470 BUSINES STATISTICS

or=1, (1H 25+. + 4) =

((1.8 3M2.5-3)+(2.53)(31- 31+ 3-
+(3-3)(12-)+(12-h(3.1-))
8-3 +(2.5-3+ (31-3+(3-3 +(1.2- 3
(-12)-0.5)+(-0.5) (0.1) +(0.) 0)+ () (1 2) +(1,2h01
1.41+0.25+0.01+0+1,410.16
)6- 0.5+ 9,4) 0.312
3.3
For k = 2

(1.8 3)(3.1 3) +(2.5- )(3-+(3.-31.2- 3

-

(18-3 +(2.5-3+.. +(3.4 3

(-1.2x0.1) + (-0.5 0) +(0.11 2)+ (00,4)
3.3 -9.12012
Since the value of r, is 33
positive, it implhes thar there is a
seasonal
months duration and r, 0 implies that here is no
atterm
sign1firant
rant rh
rhangr

Self-P ractice Problems 13B

13.11 The coefflicient rank correlation of the marks
of
the
obtained by 10 students in statistics and accountancy applicants in tie accointancv and
statistics
was found
be 0.2. It was later
to compute the rank correlation p

difference in ranks in two discovered that the coefficient,

of the students was subjects obtained by one Applicant A
B C DEF
Find the correct wrongly taken as 9 instead of 7. Marks in
coefficient of rank correlation.
accountancy: 15 20 28 12 0
13.12 The
Delhi Univ, BCom, 1996] Marks in 60 2
ranking of 10 students in
accordance with their statistics
performance in two 40 30 50 30 20
subjects A and B are as follows: 1 3
13.15 Seven methods of
A: 6 5 3 10 2 49 7 8 ranked inparting
by the MBA studentsbusiness educaton
B:3 8 of two
4 universine
9 1 6 10 7 5 2
follows:

Calculate the rank correlation coetficient Method of Teaching I

on its value.
and comment 3 45 6
Rank bv students
13.13 Calculate
Spearman's coefflicient
marks assigned to ten students of correlation
between
of Univ. A
13 3
by judges x and y in Rank bv students
a certain
compeitive test as shown below: of Univ. B
Student Marks by Judge x
I32 7
Marks by Judge y
Calculate the rank correlation coethcient and comn
on its value.
52 65
2 13.16 An investigator collected the
53 68 the
tollowing data wiuh resy
3 42
to
soo-cconomme tatus and severty of
43 illness. respt
60 38
45 77 Paltent 3 3 6
41 Socio-ccononn
48
37 st.atus (rank)
35
8 38 Sevemv of dloess
30
9 25 rank)
25 37
10
50 Calulate the 1ankmtetatO1 Letli
13.14 An examination of eight 01 SValie
ient and come
applicants for a clerical post
was 1aken by a firin. From the
marks obtained bv
 CHAPTER 13 CORRELATION ANALYSIS 471

a of marks obtamed coclicent to

llowingdata ilculate appropriate corelaton
an
ww e nhe
Iwo tests, one before ncasure the consisteney in the ranking.
tatistics

al coaching 3 . I 9 In an office some kevboard operators, who werc

Second Tst aleady ranked on their speed, were also ranked on

(4flr coaching) uracy by their supervisor. The results were as

lollows:
Hiny

24
19
Operator ABC DE F GHIJ
22
Speed : 2 3 45 6 78 910
18
20 Accuracy 79 3 4 68 2 10
22
20 Calculate the appropriate correlation coefficient
22 between speed ancd accuracy.
13.20 The personnel department is interested in
20
20 comparing
20 the ratings ol job applicants when measured by a
23 variety of standard tests. The ratings of 9 applicants
20 on interviews and standard psychological test are

b 17 shown below:
19
indicate that
the special coaching has Apphicant AB CDE F GHI
Delhi Unin, MCom, 19891 7
5 2 9 43 6 1 8
marks

the students?
Interview
ranked a lew trainees 8175 3 42 9 6
enehied
managers Standard test
depart1men

abilities. The ranking

their perceived and
Calculate Spearman's rank correlation coelficient
o r d n gt o

a eg r v e n below:

commen on its value.

A BCDEF G H I
natmer

96 2 5 87 3 10 4
A:1
Manager

3 10 8
I75 6 9 4
Manager B

Hints
and Ans wers

I-02 or
=
10: R =

1
=
0.2, n
n(n -1) 13.14 R = - .

l Given R n(n-1)

02
6
og00-1)
d or 2d = 100

=1-
153-3 8(64-1)
=

Correct value of R
= l. 6x100=
10x 99
0.394 13.15R = 0.50
13.16 R = 0.477

13.18 R 0.842
13.17 R = 0.71
6x 36 13.20 R 0.817
6d l
n(n - 1) 10(100-1)
= 0.782
13.19 R = 0.006

6x 76
13R =1 -
62 =
= 0.539

n(n1) 10(100-1)

CORRELATION COEFFICIENT
HYPOTHESIS TESTING FOR
to test whether the
estimator
r as an
correlation coefficient in the population
LSe the sample random
variables
Aible association between two
null and
alternative

n h ol r as an
estimator in testing
letter rho). When
ET words, we
use

Potheses about true population correlation coefficienl p(Greek two random

distribution
of
hypotheses
a rare tested, the
e tested, ne assumptions of normal
4Dles, say x and y is
required
498 BUSINESS STATISTICS

departent. In order to do th s
Delhi Mumbai Sample l J0 applie ants. Th
lCr assigncd a
Prodution atin
rt a
Average price pcr kilo (Rs) 2.463 2.797 e
lollows:
Standarcd deviation 0.326 0.207
A

Correlation cocflicient Worker A .

betwecn prices at Deli Testscore5 88 84 86
and Mumbai = 0.774 P'roduction
rating 45 13 89
Estimate rom the above data the most likcly price (a)a 79 84 G6A
Dchi corresponding to the price of Rs 2.334 per kilo at Fit
lncar east Squares
a

Mumbai (b) at Mumbai corresponding to the piice

ol reglessi
rating on tcst score. [Delhi
production
3.052 per kilo at Delhi 14.9 Find the regression equatinn sheqwi.Unu., MRA
14.5 The tollowing table gives the aptitude test scores a
utilization on
production irom the fole
productivity indices of 10 workers selected at radom
ollowing da
Aptitude
scores (x) 60 62 65 70 72 48 53 73
65 82

Production
AueTageani
Jevulon
Productivity
60 8I
index () 68 60 62 80 85 40 52 62 (in lakh units)
esumale (a)
Calculate the two regression equations and Capacity utilization
the productivity worker whose test score is
index of a
(111 percentage)
92, (b) the test score of a worker whose productivity 84.8
Correlation coellicient r = 0,62
index is 75. [Delhi Unin, MBA, 2001]
14.6 A company wants to assess the impact of R&D Estimate the producuon when the capacity uli
annual profit; (Rs in
expenditure (Rs in 1000s) on its 70 per cent.
inlormatüon
1000's). The following table presents the (Delhi Univ., MBA, 1997; Pune Univ,
for the last eight years: 1BA
14.10 Suppose that you are interested in
expenditure R&D by a lirm to
on Using p
Year RGDexpendiure Annual pro/it predit
expenditures on R&D. You got the followin d
c

9 45 random
1991 taking a sample of lirms, wherex is the am
1992 42 spent on R&D (in lakh of rupees) 5 years ago

1993 41 the amount spent on R&*D (in lakh of rupesl

60 current year:
1994
* 3020 50
80 10 20
1995 4 30 20
1996 5 34 50 80 30 110 20 20 4
25 (a) Find the regression equation ofy on x.
1997
20 (b) Ifa firm is chosen randomly andx = l0, can a
1998
the regression to predict the value of y: Dueu
Estimate the regression equation and predict the annual Madurai-Kamraj Univ, MB
profit for the year 2002 for an allocated sum of Rs 14.11 The
1,00,000 as R&D expenditure.
following data relates to the scores obtainei
salesmen of a company in an intelligence test ant
Jodhtur Uniu, MBA, 1998) weekly sales (in Rs. 1000's ):
14.7 Obtain the two regression equations from the following
bivariate frequency distribution: Salesman
intelligence A BC DEFGH
Sales Reveue Adverlising Expenditure (Rs in thousand) Test score 50 60 50 60 80 50 80
(Rs in lakh) 5-15 15-25 25-35 35-45 Weekly sales 30 60 40 50 60 30 70
75-125 8 (a) Obtain the regression equato
of

125-175 mtelligence test scores of the salesinen

175-225 4
(b) If the intelligence test score of a sur
slire
225-275 what would be his expected weekly
2 2 H P n e , Mo,

Estimate (a) the sales Son eyui

corresponding to
advertising 14.12 Two random variables have the regre
expenditure of Rs 50,000, (b) the + 2y -26 =0 and iv +y-
expenditure for a sales revenue of Rs
300
advertising
(oefficient of correlation. lakh, (c) the andi
Delhi Uni., MBA, 2002| (a) F'ind the values ol x andy
mean
14.8 The personnel manager of Correlauon betwecn x and y.
an electronic lind the
company devises a nanual test lor
manufacturing (b) If the varaince of x is 25,
hcu
job applicants to
predict their production rating in the deviation of y from the data.
assembly /MD Univ, M.Com., 1997: Kumau