•

-so Correlation
---
t
By su bs itu tin g values in
the ab ov e fo nn ula , we ge
I 3 J
I 3 -2 )+ 12 (2 3- 2) ]
)+ 12 (2
R = I- 6[ 45 +i z( J -3
2
8(8 - J)
6( 48) 288
= I - 6( 45+ 2+ 0.5+ 0.5)
8(63) - = 1- 50 4 = 1- 504
= I - 0.57 I == 0.429
EXAMPLES
IMPORTANT TYPICAL
s,
ati cs and Statistics were as follow
e 3 7. The ranks ofthe sam
e 8 students in tests in Mathem
Ex am pl ke ts de no ting the ranks of the same _students in
ac
bers within ~r pectively:
the two ~uman d Statistics res
Mathem ati cs
7)
(2 , 2) , (3 , I), (4 , 6), (5, 8), (6, 3), (7, 5), (8,
(1, 4) , s.
rre lation for pro fic ien cie s ofthis group in Math's and Statistic
(,) Calculate the rank co tes?
W ha t do es th e va lue of the coefficient obtained indica
( iz) een
ou t· Ka rl Pe ars on 's sim ple coefficient ofcorrelation betw
(iii) If yo u have foun d same as
es e 16 stu de nts . W ou ld your results ·have been the
th e rank•s of th ent?
obtained in ( i) or any differ
elation Coefficient -
So lu tio n: (z) Ca lCU·18 t·10 D 0 r RaD k Corr
- D= R1 -R 2
D2
Ranks in Maths Ranks in Statistics
( R1 ) - ( R2 )
9
I 4. 3
0
2 2 0
. 4
I ·- +2
3
.. 4
6 ' 2
4
9
8 3
5
9
·3 +3
.6 . 4
5 +2
7 J
7 +I
8 . - w2 40
:

N 8 -~

Ap ply ing the formula

24_0
6x 40- __ 1__
6W 2 -- -- -1
R I ... ... -- --
= - N3 _ N - s3 - 8 504

_ 50 4 - 240 = 264 =+ 0.5

23
- 504 504
 Correlation 51

(i,) TI1e value of rank correlation coefficient indicates that there is moderate degree
of positive correlation.
(iii) Calculation of Karl Pea non 's Coefficieht of Correlation
Ranks in A•4
Maths dx
_£
d.xl Ranks in A:114 rty1 dxdy
(X) Statistics dy
(Y)
I -3 -
9 4 0 0 0
2 -2
-
4 2 2. 4 4
3 -1 -
1 1
"- 3 9 3
4-A 0
I-=- -
- 0 6 +2
--
-- 4 0
I·~ .5 I-~• +1 1 8
'
+4 '"
16 4
6 +2 -
. 4 3 1 1 2
7 +3 9 5 +1 - 1
.. , 3
8 +4 16 7 +3 9 12
N=8 l:dx= 4 2
I.dx = 44 uiy=4 !.d/=44 Ll.xdy= 24
- -
Applying the formul~
r= • i-: :, N~Ldxdy- I.dx,'uly .·.
. ✓ N ·uJx.2 :-(Idx)2 ✓ N .rdy2 -(Idy)2
J

"' • ..i-,, •

_ 8 X 24 -:- ( 4) {4) .
·.Js x 44-:._{ 4)2 .Js.. x 44-( 4)2 _
... . .
192 - 16 ·_ ,, ' 176" • :• •·~ · ,
=-
✓336----36_✓
336_3_3-6 = .= 0.5 23 336
It is evident that the value of correlation coefficient co~puted by using Karl
Pearson is the same as obtained by rank correlation method. The r ason is that
1 give the
when the ranks pf the students are not repeated, then the two methods
same answer. t ••
I ,\
•\

E11mple 38. Calculate rank correlatidn coefficient from the following data:
'
Serial No.: 1 '2-· 3 4' 5 6 7t 8 9 10

Sol■tion:
Rank Difference: -2 ? -1

The total of rank differences (ID).is always equal to zero and on this basethemissin-g '
+3
•
+2
•
0 -4 +3 +3 -"2 .
-
rank difference will be calculated. Let the roissing item be 'a'.
. . . \
As W=·o·⇒ -2+a-1+3+2+0-4+3+3-2 1 0?
'· \
. ! t '
:. a=-2 --
•. j
I '•J

•
 _ _ _ _ _ _ _ _ _ _ _____ --.;;;C.=:orr~elatio

• t_ _ _ _ _ ______ ~

!2
icient of Ran k Correlation •
Calculation of Coeff Dz
r
Rank Difference
Sr. No. D ,
4
-2
I 4
2
-2
_, l
-- ----- ,.
3 -
~

9
---
.
+3
4 4
+2
5 0
0
6 -
16
-4
7 9
+3
8 .

9
+3
9 ' 4
IO -2
ID=O I.D 2 =60
N= IO
2 60
6 !D = 16x- - -
R=l- - - 3
3 10 -1 0
N -N
360 1-0.364 ='+ 0.636
= =1 -
990 h and
m arks ob tain ed by 10 students in Englis
tion of in ranks
ficient of rank correla ed that the difference
Example 39. The coef d to be 0.5. It w as later di scov er
stead of 7.
Mathematics was foun the stu dent s ~a s w rongly taken as 3 in
ed by one of
in two subjects obtain
the co rrect coef ficient of rank correlation.
Find ) =3
N = 10, Inco rrect difference of ranks (D
Solution: Given, R =
0.5,
nks (D) = 7
Correct difference of ra
We know that:
R = l- 6 W 2
3
N -N

0.5=1- 6 W 2
3
10 - 10

0.5=1- 6 ~ 2
990
2
6W
990 = 1- 0.5 =0.5
2 = 82 .5
~ Incorrected l: D
Corrected l: n• ·-- 82 .5 -(Incorrect value)2 + (Cor
2 I rect value)2

! = 82.5 _ 32 + 72
= 122.5
 Correlation 53

Corrected Coefficient of Rank 6t D 1

Correlatio n (R) = 1- N1 _ N

=l - 6x122.5 =l - 735 =O.lSS

10 3 -1 0 990
Thus, the correct value of ran
k correlation
coefficient is 0.258.
Example 40•. ,The. r~nk, corre!
ation coefficient between ma
.~tat1st1cs and Accountancy' rks obtained by some students
is fou in
differences is 33, find the number nd to be o.& . If the total of squares of rank
of students.
Solution: Given, R = 0.8, w· 2 =33

Now, R= l- 6W 2
·- N 3 -N
.. _· ·o. 8= 1- 6x33
. . •. N 3 -N
..
• • f,.

198 · • · ·
• • J

·
3
N -N · · • = 1 - 0.8 = 0.2
•
.
• . N - N= 198 ~990
3
,0.2
' .
N ( N2 - l) = 990 . •
[": a 2 - b2 =( a+ b) (a - b)]
.
+
'
'
• • • t
.. - - - · -

N (N l)( N -- i)-.990 ~ ~---~ ·-

f ---

•
(N -l) (N )(N + l)=,9 x 10 Ix 11•
; i

Comparing both sides, we get:

:. N- 1 ·= 9 · - . · . ·-·
. - : ·· · - => . (·
· • -N =· 10 -· · · · · - -
Eumple 41. The rank correlation coe ,
fficient between marks obtained by 10
and Economics was found to be O.S. students)n,'i<~fu~~cs
Find the sum of squares of differences
Solution: Given, R ~ 0.5, N =10 of ranks.
• •
• · 6ID 2
·N ow ,· · R = l~ --
. .N 3 -N ·
2 . -. - -- -
6 ID = 1- R - - -· . . - - -.
3
N -N
6I D 2 •
- - = 1- 0. 5 =0.S
103 - 10 \ .
• , I.

_ 6ID~=0.Sx990 .,
- w 2 ,.;o.s'x990 = 82.s •.
'6
§54!__ __:,__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~C~o~rreiat··
-=- ~

, ► 1\feriu and Demerits of Rank Correlation Method

Merits
(I) This method is simple to understand and easy to apply as compared to Karl Pear500 , 8
method.
(2) When the data are of qualitative nature like beauty, honesty, intelligence, etc., this is the
only method to be etnployed.
(3) When we are given the ranks and not the actual data, this method can be usefully employed.
Demerits
( 1) This method cannot be used for finding correlation in a grouped frequency distribution.
(2) When the number of items exceed 30, the calculations become quite tedious and require a
lot of time.

EXERCISE 1.9
1. Ten commerce graduates appeared before a sele~tion board consisting of two members X
and Y for the post of probationary officer in a certain bank. If the rank order of each of two
members i_s given below, find out the coefficient of rank correlation:
Rank order by X: 3 ' 2 4 9 7 8
I 6 5 )0
·4 JO 2 ) 6 9
Rank order by Y: .- -3 5 8 - 7
. . [Ans. R=-0.212]
. \ .
2. Ten competitors in an intelligence test are ranked by., three judges in the following order:
►.

Judge I: 9 3 7 5 1
• '

6 2 4 10 8
'
..
Judge II: 9 1 10 4 3 8. 5 2 7 6

Judge III: 6 3 8 7. 2 .4 I 5 9 10

Use the rank correlation coefficient

.
to determine:
-.

. the most?.
(i)., Which pair ofjudges agree
( ii) Which pair of judges disagree the most?_
[Ans. R12 = 0.71, R23·= 0.467, Rn= 0.8'42
(i)·lst and Hird (ii) IInd and Hird]
3. Find out the coefficient ofcorrela1ion between X and Y by the method of rank differences:
X: 75 88 95 70 60 80 81 50
Y: 120 130 13.0 115 110 140 142 JOO
[Ans. R = 0.7976]
4.. • Find out
,
the coefficient ofcorrelation between X and Y by the method of rank differences:
X: 46 .56 39 45 54 58 36 -40
Y: 30 60 40 • 50 70 70 30 50

[Ans. R = 0.75]
 Correlation 55

s. ~i nd tl~e ~ank correlation coefficient

from the following marks awarded by the examiners
1
- n stat1st1cs· .
~R. Nos.:
. 1
Marks Aw d
2 3 4 s 6 1 8 9 10 ll
L
ar td by Examiner A:
24 29 19
~ Marks Awar·d td by Ex11miner B:
14 30 19 27 ~o 20 28 ti
37 35 16
Marks Awarded by Examiner C: 26 23 27 19 20 l6 11 21
30 28 20 25 2S 30 20 24 22 29 IS
6 F th
11 e:
• rom e io owing data, calculate Spe [Ans. RAs = -0.027, R8c= 0.5272, RAc
arm ' = 0.26136]
X:
an s coefficient of correlation:
80 78 75 75
Y: 68 67 60
12 13 59
14 14 14 16 IS
---- '"'-~
17
7. Theranksofthesame 1 stud .
6 . [An s.R =-0 .92 8)
two .n~mbers within bracketse~~~~i~est~1n
Stat1st1cs respectively: . Mathematics and Statistics were as follows, the
g t e ranks of the same students in Mathematic
• (1, 1), (2, 10), (3, 3) (4 4) (S S) ( s and
(~2, 9), (13, 14), o4 , 2), (ls : ), 6dl!'i~}
1 16 2)~ (8, 6), (9, 8), (10, 11), (11, 1s),
(z)
.. Calculate the ran. k- correlation~or proficie . ncie
. s of
this group of Math's and Statistics.
.. ~~~) What does the-value of the coefficient
obtained indicates?
(111) If you have found out Karl p .
· ranks of these tud ,. • •
16 earson s Slm ple coefficient of correlation between the
. ·f:c.
any d1 1erence?. s ents would you · •
r results have been the same as obtained in (a)
, , or
F th £ ll ·
8• rom e o owing data, calculat - [Ans.R=0.8,r=0.81
e Speannan's·coefficient of
correlaf10n:
Sr. No.: 1 2 3 4 5 6
• Rank differences: .- I< 7 8 9 10
-2 -4
Ii
- -1
. I

., +3 +2 0
.' ., . 'f
'
·t~- - •. :r }I,.
? +3 +3 2
:. • ; _ • . :
-~~: ;
·- l , ., .,
_ _. , •. ~-..,.-·,;·'..•

9~ . The coefficteEt of rank correlation between [Ans.R=+o.636] 1· ._ . •

~eben~r_e prices and share price~ is found to

· . O. t 43. l~the s~~ of squares of.the diff~ren be
ce_i n ranks is given to
. be 48, find the value ofN.
~

[Ans. N = 71

~I ii ) C~NCUR~EN~ ~EVIAl'.l~N
MET~OD • " • • · ,. · •• •
•. C~ncurrent deviati~n ~ethod of determi
ning the correlation is extremely simple met
method, correlation is detemlined on the bas hod. In this
is of direction of the deviations. Under this
~' _taking into consideration the direc~ion of method,
deviations, they are assigned (+) or (-) or
following steps are taken to find out correla (0) signs. The
tion iri this method:
-- (1) UOdCi this method:whatevef the seri
the series is-compared ~ith its preceding itemes.•XIfand Y ar~ to be studied for c;rrelati~n, each item
the value is more than its preceding value, the~
of
deviation is assigned (+) sign, if less than its
preceding ,value theri" (-) sign and· ~f ~qu
a\ 'to the
 56
-=-=---------- --------------- ---~C~o~r1r e1auoP\
~
. prccccding value then (0) sign is assigned. After this, third item is compared with the sec d
item is compared ,vith the third and this process goes on till the deviations of all items i on ':0 urth
,,,orkc{f out. n a series are

(2) The deviations ofX and Y series (dx) and (dy) are multiplied to get d.xdy. Prod t Of . .
signs will be positive(+) and opposite signs will be negative(-) like: uc sunilar
(+) (+) = +,
(-) (-) = +,
(0) (0) =+,
(-).(+) =-,
(+) (-) =-,
(0) (-) = -, ..
I

- '·

(-) (0) =-,

(0) (+) =-
• (3) Su~ming the positive.·dxdy signs, their nu.mber is counted. Th.is i~ kn~wn ~ the ·numb~r of
concurrent deviati~ns. It is denoted by the sign 'C'. The deviations with minus signs are excluded
from .the computation. They are ignored. If all the deviations in a series have minus signs, then
number of concurrent deviations will be zero i.e. C=0.
(4) FinaJly, the following formula' is used for determining coefficient of concurrent deviations

. r.: =+ /±(2Cn- n) . . , -·:• ),.{'

,
I '.... • ~ ._; _ .•. · "V v~~ • •
V :.,., ......_,:
: •• Here, re =Coefficient ~f con~urrent d~~iations;
C· = Number of concurrent. deviations or Number of positive .signs obtained after
multiplying dx with dy; - ··· ·· • -
. r -'l , t t

-~- ·- *n = Number of pairs of observations minus .one= N - L - - •--- -- - -·---· --

- . , t ~ ~ l i. ~·"

Note:· In this formul~±sign is used both inside and outside the radical sign. Ifthevalueof(2C-n)
in
is positive, then (+)
• ..
sign will be used both inside
' .. ' . .because
and outside the radical sign . '

-such case correlation will be positive. On the contrary; if (2C-n) has negative sign, then
minus sign will be used both inside and outside the radical sign because correlation will be
• • negative.
- The valu·e of coefficient of concurrent deviation always lies between -land +1." ~ - . .,
The following examples make the procedure of concurrent deviation method _clear.
Example 42. Find coefficient of concun·ent deviation from the following data:
X: 85 91 .S6. 72 • 95 76 ~· 89 51 59 90

Y: 18.3 20.8 16.9 15.7 19.2 , 18.1 . 17,5 I •14.9 18.9 15.4
'
..
* Since there is no signfor the first value ofX and Y, n is always taken to be one less than the actual
number ofobservations. ' • ' ;.,:, ••• I :. , ' - \
 £ orrelation 57

Solution:

X
I
,beviatlon signs y Deviation signs dxdy
(d.t) (dJ)
85
18.3
91
+
20.8 + +
56
72
- 16.9 - +
+.
95 . . +
- . 15.7 . -
76 19.2 + - + --
- 18.1 +
89
+
~

-,_
1,-
51 ,_
17.S . -
59 1 14.9 +
+ 1
18.9
90 + +
+
n-(10-1) =9
-
15.4 -
~-(10-1) 9 c-6
2
Here, C- n, i.e., 2 X 6 -9 =3 is positive, therefo~e we use positive(+) sign in the
fonnula. Thus, ·

---- -- ..-._ .. r.=+~

. -
- ----- --
; _ _
- -~ 9- - - +-
'
=+
-re=++
• - ✓ (2 X6- 9)
-9 •=0.577
r
..
_
fr --
•• - - - - Thus, there is positive correlation between X and Y.
i11mple -43_ Compute the coefficient of_ correlation for the fallowing data by· the concurrent
_____ - -deviation method: ··~--- ___ .- . ~ __ _ ~- __
Year: - 1971 - -· 1972 -1973 -
- ·1974 ' 1975 1976 1977
Demand: •c ,= 150 -= , . 154 -- -160
172 160.
• I-
165 180
__ ,__Price: _ __ ~ ·- 200 180 = _ 170
-
I
= -~ 160 '= 190 180 172
Solution: Denoting Demand and Prices by X and Y. ·
' I
Year : /
Demand Deviation· signs
.,
\ . Deviation··signs
·'P~ dxdy
X (tb:) •. : (dy) ..
1971 ISO C,
.
i 200
1972 154 .
+ 180 - -
1973 160. + 170 -
,.1 -
l 1074 172 - + 160'
. - -
1075 160 -~ 190 + -
- .
+ .... - -· 180
11
1976 165 - -
1977 •_ l72 • l . ,
. 180 . +· - -
n =(7-1) 6 .. I C 0 \
 Correlation
,
-58 ~

use negative(-) sign in the·

Here, 2C-n, i.e., 2 x 0- 6=-6 is negative, therefore we
fonnula. Thus, .
r
C
=± ±(2C -n)
n
(2 x0- 6) _
=± ± - - =- ✓-(-·I)=-1
6.
.
demand.
There is perfect negative• correlation between price and
'
concurrent deviation method from the
Example 44. Calculate coefficient of correlation by •
following data: • -
-

X 112 125 126 J) 8 li8 )2) )25 125 131 135

--
y 106 102 102 104 98 96 97 97 95 . 90
'

.
Solution: Calculation of Coefficient of Concurrent Deviation
. - . dxdy
X Deviation signs
·•
y Deviation signs
. (dx) ' I - (dy)

112 106
- '. -
125 + ' 102
0 -
126 + . 102
: I ·, ;
104 + -
118 . - -
,
I_

- -
98
.....
118 0 ~

I
.I
i ;. ., -· ' J
- -
121 + 96 ,,
V .;
' , ' ·'
~

+ +
- 125 + 97 . '

.. . -. ··97 • ,. ·- ... - - :0 - .. I +·
--.. -
.,.
;
-· 125 0,
.I
' . ...
+ :" -
.

,..,r •• ' .- \ -- .. ·-- . -.

,_
131
.. ,.Jtc, · • · - ·-

95 '
• C

... - .... .... . -· .. - ·. - -· .

- I

135 +i '
I 90 .. '

- -· .. - -. .
C=2
n=l 0-1 =9
,'_therefore ·\Ye· use _negative(-) sign in the
; --·-·.. ---, . Here, 2~..: n, ,i.e.', 2 x·2:... 9---5 i~· n_egative · · . ·
: fonnula. Thus, ··· '
,
I ••• - --· - • -·· - • •• ·• ' • • •• • • • ••
• . • " •. . • .. • - ; •

r e ' ± ~ ; : C=2, n=9 · ·• •

• -- -,· • •• • •
.. '

=±: _± -2-9
(2x -=
- .' 9 -·
.
) - - -
__- 9
I (-SJ -
·
. =-✓0.5556=".:..0.15
een X and Y.
Thu~, tJiere js high degr~ _ofnegat~ve_co~lati~n _be~ I
t
- ..... - ·- - . .. - ....... -..__ - ..,
•· ~ ~
 correlation 59
~

. UVIPORTANT TYPICAL E
.... XAMPLE
£J&lllPle ~5. During the first 9 .
00th5 changes m
. the price index of ~ of the financial year 1999-2000, the following
correlation by a 5~-:arbtls
1
A and Dwere recorded as below. Calculate the coefficient of
a e method:
Changes
ovtr the previous month
Month: April May lune
Share A: July Aug. Sept. Oct. Nov. Dec.
-4 -3 -4 0 +3 +4 +2 -3 +3
Share B: +3 -3 -2 -4 -3 -2 . -3
j

. . -4 0
solution: In this question changes are giv . .
only concurrent deviation e~m companson to preceding month and in such a case
the basis of multiplicati m~t _od may be used. The value of 'C' will be calculated on
· on ° Signs only (values will be ignored)
CaleuIation of Coefficient of Correlation
Months Share A
' • Deviation Share 8 Deviation dxdy
signs (dt) signs (dy)
April. _ -4
+3 +
- ---·-IQ.
May -3 ,.
. .& - -

-3 +
;
__ ....... - ...-·- -- June -.~.- -4
-2
.1. •
,. July 0 0 . -4
August +3 .•. ·_3
.
I
September +4
:. .•
.-. - -;;.
-4
October .
.+2 I (}
0
November -3 -2 +
December +3
........
' n=9 \ . C=3
- --
. Applying the formula, .
. ··- (2C - n)
re=+ ±-·--
n
' - . I• ..
.•

Here, C .:.. 3, ,i = 9 ,

(2x3-9) •
. -•-· . ~- - re=±
"' . ± ---
. .' . 9

>:f (~
• l

- ·_ = 3) • - ✓- (-3) ~ ::_ .j0.33 :_ 0.574 - ~

· • • 9 · - · 9- ·.· -
• ' : . • . \ ' . .

Note: - Generally, the value of 'n' is written on the basis of N-1, but in the above example,. it will
not be applicable because deviation sign of first item is also known.
 60 Correlatio
~

► l\ferits and Demerits of Concurrent Deviation Method

Mtrits
(l) This method is simple to understand.
(2) Its computations involve less time.
(3) When the numb er of items is very large, we can use this method to have a quick idea about
. ·
. the correIat10n.
(4) This method is useful in studying short term fluctuations.
Demerits
'
ion ·of correlation.
(1) By applying this method, we can get an idea only about the direct
es. ,
(2) This method is not useful, for_ finding correlation of long term chang
•
• (3) This method is less accurate than Karl Pearson's method.
.. ..
EXERCISE 1.10 , '.·
,I

__ _!,._ CalcuAlate the coefficent of correlation. by the method of concurrent deviation from the
._following data: -; • - ·• . ~- ..
I .
I
60 25 45 80 85
X: 6S . 50 ' 35' 5S '
f ~

' 70 30 · 40 , 65 80
Y: 45 35 r 55i 40 ' '
!
. ; [Ans. re= 0.707)
:, .'
'
! -

2. __ Calculate coefficie~t of concurre~t deviation from the Joliowing data:

. '
I
! 80 • • 3S 20 80- 60 50
X: 65 40 35 15 .63
56 • · 30 _,·70 .. 40 35 80 15 80
Y: 60 55 .50
;•• ' ,· ' .. · · [Ans. re=+ 0.89]
I
I !
•
d of the following data:
3. -..- Find coefficient of ~orrelation bY, concurrent deviation ~etho
- ·--- ;

Students: . 1 A F
Ii B c; D E F G H

70 45 40 80 68 85 40 25
Marks in Economics:
65 60 - 55 61 35 1S 45 40
Marks in Statistics:
-- .- - [Ans. re= +l J
regar ding changes in price
4. Obtain a suitable measure of correlation from the following d~ta
•
index of two shares A and Bduring the year: -, ,
Changes over the Previous Month
J F ·M A M J J A \· ·S 0 N D

Shares A: • +4 +3 +2 -1 -3 +4 -5 • +J +2 -7 +2 -3
Shares B: -2 +5 +3 ~2 -1 -3 +4 -I ~3 .+6 +4 +2

' I (Ans.,,=..: o.4os1

 co~~,a~tio_n______ ______ ______ ______ _____
61
;----
.
5 Find out the COefficient 0 f
• con-elation between X and Y by the method ofconcurrent deviation:
X· 26 3 38
• o 30 24 29 2~ 2S 32 32
Y: 62 58 SS 68 67 64 64 1S 81 78
(An,. ,e~ -0.5771

~OEFFICI ENT OF DETERMINATION

Th~ concept of c~efficient of determination is used for the interpretation of coefficient of
correlatton and comparing the two or more correlation coefficients. The coefficient of determination
is defi~ed ~s the sq_uare of the coefficient of correlation. It is denoted by r2. The coefficient of
detenn1nat1on ~xplains the perc~ntage variation in the dependent variable y that can be explained in
terms ~f ~e in~epe~dent variable. X. _If correlati9n coefficient (r) is 0.9 then coefficient of
detennination (r ) wtll ~e 0 -81 which implies that 81 % of the total variations in the dependent
variable (Y) oc~~rs ~ue to ~h~ jnd~pendent variable (X). The remaining 19% variation occurs due to
•outside or ~x!emal factors. Thus,. the coefficient of determination is defined as the ratio of the
explained·vanance to_the total yariance. ~ terms of formula: .
I I (

~- c~;fficient~fDe t~~in~tic;;(r2) ~- Explained Variance -~·-

• • :::- ·Total Variance ·
. . - .
• Coe~cient of Non-Determination: By div:idi~g the unexplained variation by the total variation,
the coefficient of non-~etenninati on can be determined.' Assuming the total of variation as 1, then
the coefficient of detennina~ion ·can }?e. determined by.subtracting the coefficient of determination
from 1. It is denoted by· K 2 • In terms offormula, ~ - - : - . . , ·
~ ' ! t

Coeffi~ient of non-detennina~ion (K2) =I - r

•

i .... '
; -- In the above example r. -
0.81 ;then .the-~oefficieiii of non~'detennination· will be 0.19 (1- 0.81 ).
•It indi~ates tha~ 19% of the variations are due to. other factors. - ·-- . -
. -- •C~efficient ~t Alie~tati~ii~- 11 - r 2 - ~- - ---- ~ - - ; . .- ·: •. - - •. --
:-
- .- --· ... ----· --- - - - . • . "
-·- - - •• - -- --- ' -- - -- - - . - ··-
Generally, the coefficient of determination (r)2 is widely used in practice.
- -
~;.mple 46. ·11u: ~!!_ici~i~fc'o~t;i~n <f>~~ee~ ~~~s~~Pti~I_l-~~~I_ld_i~re (C) ~~ di~posable
. .
--
i
. . - --
i ~c~me (Y) in a study was found to be +0.8. What percentage ofvanat1on in Care
explained by variation 1n Y?
-
•
•• '\ ~ - ... .. --- . -.

•Solution: --Here-:-· r - 0.8 ~-;. 2. - (0.8) 2 -0.64. It-means that" 0.64 or 64% of the variation in
• -- consumptionexp enditure are explained by varlation~ in ·income. • .
a
I

. Iumple 47. - Is it true that correlation coefficient-(r) ~ 0.8 indicatesa relationship tw~~e as close
• --·· - --asr= 0.4?.
~i --- -·_ ·- i • •-· , - .... ... _..,

. . '
• The ·statement can be verified by usi~g coeffic_ient of de~ennination, i.e~, r 2•_
1
•Solution:
- -- -- - -- - -
Now, 1st case: · r 2 = (0.8) 2 =0.64
2nd case: r 2 =(0.4) 2 -~ 0.16

•
 Correlation

of the
"~ that t.-t•,, of the \'~riRt ion is o.~plnincd in the first case and 16%
l1'is ~'
cnsc. Hence r = 0.8 does not
indicate
i\,n ,s C'J lla inc d in tJ,c sec on d 8
,-.ri~t as r = 0.4.
ma tiN ish ip t\,·ice as clo se
t of0 .5 in1 plies thnt 50% of the data are explained. Comment.
Ex -a _, . 4 A ~l at io n coctl
'icicn
in y which are
lki e.u of dctcm 1in ation ( r ~) sho,v the percentage of variation
Sohttioll: C\.'Ct
e..'l)lainro by the variation in X.
1 2
NO \\', r =( 0.5 ) =0 .25
rel ation of0 .5 sho,vs tha t 25% of the data are explained by X.
Th ~ the l'l'lefficient of cor
of the var iat ion in Y is du e to X and the remaining variation is
In oth er \\\ lrd ~ 25 %
du e to oth er factors.
ng to int po rt pri ce (X ) and im po rt quantity (Y ) in respect of a given
E x- a~ -19. Th e da ta relati as under:
c-0mn,oditv . are
'78 '79 '80 '81 '82 '83 '84
Year: 1975 'i6 tn
.
4 3 5 7 8 7
6 5
lap or tp rn : 2 3
7 10 9 7 8 9
Qn ati t), iap om d: 6 s 4 5

ffic ient of correlation.

(1) Ca lcu lat e Karl Pe ars on 's coe
d that is explained by the.
(ii) Fin d the percentag e of variation in quantity importe
variation in the import price.
rrelation
Solatio■ : (,) Calculation of Coefficient of Co
• 2 y Y= 7 ,2 XJ
X X= S .r
X- X Y- Y
. y
.r
I 3
9 6 I
2 3
4 4
4 5 2
3 2
9 3
- 1 ·4 3
6 +I
.4 0
0 5 2
5 0
0 0
I 7 0
4 I •. ~
10 +3 9
2 4
3 0
9 +2 4
0 0
5 0
7 0 0
+2 4
7 3
8 +I I
+3 9
8 4
9 +2 4
- +2 4
7 uy s
:&2 = 36 l:Y= 70 1:y=O 1:; = 36
N= 10 l:x= O
u so
X = 1: X = 50 =S y = l: Y = 70 = 7
N 10 N IO
 ~1atio~n!-_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _!!63
~
'
Since the actual means of . .
from actual means 0 f X X 80d Y are whole numbcu we should take deviations
and Y to simplify the cafcul;tions.
r == - t xy 5 S
f;.
'\JLX
2
Xl:y2
- -
..J36 Jf6
--=:----:-- - -
-J6
= 0' 1389 I

(ii) Here, r = 0.1389

⇒ r2 = coefficient of de •
It n1eans that 1.92% of th termination= (0.1389)2=0.0192 or 1.92%
. .
variations in the • e_ variations in quantity imported are explained by the
import pnce.

exeRCISE 1.11
?;::i7ii
--
• h'
The relations tp between cons •
t. b I th• umptton (C) and disposable income (Y) is expressed by
C = a+ Y: n is cont~xt, explain what the value of r2measures.
z. "A co~l~~ion coe:cj,ent of 0.3 implies that 30%.of the data are explained." Comment.
3. A corre a ton coe _tcte?t of 0.6 ind_icates a relationship hvice as close to as where r = 0.3.
Comment. • • · ·• •• . ,:
Quantity ( Y) : 69 76 52
,- 56 57 77 58 55 67 63 72 64
Price (X): 9 ' 12 6 IO ·9 10 7 8 12 6 11 8

- ••• (i)_. Calculate-t be Karl__ Pearson's· coefficient of correlation between price and quantity.
-·· ,- (ii) ._Find ~e percentage of variation in ·q~anti~· de~anded that is e~plained by variation in
the prtce of the COJ?modity. [Ans. (z)r=0.645, {il)42%)
4. Calculate from the given information:
·- '
x··: 45 70 65 30 90 - 60
40 50 75 75 85
Y: 35 90 70; 40 95 40 60 80 80 80 50
:_ ____ (z) Karl Pearson's coefficient otcorrelation: • _·- ~·-· - -- . : . I -
. -~ -_· • (iz) Probable Error and sho~ ~h~ther .'r~-_.fa signifi~ant or not?· .. -- .
(iii) Coefficient of non-determination and coefficient of alientation.
•.
•
7 [Ans~ (z) r =;= 0.904, (iz) P.E.·= 0.0390, r is"significant, (iii) 1-? = 0.183, 0.4277]
- • • j •

• "... • "- J. ~

MISCELLANEOUS SOl.2VED EXAMF?LES

Example 50. (i) Find out the coefficient ofcorrelation between X and Y from the following data:

• I ~: I ~ I : I · : I : I -: I
. , •(ii) Multiply ~a~h X value.by 2 a~d add 3: M~ltiply ~ach.val~e ·ofY by 5 and subtract
, /

' . ,
. \ i: ...4.
Find the correlation coefficient between two new sets of;values. Explain why
do or do not obtain the same result as in (i). ·• ----·
p
e6,4!_________ __:...·________ ________ __.........______~C:2°!!:rre~l!,!atio
- ---!l
Calculation or Karl Peanon's Coefficient of Correlation
Solution: ......__
y yl

-
!
X xi XY
2 4 6 36 12
2
I
4 3 9 6
.
4 16 2 4 8
-
5 I
2.S 6 36 30
5 25 4 16 20
IX• 18 u 2 =74 I:Y = 21 1:r 2
=101 Ln'=76 -
N=S
N.UY-U .1:Y
-
r=--;==== ===-r=== ====
.JN .U-2 -( !X)2 ✓N .l:Y2 -( l:Y)2
Sx76-18x21 ·
= -;==== ~ = = = = = 0 036
✓s x 74-(18) 2 ✓s x 101-(21)2 •

(ii) Let us define new variables U and V as follows:

U =2X + 3 and V = SY - 4
- We now calculate the coefficient of correlation betw~en two new sets of values
U and Vas given
,
c.. .,
X
1, ~
y V ·2x+J 'V =SY :...4 u2 vz UV
'
l -- l
"
182
2 6 7 26 49 ·- 676
!• ..

3· -121 77
2 7 11 49
-- . -. - ·-
I

66
~4 2 ~~

11 - 6 121 - 36
.
~

- 338
s - 676
~

1-
6 -- · .. 13 - 26 169
~

'
- - 169 256 208
,~
s -~
·4 -- 13 ~

16 ~

.IV= 85 .IV 2 =557 .IV 2 =1765 .IUV =871

'EU= 51
- , . . • . N .1:UV- IU .IV
r= -
·✓N .IU 2
~( IU) 2
✓N .l:V 2
~( l:V)
2
.....
S x 871-"(51X85)
= ✓s x 551-(51)2 .J5 x 116S -(85)
2

20
✓184=s=-4✓-;:1=600=
= --.
- - -----
.. - .-- - .. - - =0.036 •
t
The value of ruv is the same as that of rxy. •This is so because the correlation
coefficient is independent of the change oforigin and scale and U and V are obtained
from X and Y by change of origin and scale so that we have rxy and ruv.

cf
rrela!!.=tio: :,:.n:-----~ ---_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ___:~
~ 65

111 ple st.

Two yariates X and y wh
f:Ja given as follows: en expressed as deviations from their respective means are
I
x: 0
-4 4 -2 2
y: 1
3 -1 '
'
? 0
Find the Karl Pearson Coeffi .
ic1ent of correl t. be h
In this question one de . . . a ion tween t em.
solution: , v1at1on an " ser· . . . Le . . .
by a. We know that the sum of ., . _,es 1s m1ssrng. t us denote the m1ssmg item
So ~ dev1at1ons taken from mean is always zero.
, ~y==o
••• (1)+(3)+ a+(O)+(-l) =O
3+ a== O
.
•• a==-3

rus : : co~rte ~ies is:

Now, we find the coefficient of corre'Iation.

-4
3
4
3 0
2 2

I
-- - • t 0 f CorreIaf100
Calculation of Coefti1c1en •
X '
X
2
,I
y
(
; .xy
'-I-

0 0
~ - 0
1 . 1
-i -
-4 . • 16 3 9 -12
-
4 16 ,_ 3 .. 9 -12
-
J -~ -
' -2- '4 - 0 0 0
2 4 1 1 -2
I.x=O I.x2 =40 l:y=O 1:y2 = 20 I.xy=-26

• Applying the formula:

. -
- l:xy .- -26
r=--;::=====--===
✓u .xiy ~ ✓40x20
2 2

- •• - - 26 = - 26 = - 0.9192
· ✓80Q 28 .2843
. .
Example 52. The following table gives the distribution of production and also the -.relatively
• defective items among them, according, to size groups. Find the correlation
coefficient between size and defect in quality and its probable error. Is the value of 'r'
significant or not?
Size 1roup: 15-16 16-17 i1-··18 18-19 19-20 2~21
No. of items: 200 270 340 360 400 300
No. of defective items: 150 162 170 180 180 ~ ... 114
1
 o1rrefat·
-------------JC~ ~
------------------

In this question, as rorreJation has to be

found between size and defect in q .
Solution:
henc.c defect in quality has to be first dete
rmined as % of the defective items. uai,t}t,
C11talatfon of% of Defective Items
360 400 300
200 270 340 --...
,No. or tt,mt: i -
180 180 114
150 162 170
No. or d,fM tfve
lt•m•: --..
180 x JOO 114
162 x JOO 170 x JOO 180 x JOO -X) OO
•/4 or dtfM th1e I .SOX JOO 400 300
ttrms: 270 340 360
200 =45 =38
I
1ie75 e(,() =SO -
=SO
-
items as Y
Let us den ote the mid value of the size gro up by X and % of defective
tJy2 dxdy
X A•l 8.5 t/x2 y A=50
(MV ) dy
dx
--
25 625 1S
I JS.S -3 9 15 -
➔

' -
JO JOO 20
16.S -2 4 60
·- I SO -A 0 0 0
17.5 I J
0 0 - 0
18.S = A 0 - 0 50 -
I
-J_1
45 -' .5 25 s
19.S +1 J . -24
-12 144
20.S . +2 4
.
'J8 L I .<

Li;l -=8 94 Lixd y

uly = )8
. N= 6 . Lix =-3 LJ.x2 = 19 I•.-=•- ·-----= :
=-1 24
CJ-
-.
'
_,..w.,-

Ap ply ing the formula, - - !

- N ~.'ulxdy·- ulx . Liy • · . -:·· •

• -· .. - - -
r=
✓N. Lii 2 ~ ( 1:dx): -✓N,L iy 2 :. ( u/y j2
--
-----~
i
,., .,... ..::......... ~-
.~. - - ---·4'"
, ..;.. -.-.; • :.

•
......

__ __ ~-
·- • - ·-

,_ -~ __- _ . 6x (-1 24 )-( -3) (18 )'

\ ·- ~- ✓6x!9_::-(-~)~. ✓~_>~~~1·-:(i~)2- .-~~
•:.--:,: ,,.. 690 · : ~69 0
-74 4+5 4
= ✓1·14-9-v'5364-324- .. ✓105.Jso4o ~
727.46

_ __ _ Significan_ce ·or 'r'

Pro bab le Err or ~-E,-)
• PE =0 .67 45x l~N-,2. .- • ••• /r/ = 0.95 =35.18
, , .,_P.E... •. ,0.027 . ~·. : :..
,rJV :j

. '. ;~·- ••
:(J -(- 0.9 5) 2 ) ' . -:· • ••• . '.1: .,
--·, , .,. , -·~-- ,: .--~. • •• .· : .· : .. ·.l.rL=3S.18P.E. •. L
- - • • • -- _·". -0. 67 4~ ~ · ./6
I 1;• .._, :

- _...___-.. ....,.

~ ~ ~ I • - •• • • 0.0975 . - . .. - . '

•• As the value of I r I is more than 6 times

. . • . ~-

- •- - ---- ·· -= 0. 67 4S x- -·
: 2. 45 - - --· -..
the P.E., so_,,, is highly significant.
• j , 0.027 • ; -· . -- .... .. : - - ... -- .,,_
'• - . ..
.
. --........ . . ..... _ .
. .• .
-- •
 C

18
tion .!--- ----- ----- --.___ ___ ___ ___ ____:.._
__..- --~6~ 7
~
~

le 53. Calculate correlation COefli .

t:i•IIIP N =10, IX= 140, l:Y == ;;~cnt from the following results: • I

I.(X -10)2 =180, l:(Y - IS)l::: 215

I(X - l0)(Y -15):: : 60
For calculating correlation coem1 .1 If 2 ~vv which
5otutioD! c ent we need the values of IX 2 , , L,,/\'
•
we can detenn1ne from the values .
t 2
2
g1Ven:
(X- lO) =l: (X + I00-20X) =IX 2 + I.I 00 - 20!.X
2
==tx + Nx 100-20I.X [-:Ia=Na]
2 - - -- -
== IX + 1ooo - 20 x 140
2
2 =: tx .+ 1000 - 2800 = IX 2 - 1800
I.X -1800== 180 [-:I(X - 10)2 = 180]
•
•• tx2 == 1980 ~
2 2
!.(Y- 15) f l: (Y + 225-30Y) =U2 + I.225-30I.Y
2
=:=IY +Nx225-30IY [·:Ia= Na]
2
== IY + 2250 - 30 x 150 .
.2 =. IY2 + 22so· - 4soo = 1:r2 2-2250
l:Y -2250=215. • f:'I.(Y-15>2,_ 215]
•
••
'
•IY2 =2465 ,._/' ·
• .. t

.I(X- : lO)(Y,- 15) = I(XY- 15X- l0Y +· 150) .

- "\,..•
• .... "----~ I •' .. • •

• .=IX f- l.5U~ l0IY+ Il50 .

= fxr- ISU- l0IY + N x 150....
=.IXY-15 x 140-10 x l50 + 10-x 150
·- IXY-2100 ~ 1soo·+ 1soo . _ ..
'! 1, • ..,_ ~- :. ~ . -~ \ - ,· .:. #

=IXY-2100
IXY-2100=60'· •("/ l:(X- l0)(f- 1:5)= 60]
•
•• !XY =2160
Applying the fonnula,
N.IXY..:. IX.I f
r= ✓ N.IX2 -(IX) 2 ✓ 1'{.'fiY2 -(I;Y)2 . •.
1() X 2160 - 140 X 150
= ✓l OX 1980 - (140)2 ✓l OX 2465 :- (150)2 ~
. 21600-21000
.
= .J19800 -'l 9600.J24650 - 22500 .
'•

- = • 600
,690+ 0.915 •= I

~ .J200.x21so 65.5'.74
 C •
68- - - - - - - - - - - - - - - - - - - - - - - - - - - - . ; . . ; ; . . ; ; . . ~ t t i · t \..
-
~
Alitc-r: r can also be calculated in the following manner:
x = IX
N
=!iQ= J4. f =~= ~=1 5
110 ' N 10
Thus.dcviatkmsl:(X -lO)a ndl:(Y - IS)arefromassumedmeans(Ax=IOandAJI ===ts)
• . •
Let, uix =l:(X- 10)= IX-I lO= Lf - N .JO= 140-I 0x 10= 40
r.dy =r( 1' - 1?) = If - LI 5 =IY - N .15 =150 - 15 x l 0 = 0
I drdy =l:(X- J0)(Y -15)= 60, Llx =180, Lly =215 (Given)
2 2

Applying the fonnula,

.N. ulxdy - L1x . Idy
r=-;====:2 =====:=-r ==== ~===
JN . Lix 2 2
-- ( L1x) ✓N. 'Lly - ( !dy) 2
10x6 0-40x 0
- ✓JO X 180-(40) 2 ✓10 X 215-( 0)2
600
600
= .Jioo x 2150 =655. 744 = 0•915
• •

Example 54. ed:

~n two sets ofvariables X and Y wi~ 50 items each, the following data were observ
X=IO , ax =3,Y =6,ay =2,r= 0.3,N =50 •
e
Howe ver, on subseq uent verification it was found that one value ofX(= I0) and-on
pairs
value ofY(= 6) were inaccurate and hence weeded out. With the r~maining 49
of values, how is the original value of corr~Jation coefficient affected?
Soluti on: Given : N = 50 X = Io y = 6 a = 3 a = ' 2 r = 0 .3 'y
, ' 'X

- ll - ,
... X=· -=> il=N X
N
... Incorrected rx = N X = 50 xi 0 •500 ...(i)
- . ty -
... ·:-~ N- ⇒ l:Y=NY
-Y=

... lncorrected l:Y = N Y = 50 6 = 300

..
x ...(ii)
2
...
a2
X N -
= IX , -
(X)~

=> l:X 2 =N(a ;+X.~ ) ··•· • [Fonn ulaofv arianc eofX]

• , • 6··

2
Incorrectedl:X ~ N(a; +·x )=50( 9+100 )=~45 0
2 ...(iii)
.•

a2 = 1:y2 -(Y)2
y N .2 . .

~ _l:Y
2
=N(a; +f ) • [Fonn ulaofv arianc eofY]
2 2 ...(iv)
Incorrect~dl:Y = N(a; +Y ) _50(4+ 36)=2 000
1stion ________ ________ ________ ________ _____....=-
69
~
We know,
,. == £ov.(x ,I')
0 ... o,, -
r.o .t .o,., == Cov. (x, }').
... Cov. (X, Y) == ...!__ I.( X x- - I - -
N' - )(Y-Y)=-IX Y-XY
. ,1 N
.. r.cr,x,O ~ == :v-UY-x.f
... l:XY == N[r.o .t•o,, + X Y]
l:XY == S0[0.3 X 3 X 2+] 0 X 6)
= so [l.8 + 60]
= so [61.8]°= 3090
Incorrected l:XY =3090
Thus, we have the following incorrect values:
2
IX=SOO, I.Y=300, I.X =5450, I.Y 2 =2000 IXY=3090
d • • 49
•• . out the incorrect values' the corrected values for the remaining • pairs
After ropptng
of items are given as: .
Corrected values:
Corrected U = 500- 10 = 490
Corrected I.Y .= 300-6 == 294
2
Corrected l:X = 5450-102.= 5350 · '•
2
Corrected I.Y = 2000.•:~,,. .62. = 1964
.
• -. •. .
; ( ;

Corrected UY= 3090 - IO x 6 = 3030

N = 49 - -- . .__ ·- .
4 • • • - •• ~

- Using these corrected values, we get • • • -

•

-- ·-· • - ~ - IX.I.Y
N
r = --;======-2 ;======;;=
• fu 2 _ (I.X") 2
(I.Y) _ I:Y2 _ _
V N- ,· N
'. 490 X 294 ·-
3030 - _- • I - •

49
- --;::==~====-r2 =======:=:=~ 2
• ( 490) (294)
535{\---. 49 .. 1964_- 49

._ 3030-2940, 90 = ~ 0.3 • ··
- .J450 ✓200 300 --
• Hence the correlation coefficient is unaffected in this case.
>
10
----------------------------~Correiat·
~

' Eumple 55. "lf two variables are independent the correlation between them is zero, but th
1

e
converse is not ahvays true." Comment.
Solution: If X and y are two independent variables, then the covariance between them •
Cov(X ,Y) Th ·r X d y . I.e.
Cov ( X, Y) = 0 and hence rxy = - - - - = 0. us, 1 an are mdependent,
a x·o y
they are uncorrelated.
Th~ converse of this property implies that if rxy = 0, the~ X and Y may not necessarily
be independent. To prove this property, let the two vartables X and Y are connected
by the relation Y = X 2and consider the following data:
X -2 -I 0 1 2 3 lX=O
-3
y 9 4 1 0 I 4 9 LY=28

XY -27 -8 ' -1 0 1 8 27 l:XY=O

Here, IX = 0, LY = 28 and IXY = O

. 1 U' l:Y 1 •0 28
•• Cov(X, Y) = - IXY--·-=-·(0)- -•-=0
N ·N N 7 · 7 7
Cov (X,Y) •
Thus, rxy=----=0
ax.cry

A close examination ofthe data would reveal that although r xy = 0 but2X and Y are not
independent. In fact, the variables are related by the equation Y = X , i.e., there is a
quadratic relation (i.e., non-linear relationship) b~tween the variables. This property
implies that rxy is only a measure of the linear relationship
. between X and Y. If the
relatio~ship is non-linear, the computed_ yalue of rxy is no longer a measure of the
d~gree of relationship between the two vadables.

r--------~IMPO RTANTFORM ULAE.,__ _ _ _ _ __

A. INDIVIDUAL SERIES
► 1. Karl Pearson's Coeffici~nt of Correlation (When deviations are taken from actual mean):
r= :r.xy or 1:xy
.Jrx 2 x 1:y 2 N.cr x .a Y
Where, x = ( X - X ) y = (Y - Y )
► 2. When deviations are taken from assumed mean:
r = N."uixdy-"uix.uiy ~
✓ N.Edx 2 -(Llx) 2 ✓ N."I.dy 2 ·-(uiy) 2
Where, dx = (X -A) and dy = (Y -A)
 ~ . -_-_-_-_-_-._....!7;1
~1atiO J':c!-;-;_;_:~:-~~~~~~~~~~~~~~~---------------------------_-_-.-_-_-_-_
•'hen we use actual \'aluts or X
3. '" •nd Y:
r=--- =~ ~tx.ty
~(-C-")2 r.r .~~ =
"N.t r 2 -(I.Y )l '"-1.
.
\

ri d'hen we are gi\'en Variance and C

4. "
r=
Cov (X
' f) ·
O\'ariante of X and Y·

✓var(X)~

where, Cov( X 'y) = *' I.( X - X) ( y - f) == i-I.X Y - X. y

OUPED SERIES
S, GR - .
s. In a Bivanate or Grouped Frequency Distribution:
r= Nt/dxdy -. I/dxifdy
2
•' .J N.r,fdx - (I/ciii}2 .[N.(Ifdy)2 _ (I/dy)l
\
~
/
__,.-
_

► 6~Spearme,'s Rank CorrelatioQ -Coefficient: ·

(i) When actual ranks are given:
) 6!J)2
R = l -3~ -
• N -N
(ilj When ranks are not _repeated
R=l - 6!JJ2
3
N -N
(iii) When ranks are repeated •. . · .

. • ·.
• 6[~ 1'\ 2 +-
~
. · 12 m
•1(. 3
1 ( 3. -m• ) +-
12 m
\..1...
-m,- . ...
1
R·= 1- - - - - - -3- - - - - ~
N -·N
7. Concurrent Deviation.Method

re =+- +(_U:-_n)
-
n
.

8. Probable Error and Standard Error

l-,2
P.E.::::0.674Sx .JN
l-r2
S.E., = .JN
·N
9. Coefficient of Determination
,:Z = Explained Vatiance
Total Variance
72
- - - - - - - - . Correlat·
-------------------------- ~

• QUESTIONS
What is the signific-!l..
l. Define correlation. Explain the various methods of studying correlation. • '"""ice
of study ing correlation?
s signify cause and
l. What is correlation? Explain various types of correlation. Does it alway
effect relationship between the two variables?
3. Define Pearsons' coefficient of correlation. I~terpret r when r ~ 1, - 1and
0.
preferred to Kart
4. Define rank correlation coefficient. How is it measured? When is it
Pearson's coefficient of correlation? .
S. What is meant by coefficient of concurrent deviation? How is it measured?
6• What is scatter.diagram and how is it useful in the study of correlation?
7. Explain the followings: •
(,) Probable Error (ii) Coefficient of Detennination.
8• Explain the properties of correlation coefficient.