Catanduanes State University

Panganiban Campus

Panganiban, Catanduanes

Topic : MEASURES OF VARIABILITY & POSITION

Reporters :

Sarah Mae M. Todoc

Anne Beatrice de Guzman

Janice Velasco

Mary Rose Nicolas

April Joy Bamiano

Renan Tumanlao

Professor : Mrs. Rosalie M. Ocillos

MEASURES OF VARIABILITY

In the preceding chapters, the different measures of central tendency and point measures were already presented and

discussed. However, these measures do not give us the total picture of the sample we have measured. They do not give

us information on how the data are scattered. They do not describe a mass of data. It is, therefore, necessary to have

measures of variation or dispersion to describe better characteristics of a set of data. The possible measures used to

describe this dispersion are the range, quartile deviation, average deviation or mean deviation, variance, standard

deviation, and coefficient of variation.

THE RANGE

The simplest and easiest measure of variability is the range. It is easily obtained and readily understood. However,
since its magnitude depends on the extreme values of the set of data, it is the most unreliable.

Range may be classified into (a) absolute range, (b) total range, and (c) Kelly range. Absolute range is obtained by taking

the absolute difference between the highest and lowest scores in a set. For instance, for a set of data 21, 23, 23, 25, 27,

28, 33, 38, 40, 43, 45, 47, 48. The range is 48-21 = 27. When the data are arranged in a frequency distribution, the
absolute range is the difference between the upper limit of the class interval. For example, the highest class limit is 45.5
48.5 and the lowest class limit is 9.5 - 12.5. The range is 48.5 9.5 = 39.

Total range is obtained by computing the highest score minus the lowest score plus one or total range =

(HS - LS) + 1. Consider these data, 52, 53, 57, 60, 63, 65, 70, and 75. The total range is (75 52) + 1 = 24.

Kelly range (KR) is another measure of variability where the KR value is obtained by taking the difference between the

ninetieth and tenth percentile or KR = P90 P10. For example, consider the data below:

Class Limit Frequency Cumulative Frequency

46-48 1 35

43-45 1 34

40-42 2 33
 37-39 3 31

34-36 3 28

31-33 4 25

28-30 7 21

25-27 5 14

22-24 3 9

19-21 2 6

15-18 2 4

13-15 1 2

10-12 1 1

TOTAL 35

The ninetieth percentile or P99 value is 40.25 and tenth percentile or P10 is 17.75. The Kelly Range is 22.5. Note

that of the three kinds of ranges, absolute range is commonly used by classroom teachers.

QUARTILE DEVIATION OR SEMI-INTERQUARTILE RANGE

Quartile deviation or QD is another measure of variability which divides the difference of the third and first

qualities into halves.

Quartile deviation from ungrouped data. Quartile deviation or semi-interquartile range (QD) from ungrouped

data can be estimated by obtaining half of the distance from the middle 3N/4 to the middle of the bottom of

N/4 of the cases. For example, consider the following data in Table 14.1

Table 14.1. Scores of Twenty-six Students in Ichthyology Test

The first quartile or Q1 = N/4 is 6.5 and the value is between 46 and 47 and divided by 2 or 46 + 47 = 93/2 = 46.5.

The third quartile or O3 = 3N/4 is 19.5. The value is between 61 and 62. Find the sum of these scores divided by 2 is =
61.5. Hence, quartile deviation or (QD) is equal to 03Q1 divided by 2. To substitute the formula, 61.5 46.5 = 15/2 =
7.5. The quartile deviation or semi-interquartile range from ungrouped data is 7.5. This is

sufficiently an accurate estimate of the quartile deviation for ordinary use in the classroom.

Quartile deviation from grouped data. Quartile deviation is one-half the distance of the difference of the third

and first quartiles. To compute for the quartile deviation from a grouped data or in a form of frequency

distribution, consider the following steps:

1. Estimate the cumulative frequencies "lesser than".

2. Find the third and first quartiles.

Table 14.2. Computation of Quartile Deviation from Grouped Data

THE AVERAGE MEAN or MEAN DEVIATION

The average or mean deviation is a measure of absolute variability where it is frequently used. It is the arithmetic mean

of the absolute deviations regardless of algebraic sign.

Average deviation from ungrouped data. To obtain the average deviation from ungrouped data, we simply estimate the

deviations from the arithmetic mean, get the sum regardless of signs and divide by N or the total number of cases. For

instance, consider the following examples:

Sample A 7 8 9 14 14 18 20 20 25

Sample B 14 18 19 21 21 23 25 27 30

Sample C 20 21 23 23 23 25 27 28 35

The means of the three samples are 15, 22, and 25 respectively. The deviations from their sample means are:

Sample A - 8 -7 -6 -1 -1 3 5 5 10

Sample B - 8 -4 -3 -1 -1 -1 3 5 8

Sample C - 5 -4 -2 -2 -2 0 2 3 10

For Sample A, the average deviation is (8 + 7 + 6 + 1 + 1 + 3 + 5 + 5 + 10) = -6- =5.11. For Sample B, the average
deviation

34
is(8+4+3+1+1+1+3+5+8)= =3.78. For Sample C, the average deviation is (S+4+2+2+2+D+2+3+10)=
9
39= 3.33.

FORMULA for AVERAGE DEVIATION FROM uursnouvso DATA:

AD = g/x -x/ 114.1)

Here, the /X X/ is a deviation from the mean regardless of algebraic signs. The two vertical bars represent that the

signs are disregarded, N is the total number of cases.

AVERAGE DEVIATION FROM GROUPED DATA. if the data are grouped into frequency distribution, the average
deviation

is obtained by adopting the following steps:

Step 1. Solve for the Arithmetic mean.

Step 3. Subtract the midpoint from the arithmetic mean.

Step 4. Multiply the frequency (f) and the difference of the midpoint and mean (X-X shown in the columns 3 and 4 and

write the products in columns S f(XX).

Step S. Compute the average deviation by using the formula.

- X)/ (14.2) in which the bars // enclosing the f (XX) show that signs are ignored

AD - Z/f (X

To illustrate the calculation of the average deviation from grouped data, consider Table 14.2 data. Table 14 3 shows the

computation of average deviation from grouped data.

Table 14.3 computation of Average Deviation from Grouped Data

12l345

Class Limit Frequency Midpoint /XX/ /f (X-X)/

lfi (X)

46-48 1 47

17.83

17.83

43-45 44

I-J

14.83

14.83

|\l

40-42 , 41

11.83

23.65

37-39 38

U0

8.83

26.49

34-36 35

U)

5.83

17.49

31-33 32

->

2.83

11.32

28-30 29
\l

0.17

1.19

25-27 26

U1

3.17

15.85

22-24 23 '

U-J

6.17

18.51

19-21 20

9.17

18.34

16-18 17

12.17

24.34

13-15 14

I-\

15.17

15.17

1O12 11

l\

18.17

18.17

, TOTAL as

i 223.19

x = 29.17

AD = Z/f (X Xl/

= 223.12.

35

AD = 6.38

THE VARIANCE

it is the square of the standard deviation and also known as the square.

VARIAINCE FROM UNGROUPED DATA.

2 Methods:
0 Working Formula

0 Machine Formula

Working formula is obtained by getting the sum of the squares of the deviation about the mean and to divide by the

number of cases minus one or divided by the degrees of freedom (N-1).

SD = Z(X X):

N1

SD is the variance; Z(X - X) is the sum of squares of deviations about the mean; N is the number of cases or N1 lS

the degree of freedom.

Machine Formula is obtained by getting the difference of the sum of squares of each score and correction factor divided

by N-1. The N-1 divisor is preferable because it gives unbiased results. Thus, according to this method the variance is

given by

SD = {X2 CF

N1X

Where, {X2 is the sum of squared products of X; CF or correction factor is obtained by using this formula:

CF = QXY

Tabie 14.4 Computation of Variance from Ungrouped Data in Research and Evaiuation Examination of Ten Students in
Working Fol-rnuia.

Students Scores (X) (X X} (XX)2

Z5 2.5

F-\

6.25

21 -1.5

IQ

2.25

15 -7.5

55.25

10 -12.5

b-

156.25

Z7 4.5

20.25

20 -2.5

G
6.25

I 35 12.5

Ni

156.25

. 9 -13.5

182.25

5 i as 15.5

L0

240.25

10 25 2.5

6.25

TOTAL 225

832.50

EX = 225 X = 22.5 N = 10

sol = (x-x)1

N1

=.32Li.

s01= 92.5

Table 14.5 Computation of Variance from Ungrouped Data in Machine Formula Form

Students Scores (X)

X2

1 25

625

2 21

441

UJ

15

225

-l>

10

100

U1

27

729

O1
20

400

\-I

35

1225

O0

81

38

LO

1444

10 5 25

625

TOTAL 225

S895

so= = {x1 - cr

N1X

10--1

= 832.5

SD2 = 92.5

= 2252 Zxl = 5895

10

= 50625/10 CF = 5062.5

ct = 5052.5 N = 10

VARIANCE FROM GROUPED DATA (MIDPOINT METHOD)

2 METHODS a.) midpoint b.) Class-deviation. These methods can also be used in computing for the arithmetic mean and
Standard

Deviation.

Midpoint method:

SD2= N ZFm2(ZFm)2

N2 - N

EXAMPLE:

so2= N ZFm2- (Fm)2

N2 - N

= 35 (32159) - (1021)2
352 - 35

11255554042441

1225-35

_ sal24

1190

= 69. 8521 or 69.85

VARIANCE FROM GROUPED DATA (class deviation method)

Step 1: Choose a temporary arbitrary origin. You can choose from any class limits either at the center, bottom or top.

Step 2: Assign to the class limits coded values, starting with zero at the origin and with positive values above the zero

deviation and with negative values below it if the class limits are arranged from highest down to lowest.

Step 3: Multiply the d by the corresponding class frequencies fto get fd. These products are are shown in column 4.

Step 4: Sum the fd products algebraically. The symbol is {fd.

Step 5: Take the product of fd and d to get fd2. This is shown in column 5.

Step 6: Take the sum of the products, fd2. The sum is {fd2.

Step 7: Compute the variance by using the following formula.

SD <12

The Standard Deviation

The Standard Deviation (SD) is a measure which describes the dispersion of or spread of all scores in

the distribution rather than through extremely high or low scores. lt is the square root of the average of the

squared deviations from the mean.

2 Formulas for Computing the Standard Deviation

1. Find the deviation of each score from the mean.

The steps are as follows:

a) Calculate the mean:Y

b) Find the deviation of each score from the mean: X--Y

c) Square the deviations: (XX)2

d) Sum the squares: (XX)2

e) Divide the sum by n-1 if the variance of the sample is the measure to be computed.

SD = ,/Z(X X) 2 where; SD = standard deviation, n = sample size, X = sample mean

n -1

2. Calculate the standard deviation without using the mean.

The Measures of Variability in Various Distributions

When the SD is small, score tend to cluster toward the mean. The distribution of the scores tends to be

compressed toward the mean. It is said to be homogenous. The group tends to of almost the same ability. This

is called leptokurtic distribution.

it is said to be heterogeneous. The class has different kinds of ability. The distribution is called platykurtic

distribution.

When the distribution is moderately large. Scores tend to spread evenly from the mean. The distribution of the

scores tends to be symmetrical. It is said to be close to normal. A normal distribution is called mesokurtic

distribution because it is similar to the normal curve which is symmetrical and bell-shaped.

The normal distribution is not perfectly possible.