NormalDistribution

Distribution
9.1Concept of Normal
distribution is defined as 'a function that repr
The term normal
distribution of many
Normal distribution random variables as a symmetrical
is the most common continuous probability distribtn
bell-shaped

example of an external characteristic that is normally

Height is an personality traits) are also
but internal characteristics (e.g. theorized
normally distributed.
frequency distribution that illustrates how the heights of te
The graph is a 34 (n= 1000)
normal adult males are Male Age 25 -
distributed. As youcan 250
Frequency
see, there are few men 200

who are shorter than 150

152 cm and even fewer 100

men with the stature of 50

1 8
t01
o 8 . 9

208 cm. The

1 8t8o
1 9 1 . 9
1 6
t 8
o
1 7 1 . 9

1879
1 5
t 6
1
o 5 9 . 9

1 5
t 1
2
o 5 5 . 9

Over
1 5 1
o .
L 9
re s s

163.9
1 6
t 4
1
o 6 7 .

majority of males seem

18t4o

1 6t
0 o
19 0

to cluster around an Height in cm

average height of 178
cm.
Frequency Distributionfor Male Heights

Using Some
Male Age 25 -34 (n =1000)
basic geometry 250
and judging
roughly from the 200
requency
area occupied by G150
each bar, We
100
might
legitimately 50
estimate that
about 60% of all Height (m)
males in this age
Groun are between 172 cn and 183.9 cm (ark regions). By connecting the
nomal
midpoint of each bar, the line graph Is obtained, wnich 1s known as
distribution curve.
112
 Normal Distribution

02NormalCurve

where most occurrences take place in the middle of

distributioncurveis
Normal on either side. It is also known as normal curve or
distributionand taper off
the curve,The name comes
from the fact that it looks like bell-shaped.
pell-shaped
in statistics, it's
term
'bell curve' is usually used in the social sciences:
The
Normal Distribution and in physics, it's called a Gaussian Distribution.
called.a distribution that
they all refer to exactly the same thing: a probability
However,
certain
characteristics, including the fact it's shaped like a bell.
has
in
nomal curve 1S an important, strong, reoccurring phenomenon
The distributions.
Many behavioral measurements result in normal
psychology.
isthat bell curve can be described by 2
parameters ('parameter"
fact
Anamazing
for 'number). Thesetwo paramneters are:
being a fancy word
centered - or, the value at the peak. This
is
1. Where the distribution is
called the mean.
How wide the distribution is - or, how much variability there is in the
2.
as standard deviation.
thing you're measuring. This is called

Changing 'p'shifts the

distribution left or right

Changing '¡' increases or

decreases the spread

distribution has its own mean, denoted by the Greek letter u' (say
Each normal (say
own standard deviation, denoted by the Greek letter ' '
'mu'); and its
parameters, and completely determine the shape of the
'sigma'). These two symmetry, and the
gives the location of the line of
normal curve. The mean
much the data are spread out.
standard deviation describes how
113
I'vehological Statixtics

The mean ' ' tells about location (Slape unchanged)

|=130 g=10
0.08 0.08 100 G=10
0.04 0.04

0.00 0.00

50 100 150 50 100

150
Increase u - Location shifts right
Decrease u-
LOcation
The standard deviation 'o' tells about
narrowness or flatness
shifts lej
(Location unchanged) of the b

=100 g=15 l=100

0.08 0.08 G=5

0.04 0.04

0.00 0.00

50 100 150 50 100 150

Increase G- Bell flattens Decrease o - Bell narrows
9.3 Characteristics or
Properties of Normal Probability Curve
The characteristics of the
normal probability curve are:
The Normal Curve is Symmetrical (that is, when you cut it in halfthe
two pieces are mirror images of
each other)
The Normal Curve is Unimodel (Its distribution has a bump inthe
middle, with tails going down and out to the left and right)
114
 Normal Distribution
Standard Deviation
Nevertouches X-axis
Affected by Mean
Area under curve1
Symmetrical about mean
Unimodal Bell Shaped Curve
Moretharnmean
Continuous)
50% of
values Normal Distribution Mean
Median
Mode
Lessthanmean Also known as
Exarnples Blood Pressure
ProbabilityDistribution
Law of Chance
Marks
Height
Gaussian Distribution

The Normal Curve is Asymptotic to the X Axis (The normal curve

approaches, but never touches, the x-axis as it extends farther and
farther away from the mean)
Mean = Median = Mode (The mean, median and mode are the same
and lie directly in the middle of the distribution)
unimodal

symmetric

asymptotic

mean = median = mode

100
Area under curve equals I
may be also considered
The Total Area under Normal Curve
Percent Probability
Psvchological Statistics

The Maximum Ordinate occurs at the Center

The Nomal Curve is a mathematical model in
specially in Mental Measurement behavioral
9.4 Interpretation of Normal Curve Sciences
Normal curve has great significance in the mental
measurement and
evaluation. It gives important information about the trait being

at
As an
measured.
example, let';s
the normal
educationl
look
associated with
IQ ScCUTvorese,
The mean,
mode of a median, s and
that 50% of
Nechs
Score is 100,
whicle
h r' IO
means
-20
lQs fall at 106
Or
below and 50%
Percentage 100 or above.
A
at fall
0.1% 2.3% 15.9% S0%
Z Score
84.1% 97,7% 99.9%
person is likely to score typical
-2 -1 +2
within about 15 points of
TScore the mean, that 1s,
85 and 115. between
20 3 40 50 60 80 Since 68% of
Weschler IQs (SD = 15) SCores on a normal curve
$5 85 fall within one
deviation and sincestandard
100 I15 130 145
Weschier Subset Scores (SD = 3)
an IO
4 7 10 13 16
SCore has a standard
19

that 689% of IOs fall between 85 and 115. deviation of 15, we know

If the psychologist knows the mean and the typical deviation from the mean
(called the standard deviation), the researcher can
scores is likely to fall in any determine what proportion of
given range.
For instance, the range between one
standard deviation below the mean (about
85 for IQ scores) and one
deviation above the mean (about 115 for IQ scores),
one expects to find the scores of about two
thirds of alltest takers. Further,only
about two and a half percent of test takers will
score higher than two standaro
deviations above the mean (about 130).
Normal curve is useful tool for research data: If you are
let's say it is a sample of dataX that is exploring some au
distributed normal with mean and
116
 abot 6R percem of s aes
one standard deviafom of he mes
68%

about 95 percent of its values lie

two standard deviatoms of the men
95%

almost all of ts values (bo

percent of them) he whin three
deviations of the mean

99.7%

Some 2opie are beiw aege

95 Importance of Normal Distribution

N
Since many psycho-educational measurements (eg intelligeN
noma
achievement test scores) assume normal distribution, the cont of the
Curve is very important to psychologists.
117
Nevhntneicnl Statistts

\aret of pevchological test scores have been found to approXimately

normal distribution While the underlying causes of these
nknown, the use of the normal distribution can be
sifuntions where many small effects are added together into a SCore
that can be observed
phe nome na
theoretical y justri
oT variahle
When measurements produce a normal distribution,
and predictable and help in quickly understanding certain things are
distribution. First, the mean, median, and mode are all the scOres
equal.
wiCto
h
that the normal distribution makes math casy things like AnotSehercond, one can
predict how far from the mean most scores are likely to fall.
-
nis
ni tern the
correlations betweenvariables, and other
Thus, it is possible to determine which
calculations that
ca
are
lcu lating Cor ol a r
domain spmoneecificnts ys
Scores are
more likely
proportion of score likely to be above or below any given score. to occur and the
9.6 Divergence in Normality
In normal distribution,
most cases fall in the middle but
central tendency do not exhibit normal behaviour. there are cases in
Statistically, two numerical measures of shape-
which
used to test for normality. If either of
data set is not normally distributed.
skewness
these values is not
and kurtosis -
can be
close to
zero, then a
9.6.1 Skewness

Distribution
there is
is normal when the
Mean, Median and
a perfect
balance between the Mode coincide together and
when the Mean, Median and right and left halves of the
the center of gravity is Mode fall at different figure. Rut
points in the
measure of symmetry, orshifted
to one side it is
said to be skewed.distribution, and
more precisely, the lack of Skewness is a
Mean
symmetry.
Mode Median
Median Mode Mode
Mean
Median
Mean

() Negatively Skewed
Distribution (+) Positively Skewed
(Left skewed) Distribution
Symmetric Distribution (Right skewed)
 Normal Distribution

skeWDCssis defined As a distribution not having equal probabilities above

Ihe the mean." So, in fact greater the gap between the mean and the
greateris
the skewness and if there is no difference in mean and median
median.
skewness willbe 0.
then
mcansthat the curve is not symmetrical. In askewed distribution, the
SkewnCSS
cdian and mode do not coincide it pulls the median and mean avway
modetoeither left or right. In askewed distribution,the frequency curve is
tiom and values do not lie on
shaped both sides of measure of central
bell cqually. Here, mean, median and mode fall at different
points.
rendency
PositiveSkewness
Mode

the
positively skewed curve, Median

Mean and Median lie

In

value of
from Mode values (right
away
hand),the
values are greater than
Mean
Mode.

are skewed to the right

Data that the
longtail that extends to
have a
right, and the coefficient of
Positive skew 5 7
skewness is positive.
ofa plus sign.
starts offgoing up, like part
skews the
extreme data results are larger. This
Anositive skew means that the the
(average) up. The mean will be larger than
data in that it brings the mean
median ina skewed data set.
Skewness >0

Mean> Median > Mode

Mode
Negative Skewness Median

In negatively skewed curve, the value

of Mean and Median lie away from
Mode values (left hand), the values
are smaller than Mode. Mean

Data th¡t are skewed to the left have a

long tail that extends to the left, the
7 10
coefficient of skewness is negative. 0

119
Pehological Storistics

Negative skew starts ofl flat. like a minus Sign.

Anegative skew means the opposite: that the extreme
data
andresultmodes
This means that the mean is brought down, and the median
than the mean.
Skewness < ) are
Mean <Median < Mode

Positiye Skewness
Right Skewed Distribution Negative Skewness
Left Skewed Distribution
Adistribution has a long right tail Adistribution
has a
Positive skew is probably more Negative skew is long-left
tail
COmmn
Tail is to the Right
Common
Tail is tothe
normal l
noy
Left
Mean > Median > Mode Mean < Median <
Skewness > 0 Mode
Skewness <0
Salary data is often positively Failure rate data is often
skewed skewed
negatively
9.6.2 Kurtosis

Kurtosis means the bulginess. In statistics,

'peakedness' of the probability distribution kurtosis
is any
of a measure of the
measures the degree to which a distribution is more orrandom variable.
distribution. less peaked
than a nomal Kurtosis
() Leptokurtic (Lepto means 'thin')

(Platy means
'broad'or lat')
(-) Platykurtic (Meso means
middle )
(0) Mesokurtic
(Normal)

120
 Normal Distribtutin
P'robability (uve is moderately peaked If any frequency curve is
01 flatter than the Normal Probability (urve, we can say the
/suhutIondiverges from normality. Kurtosis is a measure of such divergence

aethree kiinds of kurtosis:

here

leptokurtic: Whenthe frequency distribution is more peaked at the center then

th Nornal
|Crve is called| as Leptokurtic This kind of peakedness implies a
distrbutton. Data sets with high kurtosis tend to have heavy tails. r
thn
Dtlers.

Mesokurtic: When the frequency distribution is normally distributed the curve

Mesokurtic.Innother words, a normal curve is also known as mesokurtic.
Platvkurtic: Nhen afrequency distribution is flatter than the normal curve it is
Playkurtic. Data sets with low kurtosis tend to have light tails, or lack
calledas
ofoutliers.
Mesokurtic Platykurtic
Leptokurtic
Positive Kurtosis Normal Distribution Negative Kurtosis
Peak is neither high nor Curve has very flat top
Curve has high peak flat for its peak
The graph is almost a
in data
Ahuge swing ** flat line, indication of
be an
values that can
perfect process
indication ofa problem Thinner or Light tails or
or
Flat or Long tail lack of outliers
outliers Mean < Median
Mean > Median Mean = Median
The observations cluster
The observations cluster
less and have shorter
more and have longer tails
tails
Value is greater than Value is 0.263 Value is less than 0.263
0.263

9.7 Measuring Divergence

distribution
calculating Kurtosis and Skewness for a
There are several ways of population.
estimatingthem from a sample of a
of data and compatible ways of referenced.
Kurtosis or Skewness are used or
However, it is very rare that either
121
Psvchological Stotistics

9.7.1 Mensuring Skewness

The coefficient of Skewness is a mneasure for

the
Variable distribution. The Skewness value can be degree of
undefined. A value of zero or a value close to
zeropositive or
right-hand sides of the distribution are balanced indicates hat he
\eft
Statistical
Sk
3(Mean- Median)
central tendency

Methods
S7=
percentile values
(P90-P1o) -Pso
2

Observation
Inspection of
frequency poygon
The direction and extent of skewness can be
measured in
charts depict the two important methods to measure the various viz.ways. BeloN
and observational method. Statistical method
can
tendency and percentile values to measure the skewness.,use
skewness,
measures statisicd
of

Where,
SK= Skewness (The skewness is
G= Standard Deviation
unitless)
P10 = 10h percentile
Pso 50h percentile
P9o =90th percentile
Skewness tells the amount If Skewness is. The distribution will be.
and direction of skew
(departure from horizontal Highly Skewed
symmetry). The skewness
value can be positive or 10.5-+1 Moderately Skewed
negative, or even undefined.
As a general rule of ±0.5
thumb: Approximately
Normal

122
 Measmring kurtosis
, 2
Krtosis measures the whole extent of the
high k
Irtosis
datais
dist ribution is evident in distrihution
a sharp peak
curve
that IS.lhe clustered atound the central
istribution has a
and
Autosts more rounded peak mean), Anegative vale krm
indeatesthatthe and ashorter. thnner tal tu
distribution
is (latter and more spread out than normal
coclicicnt of
Thc Kurtosis is a measure for the
distrubution, The Kurtosis can be measured in degree of tail in fhe vanahe
the (W0
mportant
methods to various ways. Below charts degut
ohserational method measure the kurtosis, viz. statistical
kurtosis.
Statistical method uses percentile values to
md
measure the

Methods
Statistical
Observation
Pao- P10 Inspection of
frequency
polygon
Where,
Ku = Kurtosis
Q= quartile deviation
Pi0 = 10h percentile
Po = 90h
percentile
For percentile coefficient of
kurtosis, interpretation of calculation value will be:
isDistribution
K< 0.263 Leptokurtic

K, 0.263 Mesokurtic

K, >0.263 Platykurtic

123
Psehological Stotisties

9.8 Skew ness vs Kurtosis

Skewness Kurtosis
Significant skewness and kurtosis clearly indicate that data is
Skewness is probably more important than kurtosis not norrmal
Both are unit-less
Measure of Symmetry. of more

precisely, the lack of symmetry Measure of shape, i.e. tall or fat

Skewness measures the asymmetry of In kurtosis, the values
are
the distribution
The skewness for a standard normal
spread out on agraph
The kurtosis for a
unevenly
distribution is Zero standard
If adistribution is the
distribution is Three normal
wrong shape If a distribution
because it is not symmetrical then this is because it is is the wrong shane
skew not the characteriss.
bell shape
Skewness tells about the amount and Kurtosis then this is Kurtosis
direction of skew tells about how tall and
sharp the central peak is
Skewness is either left or right Kurtosis is either platykurtosis and
Skewness is leptokurtosis
thinner as compare to Kurtosis is
Kurtosis thicker as compare to
skewness
 Normal Distribution
Suggestibility
possiblehomogeneous
possible hetrogeneous
Kurtosis
Kurtosis Interest
Attitude

is
is Skewness PositiveNegative
Skewness
and and
in
in small
small, Bias, smali
resultsresults
is e.g.
by by
is Sample Score
group group
Score
Sample being Non-Normal
isMeasured

Trait
Norma
inFDIvergence

Sample
DistribUtiOn
or Made
Poorly
Unsuitable
the
of Tests
Selection
Causes
o

and
Construction
ofAdministration
the
in
Errors

Tests KurtosisKurtosis

NegativePositive

in
in
the
in
Errors in
Errorunclear to complete results results
Theinstructions
Scoringtimings test
the
Motivation

Test Test
Hard Easy

125