Numerical Desciptive Statistics

Measures of Central Tendency Measures of Dispersion Measures of Linear Relationship between variables Measures of shape
Numerical Descriptive Statistics
Ritesh Pandey
June 21, 2016

Francis Galton
1822-1911
Karl Pearson
1857-1936

Ronald Fisher
1890-1962

Table of Contents
Measures of Central Tendency
Measures of Dispersion
Measures of Linear Relationship between variables
Measures of shape

Table of Contents
Measures of shape

The Arithmetic Mean
Pn
i=1 xi
x=
n

The Weighted Average
Pn
w i xi
x w = Pi=1
n
i=1 wi

The Arithmetic Mean: when each value occurs

with a given frequency
Pn
fi x i
x = Pi=1
n
i=1 fi

The Arithmetic Mean: when values are grouped

into frequency intervals
Pn
fi m i
x = Pi=1
n
i=1 fi
where mi is the midpoint of each interval.

The Arithmetic Mean: calculation from deviations

from any number
Pn
i=1 di
x =A+
n
where di = (xi A) is the deviation of each value from any A.

The Geometric Mean
n
" #1/n
Y
xg = xi
i=1

The Harmonic Mean
n
x h = Pn 1
i=1 xi

The Median: for ungrouped data
First arrange the values in increasing (or even decreasing)

order.
For odd n:
Median is the value at the( n+1
2
)th spot.
For even n:
Median is (conventionally) taken to be the mean of the values
at the ( n2 )th and ( n+1
2
)th spots.

The Median: when each value occurs with a given

frequency
Here, the cumulative frequency table should be drawn up first.

f = n. Now locate the value n2 under the
P
Then find
cumulative frequency column in the above mentioned table.
The corresponding x value will be the median value for x.

The Median: when values are grouped into

frequency intervals
Again, first draw up the cumulative frequency table and find

the median class.
n
2
(cf )
Median = l0 + w0
f0

The Mode: when values are grouped into

frequency intervals
The modal class is the class with the highest frequency.

Mode = l0 + w0
+ +
Here,
= f0 f
and
+ = f0 f+

Quartiles
The spread of the values of the variable is divided into four

equal parts by three locations. The values at these locations
are the three quartiles. The second quartile divides the data
into two equal parts and is obviously the median.The other
two are called the lower and upper quartile respectively.
Denoted by Q1 , Q2 and Q3 respectively.

Quartiles: For ungrouped data
Simply order the data in increasing order and find the 3

locations that divide the dataset into four equal halves. For
Q1 , find n+1
4
and round up to the nearest integer. For Q3 , find
3(n+1)
4
and round down to the nearest integer.

Quartiles: For grouped data
First find the relevant class from the cumulative frequency

table. n
(cf )
Q1 = l0 + 4 w0 .
f0
n
2
(cf )
Q2 = l0 + w0 .
f0
3n
4
(cf )
Q3 = l0 + w0 .
f0

Deciles
These divide the value range into 10 equal halves. They are
obviously nine in number. The kth decile can be calculated by
k
the usual formula.The kth decile class is given by (n + 1) 10
kn
10
(cf )
Pk = l0 + w0 .
f0
for i = 1,2, . . . , 9.

Percentiles
These divide the value range into 100 equal halves. They are
obviously ninty nine in number. The kth percentile can be
calculated by the usual formula. The kth percentile class is
k
given by (n + 1) 100
kn
100
(cf )
Pk = l0 + w0 .
f0
for i = 1,2, . . . , 99.

Table of Contents
Measures of shape

Range
Range = xmax xmin

Interquartile Range
Interquartile Range = Q3 Q1

Semi - Interquartile Range
Q3 Q1
Semi Interquartile Range =
2

(Coefficient of) Quartile Deviation
Semi Interquartile Range

Quartile Deviation = Q3 +Q1
2
Hence,
Q3 Q1
2 Q3 Q1
QD = Q3 +Q1
=
2
Q3 + Q1

The Mean Absolute Deviation
P
|xi x|
MAD =
n
Here, x can be any measure of central tendency.

The Variance
Pn
2 i=1 (xi x)2
s =
n1

The Variance: For Grouped Data
Pn
2 i=1 fi (mi x)2
s =
n1

The Standard Deviation

s= s2

The Coefficient of Variation
s
cv =
x

Table of Contents
Measures of shape

The Covariance between two variables
Pn
i=1 (xi x)(yi y )
sxy = cov (x, y ) =
n1

The ( Coefficient of ) Correlation between two

variables
sxy
rxy = corr (x, y ) =
sx sy
with 1 rxy 1.

Table of Contents
Measures of shape

The Skewness
Pn
x)3 /(n 1) i=1 (xi
Skewness =
s3
Positive skew: right tail longer; mean > median
Negative skew: left tail longer; mean < median

The Kurtosis
Pn
i=1 (xi x)4 /(n 1)
Kurtosis =
s4
Negative: Lepto- (slender) -kurtic
Zero: Meso -kurtic
Positive: Platy- (broad) -kurtic

Chebysheffs Theorem
Theorem (Chebysheff)
Within k standard deviations around the mean, will be found a
proportion equal to at least 1 k12 of the total number of
observations.
2s will contain at least 75% data points.

Chebysheffs Theorem: For a Normal Distribution

3s will contain at least 99.7% data points.

Numerical Desciptive Statistics

Uploaded by

Copyright:

Available Formats

Numerical Desciptive Statistics

Uploaded by

Document Information

Original Description:

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

Numerical Desciptive Statistics

Uploaded by

Copyright:

Available Formats

Measures of Central Tendency Measures of Dispersion Measures of Linear Relationship between variables Measures of shape

Numerical Descriptive Statistics

June 21, 2016

Numerical Descriptive Statistics

Numerical Descriptive Statistics

Numerical Descriptive Statistics

Measures of Central Tendency

Measures of Linear Relationship between variables

Numerical Descriptive Statistics

Measures of Central Tendency

Measures of Linear Relationship between variables

Numerical Descriptive Statistics

The Arithmetic Mean

Numerical Descriptive Statistics

The Weighted Average

Numerical Descriptive Statistics

The Arithmetic Mean: when each value occurs

Numerical Descriptive Statistics

The Arithmetic Mean: when values are grouped

Numerical Descriptive Statistics

The Arithmetic Mean: calculation from deviations

Numerical Descriptive Statistics

The Geometric Mean

Numerical Descriptive Statistics

The Harmonic Mean

Numerical Descriptive Statistics

The Median: for ungrouped data

First arrange the values in increasing (or even decreasing)

Numerical Descriptive Statistics

The Median: when each value occurs with a given

Here, the cumulative frequency table should be drawn up first.

Numerical Descriptive Statistics

The Median: when values are grouped into

Again, first draw up the cumulative frequency table and find

Numerical Descriptive Statistics

The Mode: when values are grouped into

Numerical Descriptive Statistics

The spread of the values of the variable is divided into four

Numerical Descriptive Statistics

Quartiles: For ungrouped data

Simply order the data in increasing order and find the 3

Numerical Descriptive Statistics

Quartiles: For grouped data

First find the relevant class from the cumulative frequency

Numerical Descriptive Statistics

Numerical Descriptive Statistics

Numerical Descriptive Statistics

Measures of Central Tendency

Measures of Linear Relationship between variables

Numerical Descriptive Statistics

Range = xmax xmin

Numerical Descriptive Statistics

Numerical Descriptive Statistics

Semi - Interquartile Range

Numerical Descriptive Statistics

(Coefficient of) Quartile Deviation

Semi Interquartile Range

Numerical Descriptive Statistics

The Mean Absolute Deviation