<br>

STOCHASTIC PROCESS AND

SAMPLING
TECHNIQUES
SYLLABUS
Stochastic Process: Introduction
of Stochastic proccss, Classification of random proccss,
random process, Stochastic matrix. Stationary and non-stationary
Markov Chain: Classification statcs,
of Classification of chains, Random walk
Sampling: Population (Universe), Sampling types and Gambler ruin.
and distribution, Sampling of mcan
Null and Alternative Hypothesis, One-tail and variance. Testing hypothesis,
and two-tails tcsts (Only Introduction), t- test
Chi-square test. and F-tcst (Only Introduction),

INTRODUCTION Consider the experiment of rolling a die.

Here S = {1, 2, 3, 4, 5, 6}.
In probability theory and relatcd
ficlds, a stochastic For each outcome of the experiment, lct us arbitrarily
process or random process
is a mathematical obijcct assign a function of time t(0 <t <o) as follows:
usually defined as a family
of random variables. Outcome
Stochastic processes are widely used as f(t)
mathematical 1

models that appcar to vary in a random manner. X(t) =-4

For
Cxample, growth ofa bacterial population, an
electrical
2
X2(t)= -2
current fluctuating due to 3
thermal noise, etc. and also X3(t) =2
in biology, chemistry, ccology,
ncuroscience, jmage
processing. signal processing, computer
4
X4(t) = 4
science and
cryptography ctc. xs(t) =
2
STOCHASTIC PROCESS OR RANDOM:PROCESS 6
=t
xg()
A stochastic process or a
random process is. a
collcction (or ensemble) of.random The set of functions (x1(t),
{X(x, t)} that are functions a
variables (RVs) x(),.,.x6(t)}
of real variable, namely represents a random process.
time t where Notation:
seS (S the sample space) and
is
teT(T is the parameter set or index set). As the dependence
of a random process on S is
obvious, we omit S in
The set of possible values of any individual the notation of a random
member of process. If the parameter
the random process is
called state space. Any set T is discrete, the random
individual member itself is process is denoted
called a sample function or by {X(n)} or
a rcalization ,}. If T is
of the process. continuous, the random process
is denoted by {X(t)}.
Notes: Clussificationof Random Processes:
i) If's and t are fixed, (X(s, 1)} a
is number. Depending on the discrete or
continuous naturc of the
) Ift is fixcd, {X(s, stuto spaco S and parameter
)} is a RV. set T, a random process
) ifs is fixed, (X(s, t)} is a single can be classitled into tour types:
(iv) time function.
1fs and t are variables, () r both T an S are
(X(s, 1)} is a collection of RVs disCrete, the random process
is
that arc timc functions. called a diserete vnd sequence.
For example, if
A
RV is a rule X
(or function) that assigns a timu teresenis the Quteone of the nth toss
function of a fair
to every. outcome of a random experiment.
<br>

dic, then (Xn, is a discrete random sequencc (2) The auto-corrclation of (X()} is denotcd by
n}
R(ty. tz)or R(t, t) and is defincd as the cxpccted
......, and S = {1, 2, 3, 4, X(t;) and
T={1,2, 3, 5,6}. value of the product of two members
(2) If T is discrete and S is continuous, the random
X(t) of
the random proccss.
process is callcd a continuous random sequencc. For
i.c. R(t;, tz)= E(X(t). X(12)}
Cxamplc, if X, represents the temperature at the n"
(3) The auto-covariancc of (X()} is denotcd by
hour of a day, then {Xn,lsns 24} is a continuous as
C(l1, t2) or C(tj, t2) and is defincd
random sequence, since tempcrature can take any
C(t1, 12) = R(t1, t2)-(t). u(,)
value in an interval and hencc is continuous.
(4) The corrclation cocfficicnt of {X()} is denotcd by
(3) If T is continuous and S is discrete, the random
proccss is called a discrcte random process. For Px(tj, t2) or f(t;, t2) and is defincd as

cxamplc, if X() represents the number of tclephonc C(tj, t2)

p(t1, t2)= x
calls rcccivcd in the interval (0, t), then (X(t)} is a
Jct,, t) C(t2, l2)
discrctc random proccss as
S= {0, 1, 2, 3, ..... where, C(t1, t;)
variance of X(1}).
is the

(4) If both T and S arc continuoüs, the random process is (5) Cross-correlation of two random processes {X(0)} and
callcd a continuous random proccss. For cxample, if (Y()} isdefined by
Rgy
X(1) represents the maximum temperature in the (t1; tz)=E(X(t|). Y(4)}
interval (0, t), {X()} isa continuous random process. (6) Cross-Covariance of two random proccsses {X(t)} and
Notc: {Y()} is defincd by
Hereafter, the words discrete' or continuous' will be
uscd to refcr the nature of S and the words 'sequence
Cxy
(t1.42)=Ry(i,t2)-H(t). Hy2)
(7) Cross-corrclation cocfficicnt of tWo random proccsses
and "process' will be.used to refer the nature of T.
(X(1)} and {Y(1)} is defincd by.
Stationary Random Proccss:
A random proccss {X()} is said to be stationary or Cxy (t;, t2).
Strict Sensc Stationary (SSS) if the Probability
Pxy (t1s tg)=
JC (1 t})x Cy (t2, 12)
Distribution Function (PDF)` of any set of. samples
Wide-Sense Stationary Proccss (WSS Proccss):
docs not vary with time.
A random process (X(t)} with finite first and sccond
Let (X(t)} be a random process,
order moments is called a weakly stationary process or
cOvariance stationary process or widc scnse
stationary
rcprcsent the Cumulative Distribution Function (CDF) proces mean
if its is a constant and the auto
at times ty +, ty +T,......t, +,then (X(t)} is said correlation depends only on the time difference.
to be' strictly stationary, Strongly stationary or i.c. E(X()} = u Constant and
strict
Scnse stationary if E(X (t).
Fx (%1,
+r, X,+)=E,(*1, x,),
.., for all
X(t-)}=R(t)
A random process which is not stationary in any sense
T, t1, e
t2,.... t, R and for all neN. is called an evolutionary process.
A non-stationary random process is
characterized by a
joint PDF or CDF that depends on IMPORTANT FORMULAE:
time instants
t1t2,.,tn: (1) Mean u(t) ofa random process {X()} is
Average Valucs of Random Proccss: = E{X(t)}
(t)
Let (X(t)} be a random process,
then (2) Auto-correlation
(1) The mean of {X(t)} is the cxpected value R(tj, t2) of arandom proccss
of a typical
member X(0) of
the process i.e. H(t) =E{X()} . {X()} is
R(t1, t2) = E{X(t) .X(t2)}
<br>

(3) Auto-covariance C(tj, t2) of a random proccss X(t) = n Pn

(X()} is 0 at

C(l1, l2) = R (t1, t2) 1+at

-u(t). (t2) 1

(4) Correlation cocfficicnt p(t1, t2) of {X(t)} is (1+ at)2

C(t1, at
p(1, l2) = t2) (1+at)
Jc(t,, tj) × C(l2, l2)
3 (at)?
(5) CrosS-correlation
(1+ at) 4
Ryy
(lj, t2) of {X()} and (Y(t)} is.

Rgy(ti, tz)= B(X(). Y(tz)}

.:: E(X()}- npn
(6) Cross-covariance n=0
2at 3(at)2
Cxy(l, tz) of (X(1)}and (Y()} is
(1+ at) (1+at) (1+at)"
Cy(li t2)= Rxy(!1: t2)-H() y(t2) at 3
(aty2
1+2- at
+
(1+at)2
(11) Cross-correlation cocfficient (1+ at)| 1+
Pxy (t1, iz) of {X()} and {Y()} is 1

[1+2a+ 3a+....
Cxy(1 t2) (1+ at)
Pxy (t t2)= at
/Cxx (t, t) x Cyy (t2 t2) where, a=
1+at
(8) * = lim [R, ()] (-c)-2
(l+ aty
(9). Var(X()) = E[X (0)-{E[X(O}? (l+at) =1
(1+at)?
(10) If O is uniformly distributed in (0,21), then E (X()} =1
fo(0) =, 2T
0 <0< 2, Now, E{X()} = n'Pn
(U) Elg(0)]=ffo(0).g(0).de
n=0.i
(at)n-]
n=l (l+ atyn+|
xk The process (X(O} whose probability distribution 1
n?(at)n-]
under certain conditions is given by (1+ at) n=l (1+ at)-l
(at)-1 n-1
n+1:=l,2, 3,... (1+at)2
at
P{X() =
n}=(1+ at)" n=l
at
n=0 1
yn-1
(1+at
(1+ at)
Show that it is not stationary. Ln=l
Soln. Given:
n
at
(at)n-! n=1 |1+at
n+1 ,n=1,2,3,..... 1.
P{X(1) = n}=(1+at)
at
(*a)2+6/ 1+at at at 2
+121+a)
, n =0
1+at
+20 at
The probability distribution of X() is
1+ at
<br>

lim
25 +| 25
1+6p2

.:. Mcan =R=/25 =5= E|X()]

Also, E(X*())=R(e-0) 25+ 4 =29
.:.
Var(X ()] = E
(X()-|E(X()|
.Var[X()]
=
29-(5) =4 Ans.
3

Ly3. Find the auto corrclation function of the random

process X(0) k .cos(o,t+ 0), wherc 0 is a
2 random Variable over the enscmble, and is
uniformly distributed over the range (0, 2).
3
at
(1+ at)
Soln. Given: X() = k.cos(o,! +0) ....(1)
We know that R,(tj, t2) = E[X(L). X(t)]
=
|+N n(u+)2 .=
E[k. cos(00l + 0). k cos(o0l2 +0)]
1.2
n
(n+ 1) (in +2)
E[2.cos(0olj +0). cos (00l2 +0)]
1,2.3 2
k?
2(1+ at) (1+at) -.E[cos (ool - 00l2)
2
(1+ at) 1

+cos, (oot + @gl + 20)]

(l+at)²
[2(1+ at)-1] k? 27
(l+ at) cos (o0t
=2+2at - 1 2 D
2T -wl2)
=1+2at cos
:: E{x*(0}l+2at
+
(ool1 t 00l, + 20)]d0

Also,
(:EIg (0)l =|fo (0) g (0).d0 and
4

Variance (X(t)}
fo (0 )=,0<0<2n for uniforn distribution)
- 2
E(X*(0}-[E(X())=1+2at-(1)²
k2
: Var{X()}=2 at sin (ooly + optg + 20)1
Since Var {X()} is a function of time t,
dcos (0ol-0t2)04+: 2
hence the Jo
given process'is not stationary. k?
..Ans.
lcos (0gt - o0t2).2n + sin (0ol| + 00l2))
,Ey2. Given the auto-correlation function for a
stationary
-[0 + sin (ool + ool2)1]
process 4
R,(r)= 25 4+
Find the mean and (" sin(720º +x)=
(1+62| sin x or sin (4n +x)= sin x)
variancc of the proccss k2
{X()}. -. cos (0ot| - 00l2)
Soln. Given: 4
R,(t)=25+ k2
1+6r? ..(1) cos
2 [o (t -12)]
We have, z2 im
R())
cos [oo
2 (t -l2)] Ans.
<br>

td. Show that the random process

X() sin t
2

X(t) = A cos (opl + 0) is n wide-scnse stationary, if 1

P|X()|
A and Onare constants and 0 2
is uniformly 2
distributcd RV in (0,2r). .. Mcan
of X(t) = E (X(0}
Soln. Given: X(1) =A cos (ogt +0)
= sin t
Since O is uniformly distributcd in (0, 2n)
sin t
fo(0)
=0<0<
2T
2r
Now E{X(1)} = E{A cos(oot +
(ii) Put
t= 0,25
0)}
for II
27 ..
-cos(ogt + 0).d0 X(t) =
27t forT
0

I Elg(0)l=| So(0).g(0).d0|
and
12n
<|i |sin (ogt + 0)l
2T

<|i (ogt + 2r)-sin (o)]

=sin
=(sin (oot) -sin (oot)] = 0= a constant
.. : F(x,0.25) = o... ns,
E
{X(t)} =a constant 2
1

Now,
R(1, t2) = E{X(t). X(t2)}
MARKOV CHAIN
2
.E [cos[(tj + tz) @0 +
20)]+cos[(t -12) oo]])
A Markov Chain or Markov proccss is a stochastic
(: By Ex.(3) modcl describing a scqucncc of possiblc cvents in
.: Ry(t1, = which the probability of cach cvent depcnds only on
t2) cos 0o (t -12)
2 the state attaincd in the prcvious evcnt. Markov Chains
(": Fron Ex.(3)) .have wide range of applications in the study cruise of
.. control systcms in motor vchicles, qucucs
R(tj, t2)=a Function of t -t2 passengcrs arriving at an airport, curTency
of
.". {X(t)} is a WSS process. cxchangc
....Ans. ratcs, statistical mcchanics, thermodynamics,
signal
proccssing, information theory ctc.
the fair coin experiment, the random process.
{X(0)} is defincd as follows: Definition:
If for t, <l2 <I3<nt,
= sin nt, for Head
X() P{X()<
| 2t, for Tail x|X() =N, X(t2)
Find: (i) Mcan of X(t)
(ii) Distribution function F (x, t) at t =0,25. P(X() sx| X(1,) =
Xn},
then the process (X(0)} is called a Markov proccss.
X(1) =3sin zt,
Soln. Given: for Head
In othcr words, if the future bchaviour of a proccss
2t, for Tail
lfonc coin is tossed, then S depcnds only on lhe present state, but not on the past.
={H, T} the proccss is a Markov process.
. P(H) = P(T) = A discrcte paramcter Markov process is callcd a
2
hdnrkou Chni)
 stochastic matrix.
<br>

0.3
0.02

0.7 Company:
Company 1

0.3
0.4

0.005
0.4
0.6 0.02
0.2
process
Fig. Diagram Representing two State Markov
a
Company3
Herc cach number represents the probability of
Markov process changing from onc state
to another
Fig. Graphical Representation of a CTMC
satc. For cxample, if the Markov proccss is in state A,
a CTMC with statc spa
is 0.4, whilc The abovc graph describcs
then the probability it changes to state E and transiti
(Company 1, Company 2, Company 3)
the probability it rcmains in state A is 0.6.
ratc matrix
Classification of States:
-0.025 0.02 0.005 1
Countable State Continuous or 0.2
Gencral State Q= 0.3 -0.5
Space
0.02 0.4 -0.42|
Space t
Discrctc-Time Discrete-Time Markov Chai on Stochastic Matrix:
a measurable state A stochastic matrix is a squarc matrix uscd
to descri
Markov Chain on
a countable or space. the transitions of a Markov Chain. It is also calI
finite state space. probability matrix, transition matrix, substitutie
Continuous Continuous-Time |Any continuous matrix or Markov matrix. It has widc range
Time Markov Process. |stochastic process applications in probability thcory, financc, comput
with Markov science and population genctics.
property.
Definition:
The above table gives an overview of the different A stochastic matrix is a squarc matrix whosc colum
instances of Markov processes for different levcls of arc probability' vcctors. A probability rector is
statc spacc generality for discrete vs Continuous time. numerical vector whosc cntries arc rcal numb
Classification of Chains: between 0 and 1 whose sum is 1,
(1) Discrete-Time Markov Chain (DTMC): A right stochastic matrix is a squarc matrix who
A DTMC is a sequence of random variables X1, cach row sums to 1.
A left stochastic matrix is a squarc matrix
X2, X3,.. with Markov propcrty, namely that the whosc cac
column sums to 1.
probability of moving to the next state depcnds only
squarc matrix of nor
A doubly stochastic matrix is a
on the present statc and not on the previous statc.
ncgativc rcal numbers whosc cach row and colum
P(Xn+| =x|X = X X = X2, Xn = Xn)
sums to 1.
= P(Xn+|
=x| X, = X,) 0.5 0.5
The possible values of X; forms a countablc sct S and Let A
=|0.5 0.25 0.25
is callcd the state spacc
of thechain.
(2) Continuous-Time Markov Chain (CTMC):
fo.5 0.5 fo.5 0 0.5
A CTMC is a continuous stochastic proccss
in which, 0
B=| 0.25 C=|0.5 0.25 0.25
for cach state, the proccss will changc
according to an
Cxponential random variable and |0.5 0,25 0.75 0.25
then move to a
different state as. spccified by probabilities Herc.A is a right stochastic matrix, B is a lel
a of
stochastic matrix, Cis a
doubly stochastic matrix.
<br>

Random Walk:
A random walk is a Gambler's Ruin:
mathcmatical objcct known as conccpt, mOSt
stochastic or random proccss, The ternm gamblcr's ruin is a statistical
that describes a path that a gambler playing
consists of a succession commonly cxpresscd as the fact that
of random steps on somc a game with ncgative cxpcctcd valuc will
eventually
mathematical spacc.
go brokc, regardlcss of their betting systcm.
Supposc we toss a fair coin cvery T scconds a persistent
and aflcr The tcrm's original mcaning is that
cach toss we movc a distance
'd' to the right if lhcad a
gambler who raiscs their bet to fixcd fraction ofthcir
comes and to thc
lelt if tail comcs. If thc proccss starts a
bank roll aftcr a win, but docs not rcducc it aftcr loss,
at t = 0, our position at time t cven if cach
squencc X(nT). The proccss {X(nT)}),
=nT is a random will cvcntually and incvitably go broke,
is called a bet has a positivc cxpcctcd valuc.
random walk.
Another common mcaning is that a persistent gamblcr
Suppose that r heads and (n-) tails have occurrcd in witlh finite wcalth, playing a fair gamc will cvcntually
the first n tosscs of the coin. Then the and inevitably go broke against an opponcnt with
random walk
-consists of r steps to the right and a
(n- ) steps to thc infinite wcalth. Such a situation can be modclcd by
lcft.
random walk on thc rcal number line. In that contcxt it
..
X(nT)= rd-(n-r)d is probablc that the agent will go brokc and is ruincd
:. X(nT)=(2r -n) d= md say, m = infinite number of timcs if the random walk continucs
2r-n
Also, P{X(nT) = md} = P{getting r hcads in n tosses} forcvcr.
Consider a coin flipping game with two players wherc
each playcr has cqual chancc of winning with cach flip
of the coin. After cach flip of the coin, the loscr
r= m +n
where transfers onc dollar to the winncr. The gamc cnds
2
X(nT) can also be expressed as when onc playcrs has all dollars.
X(nT) = X + X
Xn t..... Supposc player onc has n
dollars and playcr two has

where X; represents the distance movcd in i" step. ng dollars. The probabilitics P and P
that players
onc and two respcctivcly will cnd dollar less are
The RVs X; are independent, taking the valucs td
with cqual probability. P=
X d d
P(X)
Let n = 8$,
n =53
2 2 =0.3846 or 38.46%
8+5
8
=0.6154 or 61.54%
8+5
It can bc concludcd that, the player
Also, E{X2(nT)}= E[x?] that starts with
fewer dollars is morc likely to fail.
i=1

IMPORTANT FORMULAE:

distribution with (1) P{X(nT) =md} = S.cm²/2n

Since the limiting form of binomial Vnn/2
a normal
mean np and variancc np as n-> is-
o

(2) In a Markov Chain, the mcan rcturn. time to

distribution, hence statc 1 is
=c-(I-p)/2 npq r =l+2 t Pik, t is the expccted time until
"C,p'q= J2nnpq k
the
chain hits 1.
1
m+n
By putting p=q=,2' = we get (3) IfV = (a, b, c, d) wherc a
+b+
fixcd probability vcctor a c+d=lis a
uniquc
1
.e-m²/2n
of stochastic matrix P, then
P{X(nT) = md} VP= V.
/n/2
<br>

Ex6. An insect zooms around the verticcs of the (ii) We first find P(X1 =3)
transition diagram as given below: =
P(X, =3)
=
1-P(X l)-P(X =2)
Find: () Transition Matrix (ii) P(X, =3| X = 1)

4 4 2
0.6 Now, P(X =3, X2 =2, X3 = 1)

=
P(X =3)-p32 P21
0.

/ P32 is elenent in the 3" row and

2nd
colun ofpl
0.2
** Ans.
0.6
0.2 12
Soln.
Transition Matrix: Ex.8. ^ transition matrixP is givcn by
23
f0.6 0.2 0.2
1
1
4 2 4
P= 20.4 0 0.6 2
0.8 0.2
P=
3 3
P(X, =3|X = 1) =
(i) (p)13
2
: Multiplication of 1" row of 1" matrix
with 3r column of 2 matrix] Assumc X =1and let R be the first time that the
0.6 0.2 0.270.6 0.2 0.2 Markov, Chain rcturns to state 1, i.c.
0.4 00.6||0.4 0.6 R=min{n 1| Xn =1}. Find E[R| X =1].
0.8 0.2 0
0.8 0.2 Soln. Lct r bc thc mcan rcturn time to statc 1.
= (0.6 x 0.2) + (0.2 x + x
0.6) (0.2 0.2) i.e. r = E[R = 1]
=0.28 .... Ans. X
|

Ex1. Consider the Markov Chain with

then r =1+ tkPIk .(1)
thrcc statcs, k
S= {1,2, 3}, that has the following transition matrix: wherc t is the expccted time until thc
chain hits state
1
1, given Xo
2 4
=k.
i. t =0
3
0
t =1+t; Pkj. k z1

(i) Draw the state transition 3

diagram.
(1) If
h=1+ tjP2j
P(X;=)= P(X, = 2) =
j=l
find P(X1=3, X =2, X3 = 1) =1+P21t + P22t2 +
P23 3
Soln.
2

.... (2)

1 Now, t3 3
=1+tj P3j
j=l
23 =1+ p31t +P32
t + P33 3

Fig. State transition

diagram =1++0t2+13
<br>

Soln. Trajcctory:
a scqucnce of valucs of
A trajcctory is
»t;-3=l» ;=2 Xo, X, X2 ,... *

Hcrc P(Xn = 1) = (The first valuc of X,)

2 4
P (1, 2, 3, 2, 3,4)
X P32 X
P23 XP34
= P(Xo =1)xp12 XP23
k Pik
3 3 2
k=|

=1+1| pP|+l2 P12

+l3 P13 1
Ans.

ao)6)-() Ex.11.
10

Considcra population movemcnt betwcen city and

transition
suburbs in a metropolitan region by the
8

3
matrix:
8 From
E(R | X = 1]=* •. Ans.
Suburbs To
City
M= 0.95 0.03 City
0.05 0.97 Suburbs
Ex:9. In a Markov Chain with finite number of states, the
in
3 Suppose that the population in 2021 is 6,00,000
fundamcntal matrix is F= Find thc the city and 4,00,000 in the suburbs. Find the
3/2
population distribution in the ycar 2022 and 2023.
=
min{n e
Cxpcctcd valuc of T N
|X,EC}, Soln. Given: Initial approximation
wherc C is the sct of all rccurrent statcs. (Assume fo.6
initial distribution to be uniform). Xo = 0.4

Soln. We use the reward g=1

0.95 0.03
The vcctor V of cxpcctcd valucs of Tc, where cach Also, M=
|0.05 0.97|
cntry corresponds to a differcnt transicnt state is
The population distribution in 2022 = x = Mxo
V=Fg fO.95 0.5821
.. X =
Sincc the initial distribution has cqual probabilitics
0.05 0.97|0.4 0.418
1 i.e. in city in 2022
58.29% of the population will live
(cach with probability ), hence
2 and 41.8% of the population will live in suburbs in
25 2022.
E[T¢]=7x+ •. Ans.
The population distribution in 2023 = xX) = Mx
2 2 2 4
0.9 0.03|0.5821 f0.5651
Ex.10. Dcfine trajcctory. 0.05 0.97]|0.4180.435|
i.e. 56.5% of the population will livc in city in 2023.
Let Xo
43.5% of the population will live in suburbs in 2023.
1, 2, 3, 2, 3, 4
Find the probability of the trajcctory Ans.
for the following transition diagram.
Ex42. Find the unique fixcd probability vcctor of the
1
stochasticmatrix
1/2 1/4 1/4

6 1/2 0 1/4. 1/4

5 P=
1/2 1/2 0
1/2
| 1/2 0
Soln. We have to find V = (a, b, c,
d),
 -a-
<br>

wherc a +b+c+ d=1 *..... () la.bel

Such that VP = V
Consider VP V

1/2 1/4 1/4|

-=a =c=2a
2
1/2 1/4 1/4
: [a,b,c, c]
1/2 1/2 0
= [a, b,c, c]
+=b
a C

2
|1/2 1/2 2a
bc=
d a a
b b Now,(1)»a+bt c=1
- [a, b, c, d] a+2a +
2a = 1a=

(b+c+d) =a or b +c+d=2a ....(2) Now, b= 2a =

-(a+c+d) = b or a +c+d=2b ..... (3)

and c= 2a
2
5
a+b)=e or a+b4c ....(4) :. The requircd unique fixcd probability vector

+ b) = d or a +b= 4d
(a
4
*... (5) Ans.
()»btc+d= -a ... (6)
(1)a+c+d=1-b (7) SAMPLINGTECHNIQUES
Now, (2) and (6) 2a = 1-a a
3
Population (Universe):
(3) and (7)»2b
=1-b» b=
3
A group of individuals under study is
callcd
population or universe. It may be finitc or infinitc.
:: (4)
=4c3 3
c
6
Population sizc is usually dcnotcd by N.
1 Sampling:
(5) -tz4d
33
d=
6 A part of individuals selccted from
the population is
The required unique fixed probability vector called a' sample. The proccess
of sclcction of a sample
is called sampling.
|1 11 1|
Ans. Random Sample:
A random sample is a sample in which cach member
Ex.13. Find the uniquc
fixcd probability vector of population has an cqual chance
of the it.
of being includcd in
rcgular stochastic matrix Sample sizc is usually
denotcd by n.
1 There arc C (N, n) different
samples of sizc
can be pickcd up 'n' that
01 from a population size
of 'N'.
1/2 0| Parametcrs and Statistics:
Soln. Wc have to find The statistical constants
of the population such as
V=(a, b, c), a +
b+c=1 mean (4), standard
(1) deviation (o) arc
Such that VP = V callcd
paraméters. Paramcters arc
Consider VP =V usually' denoted by Grcek
letters.
1
0 The mcan (x), standard
» la,b,c]0 knows as statistics.
deviation (s) of a sample are
|1/2 1/2 Statistics arc usually
Roman letters. dcnoted by
 an clemcnt
<br>

Population starts by sclccting

Characteristic Sample Iramc. The sampling clemcnt in
Paramcter Irom thec list at random
and thcn cvcry k"
Statistic intcrval
Sybols Population size = N the framc sclectcd,
where k is the sampling
Sample
givcn by
size n
Population mcan
=| Sample mcan kN
Population standard | Sample Sampling Distribution:
deviation =o A samplimg distribution
is the probability distribution
standard
statistic. If an
deviation=s of a given random, samplc bascd
Population involving
proportion Sample arbitrary large number of samples, cach
p in order to
proportion multiple obscrvations, werc scparatcly uscd
cach
computc samplc mcan or samplc variancc for
. is
sample, then the sampling distribution the
Aims of a Samplc:
The estimate of mean and standard deviation of thc probability distribution of the valucs that the statistic
population is a primary purposc of all scicntific 1akes on.
experimcntation. The logic of sampling theory is the Standard Error:
.
logic of induction. In induction, we pass from a Standard crror (S.E) is thc standard deviation of the
particular (sample) to:general (population). This typc sampling distribution. S.E. is uscd for asscssing the
of gencralization herc is known as statistical infercncc. differencc betwcen cxpcctcd valuc and öbscrvcd valuc.
The conclusion in the sampling studics are bascd on
Reciprocal of S.E. is callcd prccision.
probabilities and not on certainties.
Types of Sampling: S.E. = Gg =
(1) Purposive Sampling: n
Purposive sampling (also known as judgemental, Sampling Distribution of Mean from Infinite
sclective or subjective sampling) is a form of non Population:
probability sampling in which rescarchcrs rely on their Let the population be infinitcly large having a
own judgement when choosing mcmbers of the
population to participate in their surveys.
population mean and population variancc o. 1fx is
a random variable denoting thc mcasuremcnt
(2) Random Sampling: of the
charactcristic, then
Random sampling (or probability, sampling) is a E (*) =
sampling method that allows for the randomization of
šample selection i.e. cach sample has the same Vax (x) =g²
probability as other samples to be selccted to serve as a The samplc mcan X is thc sum of n random
variablcs
representation of an cntire population. X, X2, ..,
Xn cach being divided
byn.
(3) Stratificd Sampling:
Stratificd sampling is a method of sampling from a Thus E(x)= 4, Var (®) =
;S.D.(x) =
population which can be partitioncd into
subpopulations. In statistical survcys, when If the samplc size 'n' is incrcascd
thcn Var (x) gets
subpopulations within an overall population vary, it rcduccd. By taking large
valuc of n, Var (X) can
could be advantagcous to sample cach subpopulation be
made as small as desircd.
(stratum) independently.
Sampling from Normal Population:
(4) Sy'stematic Sampling:
Systcmatic sampling is a statistical method involving If x-N(H, ~
o), then
N,
the sclection of clements from an ordered sampling
<br>

IMPORTANT FORMULAE: E.IS, A sample size 25 is piclked up at random from a

populatiòn which is normally distributed with
() Standard error (S.) mean 100 and variance 36. Find:
(i) P{TS 99}
(2) i(x), Var(x) o, SD.(x) o, x is a R.V.
(ii) P{98 s*S 100}
(3) ) Var() SD(): Soln. Ilere n 100, g => G
25, =36 =6
is sample mean. Also, x
~N (100, 36)
(4) lfx~N(u, o*) then ~
R~N00,
(t)
Ex14The diameter a
component produccd on a scmi
of

automatic machìnc is known to be distributed

normally with a mean of 10 mnm and standard
deviation of 0.1 mm. If we pick up a random
sample of size $, what is, the probability that the
same mean will be betwecn 9.95 mm
and 10.05 mm.
Soln. Lct x be a random variable represcnting the diamneter -0.83333 0
of one component picked up at random.
Here x~ N(10, 0.01) 99
P{RS99} =P

Vn

and Z= - 6/5
=P{Zs-0.83333}
=0.5 - Arca bctwcen Z=0 and Z=0.83333
=0.5 -0.2967 =0.2033
P{X S99} =0.2033
Ans.
(ii) P{98 < S 100}

9.95 =10 10.05 =P

2-X-, 100

Now, P{9.95 <

Xs10.05}
=2x P{10 s s 10.05}
98-100
-2xP 10-! 10.05 = P.
-sZs0
Vn
=P-1.6666 < Z s0} =0,4515

=2xP}osZs l0.05-10
0.1

*=2x P{0sZsl.12}
=2 x 0.3686 = 0.7372
P{9.95 <x s 10.05}
=0.7372 Ans.
-1.6666
P{98s x
s100}=0.4515 Ans.
<br>

Sampling Distribution APPLIED MATHS-II

A sample statistic
of the Variance: rejcct thc
called the sample variance inconsistent with thc null hypothesis, then
estimate the population uscd to null hypothesis and conclude that
the alternative
variance. The sample variance
is usually denoted by hypothesis is truc.
S. Errors:
Therc arc two types of crrors:
(1) Type I Error:
is rcjccted whcre it should havc bcen acccptcd.
n-1
Testing a Hypothesis:
If H
(2) Type II Error:
A hypothesis
is. a proposed explanation for a If H is acceptcd where it should have been rejcctcd.
phenomenon. A statistical
hypothesis is a hypothesis Level of Significancc:
that is testable on the basis
of observed data modellcd In statistics, significancc mcans "probably truc." "The
as the recalizcd values
taken by a collection of random level of significance is stated to be the probability of
variables.
typc I crror and is presct with the outcomcs of crror. It
On the basis of samplc information,
certain decisions defincs whcthcr the null hypothcsis is assumcd to be
are madc about
the population, by making ccrtain accepted or rcjcctcd. It is denotcd by
.

assumptions. Thesc assumptions are known as

statistical Probability Valuc (p-Valuc):
hypothesis. Assuming the hypothesis correct, we
In null hypothcsis significance testing, the p-valuc is
calculate the probability of getting the observed the probability of obtaining test rcsults atlcast as
sample. If this probability is less
than. a certain extreme as the results actually obscrvcd, under thc
assigned value, the hypothesis is to be rejccted. assumption that null hypothesis is acceptcd. A vcry
Null lIypothcsis: small p-valuc mcans that such an cextreme observcd
The statement bcing tested in a test of statistical outcomc would be very unlikcly under the null
significance is called Null Hypothesis. Usually, the hypothesis.
null hypothesis is a statement of."no cffcct" or "no Examplc:
difference." It is often denoted by Ho: The valuc significant at 5% refers to p-valuc less than
In a null hypothesis it is presumed that there is no 0.05 i.e. p<0.05, similarly the value significant at 1%
means p <0.01.
significant difference between the observed value ánd
cxpected valuc. Then is tested whether this Area of Acceptance
hypothesis is satisfied by the data or not. If the
hypothesis is not approved the difference is considercd
significant. If it is approved then the difference would
be describcd as due to sampling fluctuation.
Alternate Hypothcsis: o.5%
o5%8
The statcment that is bcing testcd against the null
Z=-1.9 0
Z= 1.9
hypothesis is callcd Alternate Hypothesis. It is denotcd
by Area of Acceptance
or
H
H.
In statistical hypothesis testing, the alternatc hypothesis
a new
is a position that states somcthing is happening,

thcory is prcferred instead of an old theory. It usually

is
it is
consistent with the rescarch hypothesis because 25%1 2'5%9
ctc.
constructcd from literaturc revicw, previous studies Z=-0.6 Z= 0.6
Example: Therc arc two critical regions which cover 5% and 1%
sample data
Take a samplc fromn the population. If the arcas of the normal curvc.
do not The shadcd portions arc the
is consistent with the null hypothcsis, then
critical regions. Thus, the probability
samplc data is of the valuc of
rejcct the null hypothesis. If the
 the variate falling in the critical region
<br>

is the levcl of of a distribution is two-sided and tcsts whether a

significancc. If thc variate falls in the critical area, then
Sample is grcater or less than a range of valucs. If the
hypothesis is to be rcjected.
sample bcing tcsted falls into cithcr of thc critical
Note: arcas, thc alternate hypothesis is acceptcd. By
To mcasure the lcvel of significance of thc result, onc convention, two-tailed tests are uscd to dctcrmine
nccds to calculate the p-value. When p-value is less than significancc at 5% level, mcaning cach side of the
a., the null hypothesis is rejected. If p-valuc is not less distribution is cut at 2.5%.
than a,then theoretically nullhypothcsis is accepted. Example:
Test of Significancc: Let a company manufacturcs bolts with ncan mass
Thc tests which cnable.to decide whcthcr to accept or 100 g. i.c. H,.u =100g. Supposc an employcc
rcjcct null hypothesis arc
callcd the tests of bclieves that mcan is not cqual to 100 g. i.c.
significancc. If the differcnce betwccn the 'sample Whenever H, docs not cqual a value,
H,:100g.
valucs and the population values is so large, then the
wc usc two-tailcd test. Thc shadcd arca rcprescnts
null hypothesis is rejected.
rejected region and FTR is Failed to Rejcct Rcgion.
Confidence Limits:
The -numbers 1.96, 2.58 arc callcd confidence
cocfficicnts. 0.95
FTR

=0.025 0.025

Accept H, Reject H,

The Z valucs that separate the rejcctcd rcgion and FTR

l-1.96 l+ 1.960
region are known as critical valucs. Let therc be a
confidence level of 95% i.c. C = 0.95.
Then a=1-C=10.95=0.05
In order to accept or rejcct Ho, we have to calculate
the value Zç.If Zç> Z, then H, is rcjcctcd.
There are two types of onc-tailcd tests viz. left tailcd
l-2.58g test and right-tailcd test.
+2.586

itl.96 are 95% confidence limits as the arca

interval utl.96
interval. Similarly
,
betwccn them is 95%. If a sample statistics lies in the
then we call 95% confidence
ut2.58 o are 99% confidence
limits.
Notc: H, :u <100 g
If C
is the confidence levcl and a is the significancc
level, then C + a =1.
Onc-Tailed and Two-Tailed Tests:
(1)
A
onc-tailed test is a statistical hypothcsis test set up to
show that the sample mean would be higher or lower
than the population mean, but not both. It tests the arca
under one of the tails of a normal distribution. H, : u >100 g
(2) A two-tailed test is a method in which the critical arca
<br>

THE F-DISTRIBUTION (F-TEST) Decision Criteria: thcn

is grcater than the fcritical valuc,
An F-tcst is a statistical test in
which thc test statistic If the fstatistic
has an F-distribution under the rejcct the null hypothesis.
null hypothesis. It is
mostly uscd when comparing statistical models F-statistic (F):
that
havc bccn fitted to a data set, in order to identify is the variancc
the
modcl that best fits the population from which the data (1) For large samples, F=. wherc of
is sampled. Common examples
of the use of F-tcst and
g is the variance of the
includc the study of following cascs: of the first population
The hypothcsis that the means of a given sct sccond population.
(1) of
normally distributcd populations, all having the samc
For small samples F=, wherc Sf is the variancc
(2)
standard deviation, are cqual. This typc of F-test plays
an important rolc in the Analysis of Variancc s is thc variance of the
(ANOVA). of the first population and

(2) The hypothesis that a proposcd regrcssion model fits second population.
the data well. Notc:
(3) The hypothesis that a data sct in a regression analysis (1) For right-tailcd and a two-tailcd F-tcst, thc variance
a
numcrator. Thus, the
follows the simplcr of two proposcd linear modcls that with grcatcr valuc will be in the
arc ncstcd within cach othcr. will become the first
samplc corresponding to of
Definition: be the
sample. The smaller valüc variance will
F-tcst can be defincd as a test that uses the test statistic
to check whether the varianccs of two samples (or denominator and belongs to the sccond samplc.
(2) For a left-tailcd test, the.smallcst variancc bccomcs the
populations) are cqual to the same valuc. To conduct an
F-tcst, thc population should follow a F-distribution numerator (sample 1) and the highest variancc gocs in
and the samples must be independent cvents. On the denominator (samplc 2).
conducting the hypothesis test, if.the results of the F F-Critical Value:
test arc statistically significant, then the nullhypothesis is a a
tcst statistic is
point that
A critial value
can be rejectcd, otherwise it can not be rejected. comparcd to in òrder to decide whcther to rcjcct or not
F-test Formula: to reject the null hypothesis.
The F-test formula for different hypothesis tests is Graphically, the critical valuc divides a distribution
given as follows:
into the acccptance and rcjcction regions. If the test
(1) Left-Tailed Test: statistic falls in the.rejection region, then the null
Nullliypothesis: H,:oo
hypothesis can be rejcctcd otherwisc it cannot be
os
Alternate Hypothesis: H,:of< rejccted. The steps to find the f test critical value at a
Decision Criteria: specific alpha level (or significance levcl) a arc as
If the f statistic is less than the f critical value, then follows:
rcjcct the null hypothesis. (1) Find the degrccs of frccdom of the first samplc. This is
(2) Right-Tailed Test: done by subtracting 1 from the first sample sizc. Thus
Null lIlypothesis: H,: oi =o; X

=n-I.
Alternate Hypothesis: H}:of>o (2) Find the degrees of frecdom
of the sccond sample by
subtracting 1
from the samplc sizc. Thus y =
Decision Criteria: n2-1.
If the f statistic is grcater than the f critical value, then (3)
rejcct the null hypothesis.. Test Significance Levcl
(3) Two-Tailed Test: Right-tailed test

Null lHypothesis: H,: oi=o Left-tailed test

Two-tailed test
Alternate Hypothesis: H;:o #o
2
<br>

(4) The F-tablc is uscd to find thc critical valuc at thc

dí; =8-l=7, df, =6-l=5
rcquircd alpha level.
(5) Thc intcrscction of x column and y row Also Sf = 600, S = 400
the F-tablc
will give the F-test critical valuc. The ftcst formula is
600
IMPORTANT FORMULAE: =15
400
(1) Sample variance, From F-distribution table, thc f criical valuc for
=
significance level a=0.05 is F (0.05,7,5) 4.88

n-1 We obscrve that 1.5 <4.88

llence the null hypothcsis cannot be rcjcctcd and therc
(2) If C is thc confidcncc levcl and a is thc is not cnough cvidcnce to conclude that therc was an
significancc level,then C+a =1. improvcment in insomnia aftcr thc ncw drug. ...Ans.

(3) F-statistic for large samples, F= Pizza delivery times of two cities are given below:
ExJ
City 1: Number of delivery times obscrvcd = 28,
wherc f and os are the variance of first and Variance = 38.
sccond population respcctively. City 2: Number of delivery timcs obscrved = 25,
Variance = 83.
(4) F-statistic for small samples, F =
s Check if the delivcry timcs of city 1 arc lesser than
city 2 at 0.05 alpha levcl.
where Sf and S5 arc the variance of first and Soln. This is an cxample of a left-tailcd F-tcst.
sccond sample respcctivcly. Hence alpha levcl=|-0.05 =
0.95
(5)
F-tcst |Significancc Levcl Hj:s; <s
Right-tailed test Since 38 < 83, hcnce city l will be sample 1l
and city 2
Left-tailed test will be samplc 2.
Two-tailed test Here n
=28, n =25
=
df 28-]=27
(6) Degrce of
frecdom in F-test df=n-1, n is
the df,
=
25--l=24
number of samplcs.
Also, Sf = 38, S = 83

Ex0. A rescarch tcam wants to study the cffccts of a The F-tcst formula is
new drug on insomnia, 8 tests werc
a variance of 600 initially, After 7
conductcd with F=S38 =0.4578
were conducted with a
months 6 tests S
83
variance of 400. At a f-critical value is
significance lcvcl of 0.05, Was there any
improvement in the results after 7 months? F(0.95,27,24) =
Soln. This is an example of a right-tailcd test. F(0.05,24,27)
Herc F-test nceds to be uscd as we
have to compare =0.5181
the variancc. 1.93
Since 0.4578 < 0.5181, hence
= S the null hypothcsis can
H,: S be rejected and it can bc concluded that
therc is
H,:S >s3 enough cvidence to support the claim that the dclivery
times in city l are less than the delivery
Herc n
=8, n2 =6 times in city 2.
e
ns.
 Ex.18. A toy
<br>

manulacturer wants to get batterics for toys. S =50

A team collected 41 samples
fröm supplier
Also S= 200,
the variancc was 110 hours.Thc team A' and The F-tcst formula is
also collectcd
21 samples from supplicr
B with a variance of 65 F=: 200
=4
hours. At 0.05 alpha level, detcrminc s 50
if there is a
difference in the variances. is
Using F-distribution table, f-critical valuc
Soln. This is an example of a two-tailcd
F-tcst. F
(0.05, 10, 50) =2.026
Hence alpha levcl is = 0.05 =0,025 Since 4> 2.026, hencc thc null hypothesis cannot be
2 rejccted.
=
H,:S S; .:. It can be concludcd that thcre is cnough cvidcncc
arc
s that premiums paid by cithcr of insurancc scgmcnts
H:S; morc variablc as comparcd to othcr. Ans.
Here, n=41, n, =21
df =41-|=40 Ex.20A bank has hcad office in Dclhi and branch in
Mumbai. The variancc of Dclhi hcad office is 31 for
df, =21-1=20
a sample size of 11, while the variance of Mumbai
Also, Sf =110, S5 = 65 branch is 20. for a sample size of 21. Carry out a
The F-tcst formula is two-tailed F-test with a lcvcl of significance of 10%
S 110 and find whether the number of customers at Delhi
F= =1.69
65 branch arc morc variable than the number of
customcrs at Mumbai branch.
Using F-distribution table, f-critical value is
Soln. Givcn significance levcl = 10% =0.1
F (0.025, 40, 20) = 2.287
Since 1.69 < 2.287, hence the null hypothesis cannot Alpha level a 0.1
=0.05
be rejcctcd. .. Ans. .2 2
H,:S -s
Ex.19% An insurancc company sclls health insurance and Hj:s; S;
motor insurance policics (sce the below table).
Here, n
=11, n, =21
Determine whether the premiums paid (by
.:. df
customers) by cither of insurance scgmcnts (health =11-1=10 and df, =21-1=20
insurance and motor insurancc) are more variable Also S
=31, S = 20
as comparcd to other. Conduct a two-tailed F-test
The F test formula is
with a level of significancc of 10%.
S
Health Motor F=I3I20 =1.55
S
Insurance Insurance
Variace ($) 200 50 Using F-distribution tablc, f-critical
valuc is
Sample size 11 S1 F (0.05, 10, 20) = 2.348
Since 155 < 2.348, hence we cannot
Soln. Given significancc level 0.1 rejcct the null
hypothesis.
.:.
Alpha level = a_0.0.05 It can be concludcd that thcre is
cnough evidence that
2 customers at Dclhi branch arc more
variable than that
H,:S =S at Mumbai branch.
Ans,

THE (-DISTRIBUTION (t-TEST)

Here, n
=l1, n, =51
In statistics, the Student's
df=l|-1=10 distribution) is any
t-distribution (or simply
t
= member a of family of continuous
df, =51-l 50 probability distributions
that arise when cstimating
the
<br>

mcan of a normally distributcd population in situations Given that t = 2.539

wherc the samplc size is small and the population's ..
standard deviation is unknown. t<t
Hencc the batch is up to the standard. ...Ans.
The t-distribution is symmetric and bell-shaped like
the normal distribution. However, it has heavier tails,
Ex.22. The mcan of a random sample size of 16 valucs
mcaning that it is more prone to producing values that sum of
fall far form its mean. drawn from a population is 55 and the
squarcs of deviation from the mean is 150. Can this
The t-distribution formula is
sample be regarded as taken from thc population
t= having 58 as mean?
S/Nn
For a =0.05, t=2.131 and for a= 0.01, t=2.947
wherc x is the sample mean, u is the population
mcan, S is the standard deviation and n is thc samplc Soln. Given: u = 58, n = 16, X =55

sizc. Also E(x-x)' =150

Working Rule:
To calculate significance of sample mcan at 5% levcl,
E(x-)? 150 =
10
n-J 16-1
wc find t= and comparc it to the valuc of t
S/n
.:.
S=/10
with n- 1 degrces of
frccdomat 5% significance levcl. 55- 58
Now, t = =-3.79
We denote this tabulatcd value by
t. S/n /10/16
..
(1) If t <tj, then we accept thc hypothesis i.c. wc can say t=3.79
that the sample is drawn from the population. When a =0.05, then t =2.131
we rejcct the hypothesis and the sample is .:.
t>t,
(2) If t>t,
a
not drawn from the population. When =0.01, then t, =2.947
..
t> t
IMPORTANT FORMULAE:
Hence, the sample cannot be rcgardcd as taken from
the population. o.... Ans.
(1) The t-distribution formula is t=
S/n
where X is the sample mean, u is the A fertilizer mixing machinc is set to given
12 kg of nitrate for quintal bag of fertilizer. Tcen
populaion mean, S is the S.D. and n is. the
100 kg bags arc examincd. and
sampfsizc. the percentages of
nitrate per bag arc found to be as given below:
12, 14, 13, 13, 12,.12, 14, 11, 12, 13
Ex.21/A manufacturer intends that his LED bulbs have a
Is there any reason to belicvc
lifc of 1200 hours. He tests a samplc of 20 that the machinc is
bulbs defcctive? Value of t
drawn at random from a batch and finds that the fo 9 degrccs of frccdom is
mcan life of the sample 2.262.
bulbs is 1190 hours with an Soln. Given: We choosc
S.D. of 22 hours. Does this signify
that the batch is
not up to the standard? (Given: The value Mcan u = 12, n= 10
of t at ..
1% level of significance with 19 degrccs
of frccdom d=x-=x- 12
is 2.539). We calculatc X for the givcn
data
Soln. Given: X
=|190, S=22,
= 1200, n d=x- 12
=20
The t-distribution formula is 12
0
14
t=-#_190-1200 =-2.033 13
2 4

S//n 22//20
13
.:.t=2.033 (ignoring negative sign)
12
0 0
<br>

14
0 67-65 =
2.02
2 Now, t: 3. 13/10
11
-1
4
S/n mean height is 65)
12
(.: When
13 Given t; =2.262

Ex = 126 .. t<tj
2d =6
sa =12 Hcncc thc hypothcsis is acccptcd and the
mcan hcight

Ex 126 ....
=12,6 of univcrsc is 65 inches. Ans.
n 10

Now, s2 L? Ex.25. samplc of 6 pcrsons in an officc revcalcd an

A

10
=0.84 average daily smoking of 10, 12, 8, 9, 16, 5
.. S=0.9165 cigarcttcs. Find the avcrage levcl of smoking in the
Now, t= X-4 12.6-12 officc cstimated at 90% levcl of significancc. Usc
S/n 0.9165/W10 =2.070 t=#2.015 for 5 degrces of frccdom.
Given t =2.262
Soln. We choose mcan =
10
..
t<t d=x==X- 10
Hencc therc is no rcason to belicve that the machinc is
d=x 10
defcctivc. ..Ans.
10 0
Ex.24. Ten individuals are ch0scn at random from a 12 2 4

population and thcir heights arc found to.be (in -2 4

inches) 63 63, 64, 65, 66, 69, 69, 70, 70, 71. Discuss 9

the suggestion that the mean height of universe is 16 6 36

65. For 9 degrees of frccdom the valuc of t at 5% 5 25
level of significance is 2.262. Ex =60 La' =70
Soln. Given: n = 10
2x 60
We choose mcan =67 X= =10
6
d=x-u=x- 67

d=x-67 S
70
=3.74
16
Vn-1 V6-1
63 -4
63 -4 16 At 90% levcl of significance, t=+2.015
64 -3 9

-2 4
We have, t=
64
1
S/n
66
69 2 4 > 12.015=0-u 3.74V6
69 4

9 x 3.74
70 3
|=t2.015 +10
9
70
16 =6.92 or 13.08
71 4 Ans.
Žx=670
d
=88
CHI-SQUARE TEST
670
Ex -=67
10 A Chi-Squarc test is a statitical hypothesis
test that is
valid to perform whcn the test statistic is
88 Chi-Squarcd
Also, S = =3.13 distributcd under the null hypothesis.
Vn-1 V10-1 If the null
<br>

hypothesis that there arc no differcnccs betwccn thc Degrccs of Frccdom:

a variables
classes in the population is true, the test statistic (1) If the data is given in the form of serics of
degrccs of frccdom No.
a Chi in a row or column, then thc
computed from the observations follows
of items in the scrics -1.
Squarcd frequcncy distribution. The purpose of thc test are put in cclls in a
(2) When thc numbcr of frcqucncics
is to cvaluate how likcly the obscrved frcqucncics of frccdom (R
contingcncy tablc, thc degrccs
would be assuming the null hypothesis is truc.
In crypt analysis, the Chi-squarcd test is uscd
to (C- 1).
where, R= No. of rows
compare the distribution of plaintext and decryptcd
C= No. of Columns
Cipher tcxt. The lowest value of the test mcans that the
decryption was successful with high probability. IMPORTANT FORMULAE:
Chi-squarc Variatcs:
The square of a standard normal variatc is called
x
Chi (1) Chi-squarc variatc,

Squarc variate and is denotcd by x

(2) In Chi-squarc test,
x = (0; -E;)'
Let X
be the standard normal variatc. i=l
E;
wherc O; is the obscrved frcqucncy and E; is thc

Then y' -z is a Chi-square variate.

Cxpected frequcncy.
a
(3) Ifthe data is given in the form of scrics of variables
= No
Let X,X2, X3y . *n be a random sample of n in a row or column, then thc degrcc of frccdom
of items in the series - 1.
valucs from a population and x be a normally
(4) If the number of frequcncics arc put in cclls in
distribution variable.
a contingency table, then the degrcc of frccdom
Then ) = x +x +x t.... * has x
distribution = (R- 1) (C- 1), where R and C arc no of rows and
columns respcctivcly.
.with n-degrees of frecdom.
(5) In Chi-squarc test as a test of independence,
Chi-Square Test of Fit:
of Goodncss

The goodness of fit test is used to test if sample data Expected Frequcncy Row Total x Column Total
Grand Total
fits a distribution fróm a certain population. In other
words, it is used to test the significance of the
Ex.26/ In a random sample of 10 bags, cach bag has 100
discrepancy between theory and experiment. It helps
picces of candics and five flavors. The following
us to find whether the deviation
-of the experiment table gives the combincd flavor count from all 10
from theory is just by chance or it is due to the
bags.
inadcquacy of the theory to fit the observed data. Flavor Apple Lime| Cherry
The value of x is calculated using the following
GrapcOrange|Total
No. of 180 250 120 225 225 1000
formula: picces
Test whether the proportions of
(0, -E,)² five flavors in cach
bag are same. The Chi-Square valuc with a=
0.05
and 4 dcgrecs of freedom is 9.488.
Herc O; is thc observed frequcncy
and E; is the Soln. Here cach bag has 200 picces
cxpccted (thcorctical) frequcncy. of candics and five
flavors.
Notcs:
Hence we cxpcct 100 =
(1) The valuc of x is always positive. 20 picces of candics in cach
flavor from cach bag.
(2) x will be zero if cach pair is zero. Sincc therc arc l0 bags, hencc cxpccted
valuc = 20 X
10 = 200i.c. we cxpcct 200
3) 0sx<o flavor.
picccs of candies in cach
<br>

O; O;
-E; (0;-E) (;-E;) .:.
x=3.41 of
a=0.05 and 6 dcgrccs
The Chi-Sçuarc
valuc with
180 | 200 -20 400 2 frccdom=12.592.
250 200 +|+ 50 hencc wc acccpt the
null
2500 12.5 Since 3.41 < 12.592,
arc uniformly distributcd
hypothesis that the accidents
120 200 80 6400 32
225 200 25 625 3.125 over the weck.
225 200 25 625 3.125 Chi-Squarc Test as a Test of Indepcndence:
52.75 Row Total x Column Total
Expcctcd Frcqucncy= GrandTotal
:. x= 52.75

The Chi-Squarc value with a= 0.05 of two

and 4 degrccs of
Ex.28, In an invcstigation of hcalth and nutrition
frccdom=9.488. status, the
groups of children of differcnt social
Since 52.75 > 9.488, we rcjcct the null hypothesis that
following results arc obtaincd:
the proportions of flavors of candics, arc cqual. ...Ans.
Social status Poor Rich Total
.Ex.27. The following tablc gives the number of car Hcalth
130 20 150
accidents that occurrcd during various days of the Below Normal
102 108 210
weck. Test whether the accidents arc uniformly Normal
Above Normal 24 96 120
distributed over the weck.
256 224 480
Days No. of Accidnts Total
Monday 18 Discuss the rclation betwecn hcalth and social
Tuesday 17 status.
Wednesday 18 Critical valuc of y at 59% significance levcl is 5.99.
Thursday 17 is no
Soln. Let H,: Null Hypothesis: Therc association
Friday 15
betwccn health and social status.
Saturday 20
H,:Alternate Hypothesis: There is an association
Total 105
between health and social status.
Chi-Squarc valuc at 0.05 significance level for 6
RowTotalx ColumnTotal
Expccted Frequency=
degrccs of frccdom is 12.592. Grand Total
arce
Soln. Let Null Hypothesis: The accidents
H,: Social status Poor Rich Total
uniformiy distributed over the weck. Hcalth
The expccted frcqucncies of accidents on each day
Below Normal 256x 150 224 x150 150
10S15. 480 480
7
=80 =70
O;-E; (o; -E)* (O- E)² Normal 256 x210 224 x 210 210
480 480
18 15 3 9 0.6 = |12 =98
17 15 2 4 0.27 Above Normal 256x 120 224 x 120 120
18 15 3 9 0.6 480 480
17 15 2 4 0.27 =64 = 56
Total
0 0 256 224 480
15 15
20 15 25 1.67 Total number of obscrved frequcncics =
Total numbcr
3.41 of expectcd frequcncies = 480.
<br>

Degree of frecdom = (R - 1) (C- 1) Total number of obscrved frcqucncies

= Total numbcr
= (3 - 1) (2-1)=2 of cxpcctcd frequcncics
= 60.

E; O; -E; (O; -E) - Degrce of frccdom =(R- 1) (C- 1)

(0; E)² 1) (2 - 1)= I

E =(2-
130 80 50 2500 31.25 E; O; - E; (0; - E) (0; -E;)
102 112 - 10 100 0.89 E;
24 64 40 1600 25
12 -4.67 21.81 1.31
20 70 - 50 2500 35.71
16.67
13 8.33 4.67 21.81 2.62
108 98 10 100 1.02
28 23.33 4.67 21.81 0.93
96 56 40 1600 28.57
7 11.67 -4.67 21.81 1.87
Total 122.44
Total 6.73
y
Here =122.44
y
Herc =6.73
The Chi-Squarc valuc with a = 0.05 and 2 degrcs of
The Chi-Squarc valuc with a = 0.05 and 1
dcgrccs of
frecdom =5.99
Clcarly 122.44 >
frecdom=3.84
5.99
Clearly 6.73 > 3.84
Hence, the null hypothesis is rejcctcd and we conclude
Hence, thc null hypothesis is rcjcctcd.
that there is an association betwecn hcalth and social
We conclude that thcre is an cffcct of vaccinc in
....
status. Ans. controlling the discase. Ans.

Ex.29. In an cxperiment of immunization of cattlc from a Ex.30. A survey amongst women was conducted to study
discase, the following rcsults are obtaincd: the family lifc. The obscrvations arc as follows:
Affccted Family lifc
Unaffccted Total
Inoculatcd 12 28 llappy Not Happy Total
40
No Inoculated Educated 70 30 100
13 7 20
Uncducated 60 40 100
Total 25 35 60
Total 130 70 200
.
Examine the cffcct of vaccine in controlling the Test whether thcre is an associating betwcen family
incidence of thc discase. lifc and cducation.
Soln. Let Ho: Null Hypothesis: There is no effcct of Soln. Lct H: Null Hypothcsis: Therc is no association
vaccine in controlling the discase. betwccn family life and cducation.
H,: Alternate Hypothesis: There is cffect of H,: Altcrnatc Hypothesis: Therc is an association
vaccine in controlling the discasc. betwecn family lifc and cducation.

Expected Frequcncy=: RowTotalxColumnTotal Expcctcd Frcquency = RowTotal x ColumnTotal

Grand Total Grand Total
Affccted Unaffccted Total IHappy Not Ilappy Total
Inoculated Educatcd 130 × 100 70 x 100
25 x 40 35 x 40 40 100
60 60 200 200
= 16.67 =65 =35
=23.33 Uncducatcd 130x100 70 x 100 100
No Inoculated 25x 20
35 x 20 20 200 200
60 60 = 65
-35
= |1,67 Total 130 70
=8.33 200
Total 25 35 60 Total number of obscrvcd frcqucncics = Total number
of cxpcctcd frcqucncies 200.
 Degrcc
<br>

of frecdom = (2- 1) (2- 1) (2- 1) =1

=1 Degrcc of frccdom= (2- 1)
E; O; - E; (0; - E) - (0, - E)
(; E) O; E; O; -E; (0; -E)?
E
E;
70 65 5
15.38
25 0.385 216 165.54 50.46 2546.21
60 65 -5 1I

16.38
25 0.385 105 155.46 50,46 2546.2
30 35 --5 25 0.714 145 195.45 - 50.45 2546.21 13.03
40 35 5 25 2546.21 13.87
0.714 234 183.55 50.45
Total 2.198 Total 58.66
y
Here =2.198 x = 58.66
Hcre
The Ghi-Squarc valuc with a = 0.05 and 1
degrccs of =
The Chi-Squarc value with a 0.05 and
l
degrccs of
frccdom=3.84 frccdom=3.84
Clcarly 2.198 > 3.84 Clearly 58.66 > 3.84
Hencc, null hypothesis is acceptcd. Henco, null hypothcsis is rcjcctcd.
We conclude that there is no association betwecn We conclude that macugcn injcction can improvc thc
family life and cducation. Ans. cye sight. Ans.

Ex.31/Thc following table gives the number of persons

COMPLETE LIST OF FORMULAE:
whose cye sight is attacked and an injection of
macugen is injccted. (1) Mcan u(t) ofa random proccss {X()} is

Eyc sight Eye sight not Total u(t) = E{X(1)}

improved Auto-corrclation R(t1, t5) of random proccss
a
improved (2)
Injected 216 145 361
{(X()} is
Not Injcctcd 105 234 339 =
R(.1, t2) E{(X(1|).X(12)}
Total 321 379 700
(3) Auto-covariance C(tj, t2) of a random proccss
Do you think macugc injcction can improve the
.{X()} is.
cyc sight? = R
C(t,, t2) (t1, 1g)-(t;)(lz)
Soln. Let Ho: Null Hypothesis: There is no improvement in
(4) Correlation cocfficient p(t1, tz) of {X()} is
eye sight due to macugen injection.
an improvement . C(t1, t2)
H,: Alterate Hypothesis: There is p(t1, tz)
Jc(t, t) x C(t2, t2)
eye sight due to macugen injection.
in
Row Total x Column Total (5) Cross-corrclation Rxy(t1, t2) of (x()} and {Y(0} is
Expected Frequency Grand Total Rxy (t1. tz)= E{X(1). Y((2)}
Eye sight Eye sight not Total (6) Cross-covariane Cay(ti, t2) of {X()} and {Y()} is
improvcd improvcd
Injccted 321x361l 379 ×361 361
Cxy(t1: t2)= Ry( tz)-() y(2)
700 (7) Cros-corrclation cocfficient pyy(t), t2) of {X()}
700
= 165.54 = 195.4S and (Y()} is

321× 339 379 x 339 339

No Injccted Cxy ('}, t2)
700 Pxy('1, t2) =
700 JCu (i) x Cy
= 155.46 = 183.55 (l2,t2)
(8)
Total 321 379 700
*= lim [Ry ()]
frequencics= Total number
Total number of observed (9) Var(X())- E[X²(0]-(E[X()}?
frequcncies = 700.
of expected
<br>

(10) 1f 0 is uniformly distributcd in (0,2), then (24) Degrcc of frccdom in F-test df

=
n- 1, n is the

number of samples.
fo(0) = 0<0<2n X
2T (25) Thc t-distribution formula is t= wherc X is

(11) E[g(0)) =
(fo(0) g(0).dO S/n
the sample mcan, u is the population mcan, S is the
= md} = S,e-m'/2n
(12) P{X(nT) S.D. and n is the samplc size.
/nn/2
(13) In a Markov Chain, the mcan return time to (26) Chi-squnre variate, y2² =

state is 1

r =l+)
k
PIk
t is the cxpcctcd time until the
(27) In Chi-square
test, y'=
k O;-E)²
E

chain hits 1.
i=l
(14) IfV = (a, b, c, d) wherc a +
b+c+ d=1is a uniquc where O; is the obscrved frcqucncy and E; is the
fixcd probability vcctor of a stochastic matrix P,
expccted frequency.
then VP = V.
(28) If the data is given in thc form of a scrics of
(15) Standard Error (S.E) = Oz variablcs in a row or column, thcn thc dcgrcc of
=Fn
frecdom = No of itcms in the serics 1.
--

(16) E(x) =4, Var(x) = o, S.D.(x) = o,x is a R.V. (29) If the number of frcqucncies arc put in cclls in a

(17) E(x)=#, Var(x) =

, sD.() = contingency tablc, then the degrcc of frcedom
Vn = (R- 1) (C- 1), wherc R and C arc no of rows and
is samplc mean. columns respcctively.
(30) In Chi-square test as a test of independcncc,
(18) If x ~N(4, o), then ~ N, x
. Row Total ColumnTotal
Expected Frequcncy
n Grand Total
=
(19) Sample variance, S²
n-lj1
(20) If C is the confidence level and a. is the
significance level, then C+a=1.

(21) F-statistic for large samples,

F=:
where and os are the variance of first and
of
sccond population respectively.

(22) F-statistic for small samples, F =

.

where Sf and S are the variance of first and

sccond sample respcctively.
(23)
F-test Significancc Levcl
Right-tailed test
Left-tailed test
Two-tailed test
2
<br>

Table 1: Chi-square Distribution Table

The shaded area is equal to a for X=La 2

Area to the Right of Critical Value

Degrees of
Freedom 0.995 0.99 0.975 0.95 0.90 0.10 0.05 0.025 0.01 0.005

0.001 0.004 0.016 2.706 3.841 5.024 6.635 7.879

2 0.010 0.020 0.051 0.103 0.211 4.605 5.991 7.378 9.210 10.597
7.815 9.348 11.345 12.838
3 0.072 0.115 0.216 0.352 0.584 6.251
9,488 11.143 13.277 14.860
4 0.207 0.297 (0.484 0.711 1.064 7.779
.9.236 11.071 12.833 15.086 16.750
0.412 0.554 0.831 1.145 1.610
10.645 12.592 14.449 16:812 18.548
0.676 0.872 1.237 1,635 2.204
0.989 1:239 1.690 2.167 2,833 12.017 14.067 16.013 18.475 20.278
1.344- 2.733 3.490 13.362 15.507 17.535 20.090 21.955
1.646 2.180
2.700 3.325. 4.168 14.684 16.919 19.023 21.666 23.589
9 1.735 2.088 23.209 25.188
2.558 3.247 3.940 4.865 15.987 18.307 20.483
10 2.156
5.578 17.275 19.675 21.920 24.725 26.757
13 2.603 3.053 3.816 4.575
5.226 6.304 18.549 21.026 23.337 26.217 28.299
12 3.074 3.571 4.404
7.042 19.812 22.362 24.736 27.688 29.819
13 3.565 4.107 5.009 5.892
6.571 7.790 21.064 23.685 26.119 29.141 31.319
14 .4.075 4.660 5.629
8.547 22.307 24.996 27.488 30.578 32.801
4.601 5.229 6.262 7.261
15
9,312 23.542 26.296 28.845 32.000 34.267
16 5.142 5.812 6.908 7962 33.409 35.718
7.564 8:672 10.085 24.769 27.587 30:191
5.697. 6.408 34.805 37.156
17 7.015 8.231 9.390: 10:865 25.989 28.869 31.526
18 6.265 11.651 27.204 30.144 32.852 36.191 38.582
7.633 8.907 10.117
19 6.844 10.851 12.443 28.412 31.410 34.170 37.566 39.997
7.434 8.260 9.591
20 13.240. 29.615 32.671 35.479 38.932 41.401
8.897 10.283 11.591
21 8.034 12.338 14.042 30.813 33.924 36.781 40.289 42.796
8.643 9.542 10.982 44.181
22 13.091 14.848 32.007 35.172 38.076 41.638
10.196 11.689 42.980
23 9.260 13.848 15.659 33.196 36.415 39.364 45.559
10.856 12.401 46.928
24 9.886 14.611 16.473 34.382 37.652 40.646 44.314
10.520 11.524 13.120
25 17.292 35.563 38.885 41.923 45.642 48.290
12.198 13.844 15.379
11.160 16.151 18.114 36.741 40.113 43.194 46.963 49.645
26 11.808 12.879 14.573 41.337 44.461 48.27S 50.993
27 15.308 16.928 18.939. 37.916
„12.461 13.565 19.768 39.087 42,557 45.722 49.588 52.336
28 14.257 16.047 17.708
13.121 18.493 20.599 40.256 43.773 46.979 50.892 53.672
29 14.954 16.791
13.787 SS.758 63.691 66.766
30 26.509 29.051 51.805 59.342
22.164 24.433 76.154 79.490
20.707 34.764 37.689 63.167 67.505 71420
40 29.707 32.357 83.298 88.379 91.952
27.991 43.188 46.459 74.397 79.082
50. 37.485 40.482 100.425 104.215
35.534 51.739 55.329 85.527 90.531 95.023
60 45.442 48.758 101.879 106.629 I12.329 I16.321
43.275 60.391 64.278 96.578
70 53.540 57.153
51.172 113.145 118.136 124.116 128.299
80 69.126 73.291 107.565
61.754 65,647 135.807 140.169
59.196 74.222 77.929 82.358 118.498 124.342 129.561
90 70.065
67.328
100
<br>

numerator right freedom. 188564| 1.65671|

10) 63.32812|
49122 5.13370
3.76073|
10500
2.72216| 2.47079|
2.29257 2.15923
2.05542|
1.97211 1.90361
L84620|
79728 1.75505
1.71817

Fu2.
The 9 3

1.69099|

equal 63.06064|
2.31618| 2.08176| 1.99965 193228
82800|
178672
175075| 171909
3.12279 249279 187591
freedom.
of 948289 5.14251 3.77527 2.74229 2.18427

the degrees 120

1

F-distribution, not 62.79428| 1.95973| 1.81676| 1.75063

2.10716| 190429| 1.78156 1.72322
5.15119 2.33910 2.20849 2.02612 85723
47456 3.14023 276195 2.51422
3.78957

does
of 12 60
9

degrees
175371|

12) and
62.52905|

9.46824 515972
3.80361
3.15732|
2.78117
2.53510|
2,36136|
2.23196
2.13139| 2.05161| 198610 193147
88516 1.84539 81084 178053

Fuo, 40
10 1

the numerator 26497|

2.38302| 201149|
(e.g. with 9.45793|

5.16811
81742 3.17408 2.79996
2.55546|
2.25472
2.15543| 2.07621|
1.95757
1.91193
1.87277
83879 180901 1.78269

referencing
2.7534. 30
1
distribution F-distribution
62 3

62.00205| 1.86556| 181035|

2.03599|
944962 5.17636| 2.81834| 2.17843 2.10001 193272 1.89904 183624
3.83099 2.57533 2.27683 1.93766
3.19052 49410

represent
24
2

When
12) 2.05968| 1.89127|
18448 84434| 2.29832| 2.20074| 1.96245 86236| 1.83685|
61.74029 944131 2.59473 2.42464| 2.12305| 2.00698 1.92431
20665 83634

the 10.
20

Variance. changes
columns an F(o.05,
5 3 3 2 1

for 61.22034
42471 5.20031 3.87036 23801 2.87122
2.63223 2.46422 2.33962 2.24351 2.16709 2.10485 2.05316 2.00953 1.97222 1.93992 1.91169
1.88681

dtz)
value 0.05. 15
F(df4, 9 3
Table
freedom
the 60.70521|

9.40813 5.21562 3.89553 3.26824| 290472|

2.66811 2.50196| 2.37888| 2.32260.2.28405 2.20873| 2.14744| 2.09659| 2.05371 2.01707 1.98539| 1.95772 1.93334

of critical =
F-Distribution
and alpha 12
Analysis

freedom 60.19498
9.39157 3.91988| 3.29740o
2.93693 2.18776 2.0954o

of
5.23041 2.70251 2.53804 2.41632 2.24823 2.13763 2.05932 1.97698
2.02815 2.00094

degrees
0.05 for
F-table

in 59.85759
9.38054 5.24000 3.93567
331628 2.95774| 2.72468
56124 2.44034
2.34731|
2.08621|

of the
2.27350 221352 2.12195 2.05533]
16382 2.02839
2.00467
commonly

determine 589352
degrees

2: of the 58.90595'59,43898|

95494|
2

Table order =0.10

936677 25167 33928 98304
|
2.75158
246941
2.37715|
2.30400 2.24457
2.19535|
2.15390
2.08798| 2.03789|
2.11853 2.08134

of 8
denominator (denominator) 3
5 3 2
2
most
the
to
284735.26619|
9.34908 3.97897| 3.36790 2.62413| 2.34157|
3.01446 2.78493 2.23410|
2.50531 2.41397
2.28278 19313 2.15818
2.12800
switching 10169 2.07854
example, a
used
for 58.20442
2
2
32553 4.00975 3.40451 3.05455| 2.82739 2.66833 2.55086| 2.46058|
238907 2.33102 2.24256
2.28298 2.20808 2.12958|
distribution F-Table 2.17833
15239

reprsent 6
9 5

For row
as 57.24008 9.29263| 5.30916|
1.05058
3.45298|
2

first, 3.10751
2.88334 2.72645
2.61061 2.52164 2.45118
2.39402| 2.24376|
table. 2.30694|

12 5
34672 2.27302 2.21825|
2.19583

given rows
and 2
right-skewed
55.83296 9.24342
5.34264 4.10725 3.18076|
3.52020 2.96053

the (numerator)
2.80643
2.69268 2.60534

253619
2.48010
2.43371
239469|
2.36143 233274|
230775 2.28577

always the of
53.59324|
name
9.16179
below, 5.39077
4.19086 3.61948| 3.28876|
3.07407
2.92380 2.81286
2.72767
2.66023 2.52222|
260552 2.56027 2.48979 2.43743
2.46181 2.41601
3
a are
the column

is Freedom
F-tables 49.50000
9.00000 5.46238 4.32456 3.77972 346330
3.25744 311312

F-distribution
in 2
3.00645
2.92447 2.85951 2.80680
2.76317 2.72647 2.69517|
2.66817|

2.64464
2.62395

given

---
.10 39.86346
8.52632 5.53832
4.06042|
4.54477
four 3.77595 3.58943 3.45792 36030|
3.28502|
df;=1 3.22520 3.17655|
3.07319 3.00698|
3.13621

of is the
3.10221 3.04811 3.02623

degrees area 3
the
in
look
dí;=1

The
For tail 4
9 10 11 12 13 14 5
16 17 18
<br>

1.29146| 1.19256| 1.00000

1.63077| 1.58615 1.54903| 1.49057| 1.45636|
1.56678 L53270 1.37691
1.60738 1.47841 1.46704
1.51760 1.50360

1.16860|
1.64326| 1.62278| 1.42476 1.34757 1.26457
1.66587| 1.60415| 1.57146| 1.54368| 1.53129
1.58711 1.55703 1.51976 1.50899 1.49891

1.23995
1.58050| 1.39520 1.32034
1.69876 1.67678 162237 1.60728 59335| 1.56859 1.55753 1.54721 1.53757 1.46716
1.65691 1.63885

1.43734|
1.67138 1.64067 1.57323 1.50562 1.36760 1.29513
1.72979| 1.65535 1.62718 1.61472 1.60320 1.59250| 1.58253
1.70833 1.58896

1.64682|
1.62519 40932 1.34187
1.67210|
1.60648 54108 1.47554
1.75924| 1.73822 1.70208 1.68643 1.65895 1.63560 151551
1.71927

1.38318|
1.65600! 1.63774| 1.57411 1.51072 1.44723
1.68898 1.64655|
1.76667 1.74807 1.75899|.1.73122
1.71588 1.66616
1.70185 1.67712
1.78731

1.54349 1.42060
1.73015| 1.70589| 1.48207
1.74392| 1.69514 1.68519 1.67593 1.63731 1.60515
1.81416| 1.77555 1.71752
1.79384

48714|
1.60337| 54500|
1.79643| 1.77083| 1.66241
1.86471| 1.82715| 1.74917 1.73954 1.73060 1.72227
1.84494| 1.81106| 1.78308 175957

1.54578|
1.77270| 1.71456| 1.60120|
1.65743
1.79889| 1.78081
1.80902
1.87497 1.85925 1.84497 1.83194
1.82000 .78951
1.91170 1.89236

1.59872|
1.70701 1.65238|
1.83593| 1.82741 1.81949 1.76269
1.93674: 1.91967| 1.89025| 1.87748| 1.85503
1.84511
1.95573 1.90425 1.86578

1.79290| 1.73802| 1.68425

.86520| 1.84896| 1.63152
1.88407 1.85679
1.93273| 1.91888 1.90625
1.89469 1.87427
1.98364| 1.94797
I.96485

1.72196|
1.90909| 1.89184| 182886 1.77483 1.67020
1.91876| 1.88412
1.90014
.95312| 1.94066 1.92925
1.99853 1.98186 1.96680
201710

1.81939| 1.76748|
1.926892 1.71672
1.93452 1.87252
2.00840| 1.99492| 1.97138 1.94270
1.98263 1.96104 1.95151
2.03970| 2.02325|
2.05802

1.98033,

1.87472| 1.77411|
2.00452| 1.92688
| 2.01389| 1.99585 1.98781 .82381
2.04723| 2.03513 2.02406
2.09132 2.06050
2.10936 2.07512

.
1.99682| 1.84727
2.04925| 1.94571 1.89587
2.08218 2.07298 2.06447 2.05658
2.15823| 2.10303 2.09216
2.17596| 2.12794 2.11491
2.14231

1.94486|
1.99230|
2.14223 2.09095 2.04099
2.19488| 2.17447 2.16546 2.15714 2.14941
2.23334| 2.18424
2.26630 2.20651
2.24893 2.21927

2.27607| 2.12999| 2.08380|

2.29871| 2.29060| 2.22609
2.32739| 2.30749 2.28307 2.17741
2.36489| 2.33873 2.31702
2.39702| 2.38009| 2.35117

2.39325| 2.34734|
2.30259|
2.51910| 2.44037
2.54929| 2.50276 2.49548 2.48872
56131| 2.53833 2.51061
2.57457| 2.52831
2.58925
2.60561

2.89385| 2.83535|
2.88069| 2.79107 2.70554
2.90913| 2.90119 2.74781
2.88703
2.93736
292712 2.91774
2.98990| 2.96096 2.94858
2.97465

120
25 26 27 28 29 30 40 60
19 20. 21 22 23 24
<br>

8.5264| 2298|
3144 19.4957 5.6281 4.3650 3.6689
2.9276 2.7067 2.5379| 2.2064|
2.0096|
2.4045 2.2962 2.1307 0658 9604

254 3
2 1

253.2529 4874| 5494|

120 5.6581
4.3985 3.7047 32674|
2.9669 2.7475 2.5801| 2.3410|
24480 2.2524 2.1778 2.1141 2.0589 2.0107

8
19
252.1957 8.5720 4.4314
S0 19.4791 5.6877 3.7398 3.3043 3.0053
2.7872
2.6211
24901 23842
2965| 2.2229|
2.1601
2.1058)
2.0584

8:5944| 3.7743|
40 251.1432 19.4707 5.7170 4.4638 3.3404|
3.0428 2.8259 2.6609
2.5309
2.4259| 2.3392) 2.2664| 2.2043 2.1507
2.1040|

30 250.0951 19.4624 86166|

5.7459 4.4957 3.8082 3.3758 3.0794
2.8637
26996 2.5705
2.4663
2.3803
2.3082
2.2468
2.1938|
2.1477

249.0518|
5.7744|
8.6385
24 19.4541
4.5272
3.8415 3.4105 3.1152 2.9005| 2.7372 2.6090
2.5055| 4202
2.3487|
2.2878 22354| 1898|

2 2
248.0131|

201 19.4458
8.6602 3.1503| 2.5436|
8025 4.5581 3.8742 2.9365| 2.7740| 24589|
2.3379|
2.3275
3.4445 22756| 2.2304
26464

5 245.9499
19.4291 87029| 5.8578 4.6188 3.5107 2.3522| 2.3077|
3.9381 3.2184 3.0061 2.8450 2.7186| 2.6169 2.5331 2.4034
2.4630
dtz)
F(df4,

243.9060l

12 19.4125 8.7446 5.9117| 4.6777 3.9999 3.5747 3.2839| 3.0729|

2.9130 2.7876 2.6037 2.5342 2.4753 2.3807
2.6866 2.4247
-

19.3959
10 241.8817
8.7855 5.9644
4.7351
4.0600| 3.6365| 3.3472|
3.1373 2.9782 2.8536| 2.7534|
2.5710 2.6022 2.5437 24935 2.4499

240.5433|
19.3848|

9 8.8123 5.9988 4.7725|

4.0990 3.1789|
3.6767 3.3881 3.0204 2.8962 2.7964 2.7144| 2.6458 2.5876| 2.5377| 2.4943|

0.05
238.8827 8.8452|
8 19.3710 6.0410
4.8183 4.1468 3.7257 3.2296|
3.4381 3.0717 2.9480 2.8486 2.7669 2.6987 2.6408 2.5480|
2.5911
=
a
for 236.7684 19.3532
8.8867 6.0942 4.8759 4.2067 3.7870|
3.5005 3.2927 3.1355|
2,9134|
3.0123 2.8321 2.6143|
2,7642 2.7066 2.6572

Table

233.9860 6.1631|
19,3295 8.9406| 4.9503| .3.866O
4.2839 3.5806 3.3738
3.2172 3.0946| 2.9153| 2.7905|
2.9961 2.8477|
2.7413
2.6987

5 230.1619
19.2964 9.0135 3.4817|
5.0503 4.3874 3.3258|
6.2561 3.9715 3.6875 3.0254|
3.2039 3.1059
2.9582 2.9013 2.8524|
2.810

224,5832 19.2468|
9.1172 6.3882| 5.1922
4.5337 4.1203|
3.8379 3.6331 4780 3.3567|
3.2592
3.1791 3.1122 3.0556|
3.0089 2.9647

215.7073
9.2766| 6.5914|
3 19.1643
5.4095
4.7571 4.3468 3.8625|
4.0662 3.4903|
3.7083
3.5874 3.4105 3.3439 3.2874 3.2389|
968|

3.1
199.5000
19.0000
9.5521 6.9443
5.7861 5.1433 4.7374 4.4590 2565 4.1028
3.9823 8853|
8056 3.7389 3.6823
3.6337 3.5915

3 3

161.4476 18.5128|
df;=1 1280 7.7085
6.6079 5.9874|
5.5914| 5.3177
5.1174 4.9646 4.84431
47472| 4.5431|
4.6672
4.6001 ,4.4940 4.4513

10

dfz=1 .
2 3 4 5 6 8 10 11 12 13 14 15 16 17
<br>

1.2539|
1.8432| 1.8117| 1.7831} 13893|
1.8780 17570| 1.7330| 17110} 1.8376| 1.5089 1.0000
9163 1.6906 1.6717 1.6541 1.622

1.9302|
1.9681 1.8963 18657 1.8380 1.8128 1.7896| 1.7684| 1.7138| 1.6835| 1.4673| 1.3519| 1.2214
1.7488 1.7306| 1.5766
16981

20166| 1.8424|
1.9795 1.9165 1.8894 1.8648| 1.3180|
1.9484 1.8217 1.7537 1.7396 1.6373 1.5343 1.4290
1.8027 1.7851 1.7689

2.0629 1.9938 1.9645 1.9380 1.8718]| 1.7913|

2.0264 1.9139 1.8920 1.8533| 1.8055| 1.4952| 1.3940|
1.8361 1.8203 1.6923 1.5943

2.1071 2.0712 1.7444|

2.0391 20102 9842 9605 1.9192 1.9010
|9390 18687 8543 1.8409 16491 1.5543 1.4591
8842

1 1

2.1497| 2.1141|
2.0825 2.0540 20283| 2.0050| 1.9638 1.9643 1.9147 1.6084|
1.9464 1.9299 9005 1.8874 1.7929 1:7001 1.5173

2.1906| 2.1555 2.1242| 2.0476| 2.0075]

2.0960 2.0707 2.0267 1.9898 .1.9736 1.9586)
19446 1.9317 i.8389 1.7480 1.6587 1.5705

2.2686 2.1757| 2.1508]

2.2033 2.1077| 20716|
2.2341 2.1282 2.0889 2.0275 2.0148 1.9245
2.0558 2.0411 1.8364 1.7505
6664

2.3421 2.2776 2.2504| 2.2036| 2.1479|

2.3080 2.2258 2.1834| 2.1649 2.1179|
2.1323 2.1045 2.0035 1.9174
2.0921 1.8337 1.7522

2.4117| 2.3779| 2.3479|

2.3210| 2.2547|
2.2967 2.2747 2.2365| 2.2197 2.0772|
.2.2043 2.1900 2.1768| 2.1646|
1.9926 1.9105 1.8307

2.4563 2.3419|
2.4227 2.3928 2.3201 2.3002 2.2655| 22107|
2.3660 2.2821 2.2229 2.1240|
2.2501 2.2360 1,9588 1.8799
2.0401

2.5102 2.4768 2.4471 2.4205 2.3965 2.3748 2.3371 2.3205| 1.9384||

2.3551 2.3053 2.2913 2783 2.2662 2.0970| 2.0164|
2.1802

2.5767|
2.5435| 2.5140 2.4876|
2.4638 2.4422 2.3883| 2.3593|
2.4220 2.4047 2.3732 3463| 2.3343 2.2490 2.1665|
2.0868 2.0096|

2.6613 2.5277|
2.5990 2.5727 2.5491 2.5082 2.4904 2.4324|
2.6283 2.4741 2.4591 2.3359| 2.0986|
2.4453 2.4205 2.1750|
2.2541

2.7729| 2.7109|
2.7401 2.5719|
2.6848 2.6613 2.6400 2.6030
2.6207 5868 25454| 2.5336|
5581 2.3683| 2.2141|
24495
2899

2 2
2

29277 2.7763| 2.7278|

2.6951 2.8661 2.6401 28167 2.7955 2.7426|
2.7587
2.7141 2.7014 2.6896 2.4472|
2.6060 2.5252 2.3719

3.1599| 3.0984|
3.1274 3.0725 3.0280
3.0491 2.9912 2.9604|
3.0088 2.9752
29467 9340 2.9223 2.7581| 2.6802| 2.6049|
2.8387

3.5546 3.5219 3.4221|

4928 3.4668 3.4434 3.4028 33852|
3.3590 3.3404| 3.3277| 3.3158 3.1504!
3.3541 3.2317!
3.0718 2.9957

4,4139 1709|.
4.3807 4.3512 4.3248|
43009 4.2793 4.2597 42417
4.2252 4.2100 4.1960 4.0847|
4.1830 4.0012|

39201 3.8415

18 19 20 21 22 23 24 25 26 27 28 29 30 120
40 60
<br>

1018.258 39498 902 257| 6.015 4.849|

4.142 3570|
3080 883| 2.316|
333 2.725 2.595 2487 2.395 2.247

13 8

3 2

120 020 39.490 13.947 8309 6069| 4.199|

3.392| 3.140| 2.787| 2.659|
4.904 3.728 2.383|
2944 2.552 2.461 2.315

1014

800 481 992 360 123 4.959| 4.254| 3784| 449| 3.193) 2614|
2.848
004 2.720 2.524 2.447 2.380

1009
39 13 a 6
3
3

1005.593 39.473
40 037 8411 6.175|
5.012 4.309| 840| 3.505| 3.255|
3.061
2.906|
2.780 2.674| 2.509|
2442
2.585

14 3

414| 14.081|
455 6.227|
30 8461
065
4.362|
894 3.560 3.311 118 2.963 2.837 2.732| 2644 2.588 2502|

1001 39 5. 3
3

39.4562|
997.2492
24 14.1241 8.5109 2780 5.1172|
44150 36142 3.1725| 3.0187| 8932| 70C6|
9472 3.3654 2.7838 26252 5598

6
3
2 2 2

993.1028|

39.4479 14.1674
20 8.5599 63286|
5.1684 4.4667 39995
3.6669
3.4185| 2261| 3.0728| 2.7559| 2.6803| 2.6158|
2.9477 8437

3 2

15 984.8668
39.4313 14.2527| 86565|
6.4277
5.2687
4.5678| 4.1012|
3.7694 3.5217 3.3299 3.1772 3.0527 2.9493 2.7230
2.8621 2.7875
dfz)
F(df4.

7079|
12 39.4146 143366
8.7512
6.5245| 5.3662| 4.6658|
4.1997 3.8682 36209 3.4295| 3.2773 3.1532 3.0502 2.9533 2.8890 2.8249

976

968.6274| 39.3980| 144189|

8439| 4.7611| 3.9639|
6.6192 5.4613 4.2951 3.7168| 3.3736|
3.2497 3.1469| 3.0602|
5257 2.9862|
2.9222

8
3

963.2846|
39.3869| 6.6811|
14.4731
8.9047 4.8232 4.3572| 3.5879|
55234 4.0260 3.7790 3.3120 3.0488|
3.4358 3.2093 3,1227 2.9849|
0.025

956.6562 39.3730 14.5399

8.9796 5.5996| 4.8993| 4.4333 4.1020| 3.8549|
6.7572
3.6638| 3.5118
= 3.3880 3.2853 3.1987 3.1248|
3.0610

for 948.21691
39.331539.3552
14.6244
9.0741 6.8531
5.6955| 4.9949|
4.5286 4.1970| 3.9498| 3.7566| 3.6065| 3.4827|
3.3799 32194|
3:2934 3.1556|

Table

937.1111|
14.7347|
6.9777|
9.1973| 4.6517|
5.8198
4.3197 3.8807
5.1186 4.0721 3.7283 3.6043| 3.5014| 3.4147|
3.3406|
3.2767
F

921.8479|
14.8848|
39.2982 7.1464|
9.3645| 5.9876|
5.2852 4.8173 4.4844 4.0440|
4.2361
3.8911 3.6634|
3.7667 3.5764 3.5021|
3.4379

899.5833 39.2484 15.1010 9.6045|

7.3879 6.2272 5.5226| 5.0526|
4.7181 4.4683 4.2751|
4.1212 3.8919|
3.9959
3.8043
3.6648|
3.7294

864.1630 39.1655| 9.9792|

15.4392
7.7636 6.5988| 5.8898
5.4160 4.8256 4.6300|
5.0781 4.4742| 4.2417|
4.1528|
3472
4.0768 4.0112

799.5000 39.0000 16.0441|

10.6491
8.4335 7.2599 6.5415 6.0595 5.4564|
5.7147
2559 50959|
4.7650|
4.9653 4.8567
4.6867
4.6189

df;=1 647.7890 38.5063|

12.2179
17.4434 10.0070
8.8131 8.0727 7.5709 7.2093|
6.9367 6.7241| 6.5538| 6.4143| 6.2979|

6.1995
6.1151 6.0420

df2=1

2 3 4 5 7 8 9 10 11 12 13 14 5

15 17
<br>

107 2.133 2085|

2.042
003 1.960 1.935 1.906 1.787|
1878 1.853 807 1.000
829 1637 1,482 1310
2
2
1 1

2.203)
260 2156| 2114| 070 2010)
041 1981 1.954 930]
907 886 866 1.724 1.581 1433 268
2 2 2
1 1 1 1 1

2.270|
2.223 2.182|
2321 145| 2111 2080 052 1.530|
2.020 002 980 959 1.940 667 1,388|
1803
2
2 2 1 1 1

2.304| 2.333 2.287 2.176)

246 2,210 2.140|
118| 2.069|
2.093 049 2.028|
2.009
1.875|
1744 1.614
484
2
2 2
1

445| 394| 2340 2.308| 2.272|

2.239 2.209|
2.182 2.157 2133) 2112| 2074| 815|
2092
943 690
2 2

1 1
1

25027 24523| 4070 2.3075| 2.3315|

2.1940)|
2.2080 2003 2422 2.2174 1735|
2.1540 21350 2.0009 8817 17597 1.6402

2
2 2 2
1

2,5009| 4045| 2,4247|

5590 3890| 23507|
2.3273 3005|
2.2759 2.2533 2.0677 17085|
2.2324 2.2131 2.1952 1.9445 1.8249

2 2 2
2

2.5338|
0607 2.6171 2,5731 2.4984|
2.4005 2.4374| 2.3248|
2.-4110 3867 23044 2.3430| 2.3072| 2.1819| 1.9450| 1.8320)
0613

2.
2
2

2.7000 2.7196 2.0750 2,0308 2.0017 .2.4908| 24120|

2.5099 2.5411 2.5140 4686| 4484| 4295| 22882| 2.1692
1.9447|
0548

2 2 2 2

2.8172|
2.7737 2.7348| 2.5806|
2.8604 2,0098 2.6682 2.0390 2.0135 2.5280| 2.5112 2.3882| 2.1570|
2.5670 25473 2.2702
2.0483

2.7077| 2.7020 2.0528|

2.8801 2.8365 2.7313 2.7027 2.0309| 2.4519)
2.9201 2.6706 26106 2.5919
2.5740 2.3344 2.2217
2.1136

3.0053 2.9128| 2.8740| 2.8392 2.6872|

2.8077 2.7701 2.7531 2.7293 2.7074 2.6686 2.6513
2.9503 2.5289 24117 22094 2.1918

3.0999 2.0680 2.8478 2.7820| 2.3948|

3.0509 3.0074
2.9338 2.9023| 2.8730 2.8240 27460
2.8021 2.7633 2.6238 2.5068 2.2875

3.1718 3.1283 3.0895 3.0546 2.9685| 2.9228 2.9027 2.8840

32209 3.0232 2.9946 2.9447 8667 2.7444 2.6274 2.5154 24082

30626|
3820| 3.3327
3.2151
3.1835 3.1548 3.1287 1048 3.0828 0438| 3.0265
3.2891 32501 2.9037 2.7863 2.6740 25665

3 3 3

35147| 3.4401|
2863| 3.2499| 3.1261|
3.6083 4754| 3.3794 33530| 3.3289 3.3067
3.5587 .3.4083 2874 7852|
0077 28943

3 3 3 3
2

36472| 3.5894|
3.9539 39034 8587 3.8188 3.7829 3.7505 3.7211 3.6943 3.6697 3,6264 4633 33425| 2289|
36072 3.1161

3 3 3

4.4199
A.4613 43828 4.3492 4.3187 4.2909 4.2655 4.2421 3.8046|
4.5597 4.5075 4.2205 2006 1821 0510 3.9253
3.6889

4 4 4

5.9216| 58715| 5.8266| 5.7498 5.5878|

5.9781 7863 5.7166 5.6864 5.6586 5.6331 5.6096 5.5675 5.2856) 1523| 5.0239|
5.4239

5
5

18
eaza*a
19 20 21 22 23 24 25 26 27 28 29 30 40 60 120
<br>

653|
5.650| 3.909 3.361|
165 3.004 863
2753 568
6365.864 99.499 125! 13.463 9.020 6.880 359 4311 3.602

2
4
3
2 2
26
845| 746| 2660
4.946 398| 3.996| 3.690| 449 255 094 2.959
6339.391 99.491 13.558| 9.112 6.969 5.737
25.221
120
2 2
3 3
4 3

2.933| 2.835|
4.082| 3.776| 535/ 2749
6313.030| 5.824| 5.032 3.341 3.181 3047
13.652| 9.202
99.482 26.316 7.057 483

4 3

3.425| 3.266| 3.013 2.920 2.835

6286.782| 7.143| 5.116| 4.165| 3.850| 3.619| 3.132
5.908|
99.474 26.411 13.745 9.291 567
40
4

2.919|
3.214|
6260.649|
4247| 941 3.701 507 3348 3.101 003
25.505| 13.838
9379}
7229 5.992 5.198 649
30 99.466
3
3 3
4

3.181| 3.084 2.999

6234.631| 3.780| 3.587| 3294
99.458|
26.598
13.929|
9.458 7313
6.074 5.279 4.729 4.327 4.021 427
24 3

3.259| 3.162| 3.077|

3.372
6208.730|
6.155| 5359| 4.808| 4.099 3.858 3.665 505
99.449
26.690
14.020|
9.553 7.396 405
20 3
4

3.522| 3.409 3.312 3.227

4.010| 3.815|
6157.285| 6.314| 4.558 3656
99.433 26.872 14.198 9.722| 7.559| 5.515 4.962 251
15
4

3.553| 3.455| 3.371|

dfz) 5.111| 4.706|
3.960| 3.666|
6106.321| 14.374| 6.469| 5.667 4.397 4.155 600
•99.416 27.052 9.888 7.718
\F(dty. 12 3

3.508|
3.939| 3.593
4.296| 4.100| 3.805 3.691
6055.847| 5.257| 4.849| 4.539
27.229| 14.546| 10.051 7.874 6.620 5.814
99.399
10

3.780| 3.682|
4.388 3.895 3.597
5.911| 4.030
6022.473| 7.976| 6.719| 5.351| 4.942| 4.632 4.191
99.388| 27.345| 14.659| 10.158

0.01

4.140| 4.004 3.890| 3.705

4.744| 4.499 4.302 3.791
6.840 6.029| 5.467| 5.057
10.289|
5981.070| 14.799 8.102
= 8
99,374| 27.489

a
for 5928.356| 8.260| 6.993| 6.178 5.613 5.200 1.886 4.640
4.441|
4.278 4.142 4.026| 3.927
3.841|

99.356) 27.672 14.976| 10.456

7
Table

4.456| 4.202 4.102| 4015

5.802| 5.386| 5.069 4.821 4.620 4.318
5858.986| 10.872| 8,466 7.191 6.371
99.333| -15.207
27.911
6
F

4.336| „4.248|
6.057| 5.636| 5.316| 5.064 4.695 4.556 4.437
10.967| 8.746 7:460 6.632 4.862
5763.650 28.237| 15.522|
99.299
5

4.579|
9.148| 5.412| 4.893| 4.773 4.669
11.392| 7.847 7.006| 6.422 5.668 5.205 5.035
5624.583| 28.710| 5.994
15.977
99.249
4

5.953| 5.739| 5.292| 5,185 5.092

12.060 9.780 .7.591| 6.992| 6.552 6.217| 5.417
5403.352 99.166| 16.694 8.451 5.564
29.457
3

6.013|
4999.500| 6.226| 6.112
7.559| 1.206| 6.701 6-515| 6.359|
99.000
30.817 18.000 13.274 10.925 9.547
8.649 8.022 927

'6
8.683| 8.400|
4052.181| 9.074| 8.531| 8.285
98.503| 34.116 12.246 11.259 9.330| 8.862|
21.198 16.258 13.745 10.044 9.646
10.561
df1=1

df2=1
6.
11 12 13 14 15 16 17 8
2 3 4 5
<br>

2.489| 2.360 2.211| 2064| 2.006| 1.381 1.000

2421 2.305 2256 2169| 2.097 2034
1.805 601
2.131

1.726| 1.533| 1.325

2.517| 2.457 2403 2.310| 2.270| 2.198 2.138 2.111| 1.917|
2.584 2.354 2233 2.167

2.548| 2.403| 2.234|

2.674 2495 2.364| 2.294| 2.263| 2.208| 2.019 1.836 1.656 .1.473
2.608 2.447
2327

2.636| 2.583| 2.114| 1.936|

2.761 2.695 2.535| 2.492 2.453 2.299 1.763 1592
2.417 2384 2.354 2.325

1.696|
2.844 2.778| AAZ 2.620| 2.538|
2.503| 2.470| 2.440| 2.028|
2.577 2.412 2.386| 2.203 860
2.720

2.859| 2.801|
925| 2.702| 2.659| 2.620|
585| 2.552
2.522| 2495| 2.469 2.288 115| 950| 1.791

2 2 2 1

2.035|
3.003| 2.938 2.880| 2.781 2.738 2.699|
632 2.602 2.574 549| 369| 2198 1.878
664

2 2 2 2

3.030| 2.978 2.850 2.039|

3.088 2.931 2.889 2.815 2.783 2.753 2.726 2:700 2.522 2.192
3.153 2,352

3.074 3.032| 2.993| 2.926| 2.896| 2.496| 2336|

3.297 3.231 3.173 3.121 2.958 2.868
843
2.665 2.185

2.979| 2.632|
3.258| 3.211| 3.168| 3.032 2.801
3.434| 3.368 210 3.129 3.094 3.062 3.005 2.472 2.321

3,346| 3.299 3.256| 3.217| 3.120| 2.888 2.718 2.407

3.523| 3.457| 3.182 3.149 3.092 067 2.559

3.363| 3.256| 2993|

3.564| 3.406| 3.324| 3.288 3.226| 3.198 3.173 2.823 2.663
3.453 2.511
3.631

3.496| 3.457| 3.124|

3.388| 3.330| 3.304| 2.953| 2.792 2.639|
3.765| 3.587 3.539 3.421 3.358
3.699

3.558 3.528| 3.119|

3.939| 3.812 3.758 3.710 3.667 3.627 3.591 3.499 3.473 3.291 2.956 2.802
3.871

3.895| 3.818| 3.754| 3.725| 3.699| 3.339| 3.174|

4.103| 3.988 3.855| 3.785 3.514 3.017|
4.171 4.042 3.939

4.313| 4.264| 4.045| 4.018| 3.649|

4.431| 4.369| 4.218 4.177 4.140 4.106 4.074 3.828 3.480 3.319
4.500

4.538| 4.126|
5.010| 4.675 4.510 3.949|
4.938 4.874 4.617 4.765 4.718 4.637 4.601 4.568 313 3.782

5.526| 5.179| 4.787|

5.926| 5.719 5.614 5.568 5.488 5.453 5.390| 4:977 4.605
5.849 5.780 5.664 5.420

8.096 8.017| 7.945| 7.770| 7.721| 7.598| 7.562| 7.314|

7.823 7.536 7.077 6.851 6.635|
8.185 7.881 7677

19 20 21 22 23 24 25 26 27 28 29 30 40 60
120