Module 2.

Discrete Distributions
Student Notes 2.3: Mean, Variance and Chebyshev’s Theorem

Mean, Variance and Moments

Mean: Expected value or mean of a discrete random variable is denoted by µ or E(X) and
defined as µ=E ( x )=∑ xP ( X=x )=∑ x f ( x).
x x

Mean is considered as weighted average.

Variance: Variance of a discrete random variable is denoted by V(X) or σ 2 and defined as
σ 2 = ∑ (x−μ)2 f(x) .
Note that V(X) is nonnegative. The standard deviation () (s.d.) of X is its nonnegative square
root, that is, ¿ √ V (x) .
Variance and standard deviation gives us the spread of X about µ. That is is large if and only if
the values away from µ has large probability of occurrence.

Moments about origin and mean: The rth moment of a discrete random variable X about the
µr=∑ x f (x )
' r
origin is denoted by µr’ are defined as
x

The r moment of a random variable X about the mean µ is denoted by µr and defined as
th

µr=∑ ¿ ¿
x
'
Note: Mean = µ and variance = µ2.
1

Computing formula for variance:

Theorem :The variance V(X)= σ 2 = E(X2) – (E(X))2 = µ 2’ - µ 2.

Proof: By definition of the variance, we have
σ 2 = V(X) =∑ (x−μ)2 f(x) = ∑ ¿¿ ¿ -2µx + µ2) f(x)
=∑ ¿¿ ¿ f(x)- 2µ∑ xf ( x) + µ2∑ f ( x )
= µ2’ - 2 µ(µ) + µ2(1) = µ2’ - µ2
The mean  = 1 measures the average value of X. Var(X)= 2 measures the spread of X
about mean. The variance, hence also standard deviation, is large if X takes values away from
mean with large probability. The third moment about mean 3 describes the symmetry or
skewness about mean whereas 4 measures peakedness or kurtosis of the distribution.

Remark: The mean or variance of a probability distribution is the mean or variance of its
random variable respectively.
Mean and Variance of Uniform distribution:
If X takes values 1, 2, …, n, all equally likely, then X has (discrete) uniform distribution with pdf
f(x)=1/n, x=1,2,…,n. The mean and the variance of the uniform distribution is
2
n+1 2 n −1
μ= and σ =
2 12
Example : An unbiased die is tossed and X = the number which comes on the top of the die.
The probability distribution is given by the following table.

X 1 2 3 4 5 6
f(x) 1/6 1/6 1/6 1/6 1/6 1/6
Find the mean and variance of the given distribution,

Solution: µ=∑ xf ( x)=(1/6 x 1) + (1/6 x 2) + (1/6 x 3) + (1/6 x 4) + (1/6 x 5) + (1/6 x 6) = 7/2.

µ2’ = E (X2) = ∑ x 2 f(x) = 1/6 [1 + 4 + 9 + 16 + 25 + 36] = 91/6.
σ 2 = V(X) = µ2’ - µ2 = 91/6 – 49/4 = 35/12.

Mean and variance of Binomial Distribution

Theorem : If X has binomial distribution with parameters n and p, then
E(X) = np and V(x)= np(1-p).
Proof: If X ~ b(x; n, p), then

()n
P(X = x) = x px(1 – p)n-x , x = 0,1,2,3,……..n

n n
n! p x (1− p)n− x
x=0
n
p ()
µ = ∑ x p (1 – p) =∑
x n− x

x=1 ( x−1 ) ! ( n−x ) !

n
( n−1 ) ! p x−1 (1− p)n−x
= np ∑
x=1 ( x−1 ) ! ( n−x ) !
n

x=1
x−1
( x−1)
=np ∑ n−1 p (1− p)
n−x

=np ¿ (by putting x-1 = i)

= np ……………(1) (by binomial theorem)

n n

µ2’ =∑ x2 f (x) = ∑ x2 P( X =x)

x=0 x=0
 n n x n−x

x=0 x ()
= ∑ x n p (1− p) = ∑ x 2
2
x n−x

x=1
n! p (1− p)
x ! ( n−x ) !
n x n− x n
x n! p (1− p) ( x−1+1 ) n ! px (1−p)n− x
=∑ = ∑
x=1 ( x−1 ) !(n−x !) x=1 ( x−1 ) ! ( n−x ) !
n x n−x n x n−x
n ! p (1− p) n ! p (1− p)
=∑ + ∑ ( x−1 ) ! ( n−x ) !
x=2 ( x−2 ) ! ( n−x ) ! x=1

n
( n−2 ) ! p x−2 (1− p)n−x
= n(n−1) p 2 ∑ + np (from 1)
x=2 ( x−2 ) ! ( n−x ) !
n

= np+ n ( n−1 ) p
2
∑ ( n−2
x−2 ) p x−2 (1− p )
n−x

x=2

n−2

= np+ n(n−1) p
2
∑ ( n−2
y )
p y ( 1− p )
n−2− y
(by putting x-2 = y)
y=0

= np +n(n-1)p2 (by binomial theorem)

Example: Find the mean and the standard deviation of the distribution of each of the following
random variables (having binomial distributions):
a) The number of heads obtained in 676 flips of a balanced coin.
b) The number of 4’s obtained in 720 rolls of a balanced die.
c) The number of defectives in a sample of 600 parts made by a machine, when the
probability is 0.04 that any one of the parts is defective.
Solution:
1 2 1 1
(a) µ = np = 676. = 338 σ =np ( 1− p ) =676 . . =169 σ=13
2 2 2
1 2 1 5
(b) μ=np=720 . =120 , σ =720 . . =100 , σ=10
6 6 6
2
(c) μ=600 ( 0.04 )=24 , σ =600 ( 0 .04 )( 0.96 )=13.04 , σ =4.8

Mean and Variance of Hypergeometric Distribution

Theorem: For the Hypergeometric distribution with parameters n, a, N;

the mean is µ=n

a
N
2
and Variance is σ =n
a
N
1−
a
N ( )( NN −n
−1 )
Remark: When hypergeometric distribution is approximated by binomial distribution, that is if
n<<N, then p = a/N, so µ is approximated by np and σ 2 is approximated by np(1-p), since

(N−n
N−1 )is approximately 1.

Chebyshev’s Theorem
Chebyshev’s theorem gives us an upper bound (or lower bound) of certain probabilities.

Chebyshev’s theorem: Suppose X is a random variable with mean µ and variance σ 2. Then for
1
any positive number k, P (|X −μ|≥ kσ ) ≤ 2 .
k

1
Equivalently P (−kσ < X−μ< kσ ) ≥1− 2.
k

Thus the above theorem tells us that the value of a random variable are ‘close’ to its mean. The
probability of getting a value for X far from its mean µ is small.
Chebyshev’s theorem provides a lower bound for probability that a random variable lies within
given distance from mean. On the other hand it also provides an upper bound that random
variable is away from mean by distance more than a specified number. Chebyshev’s theorem
holds for any distribution, discrete or continuous. Similar proof for continuous distribution can
be given.

Example on Chebyshev’s Theorem:

Example 1: A coin is tossed 400 times. Using Chebyshev’s theorem find lower bound for the
probability that number if heads lies between 180 and 220.
Solution: Let X be the number of heads. Then X has Binomial distribution with parameters
1 1
n=400 , p= , μ=np=400 . =200.
2 2
2 1 1
σ =np ( 1− p ) =400 . . =100 . σ=10
2 2
1
P [|X −μ|<20 ] ≥ 1− 2 … … … . ( 1 )
K
Kσ=20 10 K=20∨K=2.
1 3
Hence from (1) P ( 180< x <220 ) ≥1− =
4 4
3
Hence required lower bound or limit is .
4

Exercise 2: How many times should we toss a balanced coin How many times should we toss a
balanced coin so that we can assert with probability at least 0.99 that the proportion of heads
occurs between 0.45 and 0.55?
Solution: Let n be number of trials. Now on each trial, p = probability of head = 0.5.
X = number of heads in n trials has binomial distribution with parameters n and p=0.5, thus
=n/2 and 2=n/4.
Proportion of heads =X/n is between 0.45 and 0.55 if and only if 0.45n < X < 0.55n.
Since X has binomial distribution with p=1/2, =0.5n, 2 = 0.25n.
Thus 0.45n < X < 0.55n ↔ -0.05n < X - < 0.05n ↔ |X- | < 0.05n.
Now let k =(k n1/2)/2 =0.05n. This gives k=0.1 n1/2.
Thus if k is such that
1 –1/k2 = 1- 100/n > 0.99 then P(0.45n < X < 0.55n) ≥ 1 –1/k2 > 0.99.
Now if n > 10000, then 1 – 1/k2 = 1-100/n > 0.99, hence the probability that proportion of heads
lies between 0.45 and 0.55 is at least 0.99.

Law of large numbers : As suggested by above example, by repeating experiment large number
of times, we can assert proportion of successes is away from p within any given distance with
probability close to 1, here p = probability of success in any trial.
This justifies the frequency interpretation of probability that probability of an event is the
relative frequency of occurrence of that event in large number of repetitions of that
experiment.

Learning Outcome:
Mean, Variance and Moments of discrete random variable.
Mean and variance of uniform distribution, Mean and variance of binomial distribution and
Mean and variance of hypergeometric distribution. Chebyshev’s Theorem and its examples.

Questions to consider:
Q1. What is mean, variance and moments of discrete random variables ?
Q2. What are some special types of discrete probability distribution?
Q3. What is mean and variance binomial distribution?
Q4. What is mean and variance hypergeometric distribution?
Q5. What is mean and variance discrete uniform distribution?
Q6. What is Chebyshev’s theorem?