
 Area Under the Curve
 pdfs should be viewed
almost like a histogram
 Top Figure: The darker
bars of the histogram
correspond to ages ≤ 9
(~40% of distribution) f ( x) 
1
 x  
 12 
e   

2

2 
 Bottom Figure: shaded
area under the curve
(AUC) corresponds to
ages ≤ 9 (~40% of area)
24
 Parameters μ and σ
 Normal pdfs have two parameters
μ - expected value (mean “mu”)
σ - standard deviation (sigma)

μ controls location σ controls spread

25
Mean and Standard Deviation of Normal
Density

μ
26
 Two types of means and
standard deviations
• The mean and standard deviation from
the pdf (denoted μ and σ) are
parameters
• The mean and standard deviation from
a sample (“xbar” and s) are statistics
• Statistics and parameters are related,
but are not the same thing!

27
t
y

D
i
s
t
r
i
b
u
t
i
o
n
s

68-95-99.7 Rule for

Normal Distributions

 68% of the AUC within ±1σ of μ

 95% of the AUC within ±2σ of μ
 99.7% of the AUC within ±3σ of μ

28
 Example: 68-95-99.7 Rule
Wechsler adult  68% of scores within
intelligence scores: μ±σ
Normally distributed = 100 ± 15
with μ = 100 and σ = 15; = 85 to 115
X ~ N(100, 15)  95% of scores within
μ ± 2σ
= 100 ± (2)(15)
= 70 to 130
 99.7% of scores in
μ ± 3σ =
100 ± (3)(15)
= 55 to 145
7: Normal Probability Distributions 29
 Symmetry in the Tails
Because the Normal
curve is symmetrical
and the total AUC is
exactly 1…

… we can easily
determine the AUC in
95%
tails
7: Normal Probability Distributions 30
 Example: Male Height
 Male height: Normal with μ = 70.0˝ and σ = 2.8˝
 68% within μ ± σ = 70.0  2.8 = 67.2 to 72.8
 32% in tails (below 67.2˝ and above 72.8˝)
 16% below 67.2˝ and 16% above 72.8˝ (symmetry)

7: Normal Probability Distributions 31

 Reexpression of Non-Normal Random
Variables

 Many variables are not Normal but can be

reexpressed with a mathematical transformation
to be Normal
 Example of mathematical transforms used for this
purpose:
 logarithmic
 exponential
 square roots
 Review logarithmic transformations…

7: Normal Probability Distributions 32

t
y

D
i
s
t
r
i
b
u
t
i
o
n
s

Example: Logarithmic
Reexpression
 Prostate Specific Antigen Take exponents of “95% range”
(PSA) is used to screen  e−1.9,1.3 = 0.15 and 3.67
 Thus, 2.5% of non-diseased
for prostate cancer
population have values greater
 In non-diseased than 3.67  use 3.67 as
populations, it is not screening cutoff
Normally distributed, but
its logarithm is:
 ln(PSA) ~N(−0.3, 0.8)
 95% of ln(PSA) within
= μ ± 2σ
= −0.3 ± (2)(0.8)
= −1.9 to 1.3

33
 7.4 Assessing Departures
from Normality
Approximately Same distribution on
Normal histogram Normal “Q-Q” Plot

Normal distributions adhere to diagonal line on Q-Q

plot
7: Normal Probability Distributions 34
 Negative Skew

Negative skew shows upward curve on Q-Q plot

7: Normal Probability Distributions 35
 Positive Skew

Positive skew shows downward curve on Q-Q plot

7: Normal Probability Distributions 36