A PRESENTATION ON MEASURES OF CENTRAL TENDENCY

ByByDr. Alok Mittal Dean MBA (Power & Engineering) Program Indore Indira Business School, Indore.

INTRODUCTIONINTRODUCTION

Mass data, which are collected, classified, tabulated and presented systematically, is analyzed further to bring its size to a single representative figure. figure. The tendency of data to cluster around a figure is known as Central Tendency. Tendency. Measures of central tendency or averages of first order describes the concentration of large numbers around a value. value. It is this single value which represents all units. units.

OBJECTIVES OF STATISTICAL AVERAGESAVERAGESTo present mass data in a concise form. form.  To facilitate comparison. comparison.  To establish relationship between sets. sets.  To provide basis for decision making. making.


REQUISITES OF A GOOD AVERAGEAVERAGE

 

It should be simple to calculate and easy to understand. understand. It should be based on all values. values. It should not be affected by extreme values. values. It should not be affected by sampling fluctuations. fluctuations. It should be rigidly and unambiguously defined. defined. It should be capable of further algebraic treatment. treatment.

STATISTICAL AVERAGESAVERAGES 

Most commonly used measure is Mean or Arithmetic Mean. Mean. Mean of Ungrouped Data is given bybyA.M. (X)= X / N Mean of Grouped Data is given bybyA.M. (X)= f X / N For simplifying the calculations another formula is also used for calculating A.M.d=x A.M. (X)= A + f d x i and d=x-A N i where A= Assumed Mean, i=width of class interval and x is the mid-point of class interval midand N= f

PROPERTIES OF A.M.A.M.

Algebraic sum of deviations of a set of values taken from their Mean is always zero i.e. ( X- A.M.)=0 )=0 Sum of squares of deviations of a set of values from their mean is always minimum i.e. (X-A.M.) is always min. (Xmin. It is capable of further algebraic treatment. treatment. Suppose if X1, X2 , ----Xn are ----Xn the means of n1 , n2 -----nn sets of values,
then their combined arithmetic mean is given byby- A.M. = n1 X1 + n2 X2 + ----- + nn Xn

n1 + n2 +-------- + nn

MERITS OF A.M.A.M.

  

 

It is simple to calculate and easy to understand. understand. It is based on all values. values. It is rigidly defined. defined. It is capable of further algebraic treatment. treatment. It is more stable. stable. Permits calculation of mean values when some weights are also associated with all values (Weighted Arithmetic Mean)

DEMERITS OF A.M.A.M.It is affected by extreme values. values.  It cannot be determined for distributions with open-end class openintervals. intervals.  It cannot be graphically located. located.  Sometimes it is a value which is not in the series. series.


MEDIANMEDIAN

The median is another measure of central tendency. tendency. It represents the middle value of the data that measures the central item in the data. data. Half of the items lie above the median, and the other half lie below it. it. Median of Ungrouped Data- First array the Datadata either in ascending order or in descending order. order. If the number of observations (N) contains an odd number then the median is the middle value. value. If it is even, then the median is the mean of the two middle values i.e. Median=[(N+1)/2] Median=[(N+1)/2 item in a data array, where N=number of items. items.

MEDIAN (Continued)(Continued)

Median of Grouped Data- When the data is in Datathe form of discrete series, the median can be computed by examining the cumulative frequency distribution and using formula Median=[(N+1)/2 Median=[(N+1)/2]th item, where N= Last cumulative frequency When data is in class interval formformMedian=L + [(N/2) p. c. f] x i [(N/2 f where L=Lower limit of median class, N=Number of observations f=Frequency of median class p. c. f =Preceding cumulative frequency of median class i= width of median class

MERITS OF MEDIANMEDIANIt is not affected by extreme values. values.  It can be easily understood & computed. computed.  It can be determined graphically (ogives)  It can be used for qualitative data. data.  It can be calculated for distributions with openopenend classes. classes. DEMERITS OF MEDIANMEDIAN It is not based on all values. values.  It is not capable of further algebraic treatment i.e. if M1 and M2 are medians of two different sets of data, we cannot get median of combined data set from M1 and M2.


MODEMODE 

 

Mode is also a measure of central tendency. The mode of the data is the value that appears the maximum number of times. Mode for Ungrouped Data- It is the value Datacorresponding to highest frequency. Mode for Grouped DataDataMode = L + ___ 1____ x i
1+ 2

Where L = Lower Limit of modal class, i = width of modal class , 1= Excess of frequency of the modal class (f1) over frequency of preceding class (f0) [i.e. 1= f1 - f0 ] 2 = Excess of frequency of the modal class (f1) over frequency of succeeding class (f2) [i.e. 2= f1 f2 ]

MERITS OF MODEMODEIn many cases it can be found by inspection. inspection.  It is not affected by extreme values. values.  It can be calculated for distributions with open end classes. classes.  It can be located graphically (histograms)  It can be used for qualitative data. data. DEMERITS OF MODEMODE It is not based on all values. values.  It is not capable of further mathematical treatment. treatment.  It is affected by sampling fluctuations. fluctuations.


EMPIRICAL RELATION BETWEEN MEASURES OF CENTRAL TENDENCYTENDENCY

MEAN MODE = 3 (MEAN MEDIAN) OR

MODE = 3 MEDIAN 2 MEAN When any two measures of Central Tendency are known to us, the third can be determined by above relation. relation. This is applicable for moderately skewed distribution of data. data.

OTHER MEASURES OF CENTRAL TENDENCYTENDENCY  

Geometric Mean (GM)- It is defined as (GM)the Nth root of the product of all N values root G.M. = n product of N values The most frequent use of G.M. is to know the average rate of change of population, compound interest, growth rate, etc. etc. If one value is zero, G.M. becomes zero and inappropriate. inappropriate. Similarly if one value is negative, G.M. can be calculated, but it will be meaningless. meaningless.

HARMONIC MEANMEAN

 

 

Harmonic Mean (HM)- It is defined as (HM)the reciprocal of the arithmetic mean of the reciprocal of the observations. observations. In other words it can be defined as the ratio of number of observations and sum of reciprocal of the values. values. HM=N/{1 HM=N/{1/x1+1/x2+----+1/xn} or N/ 1/x ----+ For any set of data wherever computation is possible the following inequality holdsholdsAM GM HM HM is useful in averaging rates. rates.

QUANTILESQUANTILES- Quantiles are related positional measures of central tendency. The most tendency. popular quantiles areare Quartiles- Quartiles are those values which Quartilesdivide the total data into four equal parts, i.e. we have three quartiles Q1, Q2 and Q3 Qj = L + [(jN/4) p. c. f] x i [(jN/4 j=1,2,3 j=1


POSITIONAL MEASURESMEASURES-

f


DecilesDeciles- Deciles are those values which divide the total data into ten equal parts, i.e. we have nine deciles D1, D2 ---------------- D9 Dk = L + [(kN/10) p. c. f] x i [(kN/10) k=1,2,----9 k=1 ----9
f

POSITIONAL MEASURESMEASURES

PercentilesPercentiles- Percentiles are those values which divide the total data into hundred equal parts, i.e. we have ninety nine percentiles P1, P2 ---------------- P99 l=1 -------99 Pl = L + [(lN/100) p. c. f] x i l=1,2-------99 [(lN/100)
f

where L=Lower limit of quantile class, N=Number of observations f=Frequency of quantile class p. c. f =Preceding cumulative frequency of quantile class i= width of quantile class

  

 

The property of deviation of values from the average is called as Dispersion or Variation. Variation. The degree of variation is found by the measure of variation. variation. Various measures used to find the degree of variation areareRange (R) Quartile Deviation (Q.D.) (Q. Mean Deviation (M.D.) / Average Deviation (M. (A. (A.D.) Standard Deviation (S.D.) (S. Above measures have units of measurement attached to them (absolute measures). measures).

DISPERSIONDISPERSION-

RELATIVE MEASURES OF DISPERSIONDISPERSIONRelative units of measurement of dispersion are

Coefficient Coefficient Coefficient Coefficient

of of of of

Range Quartile Deviation Mean Deviation Variation

Prerequisites of good measure of variation areare

  

It should be easy to understand and simple to calculate. calculate. It should be based on all values. values. It should be rigidly defined. defined. It should not be affected by extreme values and sampling fluctuations. fluctuations. It should be capable of further algebraic treatment. treatment.

RANGERANGE

  

 

Range is the difference between highest and lowest value of the data. data. R= HL Coefficient of Range =(HL)/(H+L) =(H Range is used in Statistical Quality Control (S.Q.C.) (S. It can be applied when data has got no abnormal values. values. It is not based on all values. values. It uses extreme values only. only.

QUARTILE DEVIATIONDEVIATION 

Quartile Deviation- It is defined asDeviationasQ.D. = (Q3 Q1) / 2

where Q1 = First Quartile & Q3 = Third Quartile  Coefficient of Quartile Deviation is given byby Coefficient of Q.D. = (Q3Q1)/(Q3+Q1)


Quartile deviation is superior to the range as it is not based on extreme values. Quartile Deviation can be used for openend distribution. The drawback of Quartile Deviation is that it ignores the first and last 25% 25% observations. observations.

MEAN / AVERAGE DEVIATIONDEVIATION

    

The measure of average or mean deviation is an improvement over the previous two measures as it considers all observations in the given set of data. data. This measure is computed as the mean of deviations from the mean/median/mode. mean/median/mode. M.D. = | X A.M.| / N For a grouped data the formula becomesbecomesM.D. = f| X A.M.| / N Coefficient of Mean Deviation is given byCoefficient of M.D. = M.D. / A.M.

STANDARD DEVIATIONDEVIATION

    

S.D.( ) is the most widely used and important measure of variation. In Mean variation. Deviation the signs are ignored. S.D. ignored. overcomes this problem by squaring the deviations to make them positive. positive. S.D. is also known as Root Mean Square (R. (R.M.S.) Deviation. Deviation. S.D. ( )= { (XA.M.)/N} The square of S.D. is called variance. Variance = For a grouped data the formula isisS.D. ( )= { f (XA.M.)/N}

COEFFICIENT OF VARIATIONVARIATION

 




C.V. is frequently used relative measure of variation. variation. C.V. = ( / A.M.) X 100 % When C.V. is less in the data it is said to be less variable or more consistent. Karl Pearson coefficient of skewnessS.K.=(Mean-Mode)/S.D. or 3(Mean-Median)/S.D.

 

Bowleys coefficient of skewnessS.K. = (Q3 + Q1 2 Median)/(Q3 - Q1)

THANK YOU