§ 5 Distributions Associated with N (0, 1)
The Chi-squared distribution, the t distribution(often referred to as student
t distr) and the F distr are all ’associated’ with the normal distr. In fact,
as we shall see, the Chi-squared distr is associated with the normal distr
whereas the t distr and the F distr can be said to be associated with the
Chi-squared distr.
§ 5.1 The Chi-squared Distr.
Let X ∼ N (0, 1) then Z = X 2 has Chi-squared distr with 1 degree of
freedom, that is, Z ∼ χ(21) .
Similarly, let X 1 , X 2 , · · · , X n be independent and identically dis-
tributed(iid) random variables such that each X i ∼ N (0, 1), i = 1, 2, · · · , n
Pn
then Y = i =1 X i2 has Chi-squared distr with n degree of freedom, that is,
Y ∼ χ(2n) and the pdf of r.v. Y is given by
n −y
1
(
n y 2 −1 e 2 , y>0
f ( y) = 22 ( 2n −1)! (58)
0, elsewher e
Nota Bene
1. The (critical)values for χ(2n) (α), where α is the significance level, are
available in statistical tables.
2. The Chi-squared distr is NOT symmetric about the mean µ.
Degrees of freedom can be described as the number of scores that
are free to vary. Example is getting a score of 12 when throwing a die 3
times. The last throw has no ’freedom’, so that there are only 2 degrees of
freedom.
Figure 5: The typical χ(2n) (α) distribution
Examples curve.
1. Use the statistical tables to find
(a). χ(2n) (0.05), for the degrees of freedom n = 1, 2, · · · , 10 etc
(b). n for which P ( Z < χ(2n) (0.05)) = 0.9,
26 I N T R O D U C T I O N T O M AT H E M AT I C A L S TAT I S T I C S
(c). n for which P ( Z > χ(2n) (0.05)) = 0.01
2. For the sample size of n = 25 and the cumulative probability of
0.90 find the Critical Value using the χ(2n) (α) table. The meaning of this
question is that if one selects a random sample of 25 observations, then
there is a 90% chance that the chi-square statistic from that sample will
be less than or equal to 33.2.
§ 5.1.1 Distribution of the Chi-squared Statistic
A chi-square statistic is a statistic whose values are given by X 2 =
[( n − 1)s2 ]/σ2 where n is the sample size, s is the standard deviation
of the sample, and σ is the standard deviation of the population. The
distribution of the chi-square statistic has n − 1 degrees of freedom.
Example
A serobe packaging company says that the shelf life of their product is 5
years, with a standard deviation of 1 year. Assuming that the company’s
claims are true. If you test a random sample of 9 cans of Serobe, what is
the prob that the std deviation in your sample will be less than 0.95 years?
ANSWER: There is a 49% chance that the sample std dev won’t be more
than 0.95.
PERSONAL: The chi-sqauared statistic is similar to the cv.
§ 5.2 The Student t Distr.
Let X ∼ N (0, 1) and Z ∼ χ(2n) be independent r.v.s then the r.v.
X
T=p
Z /n
is t distributed with n degrees of freedom.
Nota Bene
1. For the degrees of freedom n, the values for t α , where α is the signifi-
cance level, are available in statistical tables.
2. The t distr is symmetric about the mean µ = 0.
Examples
For a certain sample a student t-distribution has 8 degrees of freedom,
find the probability and T statistic for which
(a). T > t 0.025 , that is P ( T > t 0.025 )ANS 0.025
(b). −t 0.10 < T < t 0.025
Figure 6: The typical student t distribu-
tion curve for some significance level
α.
§ 5 D I S T R I B U T I O N S A S S O C I AT E D W I T H n ( 0, 1 ) 27
§ 5.3 The F Distr.
Let r.v.s Y ∼ χ(2d ) and Z ∼ χ(2d ) be independent, then
1 2
Y / d1
F= (59)
Z / d2
has an F distr with d2 and d2 degrees of freedom.
Nota Bene
The values corresponding to P ( F > fα ) are usually availed in F -distribution
tables.
Examples
(a). Find F10,6 (0.05) ANSWER: P ( F10,6 > 4.06) = 0.05, so f0.05 = 4.06 Figure 7: The F distribution curve
showing P ( Fd1 ,d2 > fα ).
(b). Find F6,10 (0.025)
Nota Bene
Note that
1
Fd1 ,d2 (1 − α) = (60)
Fd2 ,d1 (α)
Example
Find F3,7 (.95) and F2,5 (0.90)
§ 6 Sampling Distributions
A random sample is a subset of a statistical population in which each
member of the subset has an equal probability of being chosen. A simple
random sample is meant to be an unbiased representation of a strata.
Let X i , i = 1, · · · , n be iid with pdf f ( x ), then x 1 , x 2 , · · · , x n is said to be
a random sample of size n from a population with pdf f ( x ). Note that
Qn
f X 1 ,X 2 ,··· ,X n ( x 1 , x 2 , · · · , x n ) = i =1 f ( x i ) because the X i , i = 1, · · · , n are
independent.
Any function f ( x 1 , x 2 , · · · , x n ) is a statistic. The functions of a random
sample x 1 , x 2 , · · · , x n that come to mind are the sample mean x̄ and the
sample variance s2 . The sample mean is defined
n
1X
X̄ = Xi (61)
n i =1
whereas the sample variance is defined by
n
1 X
S2 = ( X i − X̄ )2 , whil e s is cal l ed the sample st d dev. (62)
n − 1 i =1
Theorem 6
Pn 2
Pn 2
2
n i =1 X i − ( i =1 X i )
S = (63)
n( n − 1)
Proof 7
§ 6.2.1 The Sampling Distr. of X̄
Let a population have mean µ and variance σ then for a random sample
2
X 1 , X 2 , · · · , X n from the same population E [ X̄ ] = µ and Var ( X̄ ) = σn .
Proof 8
2
Therefore, X̄ ∼ N (µ, σn ).
30 I N T R O D U C T I O N T O M AT H E M AT I C A L S TAT I S T I C S
§ 6.2.2 The Central Limit Theorem (Mean)
If X̄ is the sample mean for a random sample of size n taken from a
population with mean µ and variance finite variance σ2 then as n → ∞
we have
x̄ − µ
Z= p ∼ N (0, 1). (64)
σ/ n
Example
The average life of a washing machine is (i)6 years ((ii)7 years) with
a std dev of 1 year. Supposing the lifes of these machines are normally
distributed, find the prob that the mean life of a random sample of size 9
of such machines falls between 6.4 and 7.2 years. ANSWER: (a)P (6.4 <
X̄ < 7.2) = P (−1.8 < Z < 0.6) = 0.6898.
§ 6.2.1 The Sampling Distributions of S 2
Let X ∼ N (µ, σ2 ) then
( n − 1)S 2 2
∼ χn−1 . (65)
σ2
Suppose that we have the assumption that X ∼ N (µ, σ2 ), but now with
NO knowledge of the std dev σ, then
x̄ − µ
T= p ∼ t n−1 . (66)
S/ n
Theorem 7 Let X ∼ N (µ1 , σ12 ) and Y ∼ N (µ2 , σ22 ) and X and Y are
independent then
S12 /σ12
F= ∼ F( n1 − 1, n2 − 1) wher e n1 & n2 ar e sample sizes f or X & Y r espec t ivel y.
S22 /σ22
(67)
Examples
(a). Find P (2.3 < S 2 < 22.2), where S 2 is the sample variance for sample
of size n = 6 from a normal distribution with variance σ2 = 12.
(b). P ( T < t 0.025 )