STATISTICAL METHO

178

INTRODUCTION is to get
objectives of statistical analysis
entire One
important the
One of the most describes the characterlstic of
value or an
mass o
single value
data.
that
Such a valuevartable. The
central
Is called the word average ls very or 'average
commonly
unwieldy value of the we often talk of
the expected
used in day-to-day
conversation.
average height or
For
life
example,
what it
average
of an Indian, means 1sincome,
that ett,
he ts
average
boy in a class, student' student.
an average of
When we
neither
say
very
'he
good
is
nor
a mediocre
very bad, justdífferent
type
meaning. However,
term average has a
in statistics the AVERAGE DEFINED
Some
defined differently by various authors.
has been
The word 'average'are given below:
important definitions
single figure to describe whole of fiqures."
one Clark
"Average is an attempt to find
represent
selected from a groUp of values to group, of
value
"An average is a singlevalue which is supposed to stand for whole
them in some way-a -A.E. Waugh
part, as typical of all the values in the group. to
which it is a
in the sense that it is sometimes employed
typical value
"An average is aindividual variable."
values in a series or of a -Y>Lun-Chou
represent all the
typical of the whole
is Sometimes described as a number which is Leabo
"The average
group." within the range of the data
that is used to
value is a single value somewhere within
" "An average values in the series. Since an average is
represent all of the central value."
also called a measure of
the range of the data, it is -Croxton& Cowden

value
definitions that an average is a singde
It is clear fron the above great significance
that represents a group of values, Such a value is of Since an average
the whole group.
because it depicts the characteristic oflles somewhere in between the two
represents the entire data, its value
smallest items. For this reason a1
extremes, Le., the largest and the central tendengy.
average is frequently referred to as a measure of
OBJECTIVES OF AVERAGING

There are two main objectives of the study of averages:

(0 To get singlevalue that describes the characterlstlc of the entire gro

Measures of central value, by condensing the mass of data in one sing
value, enable us to get a bird's-eye vlew of the entire data. Thus
value can represent thousands, lakhs and even milions of values.
example, it is impossible to remember the individual incomes of m
hardy any
of earningif people of India and even if one could do it there is total natlonal
use. But the average income is obtained by dtviding the represents the
income by total populatlon we get one single value that standard of
entire populatlon. Such a Agure would throw light on the
living of an average Indian.
 MEASUREs
OF CENTRAL VALUE

(0To Facilitate Comparlson Measures of central value. by reduclng

the mass of data to one single Agure, enable comparison to be made.
Comparison can be made either at a potnt of
time or over a of
timc. For example, we can compare the percentage results of the perl0d
students
different colleges in a certain examination, say,
thereby conclude which college is the best or we can
B.Com. for 2011, and
percentage of the same college for different time compare the pass
Conclude as to whether the results are improving or perlods and thereby
omparisons are of immense help in
framing suitable deteriorating.
and
Such
For
example, if the pass percentage of students in tirnely policies.
80in 2013 and 75 in 2014, the authorities have College A in B.Com.
Imvestigating the possible cause of the sufficient
However. whle makng deterloratlon in results. reason for
0onsideration the multiplicity comparisons one should also take into
of forces that míght
For Cxample, if per capita income is rising in absolute be affecting
terms the from data
one
period to another, it should not lead one to think that the
Iving is necessarily improving because the prices standard of
might be rising faster
than the rise in per capita income and so in real terms people
WOrse off. Moreover,
M the same measure should be used for might be
Comparison betweern two or more groups. making
For example, we should not
Compare the mean wage of one factory with the
factoryfor drawing any inference about wage levels.
medlan wage of another
REOUISITES OF A GOOD AVERAGE
Snce an average is a single value representing a group of
iesired that such a value satisfies the following properties: values, it is
OEasy to Understand Since statistical methods are
stmplify complexity, it is desirable that .an average be such designed
that can
to
Tadily understood; otherwise, its us is bound to be very be
limited.
Simple to Compute An average should not only be easy to
but also simple to compute so that it can be used understand
e of computation is desirable, it should not be widely. However, though
oher advantages, Le., if in the interest of sought at the expense of
greater
dicutt average is desirable, one should prefer that. accutacy, use of a more
t) Based on all the ltems. The
Sery item of the serles so that if average
any of
should depend upon each and
the
Self is altered. For example, the items is dropped the average
+20 +30 + 40 +50 150 arithmetic mean of 10, 20, 30, 40, 50, is.
5 5 = 30. If we drop one item, say, 50, the arithmetic
an would be = 25.
tu) Not be
Fery tem undulyinfluence the
should
Affected by Extreme Observattons Although each and
value of the avcrage, none of the items
ould
nduly influence unduly. If one or two very small or very large items
affect the average,
aue,ords, Le., either increase its value or reduce its
average cannot be really typical of the entíre series. In other
extremesmay distort the average and. reduce its usefulness.
Yule and Kendall : An. Introductlon to the Theory of Statistics.
180 STATISTCAL METHr
(u) Rlgidly Defned An average should be properly defined so that
has one and only one Interpretation. It should preferably be defined
algebraic formula so that If d1fferent people conpute the average from th
same flgures they all get the sane answer (barring arithmetical mistake
The average should not depend upon the personal prejudice and blas
results can be misleading.
the investigator,
otherwise the
Algebralc Treatment We should prefer to ho
(0r Capable of Furtherbe used for further statistical computations
data s0 tha
an average that could the
iis utilty is enhanced. For example,
if we are-
of
given
two or more about
factores h
Tage ncome and number of employees
combined average.
2ould be able to compute the
least, we should
Stability Last, but not the 'sampling stability.prefer
This
to e
(vi) Sampling statisticians call
value which has what the
different groups of colege students, and compute
that if we pick 10 expect to get approximately the
we should can be no difference n
average of eachnotgroup,
mean, however, that there
those sam
Value. It does
samples, There may be some differernce but is less
values of different
difference (technically called sampling isluctuation)
more.
in which this than those in whích this difierence
Considered better
TYPES OF AVERAGES
types of averages:
The following are the important and (i) weighted.
simple,
" Arithmetic mean: ()
Median
Mode
Geofntric mean
" Harnonic mean
are, less important averages like moving average
Besides these, there have a very limited field
These averages
progressive average, etc. popular.
application and are, therefore, not so
Arithmetic Mean
popular and widely used measure of representing the enti
The most
one value is what most laymen call an 'average' and bywhat
adar
data byy arithmetic mean, Its value is obtained
call the
statisticians
the items and by dividing thls total by the number of items
together all
Arithmetic mean may elther be
(0 simple arithmetic mean, or
() welghted arithmetic mean.
becaus
do not like the term 'average D
" It should be noted that the statisticians
different meanings: for example.eltheran mes
has too loose a connotatdon. It has
etc. It can refer to practi
prson, an average wage, an average height, or any other average. witho
median, mode, geometric mean, harmonic mean'mean alone wh
or average That Is,
arnthmetic mean is so popular that the word
qualificatton 1s implled to denote this particular type of average. t may.
'the mcan' or 'the average' of a series of observations, ted.
 factortes me compute sa in lesS We
what w may,stated.
so(u)
Rigidly and
average computatlons This thediference the by of an elther
1s,
In alone
defined arithmeticalprefer
(Ur data
those
is
approximately moving arnd nunber 'average'
Thatit
average.observations,
be luctuation) examnple,
preferablyprejudice more stability.
representing
obtained
is to otherwise
and but average.
average
should 'average' refer
the shouldthe more.
properly misleading. or
compute given
no
students,
be
difference like
very the term
forcanother
unless
personal
(barring statistical 'sampling is or of
two sampling averages
a by themeanings: of
should average. get can difference averages: of an value
Its It any'mean
typeseries
We least,
the
not weighted. 'have total ike
are of
etc.or 'meant
be people be Treatment college some
to there call
popular. measure mean particular
we employeescall not helght, word aof
should answer the can further expect called important
averages this different is
It
interpretatlon. 1f combined be this of (i)
laynen do
harmonic
the
average'
mean
upon of that
statisticians dividing statisticians
d1ferent results example, but groups
may which types
and so used
mean average this
that
average same Algebratc for
used
be should (technically
however, not most or mean.
has arithmetic
'the
denote
depend of
the Last,
There simple,
important These
less arlthmnetic It mearn, popular
the number in therefore,widely by be mean, an or
if the For compute different those what elther connotation.
the wage, tomean'
An thatget otherwise the we mean, samples, are and arithmetic geometricthe
not Further enhanced. Stability differenceTYPES ()
AVERAGES
OF
etc.
and arithmeticthat so fmplled
so all should couldand what group, than the mean: there
Geómetric
mean
"
is items may 'the
that
10 mean are,
average,
Harmonic the
value notedaverage is
they
formula
only notdifferent popular mean of assumed
ofCapable to
investigator, ncome Sampling
has pick better are these, Mean call mean a mode, is speaks
that able each Arithmetic and the weighted loose
qualificaton
average average
and figures does this following one simple be an arithmetic
is which Median application
progressive all
Arithmeticstatisticians should
algebraic utilitybe we of Considered Mode Besides most Arithmetic too median,
person, be
anyone
one "Crage
ould (ut) if average It of which by together rule;
values has
same value value.
that The Thedata (0 () It
has Thene aniis in "
180
 MBASURES OF CENTRAL VALUE 181

Calenlation of Simple Arithmetie Mean--Individual Observations

Theprocess of computing mean in case of individual observations (Le.,
where frequencites are not given) is very simple. Add together the varlous
values of the varlable and divide the total by the number of Iterns.
Symbolically:
Fe-1 + X, + Xg +... +X, or X X

.Ñ= Arithmetic Means, EX = Sum of all the values of the varlable

X le, Xj.
Xg. Xg.,..., Xn: N= Number of observations.
Steps. The formula involves two steps in calculating mean:
8 Add together all the values of the varlable X and obtain the total,
Le., EX.
6 Divide this total by the number of the observations, Le., N.
mustration 1. The following table gives the monthly income of 10 employees in an office:
Income (Rs.) 14,780 15,760 26,690 27,750 24,840
24,920 16,100 17,810 27,050 26,950
Calculate the arithmetic mean of income.
Solution. Let income be denoted by the symbol X.
-i, CALCULATION OF ARITHMETIC MEAN
Employees Monthly Income (Rs.) Employees Monthly Income (Rs)
X X
14,780 6 24,920
2 15,760 7 16,100
26,690 8 17,810
27,750 9 27,050
24,840 10 26,950
N= 10 EX = 2,22,650

= EX - 2,22,650, N=10
N.'

X222,650 = Rs. 22,265.

10

Henca the average income is Rs. 22,265.

Short-cut method., The arithmetic mean can be calculated by using
What is known as an arbitrary origin. When deviations are taken rom an
Ditrary orlgn, the formula for calculating arithmetic mean is
X=A+22
N
The arithmetic
arthmetc mean mean
of a
sample ls destgnated by the symbol X which is read 'X bar and the
of a populatlan is destgnated by the Greek Letter pronounced as mue'.
Thath nbol tò s the letter capita sigma of the Greek alphabet and is used tn
mathemattics denote the sum of the values.
where A the aeeumed mean and d is the devtatlon
A).
mean. ie, d (X-
seerse
asSumed mean.
0) Take an deviations of Items from the
(21 Take the
deviations by d.
assumed me an
these
sum of these devtatlons, Le., Ld nd
(3 Obtatn the 2d
Apply the formula : R-A N
(4) arithmettc mean by
Ihustratlon 1 calculate
From
assumed mean.
CALCULATION OF AAITHMETIC MEAN
taking 22.000
Solution. Income (Rs.)
Employees X
14,780
(X-22000)
d
-7220
15,760
2
26,690
-6240
+4690
3
4 27,750 +5750
5
24,840 +2840
6 24,920 +2920
7 16,100 -5900
8 17,810 -4190
9 27,050 +5050
10 26,950 +4950
N= 10 Ld = 2650

N
A = 22000, £d =2650, N= 10
2650
X = 22000 + = 22000 + 265 = Rs. 22,265.
10
Hence the average income is Rs. 22,265.
na
Note. The reader will find that the calculations here are more than what we
we used the formula

N
considerable saving in n
This is true tor ungrouped data. But for grouped data
possible by adopting the shon-cut method.

Calculation of Arithmetic Mean-Discrete Serles

applytng
In discrete sernes arithmetic mean may be computed by
(0 Direct method, or
(L0 Short-cut method.
assuned
the
assumed
as
" Any vaue whether edsting in the data or not can be taken the.
and the hnal answer would be the same. However, the nearer
to the actual nean, the lesser are the calculations.
 MEASURES OF CENTRAL VALUE 183

pirect Method

The formula for computing mean is

-here. f Frequerncy: X = The variable in questlon; N° - Total number of

observatlons, i.e., Ef.
Steps:
n Multiply the frequency of each row with the varlable and obtain the
total EfX.
ià Divide the total obtained by step (0 by the number of
i.e., total frequency. observations,
lstration 2. From the following data of the marks obtained by 60 students of a class.
calculate the arithmetic mean:
Marks No. of Students Marks No. of Students
20 8 50 10
30 12 60
40 20 70 4

Solution. Let the marks be denoted by X and the number of students by f.

CALCULATION OF ARITHMETIC MEAN
Marks No. of students
X f fX
20
160
30 12 360
40 20 800
50 10 500
60 6 360
70 4 280
N= 60 EfX = 2,460

=41
N 60
Hence the average marks = 41.
Shórt-cut Method According to this method,

Efd
X=A+ N
where A =Assumed mean: d = (X - A): N = Total number of observations,
Le., 2f.
ereader should note carefully that in discrete and continuous frequency distributions the
oal number of observations, Le. N= the sum of frequency or N= .
184
ST ATISTCAL,
Steps:
mean.
(0 Take an assumed
(0 Take the deviations of the varlable X
Xfrom the
denote the deviations by d.
(t0 Multiply these deviatlons
total fd.
with the

(tu) Divide the total obtained in third step by the

respecttve AsSUmed mean
r teoqtuaelncy and ta,
llustretion 3. Cakculate arithmetic mean by the short-Cut
distribution of illustration 2. method frequUSéninegy. tre
Solution. CALCULATION OF ARITHMETIC MEAN
Marks No. of Students (X- 40)
X
8 -20 fd
20
30 12 -10 160
40 20 120
0
50 10 +10
+100
60 6 +20
4
+120
70 +30
+120
N= 60 2fd = 60

X = A+ Efd
N
= 40 + = 40 +t l = 41

Calculation of Arithmetic Mean Continuous Series

In contnuous series, arithmetic mean may be computed by applying:
of the folowing methods:
(0 Direct, method,
(t Short-cut method.

Direct Method When direct method is used

X=
N
of each
where m° = mid-point of varlous classes: f= the frequency
N= the total frequency.
the
used
Note. Por the sake of clarity and ease of understanding We havethroughoutungr
serles dataandgrouped
fndividual observations, discrete serles and continuousgrouped
However, the reader should be familiar with the terms
observatlons whereasS
data' also. Ungrouped data refer to the individual
refer. to continuous series and the discrete serles.
Lower imit + Upper limt
MId potnt 2
MEASURES OF CENTRAL VALUE 185

Steps:
Obtain the mid-point of cach class and denote It by m.
Multiply these mid-points by the respectlve
(0 and obtain the total Efn. frequency of each class
Divide the total obtained in step (0 by the sum of the frequency.
Le., N.

Mustration 4. From the following data compute arithmetic mean by direct method:

Marks 0-10 10-20 20-30 30-40 40-50 50-60

No. of students 10 25 30 20 10

Solution. CALCULATION OF ARITHMETIC MEAN BY DIRECT METHOD

Marks Mid-points No, of Students
m f fm
0-10 5 5 25
10-20 15 10 150
20-30 25 25 625
30-40 35 30 1,050
40-50 45 20 900
50-60 55 10 550
N= 100 Efm = 3,300

3,300
N = 33.
100

Short-cut Method When short-cut method is used, arithmetic mean is

computed by applying the following formula:
X = A +
N
where A = assumed mean; d = deviations of mid-points from assumed
mean, Le., (m - A): N = total number of observations.
Steps:
(0 Take an assumed mean.
( From the mid-point of each class deduct the assumed mean.
Multiply the respective frequencies of each class by these deviations
and obtain the total £fd.
D) Apply the formula: X = A +
N
Calculate ithmetic mean by the short-cut method from the data of
lustration 4..
 ST ATTSTCAL
CALCULATION OF ARITHMETIC MEAN
Solution. No. of Students (m -35)
Mid-points
Marks
5 -30
5
0-10 10 -20
15
10-20 -10
25
25
20-30
30
35
30 40 20 +10
45
40-50
10 +20
50-60 55
N= 100

£fd = 35 200
=35 - 2 = 33
X= A+ N 100

In order to simplify the calculations, we can divide the devatlan

class intervals, ie., calculate (m - A)/i and then multiply by tin
becomes:
formula for getting mean. The formula
X= A + N

It may be pointed out that when class intervals are unequal R:

stmplify calculations by taking a common factor. In such a cag
should use (m - A)/Cinstead (m - A)/iwhile making calculations.
From the data of }!ustration 4, compute arithmetic mean by s
devlation method.
Solution. CALCULATION OF ARITHMETIC MEAN
Marks Mid-points No. of Students
m
(m - 35) (m - 35)/10
f
0-10 5 5 -15
-30 -3
10-20 15 10 -20 -2
20-30 25 25 -10 -1
30-40 35 30
40-50 45 +20
20 +10 +1
50-60 55
+20
10 +2
+20
N= 100

X= A+zfdN X|=3520
100X 10 = 35 - 2= 33.
 CENTRAL VALUE 187
MEASURES OF

clear from above that all the three methods of fInding arithmetic
IIis continuous series give us the same answer. The direct method.
meanIn
though he simplest, involves more calculatlons when mid-points and
frequencies are large in magnitude. For example, observe the following
data:
Inoome(Rs) No. of Persons Income (As.) No. uf Persons
40000-50000 368 70000-80000 567
50000-60000 472 80000-90000 304
60000-70000 969

In this case step deviation method would be simpler. In fact, step

deviation method should be adopted wherever possible because it
minimises the calculations.
uhile computing mean in continuous serles the mid-points of the
various classes are taken as representative of that particular class. The
reason is that when the data are grouped, the exact frequency with
ich each value of the variable occurs in the distribution is unknown.
We only know the limits within which a certain number of frequencies
OCCur. For example, when we say that the number of persons within
the income group 20000-25000 is 50 we cannot say as to how many
nersons out of 50 are getting 20500, 21700, 23000, etc. We, therefore,
make an assumption while calculating arithmetic mean that the
frequencies within each class are spread evenly over the range of the
class interval, Le., there will be as many itemns below the mid-point as
above it. Unless such an assumption is made the valåe of mean cannot
be computed.
This assumption is likely to lead to some error. As a result thereof the
mean of a number of observations calculated from a frequency
distribution will generally be only an approximation to the mean
calculated from the original data. However, the possibility of compensating
errors must be considered. Some of the mid-points er by being too low
and others err by being too high. In general, then mid-points of the
classes below the class containing the arithmetic mean tend to be too low
and the mid-points of the classes above the class containtng the
arithmetic mean tend to be too high. It is quite possible, therefore, that
When the errors 0ccur, those which are too low wll offset, in part at
least, those which are too high, so that the arithmetic mean for the entire
astribution
lIist of values.will be approxlmtely of the same value as is obtained from a.
lustration 5. Calculate arithmetic mean from the following data:
Marks 0-10 10-30 30-60 60-100
No. of Students 5 12 25 8

Solution. The class intervals are unequal but stil to simplifty calculations we can take 5as
ne common factor.
 and twof
o
lustration Hencethe correct
score.
the
items .. Correct But iustration
Solution:discovered for
Çorrect Since observations.correct
calprocess
culaOvertingsight
this Correctngadjuscaltmcentulating Note.
are EX is
7.
of We that 6. items is is When
correct X not 60-100 30-60 10-30 0-10 Marks
wrongly Mean = 4000 = =
Incorrect correct
are aThe y or
vermean. made.
arithmetic data
Correct
3970 X Score The Incorect
average given mean
taken N - siamnd ple. mistake However, are
100 83 2X of result, The
+EX- EX 53 then meant given
as 53
Nmarksof
in 80 20 m
observations = =
100 = was problem Valinues.
30 39.7. =
3970 wrong NX= NEX
100,
sOFrom
case
because by
the
Midponts
and -39.7 divide
misread obtained, copying,
X= 100 of X=
A+
27 item 100 incorrect is It X= A
=,
CALCULA
insteadfoundtois x 40 students mediainncluthsie ve 45, N=
50
+ 40= as the how 25 12 TION
Correct 83. 44 N
4000 wil
sometcerttoaimn es mid45-X5=45-
-poiandntsmethod 50EÍd=-44, OF
Find were correct
give find
3 be item it MEAN
40.
he
tfound the deduct mode-
Wrong adjustmentremain is N= 45)/5(m
-
out not +
7 -8
72, Xhappens whether
samethe 4.4 50, 0 5 d
If correctmean tovalue
find at e
wrong thitems necessary = C=
be by 40.6. 5 STNISCA
the of correct thatnecessary.is
40. the are
Correct time correct items teto
correspondr: Later adusi
of numb mesmtaken due to
Mean. on mes
i and IEA
 MEASURESOF CENTRAL VALUE

Sotlon N or SX NY
Here X=
40, N= 100
2 X 100 x 40
4000
Less incorrect items 57
3943
Add corect items 75
Correct total 4018

Correct mean 4018

- 40.18.
100
Calculationof Arithmetic Meanin Case of Open-end Classes
Opern-end classes are those in which lower limit of
the first class and the
upperlimit of the last class are known. In such a case we
out the arithmetic mean unless we make an assurnption cannot about find
the
unkrnowm limits. Theassumption would naturally depend upon the cless
interval following the
first class and preceding the last class. For exanple.
obserethe following data:
Marks No. of students Marks No. of students
Below 10 4 30 40 15
10-20 6 40-50 8
20-30 10 Above 50 7

nthe above case since the class interval is uniform, the approprlate
assuimption would be that the lower limit of the first class is zero and the
onmer limit of the last class is 60. The first class thus would be 0-10 and
he last class 50-60. Observe another case :
Marks No. of students Marks No. of students
Below 10 4 60-100 7
10-30 6 Above l100 3
30-60 10

h the above case since the class interval is 20 in the second class, 30
in the third class, 40 in the fourth class, Le., it is increasing by 10. The
appropriate assumnption wóuld be that the lower limit of the first class is
ero and the upper limit of the last class 150. In other words, first class
Is 0-10 and the last one 100-150.
Tthe class intervals are of varying width, an effort should not be made
to determine the lower limit of the lowest class and upper limit of the
gnest class. The use of median or mode would be better in such a case.
Dtause of the difflculty of ascertaining lower limit and upper imit in
en-end distributions it is suggested that in such distributions arithmetc
Mean should not be used.
Mathematical Properties of Arithmetic Mean
he following are a few important mathernatical properties of the
arlthmetic mean:
190
STATISTICAL METHOCA
1, The
sum of the devlations of the iterns from the arithmetic
(taking stgns into
would be clear from account)
is always zero, Le.. E(X- X) : 0. meThiann
the following example:
(X - 30)
10 - 20

20 - 10

30
40 4 10

50 + 20

EX = 150 E(X- ) - 0

Here 2X 150 30. When the sum of the deviations from

N5
actual mean, Le.. 30, 1s taken it comes out to be zero. It is because of
this property that the mean is characterised as point of balance, Le. t
sum of the positive deviations from it is equal to the sum of the nesath
deviations from it.
2. The sum of the squared deviations of the items from arithmet
mean is minimum, that is, less than the sum of the squar
deviations of the items from any other value. The following example
would verify the point:
X (X- 4) X-4
2 -2 4

3 -1 1

4
5 +1 1
6 +2 4

ZX = 20 2(X - X) =0 2 (X- X = 10

The sum of the squared deviations is equal to 10 in the above case. i

the devlatlons are taken from any other value the sum of the squared
deviations would be greater than 10. For example, let us calculate the
squares of the deviations of item from avalue less than the arithmetic
mean, say,
3. X X- 3)

3 0
4 + 1 1
5 + 2 4
6 +3
Z(X- X÷ = 15
It is clear that (X - X} 1s greater. This property that the sum of the
squares of items is least from the mean is of immense use In regressiol
analysis which shall be discussed later.
MEASURESOF CENTRAL VALUE 191

3.Since N'
NX - EX
Inotther words, if we replace cach item in the serles by the mean, then
thosumof these substitutions will be equal to the sum of the indívidual
tems. For example, in the discussion of fArst property EX = 150 and the
aithnetic mean 30. If for each item we substitute 30, we get the same
Le., 30 + 30 + 30 + 30 + 30 - 150.
otal,
This property is of great practical alue. For example, if we know the
gverage wage in a factory is Rs. i,060 and the number of workers
employed is 200, we can compute total wage bill from the relation
The total wage bill in this case would be 200 x 5,060. Le.,
10,12,000 which is cqual to 2X.
Rs.4. If we have the arithmetic mean and number of items of twO or more
than two related groups, we can compute combincd average of these
groups by applying the following formula:
X12 = Ni XË+X NË + N2
X12 = combined mean of the two groups
X1 arithmetic mean of first group
arithmetic mean of second group
NË = number of items in the first group
N2 = number of items in the second group.
The folowing example shall ilustrate the application of the above formula:
Mustration 8. The mean height of 25 male workers in a factory is 61 inches and the mean
heicht of 35 female Workers in the same factory is 58 inches. Find the combined mean
height of 60 workers in the factory. (M.A. Econ., Punjab Univ, 2011)
Solution. X12 = MX+X
NË = 25, X =61, Nh =35, X2 =58
Xio 25 x61)+(35x58)_ 1525 + 2030 3555 =59.25.
25 + 35 60
Thus the combined mean height of 60 workers is 59.25 cm.
I we have to find out the combined mean of three sub-groups the above
ertended as follows : formula can be

X123 = N XNË+
+ N2 X2+ Ng X3
Nht Ng
Merits and Limitations of Arithmetic Mean
Merits Arithmetic mean is most widely used in practice because of the
tolowing reasons:
3 the simplest
Nelther the arrayirngaverage
of datato asunderstand
required forandcalculating
easiest to median
compute.
nor
grouping of data as requtred for calculating mode is needed while
calculating mean.
t s affected by the value of every item in the series.
 " It is STATISTICAL METHO
" It is affected by the value of every item the series.
in
defined by a rigid mathematical formula with the
" everyone
Being who computes the average gets the same answer. result that
deternined by a
algebraic treatnent better tgid formnula, It lends itself to
" It is
when relatively
than the median or node.
reliable in the sense that it does not vary
repcated samples arc taken fron one and the too muh
subsequent
at least not as much as some other
kind of sarme population
The mean is typical in the sensc that it statistical descriptions
is the centre of
balancing the values on either side of it. gravity,.
" It is a calculated value, and not based on position in the series.
Limitations
Since the value of mean depends upon each
xttreme items, ie., very small and very large items,and every item of the series
the average. For example, if in a tutorial group unduly affect the value of
marks in a test are 60, 70, 10 and 80 the there are 4 students and thei:
60 + 70 + 10 + 80 220 average marks would be
4 4 =55. One single item, e., 10, has reduced the
average marks considerably. The
greater is likely to be the impact of smaller the number of observations, the
extreme value.
In a distribution with open-end
classes the value of mean cannot be
computed without making assumptions regarding the size of the class
interval of the open-end classes. If such classes
proportion of the values, then mean may be subjectcontain to
a large
error. However, the values of the substantial
where there are open-end classesmedian and mode can be computed
without making any assumptions
about size of class interval.
The arithmetic mean is not always a
good measure of central
tendency. The mean provides a "characteristic" value,
indicating where most of the values lie, only when theindistribution
the sense of
of
the variable is reasonably normal (bell-shaped). In case of a
distribution the mean is not likely to serve a useful purpose. U-shaped
Weighted Arithmetic Mean
One of the limitations of the arithmetic mean
discussed above is that it
gves equal importance to all the items. But there
relative importance of the different items is not the are cases where the
same.
So, we compute weighted arithmetic mean. The term When this is
the relative weight' stands for
importance.of the different items. The formula for computing
weighted arithmetic mean is:
EWX
EW
" The mean of a frequency distribution X= is, in fact, a mean of the (X's class
mid-points) where cach X is weighted by its importance. This is only a special case ol
the more general notion of the welghted mean.
2WX
 MEASURES
OF CENTRAL VALUE 193

Xy represents the weighted arithmetlc mean; X represents the

where
ariable values, Le.,
W represents the weights attached to variable values, Le.,
, Wh respectively.
0 Multiply the weights by the variable Xand obtain the total EWX.
fil Divide this total by the sum of the weights, Le., W.
In case of frequency distribution, if fi.h..Sn are the frequencies of
dahle values X], X, . Xn respectively then the weighted
the
arithmetc mean is given by:

EWð2)
EW
expanded form of the above formula:
Jhe
X = Wi f X) + W2 f2 Xy) + ... + W, X
W) + W, t ... + Wn
An Important problem that arises while using weighted mean is
regarding selection of weights. Weights may be either actual or
arbitrary.
estimated. Needless to say, if actual weights are available, nothing
this. However, in the absence of actual weights, arbitray or
imaginary weights may be used. The use of arbitrary weights may lead to
SOme erTor, but it is better than no weights at all. In practice, it is found
hat if weights are logically assigned keeping the phenomena in view, the
or involved will be so small that it can be easily over looked.
Tt should be noted that:
0 Simple arithmetic mean shall be equal to the weighted
nean if the weights are equal. Symbolically, arithmetic

X-X, if WË =W2.
(0 Simple arithmetic mean shall be less than the
mean if weighted arithmetic
and only if greater weights are assigned to greater values
and smaller weights are assigned to smaller
: values. Symbolically,
X< Xy if (w- w) (X1 -X2)<0.
(üs Simple arithnetic mean is greater than the
TAean if and only if smaller weight is attached toweighted
the
arithmetic
higher values
and greäter weight is attached to the smaller values.
Symbolically,
X>X, if (u - w) (X1 - X2) <0.
may be noted that weighted arithmetic mean is
roblems relating to: specially useful in
r Construction of index
(4 Standardized birth and
numbers, and
death rates.
0astratsn 9. (a) Acontractor employs three types of workers-male, female and children.
male he pays Rs. 300 per day, to a female worker Rs. 250
oker Rs, 140 per day. What is the per day and to a child
average wage per day paid by the contractor?
194 STATISTICAL METHODA

300 250 +140

Solution. The average waoe is not the simple aithmetic mean, 1e.
3
Rs. 230 per day. If we assume that the number of male, fermale and chiid workers is ho
same, this answer would be correctFor example, i we take 10 workers n eacn ase than
the mean wage would be
3000 + 2500 + 1400 6900 - Rs. 230.
(10 x300) +(10 x 250) +(10 x 140) 30
10 + 10+ 10 30

male,offemale and child workers employed is generaly differan1

we know the
If However, many of
how number workers each type are employed by the contractor in question,
nothing like this. However, in the absence of this we take assumed weights. Let as assume
workers employed is 20, 15 and 5 respectival
that the number of male. temale and childmean calculated as follows:
Tne average wage would be the weighted
No. of Workers
Wages per day (Rs.) W WX
X
20 6,000
300
15 3,750
250
5 700
140
EW= 40 EWX = 10,450

EWX 10,450 = Rs. 261.25.

ZW 40
m.p.h., another 480 miles at a speed af 80
(b) Atrain runs 240 miles at a speed of 60 for 15 minutes at a speed of 10 m.p.h. and
m.p.h., then due to repairs of the track travelsmiles at a speed of 55 m.p.h. What is the
finally covers the remaining distance of 270
average speed in miles per hour?

Solution. Time taken in covering 240 miles at a speed360 of 60 m.p.h. = 240 minues. Time
minutes. Distance covered in 15
taken in covering 480 miles at a speed of 80 m.p.h, in covering 270 miles at a speed
minutes at a speed of 10 m.p.h. = 2.5 miles. Time taken
of 55 m.p.h. = 300 minutes.
mean as
Therefore, taking the time taken as weights we have the weighted
Speed (in m.p.h.) Time taken (in minutes)
W WX

60 240 14,40G

80 360 28,800

10 15 150

55 300 16,500
EW=915 EWX = 59,850

59,850
.. Average speed 915 65.41 m.p.h.

gvert
Illustration 10. Comment on the performance of the students of the three coleges
below using simple and weighted averages:
 Coliga'A' lmple M.Sc. BSC. B.Com. BA. M.Com. MA
Courofse Collegé Solution. MSC. BSC. BCom. BA. MA.
and
Mom. Cose
:
74
Weighted 432 EX 66 65 73 83 71 X
Pass CALCULATION
Pass
MEASURES OF
66 65 74 73 83 71 CENTRAL
Students No.of
=EW
3 3 2 5 W
(000) A'
Arithmetic 20
X72: 'A'
OF Students No.of VALUE
=
1,451 EWX (000)
198 195 148 365 332 213 WX 3 2 5 3
SIMPLE
F=72; Mean
65 76 76 82 Pass
432 EX 60 73 X %
AND Pass
60 65 76 73 76 82 %
Students WEIGHTED
=2W (O00) No.of
28 7 3 7 6 3 2 W 'B' 'B
Students No.of
7
('000)
2
YWA72.55 EWX= 420 195 532 438 228 164 7 6 3
1,977 WX ARITHMETIOC
X_197-70.61EW EW
EX 70 58 74 76 Pass
73 8 X Pass
432 73 70 58 74 76 81
28 20 MEANS
Students
(000) No. of
=EW 2.0 7.0 2.0 4.5 3.5 2.0 W C'
21
Students
('000) No.of
=1,513 2 7 2 4.5 3.5 2
EWX 146 490 116 333 266 162 WX
195