By- Prof.

Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
M easur es of Cent r al Tendency
(Types of Aver age)






M at hemat i cal
Aver age
Ar i t hemat i c
M ean
Si mpl e
Wei ght ed
Geomet r i c
M ean
Har moni c
M ean
Posi t i onal
Aver age
M edi an
M ode
Quar t i l es,
Deci l es,
Oct i l es
Per cent i l es
Commer ci al
Aver age
M ovi ng
Aver age
Pr ogr essi ve
Aver age
Composi t e
Aver age

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710

1. Arit hmet ic M ean ( x )
Def i ni t i on: The ar i t hmet i c mean may be def i ned as t he sum or aggr egat e
of a ser i es of i t ems di vi ded by t hei r number
- W .I . King
a. Simple Arit hmet ic M ean ( Individual Series)
i . Direct M et hod- x = X
N

x = x
1
+ x
2
+ x
3
++ x
n

N
i i . Short -Cut M et hod
X = A + dx
N
dx = x A
A = Assumed M ean
N No. of obser vat i ons
e.g-







X = A + dx
N
= 5 + 3
6
x dx (x-A)
3 -2
4 -1
5 0
6 1
7 2
8 3
dx= 3

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
= 5 + 0.5
= 5.5
b. Simple Arit hmet ic M ean (Discret e Series)
i . Direct M et hod- x = fx
N

i i . Short -Cut M et hod
X = A + fdx
f
dx = x A
A = Assumed M ean
f sum of f r equenci es
e.g (Shor t -Cut M et hod)
x f dx (x-A) f dx
3 2 -2 -4
4 7 -1 -7
5 13 0 0
6 5 1 5
7 3 2 6
f = 30 dx = 0

X = A + fdx
f
X = 5 + 0
30
X = 5





By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710

c. Simple Arit hmet ic M ean (Cont inuous Series)

i . Direct M et hod-
x = fx
f

i i . Short -Cut M et hod
X = A + fdx
f
dx = x A
A = Assumed M ean
f sum of f r equenci es

i i i . St ep Devia t ion M et hod
X = A + fdx X i
f
dx = x A
dx = dx/ i or (x A)/ i
A = Assumed M ean
N Sum of f r equenci es
i = l
2
-l
1












By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
e.g. (Di r ect M et hod)
C.I. f x f x
0-10 2 5 10
10-20 7 15 105
20-30 13 25 325
30-40 5 35 175
40-50 3 45 135
f = 30 fx = 750

X = fx
f
X = 750
30
X = 25
e.g. (Shor t -Cut M et ho d)
C.I. f x dx (x-A) f dx
0-10 2 5 -20 -40
10-20 7 15 -10 -70
20-30 13 A = 25 0 0
30-40 5 35 10 50
40-50 3 45 20 60
50-60 10 55 30 300
f = 40 dx = 300

X = A + fdx
f
X = 25 + 300
40
X = 25 + 7.5

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
X = 32.5
e.g. (St ep Devi at i o n M et ho d)
C.I. f x dx (x-A) dx' f dx'
0-10 2 5 -20 -2 -4
10-20 7 15 -10 -1 -7
20-30 13 A = 25 0 0 0
30-40 5 35 10 1 5
40-50 3 45 20 2 6
50-60 10 55 30 3 30
f = 40 dx = 30

X = A + fdx X i
f
X = 25 + 30 X 10
40
X = 25 + 0.75 X 10
X = 32.5
Not e: A = 25
dx = dx/ i
i = 10 (30-20)






By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
2. M edian (M ) - M edi an i s descr i bed as t he numer i cal val ue separ at i ng t he
hi gher hal f of a sampl e, a po pul at i on, or a pr obabi l i t y di st r i but i on, f r om t he
l ower hal f .
Formula-
a) Individual Series-
a. When t he number of obser vat i o ns ar e o dd
M = (N+1)
t h
t erm
2
b. When t he number of obser vat i ons ar e even
M = Average of ( N)
t h
t erm + ( N + 1)
t h
t erm
2 2
b) Discret e Series-
M = (N+1)
t h
t erm or it em
2
x f cf
5 3 3
6 4 7
8 6 13
9 4 17
10 3 20
15 5 25
f = 25

M = (25 + 1)
t h
t er m
2
= 13
t h
t er m
= 8 (because val ue of 13
t h
t er m i s 8)

c) Cont inuous Series

M = l
1
+ (N/ 2) pcf X i
f

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710

W here,
l
1
- Lower l i mi t of t he M edi an Cl ass
pcf- Pr evi o us Cumul at i ve Fr equency of M edi an Cl ass
i = l
2
l
1

f = Fr equency of M edi an Cl ass
M edi an Cl ass wher e t he (N/ 2)
t h
t er m l i es






M = l
1
+ (N/ 2) pcf X i
f
N = 25
N/ 2 = 25/ 2
= 12.5
It l i es i n t he 20-30 cl ass i nt er val s; t her ef or e t he M edi an Cl ass i s 20-30

M = 20

+ {12.5 7} X 10
6
M = 20

+ 0.917X 10

M = 20

+ 9.17
M = 29. 17

3. Harmonic M ean The mean of n number s expr essed as t he r eci pr ocal of
t he ar i t hmet i c mean of t he r eci pr ocal s of t he number s.
Har moni c mean i s used t o cal cul at e t he aver age of a set of number s. Her e t he
num ber of el ement s w i l l be aver aged and di vi ded by t he sum of t he
r eci pr ocal s of t he el ement s. The Har moni c mean i s al w ays t he l ow est mean.
x f cf
0-10 3 3
10-20 4 7
20- 30 6 13
30-40 4 17
40-50 3 20
50-60 5 25
f = 25

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710

Individual Series:

Discret e Series:
H. M = f = N
(f1/X) (f1/x)

x f 1/ x f * 1/ x

Cont inuous Series:
H. M = f = N
(f1/X) (f1/x)

C.I. x f 1/ x f * 1/ x
0-10 5 3 0.2000 0.6000
10-20 15 7 0.0667 0.4667
20-30 25 12 0.0400 0.4800
30-40 35 5 0.0285 0.1428
40-50 45 3 0.0222 0.0667
f = 30 f*1/x=1.7562


H. M = f
(f1/X)
= 30
1.1762

= 25. 50



By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
4. Geomet ric M ean Geo met r i c mean i s t he n
t h
r oot of t he pr oduct of N
i t ems or val ues.
Calculat ion of Geomet ric M ean (G):

Individual Series Discret e & Cont inuous Series
G = Ant i l og log X
N

G = Ant i l og f log X
N
5. M ode (Z) M ode i s t he val ue t hat appear s most f r equent l y i n a ser i es i .e. i t
i s t he val ue of t he i t em ar o und w hi ch f r equenci es ar e most densel y
concent r at ed.
Calculat ion of M ode:
a) Indi vi dual Ser i es:
(i ) By i nspect i o n val ue r epeat ed most .
(i i ) By conver t i ng i ndi vi dual ser i es i nt o di scr et e ser i es.
(i i i ) By empi r i cal r el at i o nshi p bet ween t he aver ages -
Z = 3M 2X

b) Di scr et e Ser i es:
(i ) By i nspect i o n val ue havi ng hi ghest f r equency.
(i i ) By gr o upi ng.
(i i i ) By empi r i cal r el at i o nshi p.

c) Cont i nuo us Ser i es:
(i ) Fi r st cal cul at e model cl ass by i nspect i on or by gr o upi ng.
(i i ) Then appl y t he f ol l o wi ng f or mul a
Z = l
1
+ f
1
-f
0
X i
2f
1
- f
0
-f
2





By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
e.g.





Z = 20 + 12 7 X 10
(2x12 7 5)
Z = 20 + 5 X 10
12
Z = 20 + 4.17
Z = 24.17
6. W eight ed Arit hmet ic Average-
Ar i t hmet i c mean comput ed by consi der i ng r el at i ve i mpor t ance of each
i t ems i s cal l ed wei ght ed ar i t hmet i c mean. To gi ve due i mpor t ance t o each
i t em under consi der at i o n, we assi gn number cal l ed wei ght t o each i t em i n
pr opor t i o n t o i t s r el at i ve i mpor t ance.

Wei ght ed Ar i t hmet i c M ean i s comput ed by usi ng f ol l owi ng f or mul a:

Direct M et hod:
X

= x

Wher e:
X

St and f or wei ght ed ar i t hmet i c mean.

x St ands f or val ues of t he i t ems and
St ands f or wei ght of t he i t em

Example:
A st udent obt ai ned 40, 50, 60, 80, and 45 mar ks i n t he subj ect s of
M at h, St at i st i cs, Physi cs, Chemi st r y and Bi ol ogy r espect i vel y. Assumi ng
C.I. f
0-10 3
10-20 7 f
0

20-30 12 f
1

30-40 5 f
2

40-50 3
f = 30

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
wei ght s 5, 2, 4, 3, and 1 r espect i vel y f or t he above ment i o ned subj ect s. Fi nd
Wei ght ed Ar i t hmet i c M ean per subj ect .

Sol ut i on:

Now we wi l l f i nd wei ght ed ar i t hmet i c mean as:

X

= x = 825 = 55 mar ks/ subj ect .

15

Short Cut M et hod:

X

= A

+ . dx

Wher e:
X

St and f or wei ght ed ar i t hmet i c mean.

x St ands f or val ues of t he i t ems and
St ands f or wei ght of t he i t em

Example:
A st udent obt ai ned 40, 50, 60, 80, and 45 mar ks i n t he subj ect s of
M at h, St at i st i cs, Physi cs, Chemi st r y and Bi ol ogy r espect i vel y. Assumi ng
wei ght s 5, 2, 4, 3, and 1 r espect i vel y f or t he above ment i o ned subj ect s. Fi nd
Wei ght ed Ar i t hmet i c M ean per subj ect .


Subject s M ar ks Obt ai ned (x) Wei ght s x
M at h 40 5 200
St at i st i cs 50 2 100
Physi cs 60 4 240
Chemi st r y 80 3 240
Bi ol ogy 45 1 45
Tot al = 15 x = 825

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710

X

= 60 + (-75)
15
X

=

60 5

X

= 55

7. Quart iles-
A st at i st i cal t er m descr i bi ng a di vi si on of obser vat i ons i nt o f our def i ned
i nt er val s based upo n t he val ues of t he dat a and ho w t hey compar e t o t he
ent i r e set of obser vat i o ns.
Each quar t i l e cont ai ns 25% of t he t ot al obser vat i ons. Gener al l y, t he dat a i s
or der ed f r om smal l est t o l ar gest wi t h t hose obser vat i ons f al l i ng bel ow 25%
of al l t he dat a anal yzed al l ocat ed wi t hi n t he 1st quar t i l e
Obser vat i ons f al l i ng bet ween 25. 1% and 50% and al l ocat ed i n t he 2nd
quar t i l e,
The o bser vat i ons f al l i ng bet w een 51% and 75% al l ocat ed i n t he 3r d
quar t i l e, and f i nal l y t he r emai ni ng o bser vat i ons al l ocat ed i n t he 4t h
quar t i l e.

Subject s M ar ks Obt ai ned (x) dx Weight x
M at h 40 -20 5 -100
St at i st i cs 50 -10 2 -20
Physi cs 60 0 4 0
Chemi st r y 80 20 3 60
Bi ol ogy 45 -15 1 -15
Tot al = 15 x = -75

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
Comput at ion of Quart iles in Individual Ser ies

Q
1
= N + 1 t h val ue
4

Q
2
= 2 N + 1 t h val ue o r N + 1 t h val ue
4 2

Q
3
= 3 N + 1 t h val ue
4
Not e: Fi r st ar r ange t he obser vat i on i n ei t her ascendi ng or descendi ng or der .
Exa mple:
Cal cul at e f i r st and t hi r d quar t i l e f r o m t he f ol l owi ng dat a:
12, 15, 17, 18, 18, 19, 19, 20, 21, 22, 23, 24, 24, 24, 30, 35, 37, 38, 40

Solut ion:
Number of obser vat i ons, N = 20

Q
1
= N + 1 t h val ue
4
= 20 + 1 t h val ue = 5. 25
t h
val ue
4
= 5
t h
val ue + 0.25 (6
t h
val ue 5
t h
val ue)
= 18 + 0. 25(19 18)
= 18 + 0. 25
= 18. 25
Q
3
= 3 N + 1 t h val ue
4
=3 20 + 1 t h val ue = 15. 75
t h
val ue
4
= 15
t h
val ue + 0. 75 (16
t h
val ue 15
t h
val ue)

By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
= 24 + 0. 75(30 24)
= 24 + 4. 50
= 28. 50
Comput at ion of Quart iles in Discret e Series

Q
1
= N + 1 t h val ue
4

Q
2
= 2 N + 1 t h val ue o r N + 1 t h val ue
4 2

Q
3
= 3 N + 1 t h val ue
4
Not e: Fi r st ar r ange t he obser vat i on (x) i n ei t her ascendi ng or descendi ng o r der (i f
not ar r anged).


Comput at ion of Quart iles in Grouped Series

Q
k
= L
1
+ kN/ 4 pcf x i or (L
2
L
1
)
f

k = 1, 2, 3
f = f r equency of t he k
t h
quar t i l e cl ass
pcf = cumul at i ve f r equency of t he cl ass pr ecedi ng t he k
t h
quar t i l e cl ass

Q
1
= L
1
+ 1N/ 4 pcf x i or (L
2
L
1
)
f

Q
2
= L
1
+ 2N/ 4 pcf x i or (L
2
L
1
) (Q
2
al so kno wn as M edi an)
f


By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
Q
3
= L
1
+ 3N/ 4 pcf x i or (L
2
L
1
)
f


8. Deciles:

It i s a met hod of spl i t t i ng up a set of r anked dat a i nt o 10 equal l y l ar ge
subsect i ons.

For mul a,

Individual Series and Discr et e Series

D
1
= N + 1 t h val ue
10

D
2
= 2 N + 1 t h val ue
10


D
9
= 9 N + 1 t h val ue
10

Grouped Series

D
1
= N t h val ue
10

D
2
=2 N t h val ue
10

D
3
=3 N t h val ue
10


D
9
=9 N t h val ue
10


By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710

Cont inuous Series

D
k
= L
1
+ kN/ 10 pcf x i or (L
2
L
1
)
f

k = 1, 2, 3 9
f = f r equency of t he k
t h
deci l e cl ass
pcf = cumul at i ve f r equency of t he cl ass pr ecedi ng t he k
t h
deci l es cl ass

9. Percent iles:
A per cent i l e (or cent i l e) i s t he val ue of a var i abl e bel ow w hi ch a
cer t ai n per cent of obser vat i o ns f al l .
For exa mple, t he 20t h per cent i l e i s t he val ue (or scor e) bel ow w hi ch 20
per cent of t he o bser vat i o ns may be f ound. The t er m per cent i l e and t he
r el at ed t er m per cent i l e r ank ar e of t en used i n t he r epor t i ng of scor es
f r om nor m-r ef er enced t est s.

Individual Series and Discr et e Series
P
1
= N + 1 t h val ue
100

P
2
= 2 N + 1 t h val ue
100


P
99
= 99 N + 1 t h val ue
100

For Grouped Series
P
1
= 1 N t h val ue
100

P
2
= 2 N t h val ue
100



By- Prof. Saurabh Singh
Saurabh.singh.21@gmail.com
+919039962710
P
3
= 3 N t h val ue
100


P
99
= 99 N t h val ue
100


Cont inuous Series

P
k
= L
1
+ kN/ 100 pcf x i or (L
2
L
1
)
f

k = 1, 2, 3 99
f = f r equency of t he k
t h
per cent i l e cl ass
pcf = cumul at i ve f r equency of t he cl ass pr ecedi ng t he k
t h
per cent i l e cl ass


10. Combined M ean:

X =
N1X1+N2X2++NkXk
N1+N2++Nk