stats - School of Computing

Download Report

Transcript stats - School of Computing

Summary Statistics
When analysing practical sets of data, it is useful to be able to define a small number of
values that summarise the main features present. We will derive (i) representative values, (ii)
measures of spread and (iii) measures of skewness and other characteristics.
Representative Values
These are sometimes called measures of location or measures of central tendency.
1. Random Value
Given a set of data S = { x1, x2, … , xn }, we select a random number, say k, in the range 1 to
n and return the value xk. This method of generating a representative value is
straightforward, but it suffers from the fact that extreme values can occur and successive
values could vary considerably from one another.
2. Arithmetic Mean
This is also known as the average. For the set S above the average is
x = {x1 + x2 + … + xn }/ n.
If x1 occurs f1 times, x2 occurs f2 times and so on, we get the formula
x = { f 1 x1 + f 2 x2 + … + f n xn } / { f 1 + f 2 + … + f n } ,
written
x =
 fx /  f
, where  (sigma) denotes a sum.
Example 1.
The data refers to the marks that students in a class obtained in an examination. Find the
average mark for the class. The first
point to note is that the marks are presented as Mark
Mid-Point Number
ranges, so we must be careful in our
of Range
of Students
interpretation of the ranges. All the intervals
xi
fi
f i xi
must be of equal rank and their must be no
gaps in the classification. In our case, we
0 - 19
10
2
20
interpret the range 0 - 19 to contain marks
21 - 39
30
6
180
greater than 0 and less than or equal to 20.
40 - 59
50
12
600
Thus, its mid-point is 10. The other intervals 60 - 79
70
25
1750
are interpreted accordingly.
80 - 99
90
5
450
Sum
50
3000
The arithmetic mean is x = 3000 / 50 = 60 marks.
Note that if weights of size fi are suspended
from a metre stick at the points xi, then the
average is the centre of gravity of the
distribution. Consequently, it is very sensitive
to outlying values.
x1
x2
f1
x
xn
fn
f2
Equally the population should be homogenous for the average to be meaningful. For
example, if we assume that the typical height of girls in a class is less than that of boys,
then the average height of all students is neither representative of the girls or the boys.
3. The Mode
Frequency
50
This is the value in the distribution that occurs
most frequently. By common agreement,
it is calculated from the histogram using linear
interpolation on the modal class.
13 20
25
13
The various similar triangles in the diagram
generate the common ratios. In our case,
the mode is
60 + 13 / 33 (20) = 67.8 marks.
20
12
6
2
20
5
40
60
80
40
60
80
100
4. The Median
Cumulative
50
This is the middle point of the distribution. It
is used heavily in educational applications. If
{ x1, x2, … , xn } are the marks of students in a
class, arranged in nondecreasing order, then 25.5
the median is the mark of the (n + 1)/2 student.
It is often calculated from the ogive or
cumulative frequency diagram. In our case,
the median is
60 + 5.5 / 25 (20) = 64.4 marks.
Frequency
20
100
Measures of Dispersion or Scattering
Example 2. The following distribution has the same
arithmetic mean as example 1, but the values are more
dispersed. This illustrates the point that an average
value on its own may not be adequately discribe
statistical distributions.
To devise a formula that traps the degree to which a
distribution is concentrated about the average, we
consider the deviations of the values from the average.
If the distribution is concentrated around the mean,
then the deviations will be small, while if the distribution
is very scattered, then the deviations will be large.
The average of the squares of the deviations is called
the variance and this is used as a measure of dispersion.
The square root of the variance is called the standard
deviation and has the same units of measurement as
the original values and is the preferred measure of
dispersion in many applications.
Marks
x
Frequency
f
10
30
50
70
90
Sums
6
8
6
15
15
50
fx
60
240
300
1050
1350
3000
x6
x5
x4
x3
x2
x1
x
Variance & Standard Deviation
s2 = VAR[X] = Average of the Squared Deviations
= S f { Squared Deviations } / S f
= S f { xi - x } 2 / S f
= S f xi 2 / S f - x 2
, called the product moment formula.
s = Standard Deviation =  Variance
Example 1
f
x
2
10
6
30
12
50
25
70
5
90
50
fx
20
180
600
1750
450
3000
f x2
200
5400
30000
122500
40500
198600
VAR [X] = 198600 / 50 - (60) 2
= 372 marks2
Example 2
f
x
6
10
8
30
6
50
15
70
15
90
50
fx
60
240
300
1050
1350
3000
f x2
600
7200
15000
73500
121500
217800
VAR [X] = 217800 / 50 - (60)2
= 756 marks2
Other Summary Statistics
Skewness
An important attribute of a statistical distribution relates to its degree of symmetry. The
word “skew” means a tail, so that distributions that have a large tail of outlying values on the
right-hand-side are called positively skewed or skewed to the right. The notion of negative
skewness is defined similarly. A simple formula for skewness is
Skewness = ( Mean - Mode ) / Standard Deviation
which in the case of example 1 is:
Skewness = (60 - 67.8) / 19.287 = - 0.4044.
Coefficient of Variation
This formula was devised to standardise the arithmetic mean so that comparisons can be
drawn between different distributions.. However, it has not won universal acceptance.
Coefficient of Variation = Mean / standard Deviation.
Semi-Interquartile Range
Just as the median corresponds to the 0.50 point in a distribution, the quartiles Q 1, Q2, Q3
correspond to the 0.25, 0.50 and 0.75 points. An alternative measure of dispersion is
Semi-Interquartile Range = ( Q3 - Q1 ) / 2.
Geometric Mean
For data that is growing geometrically, such as economic data with a high inflation effect, an
alternative to the the arithmetic mean is preferred. It involves getting the root to the power
N = S f of a product of terms
Geometric Mean = N x1f1 x2 f2 … xk fk
Regression
[Example 3.] As a motivating example, suppose we are modelling sales data over time.
SALES
3
5
4
5
6
7
TIME 1990
1991
1992
1993
1994
1995
Y=mX+c
We seek the straight line “Y = m X + c” that best
Y
Yi
approximates the data. By “best” in this case, we
mean the line which minimizes the sum of squares
of vertical deviations of points fromthe line:
m Xi + c
SS = S ( Yi - [ mXi + c ] ) 2.
Setting the partial derivatives of SS with respect to m
and c to zero leads to the “normal equations”
0
X
S Y = m S X + n .c
, where n = # points
Xi
S X .Y= m S X2 + c S X .
Let 1990 correspond to Year 0.
X.X
0
1
4
9
16
25
X
0
1
2
3
4
5
X.Y
0
5
8
15
24
35
Y
3
5
4
5
6
7
Y.Y
9
25
16
25
36
49
55 15
87
30
160
Sales
10
5
Time
0
5
Example 3 - Workings.
The normal equations are:
30 = 15 m + 6 c =>
87 = 55 m + 15 c
=>
24 = 35 m
=>
150 = 75 m + 30 c
174 = 110 m + 30 c
30 = 15 (24 / 35) + 6 c
=>
c = 23/7
Thus the regression line of Y an X is
Y = (24/35) X + (23/7)
and to plot the line we need two points, so
X = 0 => Y = 23/7
and X = 5 => Y = (24/35) 5 + 23/7 = 47/7.
It is easy to see that ( X, Y ) satisfies the normal equations, so that the regression line of Y on
X passes through the “Center of Gravity” of the data. By expanding terms, we also get
S ( Yi - Y ) 2 = S ( Yi - [ m Xi + c ] ) 2
Total Sum
of Squares
SST
=
ErrorSum
of Squares
SSE
+
+
S ( [ m Xi + c ] - Y ) 2
Y
Regression Sum
of Squares
SSR
In regression, we refer to the X variable as the independent
variable and Y as the dependent variable.
Yi
mXi +C
Y
Y
X
X
Correlation
The coefficient of determination r2 ( which takes values in the range 0 to 1) is a measure of
the proportion of the total variation that is associated with the regression process:
r2
=
SSR/ SST
=
1 - SSE / SST.
The coefficient of correlation r ( which takes values in the range -1 to +1 ) is more
commonly used as a measure of the degree to which a mathematical relationship exists
between X and Y. It can be calculated from the formula:
r =
S(X-X)(Y-Y)
S
=
( X - X )2 ( Y - Y ) 2
nSXY-S XSY
 { n S X 2 - ( S X )2 } { n S Y2 - ( S Y)2 }
Example. In our case r = {6(87) - (15)(30)}/  { 6(55) - (15)2 } { 6(160) - (30)2 } = 0.907.
r=-1
r=0
r=+1
Colinearity
If the value of the correlation coefficient is greate than 0.9 or less than - 0.9, we would take
this to mean that there is a mathematical relationship between the variables. This does not
imply that a cause-and-effect relationship exists.
Consider a country with a slowly changing population size, where a certain political party
retains a relatively stable per centage of the poll in elections. Let
X = Number of people that vote for the party in an election
Y = Number of people that die due to a given disease in a year
Z = Population size.
Then, the correlation coefficient between X and Y is likely to be close to 1, indicating that
there is a mathematical relationship between them (i.e.) X is a function of Z and Y is a
function of Z also. It would clearly be silly to suggest that the indicence of the disease is
caused by the number of people that vote for the given political party. This is known as the
problem of colinearity.
Spotting hidden dependencies between distributions can be difficult. Statistical
experimentation can only be used to disprove hypotheses, or to lend evidence to support the
view that reputed relationships between variables may be valid. Thus, the fact that we
observe a high correlation coefficient between deaths due to heart failure in a given year
with the number of cigarettes consumed twenty years earlier does not establish a cause-andeffect relationship. However, this result may be of value in directing biological research in a
particular direction.
Overview of Probability Theory
In statistical theory, an experiment is any operation that can be replicated infinitely often
and gives rise to a set of elementary outcomes, which are deemed to be equally likely. The
sample space S of the experiment is the set of all possible outcomes of the experiment. Any
subset E of the sample space is called an event. We say that an event E occurs whenever any
of its elements is an outcome of the experiment. The probability of occurrence of E is
P {E} = Number of elementary outcomes in E
Number of elementary outcomes in S
S
E
The complement E of an event E is the set of all elements that belong to S but not to E. The
union of two events E1 E2 is the set of all outcomes that belong to E1 or to E2 or to both.
The intersection of two events E1  E2 is the set of all events that belong to both E1 and E2.
Two events are mutually exclusive if the occurrence of either precludes the occurrence of
the other (i.e) their intersection is the empty set  . Two events are independent if the
occurrence of either is uneffected by the occurrence or nonoccurence of the other event.
Theorem of Total Probability.
P {E1 E2} = P{E1} + P{E2} - P{E1
S
E2}
P{E1 E2} = (n1, 0 + n1, 2 + n0, 2) / n
= (n1, 0 + n1, 2) / n + (n1, 2 + n0, 2) / n - n1, 2 / n
= P{E1} + P{E2} - P{E1  E2}
Corollary.
If E1 and E2 are mutually exclusive, P{E1  E2} = P{E1} + P{E2}
Proof.
n = n0, 0 + n1, 0 + n0, 2 + n1, 2
E1
n1, 0
E2
n1, 2
n0, 2
n0, 0
The probability P{E1 | E2} that E1 occurs, given that E2 has occurred (or must occur) is called
the conditional probability of E1. Note that in this case, the only possible outcomes of the
experiment are confined to E2 and not to S.
S
Theorem of Compound Probability
P{E1  E2} = P{E1 | E2} * P{E2}.
Proof.
P{E1  E2} = n1, 2 / n
= {n1, 2 / (n1, 2 + n0, 2) } * { n1, 2 + n0, 2) / n}
Corollary.
If E1 and E2 are independent, P{E1

E2
E1
n1, 0
n1, 2
n0, 2
n0, 0
E2} = P{E1} * P{E2}.
The ability to count the possibily outcomes in an event is crucial to calculating probabilities.
By a permutation of size r of n different items, we mean an arrangement of r of the items,
where the order of the arrangement is important. If the order is not important, the arrangement
is called a combination.
Example. There are 5*4 permutations and 5*4 / (2*1) combinations of size 2 of A, B, C, D, E
Permutations:
AB, BA, AC, CA, AD, DA, AE, EA
BC, CB, BD, DB, BE, EB
CD, DC, CE, EC
DE, ED
Combinations:
AB, AC, AD, AE, BC, BD, BE, CD, CE, DE
Standard reference books on probability theory give a comprehensive treatment of how these
ideas are used to calculate the probability of occurrence of the outcomes of games of chance.
Statistical Distributions
If a statistical experiment only gives rise to real numbers, the outcome of the experiment is
called a random variable. If a random variable X
takes values
X1, X2, … , Xn
with probabilities p1, p2, … , pn
then the expected or average value of X is defined to be
n
E[X] =  pj Xj
j =1
and its variance is
n
VAR[X] = E[X2] - E[X]2 =
 pj Xj2 - E[X]2.
j =1
Example. Let X be a random variable measuring
the distance in Kilometres travelled by children
to a school and suppose that the following data
applies. Then the mean and variance are
E[X]
= 5.30 Kilometres
VAR[X] = 33.80 - 5.302
= 5.71 Kilometres2
Prob. Distance
pj
Xj
pj Xj
0.15
0.40
0.20
0.15
0.10
1.00
2.0
4.0
6.0
8.0
10.0
-
0.30
1.60
1.20
1.20
1.00
5.30
pj Xj2
0.60
6.40
7.20
9.60
1.00
33.80
Similar concepts apply to continuous distributions. The distribution function is defined by
F(t) = P{ X  t} and its derivative is the frequency function
f(t) = d F(t) / dt
t
so that
F(t) =


f(x) dx.
Sums and Differences of Random Variables
Define the covariance of two random variables to be
COVAR [ X, Y] = E [(X - E[X]) (Y - E[Y]) ] = E[X Y] - E[X] E[Y].
If X and Y are independent, COVAR [X, Y] = 0.
Lemma
E[ X + Y]
= E[X] + E[Y]
VAR [ X + Y]
= VAR [X] + VAR [Y] + 2 COVAR [X, Y]
E[ k. X] = k .E[X] VAR[ k. X] = k2 .E[X] for a constant k.
Example. A company records the journey time X
of a lorry from a depot to customers and
the unloading times Y, as shown.
E[X]
= {1(10)+2(13)+3(17)+4(10)}/50 = 2.54
E[X2] = {12(10+22(13)+32(17)+42(10)}/50 = 7.5
VAR[X] = 7.5 - (2.54)2 = 1.0484
E[Y]
= {1(20)+2(19)+3(11)}/50 = 1.82
VAR[Y] = 3.9 - (1.82)2 = 0.5876
Y =1
2
3
Totals
X=1
7
2
1
10
2
5
6
2
13
3
4
8
5
17
4
4
3
3
10
Totals
20
19
11
50
E[Y2] = {12(20)+22(19)+32(11)}/50 = 3.9
E[X+Y]
= { 2(7)+3(5)+4(4)+5(4)+3(2)+4(6)+5(8)+6(3)+4(1)+5(2)+6(5)+7(3)}/50 = 4.36
2
E[(X + Y) ]
= {22(7)+32(5)+42(4)+52(4)+32(2)+42(6)+52(8)+62(3)+42(1)+52(2)+62(5)+72(3)}/50 = 21.04
VAR[(X+Y)] = 21.04 - (4.36)2 = 2.0304
E[X Y]
= {1(7)+2(5)+3(4)+4(4)+2(2)+4(6)+6(8)+8(3)+3(1)+6(2)+9(5)+12(3)}/50 = 4.82
COVAR (X, Y) = 4.82 - (2.54)(1.82) = 0.1972
VAR[X] + VAR[Y] + 2 COVAR[ X, Y] = 1.0484 + 0.5876 + 2 ( 0.1972) = 2.0304
Standard Statistical Distributions
Most elementary statistical books provide a survey of commonly used statistical
distributions. The reason we study these distributions are that
They provide a comprehensive range of distributions for modeling practical applications
Their mathematical properties are known
They are described in terms of a few parameters, which have natural interpretations.
1 Prob
1. Bernoulli Distribution.
This is used to model a trial which gives rise to two outcomes:
success/ failure, male/ female, 0 / 1. Let p be the probability that
the outcome is one and q = 1 - p that the outcome is zero.
1-p
E[X]
= p (1) + (1 - p) (0) = p
VAR[X] = p (1)2 + (1 - p) (0)2 - E[X]2 = p (1 - p).
0
2. Binomial Distribution.
Suppose that we are interested in the number of successes X
in n independent repetions of a Bernoulli trial, where the
probability of success in an individual trial is p. Then
Prob{X = k} = nCk pk (1-p)n - k, (k = 0, 1, …, n)
E[X]
=np
VAR[X] = n p (1 - p).
This is the appropriate distribution to use in modeling the
number of boys in a family of n = 4 children, the number of
defective components in a batch n = 10 components and so on.
Prob
p
p
1
(n=4, p=0.2)
1
np
4
3. Poisson Distribution.
The Poisson distribution arises as a limiting case of the binomial distribution, where
n  , p   in such a way that n p    a constant). Its density is
Prob{X = k} = exp ( -  ) k / k ! k = , 1, 2, … ).
Note that exp (x) stands for e to the power of x, where e is
Prob
approximately 2.71828.
1
E [X]
=
VAR [X] = .
The Poisson distribution is used to model the number of
occurrences of a certain phenomenon in a fixed period of
time or space, as in the number of
5
O particles emitted by a radioactive source in a fixed direction and period of time
O telephone calls received at a switchboard during a given period
O defects in a fixed length of cloth or paper
O people arriving in a queue in a fixed interval of time
O accidents that occur on a fixed stretch of road in a specified time interval.
X
4. Geometric Distribution.
This arises in the “time” or number of steps k to the first
success in a series of independent Bernoulli trials. The
density is
Prob{X = k} = p (1 - p) k-1 (k = 1, 2, … ).
E[X] = 1/p
VAR [X] = (1 - p) /p2
Prob
1
X
5. Negative Binomial Distribution
This is used to model the number of failures k that occur before the rth success in a series of
independent Bernoulli trials. The density is
Prob {X = k} = r+k-1Ck pr (1 - p)k
(k = 0, 1, 2, … )
Note
E [X]
= r (1 - p) / p
VAR[X]
= r (1 - p) / p2.
6. Hypergeometric Distribution
Consider a population of M items, of which W are deemed to be successes. Let X be the
number of successes that occur in a sample of size n, drawn without replacement from the
population. The density is
Prob { X = k} = WCk M-WCn-k / MCn ( k = 0, 1, 2, … )
Then
E [X] = n W / M
VAR [X] = n W (M - W) (M - n) / { M2 (M - 1)}
7. Uniform Distribution
Prob
1
A random variable X has a uniform distribution on
the interval [a, b], if X has density
f (X) = 1 / ( b - a)
for a < X < b
1 / (b-a)
X
=0
otherwise.
Then
E [X] = (a + b) / 2
a
b
VAR [X] = (b - a)2 / 12
Uniformly distributed random numbers occur frequently in simulation models. However,
computer based algorithms, such as linear congruential functions, can only approximate this
distribution so great care is needed in interpreting the output of simulation models.
If X is a continuous random variable, then the probability that X takes a value in the range
[a, b] is the area under the frequency function f(x) between these points:
Prob { a < x < b } = F (b) - F (a) = ab f(x) dx.
In practical work, these integrals are evaluated by looking up entries in statistical tables.
9. Gaussian or Normal Distribution
A random variable X has a normal distribution with mean m and standard deviation s if it
has density
f (x)
=
1
Prob
1
exp { - ( x - m )2 }, -  x < 
2 p s2
2 s2
=
0,
otherwise
f(x)
E [ X]
=m
VAR [X] = s2.
0
m
As described below, the normal distribution arises naturally as the limiting distribution of the
average of a set of independent, identically distributed random variables with finite
variances. It plays a central role in sampling theory and is a good approximation to a large
class of empirical distributions. For this reason, a default assumption in many empirical
studies is that the distribution of each observation is approximately normal. Therefore,
statistical tables of the normal distribution are of great importance in analysing practical data
sets. X is said to be a standardised normal variable if m = 0 and s = 1.
X
10. Gamma Distribution
The Gamma distribution arises in queueing theory as the time to the arrival of the n th
customer in a single server queue, where the average arrival rate is . The frequency
function is
f(x)
=  ( x )n - 1 exp ( -  x) / ( n - 1)! , x  0,  > 0, n = 1, 2, ...
= 0,
otherwise
E [X]
=n/
VAR [X] = n /  2
11. Exponential Distribution
This is a special case of the Gamma distribution with n = 1 and so is used to model the
interarrival time of customers, or the time to the arrival of the first customer, in a simple
queue. The frequency function is
f (x)
=  exp ( -  x ),
x  0,  > 0
= 0,
otherwise.
12. Chi-Square Distribution
A random variable X has a Chi-square distribution with n degrees of freedom ( where n is a
positive integer) if it is a Gamma distribution with  = 1, so its frequency function is
f (x)
= xn - 1 exp ( - x) / ( n - 1) !, x  o
Prob
= 0,
otherwise.
c2 n (x)
X
Chi-square Distribution (continued)
The chi-square distribution arises in two important applications:
O If X1, X2, … , Xn is a sequence of independently distributed standardised normal
random variables, then the sum of squares X12 + X22 + … + Xn2 has a chi-square
distribution with n degrees of freedom
O If x1, x2, … , xn is a random sample from a normal distribution with mean m
and variance s2 and let
x =  xi / n
and S2 =  ( xi - x ) 2 / s2,
then S2 has a chi-square distribution with n - 1 degrees of freedom, and the
random variables S2 and x are independent.
13. Beta Distribution.
A random variable X has a Beta distribution with parameters a > 0 and b > 0 if it has
frequency function
f (x)
= G  a + b ) x a  1 ( 1 - x) b  1 / G (a) G b), 0 < x < 1
= 0,
otherwise
E [X]
=a/a+b)
VAR [X] = a b/ [  a + b)2  a + b + 1) ]
If n is an integer,
G (n) = ( n - 1 ) !
G (n + 1/2) = (n - 1/2) ( n - 3/2) …
with G (1) = 1
with G ( 1/2) =  p
14. Student’s t Distribution
A random varuable X has a t distribution with n degrees of freedom ( tn ) if it has density
f(x)
= G (n+1) / 2 )
 n p G  n / 2)
1 + x2 / n ) - (n+1) / 2
= 0,
The t distribution is symmetrical about the origin, with
E[X]
=0
VAR [X] = n / (n -2).
( -  < x < )
otherwise.
For small values of n, the tn distribution is very flat. As n is increased the density assumes a
bell shape. For values of n  25, the tn distribution is practically indistinguishable from the
standard normal curve.
O If X and Y are independent random variables
If X has a standard normal distribution and Y has a cn2 distribution
then
X
has a tn distribution
Y/n
O If x1, x2, … , xn is a random sample from a normal distribution, with mean m
and variance s2 and if we define s2 = 1 / ( n - 1)  ( xi - x ) 2
then ( x - m ) / ( s / n) has a tn- 1 distribution
15. F Distribution
A random variable X has an F distribution with m and n degrees of freedom if it has density
f(x)
= G (m + n) / 2 ) m m / 2 n n / 2 x m / 2 - 1
G  m / 2) G  n / 2) (n + m
Note
If
x>0
x) ( m + n ) / 2
= 0,
E[X]
= n / ( n - 2)
VAR [X] = 2 n2 (m + n - 2)
m (n - 4) ( n - 2 )2
otherwise.
if n > 4
if n > 4
O X andYare independent random variables, X has a cm2 and Y a cn2 distribution
X / m has an Fm , n distribution
Y/n
O One consequence of this is that the F distribution represents the distribution of
the ratio of certain independent quadratic forms which can be constructed from
random samples drawn from normal distributions:
if x1, x2, … , xm ( m  2) is a random sample from a normal
distribution with mean m1 and variance s12, and
if y1, y2, … , yn ( n  2) is a random sample from a normal
distribution with mean m2 and variance s22, then
 ( xi - x )2 / ( m - 1)
 ( yi - y )2 / ( n - 1)
has an Fm - 1 , n - 1 distribution
Sampling Theory
The procedure for drawing a random sample a distribution is that numbers 1, 2, … are
assigned to the elements of the distribution and tables of random numbers are then used to
decide which elements are included in the sample. If the same element can not be selected
more than once, we say that the sample is drawn without replacement; otherwise, the
sample is said to be drawn with replacement.
The usual convention in sampling is that lower case letters are used to designate the sample
characteristics, with capital letters being used for the parent population. Thus if the sample
size is n, its elements are designated, x1, x2, …, xn, its mean is x and its modified variance is
s2 =
 (xi - x )2 / (n - 1).
The corresponding parent population characteristics are N (or infinity), X and S2.
Suppose that we repeatedly draw random samples of size n (with replacement) from a
distribution with mean m and variance s2. Let x1, x2, … be the collection of sample
averages and let
xi’ =
xi - m
(i = 1, 2, … )
s/n
The collection x1’, x2’, … is called the sampling distribution of means.
Central Limit Theorem.
In the limit, as n tends to infinity, the sampling distribution of means
has a standard normal distribution.
Attribute and Proportionate Sampling
If the sample elements area measurement of some characteristic, we are said to have
attribute sampling. On the otherhand if all the sample elements are 1 or 0 (success/failure,
agree/ no-not-agree), we have proportionate sampling. For proportionate sampling, the
sample average x and the sample proportion p are synonimous, just as are the mean m and
proportion P for the parent population. From our results on the binomial distribution, the
sample variance is p (1 - p) and the variance of the parent distribiution is P (1 - P).
We can generalise the concept of the sampling distribution of means to get the sampling
distribution of any statistic. We say that a sample characteristic is an unbiased estimator of
the parent population characteristic, is the mean of the corresponding sampling distribution
is equal to the parent characteristic.
Lemma.
The sample average (proportion ) is an unbiased estimator of the parent
average (proportion):
E [ x] = m;
E [p] = P.
The quantity  ( N - n) / ( N - 1) is called the finite population correction (fpc). If the
parent population is infinite or w have sampling with replacement the fpc = 1.
Lemma.
E [s] = S * fpc.
Confidence Intervals
From the statistical tables for a standard normal
distribution, we note that
Area Under
Density Function
0.90
0.95
0.99
From
To
-1.64
-1.96
-2.58
1.64
1.96
2.58
n (0,1)
0.95
-1.96
0
+1.96
2
From the central limit theorem, if x and s are the mean and variance of a random sample of
size n (with n greater than 25) drawn from a large parent population, then we can make the
following statement about the unknown parent mean m
Prob { -1.64  x - m  1.)  .
s/n
i.e.
Prob { x - 1.64 s /  n  m  x + 1. s /  n }  .
The range x + 1.64 s /  n
is called a 90% confidence interval for the parent mean m.
Example [ Attribute Sampling]
A random sample of size 25 has x = 15 and s = 2. Then a 95% confidence interval for m is
15 + 1.96 (2 / 5) (i.e.) 14.22 to 15.78
Example [ Proportionate Sampling]
A random sample of size n = 1000 has p = 0.40  1.96  p (1 - p) / (n - 1) = 0.03.
A 95% confidence interval for P is 0.40 + 0.03 (i.e.) 0.37 to 0.43.
Small Sampling Theory
For reference purposes, it is useful to regard the expression
x + 1.96 s /  n
as the “default formula” for a confidence interval and to modify it to suit particular
circumstances.
O If we are dealing with proportionate sampling, the sample proportion is the
sample mean and the standard error (s.e.) term s /  n simplifies as folows:
x -> p and
s /  n ->  p(1 - p) / (n -1).
O A 90% confidence interval will bring about the swap
1.96 -> 1.64.
O If the sample size n is less than 25, the normal distribution must be replaced by
Student’s t n - 1 distribution.
O For sampling without replacement from a finite population, a fpc term must be
used.
The width of the confidence interval band increases with the confidence level.
Example. A random sample of size n = 10, drawn from a large parent population, has a mean
x = 12 and a standard deviation s = 2. Then a 99% confidence interval for the parent mean is
x + 3.25 s /  n (i.e.)
12 + 3.25 (2)/3
(i.e.)
9.83 to 14.17
and a 95% confidence interval for the parent mean is
x + 2.262 s /  n (i.e.)
12 + 2.262 (2)/3 (i.e.)
10.492 to 13.508.
Note that for n = 1000, 1.96  p (1 - p) / n  . for values of p between 0.3 and 0.7. This
gives rise to the statement that public opinion polls have an “inherent error of 3%”. This
simplifies calculations in the case of puplic opinion polls for large political parties.