Measures of Variability- Khyati

Dr.
Khyati Nimavat
Dept. of Community Medicine
GMERS- Junagadh
 Introduction
 Types of Variability
 Measures of Variability
 No two measurements in man are absolutely equal
 Cure rate in typhoid with same drug varies in different

patients
 Mean height of students of the same age group at one

place will differ from that at other place
 Birth rate, IMR, MMR of two towns are seldom equal

 There are 3 main types of variability
1) Biological Variability
2) Real Variability
3) Experimental Variability
 Individuals in similar environments differ when
compared as regards sex, weight, height, BP etc. but
the difference noted may be small and is said to occur
by chance. It is called biological variability
 It may be normal or the difference may be natural
 This occurs within certain accepted biological limits

 When the difference between two readings,
observations or values of classes or samples is more
than the defined limits it is said to be real
 Then the cause/reason may not be natural and some

particular factor might be playing role for such
difference. e.g. treatment of malaria (with and without
antimalarial) and cure rate
 Height and weight among citizens of USA are more

than those in India; perhaps due to better socio-
economic conditions. Higher rate Ca. Lung among
smokers as compared to Non-smokers
 Error of difference or variation may be due to
materials, methods, procedure employed in the study
or defects in the techniques in the experiment
 They are of 3 types.

a. Observer
b. Instrumental
c. Sampling
 Measures of variability of individual observations
1)Range
2)Inter-quartile range
3)Mean Deviation
4)Standard Deviation
5)Coefficient of variation
 Measures of variability of samples
1)Standard error of mean
2)Standard error of difference between two means
3)Standard error of proportion
4)Standard error of difference between two proportion
5)information of central tendency.
A range defines the normal limits of a biological
characteristic
e.g. SBP- 100-140mm of hg
Menstrual cycle- 21-34 days
 It is calculated by the finding the difference between
the maximum and minimum measurements in the
series
 Range is not a good measure of dispersion as
compared with standard deviation because SD uses all
the available information while range do not
 There are three different points located on the entire
range of a variable
e.g. Q1, Q2, Q3
 Q1 or lower quartile will have 25% observations on its
left and 75% on its right
 Q2 or middle quartile will have 50% observations on its
 Q3 or upper quartile will have 75% observations on its
 The range of a variable between first quartile and third
quartile is called interquartile range
 Range of variable between first quartile Q1 and
third quartile Q3 is called interquartile range. Half
of this range is called semi- interquartile range.
 Also known as quartile deviation which is a
measure of dispersion around the mean, slightly
better than Range.
 Q3- Q1
Q=---------------
2
 The semi-interquartile range takes into account
only the middle half of the data between Q3 and
Q1.
nie
 If mean blood pressure of a large representative series is
taken, some observations are found above the mean and
other are below the mean.
_
 MD= Σ | X -X|
n
 To find MD of observations from the mean, add the
difference from the mean and divide by the number of
observations.
 e.g. Height of 5 students in Centimeter

=
Observations X X- |X-| |
150 -5 5
160 +5 5
155 0 0
170 +15 15
140 -15 15
ΣX= 775 Σ |X- | = 40
= ΣX = 775 = 155cm
n 5
Mean Deviation M.D. = Σ | X - | = 40 = 8cm
n 5
Mean deviation is not used in statistical analysis being less
mathematical value, particularly in drawing inferences.
 Standard deviation is an improvement over mean deviation as
a measure of dispersion and is used most commonly in
statistical analysis.
 It is computed by following steps:
a. Calculate Mean: :
b. Find the difference of each observation from the mean: X -
c. Square the difference of observations from the mean: (X - )2
d. Add the squared values to get the sum of squares: Σ(X - )2
e. Divide the sum by the number of observations minus
1 to get mean-squared deviation called variance (σ2 ):
: Σ(X - )2
n-1
 f. Find the square root of this variance to get root-
mean squared deviation called SD.
- )2
Σ(X -X
SD= n -1
 Large SD indicates frequency distributions are widely

spread out, while small SD indicated closely spread
frequencies.
 It summarizes the deviations of a large distribution
from mean
 Indicates whether the variation of individual

observation is by chance or real due to some special
reasons
 SD helps to find out Standard Error (SE) which

determines whether the difference between means of
two samples is by chance or real
 It also helps in finding the suitable size of sample

 First find the mean of series. It is calculated as the sum of -----
observations divided by their number
= ΣX
N
 Then find the deviation or difference of the individual
measurements from the mean
x=X-
 Next find the sum of the squares of deviations or difference
of individual measurements from the mean. This can be
expressed by the formula
Σ(X - )2
 Sum of square
= Σ(X - )2
= Σ(X2 -2X + 2)
= ΣX2 -2 ΣX + Σ 2 = ΣX/ N or N = ΣX
= ΣX2 -2 (N ) + Σ 2
= ΣX2 -2N 2+Σ 2
= ΣX2 -N 2 Σ 2 =N 2
= ΣX2 –N (ΣX / N) 2
= ΣX2 –N ΣX 2
N2
= ΣX2 – (ΣX) 2
 N
 Now, find the variance (var) which is mean squared deviation,
i.e. sum of squares, divided by the number of independent
observations. This number is not total but one less than the
total number of measurements or observations (n) in the
series, therefore, divided by n-1. It is also called the degree of
freedom in statistical terms and gives an unbiased estimate of
variance.
 Hence variance = Σ(X - )2 or ΣX2 – (ΣX) 2
n-1 ______n____
n-1
or = Σx2 – (Σx) 2
____n______
n-1
-Very often variance is written as Var. or SD2 or s2 for sample or
σ2 for universe or population
 Lastly, determined the square root of the variance. That gives
the standard deviation which in fact is the square root of the
mean squared deviation. It will be in the same units as the
original measurements. We had squared the deviatio ns so
the square root has to be found.
____
Σ(X - )2
s or SD = √Var = n -1 by formula (a) or
ΣX2 – (ΣX)2
= _____n__
by formula (b1) or
n-1
by formula (b2) where assumed mean (w) is used
= Σx2--(Σx)2
_____n__
n-1
 It is a measure used to compare relative variability
 The variation of the same character in two or more different
series has to be compared quite often. It may be of interest to
know whether the weight varies more in spleen or in heart;
weight varies more among girls or among boys
 Variation of pulse rate is more among youth or among aged
or in students appearing for the examinations and in others
not doing so
 At other times, the variation of two different characters in the
same subjects can also be compared. e.g. Height & Weight of
same group of students, Blood Pressure & Height, Blood
Pressure and pulse rate etc.
 It compares the variability irrespective of the units of
measurement used in two or more distributions such as
Height in centimeter in one and inches in other or pulse rate
in beats and Blood Pressure in mm Hg
 Coefficient of variation (CV) is used to compare the variability
of one character in two different groups having different
magnitude of values or two characters in the same group by
expressing in percentage
 The coefficient of variation is calculated from standard
deviation and mean of the characteristic. The ratio of SD and
mean is found in percentage. Thus, SD expressed as
percentage of mean is the coefficient of variation.
 Coefficient of variation CV = SD x 100 = s x 100
Mean
In two series of adults aged 21 years and children 3 months old following
values were obtained for height. Find which series shows greater variation?
(can you prove the difference between these two series is statistically
significant?)
Persons Mean Ht in Cm SD in Cm
Adults 160cm 10cm

Children 60cm 5cm
CV = SD x 100
Mean
CV of adults = 10 x 100 = 6.25%

160
CV of children = 5 x 100 = 8.33%

60
Thus it is found that heights in children show greater variation than in
adults. (The ratio of this variation is 8.33/6.25= 1.3:1.0) [for calculation
of the statistical different between this two series we need to calculate
Z test ]
Thank you

Measures of Variability- Khyati

Uploaded by

Copyright:

Available Formats

Measures of Variability- Khyati

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Measures of Variability- Khyati

Uploaded by

Copyright:

Available Formats

Dr.

 Cure rate in typhoid with same drug varies in different

 Mean height of students of the same age group at one

 Birth rate, IMR, MMR of two towns are seldom equal

 It may be normal or the difference may be natural

 This occurs within certain accepted biological limits

 Then the cause/reason may not be natural and some

 Height and weight among citizens of USA are more

 They are of 3 types.

 e.g. Height of 5 students in Centimeter

 Large SD indicates frequency distributions are widely

 Indicates whether the variation of individual

 SD helps to find out Standard Error (SE) which

 It also helps in finding the suitable size of sample

= ΣX2 -2N 2+Σ 2

Adults 160cm 10cm

CV of adults = 10 x 100 = 6.25%

CV of children = 5 x 100 = 8.33%

You might also like