Statistics

STAT-111
 CONTENTS

I-STATISTICAL METHODS

1) BASIC CONCEPTS
Introduction – Definitions – Data – Types of data – Inclusive class –
Exclusive class – Frequency Distribution – Construction of frequency table –
Cumulative Frequency – Cumulative Frequency distribution - Types of
cumulative frequency – Variable – Attribute – Population – Sample –
Characteristics of statistics – Uses of statistics – Limitations of statistics – Solved
Problems-Exercise.

2) DIAGRAMS AND GRAPHS

Introduction – Advantages and Disadvantages of diagrams and graphs over table
– Difference between diagrams and graphs over table – Difference between
diagram and graphs – Types of diagrams – Types of graphs –Exercise.

3) MEASURE OF CENTRAL TENDENCY

Introduction – Definition – Advantages of averaging – Characteristics of good
average – Types of averaging – Mean – Median - Mode - Merits and Demerits
– Geometric mean – Harmonic Mean – Quartiles – Deciles – Percentiles – Solved
Problems-Exercise.

4) MEASURES OF DISPERSION
Introduction – Definition – Characteristics of good measures of dispersion –
range – quartile deviation – mean deviation – standard deviation – Relative
measures of dispersion – co-efficient of variation – Solved ProblemsExercise.

5) CORRELATION AND REGRESSION

Telegram: AgroMind Website: [Link]

Correlation – Types of correlation - Direct correlation – Indirect correlation –
Scatter diagram – Correlation co-efficient – Properties of Correlation Co-efficient –
Regression – Definition – Curve fitting - Methods of Least square – Regression
coefficient - Properties of regression co-efficient – Solved Problems-Exercise.

6) PROBABILITY AND ITS DISTRIBUTION

Introduction – Terminology – Experiment – Event – Trial – Types of Events –
Equally likely events – Mutually Exclusive events – Exhaustive events –
Compound events – Independent and Dependent events – Definition for
probability – Theorems on probability – Addition and compound theorem –
Random variable – Discrete and continuous Random variable – Binomial
Distribution – Poison distribution – Normal distribution and its properties –
Solved Problems-Exercise.

7) SAMPLING THEORY
Introduction – Population – Sample – Sampling – Sample survey –
Census – Sample survey VS Census – Sampling error – Methods of sampling –
Probability sampling – Simple random sampling – Systematic sampling –
Stratified sampling – Cluster sampling – Multi stage sampling - Non probability
sampling – Purposive sampling – Sampling distribution - Standard error and its
importance – Solved Problems-Exercise.

8) TESTING OF HYPOTHESIS
Scientific hypothesis – Statistical hypothesis – Simple hypothesis – Composite
hypothesis – Null hypothesis – Alternate hypothesis – Testing of hypothesis –
Type I and Type II Error – Level of significance – Critical region - One tailed
and Two tailed test - Degrees of freedom – Steps involved for
Testing of hypothesis – “Z” test – one sample case – two sample case - “t” test –
one sample case – two sample case(equal and unequal variances) – Paired ‘t’ test

Telegram: AgroMind Website: [Link]

– Chi square test – Contingency table – Goodness of fit – Solved
ProblemsExercise.

9) DESIGNS OF FIELD EXPERIMENTS

Introduction – Definitions – Design of experiment – Treatment – Experimental
unit – Experimental material – Experimental error – Basic principles of Design
of Experiment – Replication – Randomization – Local control – Analysis of
Variance (ANOVA) Assumptions of ANOVA – completely Randomized block
design (CRD) – Definition – Randomization – Analysis of CRD with equal and
unequal replication – Advantages and Disadvantages of CRD – uniformity trial
- Necessity – Contour map – Uses – Randomized Block Design (RBD) -
Definition – Randomization – Analysis of RBD – Advantages and Disadvantages
of RBD – Limitations – Latin Square Design (LSD) – Necessity – Layout -
Randomization – Analysis of LSD – Factorial Experiment – Necessity –
Definition – Factor – Single Factor Experiment – Factorial Experiment – Main
Effect – Interaction Effect - Advantages and Disadvantages – Split Plot Design
– Necessity – Main plot treatments – Sub plot treatments –Randomization –
Analysis - Advantages and Disadvantages – Strip plot Design – Necessity -
Randomization – Analysis – Exercise.
II-MODEL QUESTION PAPER WITH ANSWER KEYS
III – STATISTICAL TABLES
1. Probability of Normal Distribution

2. Critical Values of “t” Distribution

3. Chi-Square ( 2) Test

4. One – Digit Random Numbers

5. Two Digit Random Numbers

6. Three Digit Random Numbers

Telegram: AgroMind Website: [Link]

 7. “F-table for 5% and 1% Level”

8. Significant Studentized Ranges for 5% and 1% Level

Duncan’s New Multiple – Range Test (DMRT)

9. Critical Values of Simple Correlation Coefficients(r)

Telegram: AgroMind Website: [Link]

 AGRICULTURAL STATISTICS
UNIT-I BASIC CONCEPTS

The word ‘statistics’ comes from the Italian word ‘statista’ (meaning
“statesman”) or the German word ‘statistik’ each of which means a political state.
It was first used by Professor Gottfried Achenwall (1719-1772), a professor in
Marlborough in 1749 to refer to the subject-matter as a whole. Achenwall defined
statistics as “the political science of the several countries” to help the government
of rule, plan policies in peace and war (or) to collect taxes.

It has several meanings, one meaning is “numerical/quantitative figures”

Examples are number of crops in a farm, number of students in a class and
number of unemployed persons in a district etc. Another meaning of statistics is
“Body of scientific principles and techniques”. Based on that several definitions are
given by different authors as follows.

Definitions of Statistics given by different authors

1. HORACE : ‘By statistics we mean aggregates of facts affected

SECRIST to a marked extent by multiplicity of causes, numerically
expressed, enumerated or estimated according to
reasonable standards of accuracy, collected in a
systematic manner for a predetermined purpose and
placed in relation to each other!.’
2. KING : ‘The science of statistics is the method of judging collective
natural or social phenomenon from the results obtained from the analysis
of enumeration or collection of the estimates’
3. BUDDINGIO : Statistics is a science of estimates and
N probabilities.
4. A.L. : Statistics may be called as a science of counting.
BOWELY
5. COWDEN : Statistics defined as collection, analysis and and
interpretation of data.
CROXTEN
Telegram: AgroMind Website: [Link]
In General
It can be defined as the body of concepts, principles and methods dealing with
collection, summarisation, analysis and interpretation data.
Data
Collection of data points is known as Data.
Primary data
The data collected primarily for a particular study, called primary data.
Eg: census data.
Secondary Data
The data collected for some other purpose and it is used for another studies,
called secondary data or second hand data. Methods of collecting primary data
i) Interview method ii) Mailed questionnaire method.
Interview method
In this method, the investigator from house to house interviews the individuals
personally. He asks the questions one by one and fills up the schedule on the
basis of the information supplied by the individuals.
Mailed questionnaire method
In this method, the questionnaire is mailed to the individuals who are required
to fill them up and return it duly completed.
Qualitative data
If the observations of an experiment are in qualitative nature called qualitative data.
Eg.: Colour of a flower, taste of a fruit, shape of the seed etc.,
Quantitative data
If the observations of an experiment are in quantitative nature, called quantitative
data.
Eg.: Yield of a crop, number of unemployed persons in our nation, rain fall over a
period of time etc.,
Inclusive class

Telegram: AgroMind Website: [Link]

In a class, if the upper limit of a class is not equal to the lower limit of the immediate
succeeding class then it is called inclusive class
Eg.: 10 – 20
21 – 30
31 – 40
41 – 50
Exclusive class
In a class, if the upper limit of class is equal to the lower limit of immediate
succeeding class is called exclusive class .
Eg.: 10 – 20
20 – 30
30 – 40
40 – 50
Frequency distribution
Summarisation of data starts with classification of the data. Arranging of data
according to some common features is known as classification. Broadly
classificatin is done according to the type of data ie., qualitatiive or quantitative
data. In case of qualitative data, we count items with specified character and
put them in appropriate categories. If the data are measurable (ie quantitative),
they are classified according to classes, Sorting of data into categories or classes
will lead to formation of frequency distribution .It is a quantitative or qualitative
classification of data.
Construction of frequency distribution for quantitative data
Step 1 : The number of classes will be decided by Yule’s formula
Number of Classes = 2.5 X (n)1/4
Step 2: The class interval will be decided by the relation
(Maximum value in the data set) - (Minimum value in the
data set)
Class interval = - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Number of classes
Telegram: AgroMind Website: [Link]
Step 3: Construct a rough frequency distribution by using exclusive classes with tally
bars.
Step 4: Construct a fair final frequency table by using exclusive classes. Step 5:
A self-explanatory title should be given at the top or bottom of the frequency
table.

Cumulative frequency curve : (or Ogive)

The cumulative frequency of a class is the total frequency up to and
including that class. The table of cumulative frequency is called cumulative
frequency distribution. There are two type of cumulative frequency distribution.
The cumulative frequency of all values greater than or equal to the lower limit
of each class is called more than cumulative frequency distribution. The
cumulative frequency of all values less than to the upper limit of each class is
called less than cumulative frequency distribution.
In drawing frequency curve, if the cumulative frequencies are taken then the
curve is said to be Cumulative frequency curve (or) Ogive. If it is greater than
type then the curve is said be greather than [Link] it is less than type then the
curve is said to be less than Ogive.
Variable
A quantitative or qualitative characteristic that varies from observation to
observation in the same group in called a variable. Generally quantitative
variables are called variables.
Eg. Yield of a crop, height of the students etc.,

Attribute
The qualitative variables are called attributes.
Eg. : Crop varieties, Soil types, Shape of seeds, Colour of flowers etc.,

Population
Telegram: AgroMind Website: [Link]
 The aggregate or totality of all possible objects possessing the specified
characteristic is called a population.
Eg: If some information is required about an average fourth year [Link].(Ag.)
student in Agricultural college & Research Institute Madurai,then thr collection
of fourth year [Link]. (Ag) students in AC&RI,Madurai will be the population.
Sample
It is defined as smaller part of population.
Eg. : If collection of fourth year students in AC&RI,Madura is the population, then
the students in ‘A’ batch of fourth year will be the sample.

Characteristics of : 1. They must be numerical statements.

statistics
2. They must be aggregates of facts.
3. They must be related to each other.
4. They must have certain objectives behind them.
5. They must be affected to a marked extent by
multiplicity of causes.
6. They are estimated with a reasonable standard of
accuracy.
Uses of statistics : 1. Simplifies complexity of the data.
2. Measures the results like averages etc.
3. Studies the relationship.
4. Useful for estimations.
5. Useful for surveys.
6. Useful in Agricultural and biological sciences
medical sciences research.
Limitations of : 1. Statistics does not study qualitative data. i.e. It
statistics studies only numerical figures.
2. Statistics does not study individuals. i.e. It
studies only aggregates.
3. Statistics is liable to be misused.
4. Statistical results some times lead to wrong
conclusions unless they are properly used.
5. Statistical laws are not exact.

Solved Problems
Steps involved for construction of frequency table:
Telegram: AgroMind Website: [Link]
 Step 1 : Find out the maximum and minimum values for the given
data and calculate range(R).
R = maximum value – minimum value.
Step 2 : Decide the number of class. For this, Yule’s formula is used.
No. of class (N) = 2.5 x n1/4 , n – total number of observation.
Step 3 : Find out the Class Interval ( C.I)
C. I = R / N , R- Range; N-Number of classes.
Step 4 : Construct a rough frequency table with exclusive or inclusive
classes using tally bars or tally marks.
Tally marks are used to transfer the data to frequency table.
Step 5 : Construct a fair frequency table with exclusive and inclusive
classes and give suitable title at the top or bottom of the table.

1) In an uniformly trial (to know fertility gradient) with malvi cotton the yield
of 100 plots were as follows. The yields will record in grams. Form a
frequency table and find the cumulative frequency.
93 81 57 42 95 80 52 70 111 72
49 74 60 57 63 51 90 41 50 66
89 85 59 44 90 89 69 68 95 112
85 84 63 99 91 64 78 33 115 101
63 85 23 58 96 55 67 63 68 88
86 82 89 79 82 83 57 66 113 102
62 46 79 89 86 54 29 51 119 83
94 71 51 62 76 68 54 69 119 39
45 58 58 81 96 52 49 106 115 85
51 65 56 31 79 45 78 87 111 77

Step: 1

For the given values,

The maximum value = 119
Telegram: AgroMind Website: [Link]
 The minimum value = 23
Range = 119-23
=96
Step: 2
No. of class = 2.5x(n)1/4
= 2.5 x (100)0.25
= 2.5 x 3.162
= 7.9
Step: 3
The class interval = Range / No. of class
= 96/7.9
= 12
Step: 4 Rough Frequency table
CI Tally bars Frequency
23-35 IIII 4
35-47 IIII II 7
47-59 IIII IIII IIII 19
IIII
59-71 IIII IIII IIII 19
IIII
71-83 IIII IIII IIII 15
83-95 IIII IIII IIII 20
IIII
95-107 IIII III 8
107-119 IIII I 6
119-131 II 2

Step: 5 Fair frequency table:

Malvi cotton yield (gm)

CI F
23-35 4
35-47 7
47-59 19
59-71 19
71-83 15
82-95 20
95-107 8
107-119 6

Telegram: AgroMind Website: [Link]

 119-131 2

2) The following data on length of ear head (cms) in 100 wheat plants.
Construct a frequency table and also find cumulative frequency.
10.0 12.5 9.5 8.9 10.2 10.1 10.0 12.5 12.5 8.9
10.7 8.3 9.4 9.4 12.0 9.9 8.2 11.7 10.5 9.0
9.8 9.8 10.0 10.2 13.0 9.8 11.4 8.7 11.3 11.2
10.6 10.4 10.3 11.4 9.7 8.4 7.5 8.5 10.8 9.7
11.2 10.8 7.5 9.9 8.9 8.8 9.6 10.0 10.7 11.0
9.2 10.0 8.4 10.2 7.9 9.0 6.6 8.3 10.5 11.1
9.2 9.6 11.6 10.1 8.9 9.5 9.9 7.9 9.5 9.7
11.2 8.8 11.1 10.4 9.5 13.4 12.7 10.0 9.1 8.9
8.5 11.8 9.5 9.9 11.0 10.5 10.2 10.9 9.1 10.1
9.7 9.5 11.3 9.4 9.8 7.7 9.6 9.7 12.1 11.4

Step: 1 For the given values, the maximum value = 13.4

the minimum value = 6.6
Range = 13.4 – 6.6
= 6.8
1/4
Step: 2 No. of class = 2.5 x(n)
= 2.5 x (100)Q.25
= 3.162 x 2.5 = 7.9
Step: 3 The class interval = Range / No. of class = 6.8/7.9 = 0.9

Step: 4 Rough frequency table

CI Tally bars Frequency
6.6-7.5 I 1
7.5 – 8.4 IIII III 8
8.4 – 9.3 IIII IIII IIII III 18
9.3 – 10.2 IIII IIII IIII IIII IIII IIII 34
IIII
10.2 - 11.1 IIII IIII IIII III 18
11.1 – IIII IIII III 13

Telegram: AgroMind Website: [Link]

 12.0

12.0 – IIII I 6
12.9
12.9 – II 2
13.8

Step:5 Fair frequency table

Ear head length in wheat plants
CI Frequency Cumulative
frequency
6.6-7.5 1 1
7.5 – 8.4 8 9
8.4 – 9.3 18 27
9.3 – 10.2 34 61
10.2 - 11.1 18 79
11.1 – 13 92
12.0
12.0 – 6 98
12.9
12.9 – 2 100
13.8

Exercise
PART – I Explain
/ Define
i. Statistics ii.
Types of Data iii.
Variables iv.
Attributes
v. Methods of Collecting Primary Data
vi. Frequency vii. Frequency Distribution viii.
Cumulative frequency ix. Cumulate frequency
distribution
x. Types of Cumulative frequency distribution xi.
Ogive
xii. Inclusive and exclusive classes with examples

Telegram: AgroMind Website: [Link]

 Part – II
1. Uses of statistics
2. Definition of statistics by various authors
3. Characteristics of statistics
4. Limitation of statistics

Part – III
i. Briefly explain about steps involved for the contraction of frequency
table
ii. Explain the following
i. Definition of statistics ii.
Uses of statistics
iii. Characteristics of statistics
iv. Limitation of statistics

UNIT-II DIAGRAMS & GRAPHS

Introduction
The visual aids, like diagrams and graphs, are helpful in understanding the
facts of numerical data easily. The salient features of the data can be known at a
galance by using these visual aids. Comparison of data is also made easier by
such diagrams and graphs.
Advantages :
i) They are more attractive ii) They give complete analysis of the given
data at a glance. ie., they are more eye-catching.
iii) They are more impressive. iv) They are useful in the
comparison of two or more data sets.

Disadvantages:

i) The charts/graphs con not show as many facts as can be shown in

a table
ii) Exact values con not be given in a chart/graphs iii)
Charts require a certain amount of time to construct

Telegram: AgroMind Website: [Link]

Differences between diagrams and graphs
There is no clear cut definition about the differences between diagrams and
graphs. But some points may be remembered while drawing diagrams and
graphs.
i) Generally speaking graphs are drawn on graph paper while diagrams
are on plain sheet. ii) A graphs represents mathematical relationship of
two variables where as a diagram does not.
iii) Diagrams are not helpful in the analysis of the data but graphs are.
iv) To represent a frequency distribution, graphical representation are
preferable than diagrams.

Telegram: AgroMind Website: [Link]

Telegram: AgroMind Website: [Link]
 Important diagrams
i) Simple bar diagram ii)
Multiple bar diagram
iii) Component bar
diagram iv)
Percentage bar diagram
v) Pie-chart
Important graphs
i) Histogram ii)
Frequency curve iii)
Frequency polygon iv)
Ogive
Simple bar diagram
It consists of vertical bars of equal width. The heights of the bars are
proportional to the volume or magnitude of the attribute. All bars stand on the
same base line. The bars are separated from each other by equal intervals. The
bars may be coloured are marked.

Multiple bar diagram

It is simply the extension of simple bar diagram. For each attribute two or more
bars to be placed side by side. Each bar within a attribute will be marked or
coloured differently in order to distinguish them. A foot note has to be given
explaining the markings or colourings.
Component bar diagram
It is also called sub-divided bar diagram. Here the bars are placed one on top of
the other. These diagrams are used to represent various parts of a whole. Each
component will be marked or coloured differently. A foot note has to be given
explaining the markings or colourings.

Telegram: AgroMind Website: [Link]

 Percentage bar diagram
It is a type of component bar diagram. Here the frequencies are converted into
percentage. The total frequency of each attributed is taken as 100 percentage.
Therefore the heights of rectangular bars are same.
Pie chart
It is a circular diagram. It may be used in place of bar diagrams. Here the
frequencies are converted into angles by taking 360 degrees as whole. For a
convenient radius a circle is drawn and it will be divided into many sectors
according to the converted angles. These sectors may be marked or coloured
differently. A foot note has to be given explaining the markings or colourings.
Histogram
When the data are classified based on the class – intervals, It can be represented
by a histogram. Histogram is just like a simple bar diagram with minor
differences. There is no gap between the bars since the classes are continuous.
The bars are drawn only in outline without colouring or making as in the case of
simple bar diagram.
Frequency polygon
It is a graph of frequency distribution. There are two ways to construct a
frequency polygon.
i) In a histogram if we join the mid points of the top of these
rectangular bars we get a polygon. Such a polygon is called a
frequency polygon.
ii) A frequency polygon can also be drawn directly without
constructing the histogram. This is done by plotting the mid points
with frequencies as co – ordinates on a graphs sheet. These points
are joined by straight lines to get the frequency polygon.
Frequency curve

Telegram: AgroMind Website: [Link]

 The procedure for drawing a frequency curve is same as for frequency polygon.
But the points are joined by smooth or free – hand curve. The curve may not
pass through all points as in case of frequency polygon.
Exercise
Part – I
Explain / Define
1. Simple hour diagram
2. Multiple hour diagram
3. Component hour diagram
4. Percentage hour diagram
5. Pie chart
6. Histogram
7. Frequency curve
8. Frequency Polygon
9. Ogive

Part – II
1. Advantages of diagram and graphs
2. Disadvantages of diagram and graphs
3. Distinguish between diagrams and graphs
Part – III
i. Briefly explain about Types of diagram
ii. Explain “Types of Graphs

Telegram: AgroMind Website: [Link]

 UNIT III MEASURES OF CENTRAL TENDENCY
From the frequency distribution, we can only have a general opinion of the given
data. It is used to condense the bulk of unorganised data .To analyse the given
data some statistical constants will be defined. In a bulk of unorganised data,
there is a tendency that majority of the values will cluster around the central
value of the data. This tendency of the data is known as central tendency of the
[Link] method or measuring these constants is known as averaging.
Advantages of averaging
i) It will reduce the complexity of the data ii) It will become a useful tool
for the comparisons of two or more data sets.
Characteristics of a good average
i) It should be well defined.
ii) It should be based on all observations. iii) It should be
suitable for further algebraic treatment. iv) It should be
easy to calculate.
v) It should be easy to understand.
vi) It should bot be affected by sampling fluctuation.
Types of averaging
i) Arithmetic mean ii) Median
iii) Mode iv) Geometric mean v)
Harmonic mean vi)
Quartiles, deciles & percentiles.
Arithmetic mean
It is the value arrived at by dividing the sum of all observations by the total
number of observations.
Arithmetic Mean = ( Sum of all observation ) / ( Total number of observation )
Case 1 : For raw data

Mean ( x ) = ( Σx / n )
Telegram: AgroMind Website: [Link]
 = ( x1 + x2 + . . . . . . . +xn ) / n

where, n – Total number of observation.

Case 2 : For discrete data

Mean ( x) = ( Σ f x / n )

Where , n- total frequency = Σ f , f- frequency

Case 3 : For continuous data

1) Direct Method

Mean ( x) = ( Σ f m / n )

Where, m- midpoint of each class, n- total frequency.

2) Indirect / short cut method

Mean (x) = A + [Σ f d / n]

Where, A - assumed mean, n – total frequency

D= m – A , m – mid point.

3) Step – deviation Method

Mean (x) = A + [Σ f d / n] x i

Where, d – assumed mean, n – total frequency

d = ( m – A ) / i , i= common class interval.

Merits of arithmetic mean

i) It is well defined

Telegram: AgroMind Website: [Link]

 ii) It is based on all observations iii) It is easy
to calculate iv) It is easy to understand
v) It is suitable for algebraic treatment vi) It is least
affected by fluctuations of sampliling.

Demerits
i) It is affected by extreme values ii)
It cannot be determined by inspection
Median
It is the middle most value of the given data after arranging the data in ascending
descending order. Case – 1: For raw data

Step -1 : Arrange the data in ascending or descending order. Step–

2: Median = Size of ( N+1)/2 th item ( for odd numbers )

Where, N – total number of observation.

For even number,

Median = ( n/2 th item + ( n+2) /2 th item ) / 2

Case – 2 : For discrete data

Step -1 : Arrange the data in ascending or descending order.

Step–2 : Median = Size of ( N+1)/2 th item

Where, N – total frequency.

Case-3 : For continuous data

Step 1 : Find N/2.

Step 2 : Median = L + [ ( N/2 – c f) / f ] x i

Telegram: AgroMind Website: [Link]
 Where, L = Lower limit of median class

N = Total frequency

Cf = Cumulative frequency of pre-median class

F – frequency of the median class

i - width of the median class

Merits of median
i) It is easy to calculate ii) It is
unaffected by extreme values iii) It
can be determined by inspection
Demerits
i) It is not well – defined ii) It is not based
on all observations iii) It is affected by
fluctuations of sampling iv) It is not suitable
for algebraic treatment.
Mode
Mode is a value which has the greatest frequency in a given distribution.
Case -1 : For raw data

Value which occur more frequently.

Case -2 :For discrete data

Value which has the highest frequency is called mode of the data.

Case – 3: For continuous data

Mode = L + [ ∆1 / ∆1+∆2] x i Where, L

– lower limit of modal class.

i – width of the modal class

Telegram: AgroMind Website: [Link]
 ∆1 – Difference between the frequencies of modal class and pre-modal

class.

∆2 - Difference between frequencies of modal class and post modal class.

Merits
i) It can be easily determined ii) It is
not affected by extreme values.
iii) It can be determined by inspection
Demerits
i) It is not well defined ii) It is not
based on all observations.
iii) It is not suitable for algebraic treatment iv)
It is affected by fluctuations of sampling.
Geometric Mean: It is the nth root of the product of n observations.
Case -1 : For raw data
G.M = ( x1, x2 . . . . . . . .xn )1/n
Take log on both sides,
Log G.M = 1/n [ log x1 + log x2 + . . . . . + log xn ]
Log G.M = ( Σ log x / n )
G.M = Antilog of ( Σ log x / n )

Case – 2 : Discrete data

G.M = Antilog of ( Σ f log x / n )
Where, n – total frequency, f – frequency.

Case – 3 : Continuous data

G.M = ( Σ f log m / n )
Where, m – midpoint of each class, n – total frequency.

Telegram: AgroMind Website: [Link]

 Harmonic Mean: It is the reciprocal mean of the reciprocal observation.
Case -1 : For raw data
H.M = ( Σ(1/ x) / n ) or n / (Σ(1/ x)

Case – 2 : Discrete data

H. M = n / ( Σf (1/ x) )
Where, n – total frequency.

Case – 3 : Continuous data

H. M = n / ( Σf (1/ m) )
Where. M – midpoint, n – total frequency

Quartiles: Quartiles are the values which divide the given distribution into 4
equal parts.
Deciles: Deciles are the values which divide the given distribution into 10 equal
parts.
Percentiles:Percentiles are the values which divide the given distribution into
100 equal parts.
Solved Problems
1. If the weights of 5 ear head of sorghum plants are 100, 102, 118, 124 and
126 , Calculate mean, median and mode values and also interpret the results. i.
Mean
Sum of all observation

x
x= = -----------------------------------------
-
n
Total no. of observation

100+102+118+124+126
= ------------------------------------------ 5
570
Telegram: AgroMind Website: [Link]
 = -----
5
= 114g.
ii. Median
In ascending order
100, 102, 118, 124, 126, it is for the given data is odd

N +1
So the median is th
item
2

= 5 +1 th item
2
rd
= 3 item
3rd item = 118

iii. Mode
There is no repeated value in this data, it has no mode.

Conclusion:
The mean median mode. So the data follows asymmetrical one.

2. The following data related to yield of paddy plants collected from 14 different
plots. Calculate
Mean, Median and Mode and also interpret the results.
60, 62, 35, 58, 32, 30, 42, 48, 49, 40, 52, 55, 38.

Telegram: AgroMind Website: [Link]

 i. Mean

x=
x
n

60+62+35+58+32+30+42+48+49+40+52+55+38
= ----------------------------------------------------------------
13
601
= ------
13
= 46.2

ii. Median
In ascending orders:
30,32,35,38,40,42,48,49,52,55,58,60,62. For the given data is odd

number

N +1
So the median is th
item
2

13 +1
= th
item
2

= 7th item.

7th item = 48

iii. Mode
In this data, there is no repeated value. So there is no mode.

Conclusion:
The mean median mode. So it is also called asymmetrical one.

[Link] number of cotton bolls recorded in 23 plants are given below. Calculate
Telegram: AgroMind Website: [Link]
 Mean, Median and
Mode.
31 21 28 29 26 24 36 22 25 25 16 19

22 30 21 26 31 35 29 31 34 40 20

Solution:
i. Mean
sumof all observation
=
[Link] observation

x= xn

=
Mean=27 ii.
Median
Ascending order
16, 19, 20, 21, 21, 22, 22, 24, 25, 25, 26, 26, 28, 29, 29, 30, 31, 31, 31, 34,
35, 36, 40
+1
Median = N
2

= =
=12
the12th item=26

iii. Mode

The repeatable value is 31.

Conclusion:
For the data the mean # median # mode. So it is asymmetrical one.

Discrete type of data:

Telegram: AgroMind Website: [Link]

 [Link] of seeds germinated out of 5in each of 50 pods, calculate Mean,
Median and Mode.

No. of seed No. of pods

germinated (f)
(x)
0 4
1 13
2 16
3 9
4 5
5 3

Solution:
No. of seed No. of fx Cumulative
germinated pods (f) frequency(cf)
(x)
0 4 0 4
1 13 13 17
2 16 32 33
3 9 27 42
4 5 20 47
5 3 15 50
Mean x=
fx
f

= 2.14
Median
n +1 th

= item
2

51 th

= item
2

Telegram: AgroMind Website: [Link]

 = (25.1)th item
The nearest cumulative frequency is 33. So the median is 2.

Mode
Find the greatest frequency. In this problem the greatest frequency is 16. So the
mode is 2.

Conclusion:
Mean = Median = Mode. So it is called as symmetrical distribution.

[Link] following frequency distribution gives the number of cabbage plants on

60 cabbage plants, calculate mean, median and mode
x f
2 4
4 7
8 11
16 13
32 25

x Frequency (f) fx Cumulative

frequency(cf)
2 4 8 4
4 7 28 11
8 11 88 22
16 13 208 35
32 25 800 60

i. Mean
fx
=
f

=
=18.8

ii. Median

Telegram: AgroMind Website: [Link]

 N +1 th item
=
2

60 +1 nd
item
=
2
= 61/ 2 = 30.5
The nearest cumulative value is 35. So the median is 16.

iii. Mode
Find the greatest frequency. In this problem the greatest frequency is 25. So
the mode is 32.

Conclusion:
The mean # median # mode. So it is also called as asymmetrical one.

[Link] following is the frequency distribution of the number of mites / leaf on 75

leaves. Calculate mean, median and mode.
Number of mites/ leaf 0 1 2 3 4 5 6

Number of leaves 3 16 21 32 2 0 1

Solution:

Number of mites leaf Number of leaves fx cf

(x) (f)
0 3 0 3
1 16 16 19
2 21 42 40
3 32 96 72
4 2 8 74
5 0 0 74
6 1 6 75
Total =21 75 168 357

Telegram: AgroMind Website: [Link]

 Mean x =
fx
f

= =2.24

Median:-
N +1 th item
=
2

= = =38
The nearest cumulative value is 40. So the class is 2.

Mode:
Find the greatest frequency. In this sum the greatest frequency is 32. So the class
is 3

Conclusion:
It is asymmetrical one, because mean # median # mode.

Continuous type of data

[Link] Frequency distribution table gives the weights of 190 sorghum ear
heads, calculate Mean, Median and Mode.

Weight of ear head No. of ear head

(x). (g) (f)
40-60 6
60-80 28
80-100 35
100-120 55
120-140 30
140-160 15
160-180 12
180-200 9
Total 190
Telegram: AgroMind Website: [Link]
 Solution:

Weight of ear No. of ear head fx cf m

head (x). (g) (f)
40-60 6 300 6 50
60-80 28 1960 34 70
80-100 35 3150 69 90
100-120 55 6050 124 110
120-140 30 3900 154 130
140-160 15 2250 169 150
160-180 12 2040 181 170
180-200 9 1710 190 190
Total 190 21360 927 960
x=
fx
n

i. Mean =
=112.4g

ii. Median x =l + N /2−CF i

F
l =100
N /2=95 CF
=69
i= 20
F =55

x =100 + 95− 69 x20

55
=100 + 9.4545
=109.4545
iii. Mode
x=L+ S1 xi
S1+ S2

Telegram: AgroMind Website: [Link]

 =100
S1=(55−35)
S2=(55−30) i=20
x=100+x20

=100+ x20
=108.88

Conclusion
For the data the mean # media # Mode. So it is asymmetrical one.

[Link] the frequency distribution of seed yield of sesamum given in table.

Calculate Mean, Median and Mode.
Yield per plot No. of post
(x) (f)
65-84 3
85-104 5
105-124 7
125-144 20
145-164 24
165-184 26
185-204 12
205-224 2
225-224 1
100
Solution:
Yield per plot (x) No. of post m fx cf
(f)
65-84 3 74.5 223.5 3
85-104 5 94.5 472.5 8
105-124 7 114.5 801.5 15
125-144 20 134.5 2690 35
145-164 24 154.5 3708 59
165-184 26 174.5 4537 85
185-204 12 194.5 2334 97
205-224 2 214.5 429 99
225-224 1 234.5 234.5 100
Telegram: AgroMind Website: [Link]
 100 1390.5 15430 501
i. Mean

x=
fx
n

=
=154.3
ii. Median

(N / 2−CF)
x =l + xi
F
l =145
N / 2= 50
CF =35
i=19

F = 24 x =145+

x19

=145+ x19
=145+11.875
=156.875
iii. Mode
x =l + S1 xi
S1+ S2 l
=145 S1=(24 −
20)
S2 = (24 − 26) x

=145+ *19
=145+ 38
=183

Conclusion:
For the data the mean # median # mode. So it is asymmetrical one.

Telegram: AgroMind Website: [Link]

 [Link] following frequency distribution gives the number of tomatoes / plant on
100 tomato plants. Calculate mean, median and mode.
Number of Number of Plants
tomatoes / plant

1-3 13
3-5 53
5-7 26
7-9 8

Solution:
CI f fm cf m
1-3 13 26 13 2
3-5 53 212 66 4
5-7 26 156 92 6
7-9 8 64 100 8
Total 100 458 271 20

Mean

x = fx = 458
f 100
Mean =4.58
Themeanvalueis 4.68
Median

N −CF
2
=L+ *i
F

=3+ 50 −13 2=3+ x2=4.396=4.4

53
Mode

Telegram: AgroMind Website: [Link]

 =L+ Si
xi Si=(53−13)
Si+ S2

=3+ x2 =3+ x2 =3 + =4.19

Conclusion:
For the given data mean = median= mode. So the data is symmetrical
one.

[Link] Mean, Mode and Median for the following frequency distribution

table .
Weight (gm) No. of Onion (F)

75-108 5

108-141 9

141-174 15

174-207 13

207-240 7

240-273 1

Solution:

Weight No. of onion m fm cf

75-108 5 91.5 457.5 5
108-141 9 124.5 1120.5 14
141-174 15 157.5 2362.5 29
174-207 13 190.5 2476.5 42
207-240 7 223.5 1564.5 49
240-273 1 256.5 256.5 50
Total 50 1044 8238 189
Mean
Telegram: AgroMind Website: [Link]
 x = fm =8238 =164.76
f 50
Median

2 2
1+ N −CF xi 141+ 50

−14 x33 md = =
F 15

=141+ (11/15)x33
=141+ 24.189=165.2

Mode

=L+ S1 i
S1+ S2

=141+ x33

=141+ x33 =141+24.75 =165.75

Conclusion:
For the given data the mean = median=mode. So the data is symmetrical one.

[Link] Mean, Median and Mode for the following frequency distribution
table.
Class Interval Frequency

6.6-7.45 1

7.45-8.3 6

8.3-9.15 19

9.15-10.0 27

Telegram: AgroMind Website: [Link]

 10.0-10.85 24

10.85-11.7 13

11.7-12.55 7

12.55-13.4 2
13.4-14.25 1
Solution:

Class interval Frequency (f) m fm cf

6.6 - 7.45 1 7.025 7.025 1

7.45 - 8.3 6 7.875 147.25 7
8.3 - 9.15 19 8.725 165.775 25
9.15 - 10.0 27 9.575 258.525 53
10.0 - 10.85 24 10.425 250.2 77
10.85 - 11.7 13 11.275 146.575 90
11.7 - 12.55 7 12.125 84.875 97
12.55 - 13.4 2 12.975 25.9599 99
13.4 - 14.25 1 13.525 131.825 100
Total 100 1000

Mean

x= fm,= 1000,=10
f 100

Median

N −CF
2
=l+ x1
F

Telegram: AgroMind Website: [Link]

 100 − 25
2
= 9.15+ x0.85
27

= 9.15+ x0.85
=9.936
Mode
=l + S1 *i
S1+ S2

= 9.15 + x0.85

= 9.15 + 0.72
= 9.87

Conclusion
For the given data the mean = median=mode. So the data is symmetrical one.

[Link] Mean, Median and Mode of the following frequency distribution of

marks secured by students in statistics.
Class Interval Frequency

0-5 1
5-10 6
10-15 8
15-20 7
20-25 11
25-30 10
30-35 10
35-40 17
40-45 20
45-50 25
50-55 12

Telegram: AgroMind Website: [Link]

 55-60 7
60-65 5
65-70 6
70-75 4
75-80 1

Solution:

Class interval Frequency(f) (m) fm cf

0-5 1 2.5 2.5 1
5-10 6 7.5 45 7
10-15 8 12.5 100 15
15-20 7 17.5 122.5 22
20-25 11 22.5 247.5 33
25-30 10 37.5 275 43
30-35 10 32.5 325 50
35-40 17 37.5 637.5 70
40-45 20 42.5 850 90
45-50 25 47.5 1187.5 115
50-55 12 52.5 630 127
55-60 7 57.5 402.5 134
60-65 5 62.5 312.5 139
65-70 6 67.5 405 145
70-75 4 72.5 290 149
75-80 1 77.5 77.5 150
Total 150 650 5910 1290

Mean x =
fm
f

=
= 39.4

N
−CF Median
2
=l+ xi

Telegram: AgroMind Website: [Link]

 F

= 40 + x5

= 40 + x5
=40 −1.47
= 41.47

Mode

=l + S1 *i
S1+ S2

=40 + x5

= 40 + 3 x5
3−5

= 40 + x5
= 40 − 7.5
= 32.5

Conclusion:

For the given data the mean# median # mode. So it is asymmetrical

distribution.

Exercise
Part – I Explain
/ Define
i. Measure of central tendency ii.
Averaging iii. Advantages of
averaging iv. Types of Averaging v.
Arithmetic mean vi. Median vii.
Mode viii. Geometric mean ix.
Quartiles x. Deciles xi.
Percentiles
Telegram: AgroMind Website: [Link]
 Part – II
1. Characteristics / of a good averaged properties 2.
Uses / of mean. Median and mode application
Part – III
1. Briefly explain about different types of average its merits and demerits
UNIT-IV
MEASURES OF DISPERSION
Introduction
The variation or deviation of the different values of variable form the
average is known as dispersion.
For example,consider the yield of two crops A & B.
Crop:A(yield per plant in gms): 5,6,7
Crop: B (yield per plant in gms) 2,6,10
If mean alone is taken to explain the data, then it will be concluded that the two
crops are equal with respect to yield per plant (because the mean value of crop A
and Crop B are equal to 6). Actually the two crops are different when the
variation of observations is concerned. Therefore, measure of central tendency
alone is not sufficient to explain a data. The variation should also be considered
simultaneously.
Similarly, measure of variation alone will not explain a data completely. For
example, consider the two sets of data which relate to the yield per plant (in gms)
of two crops A & B. Crop a: 100,102,
Crop B: 1,2,3
The variation is equal to 1 gm in both the crops A and B. This means that the
yield of the two crops are equal (with respect to variation). But it is evident that
means are different. Therefore, measure of central tendency and measure of
dispersion should go hand in hand to explain a data. Characteristics of good
measure of dispersion
(1) It should be well defined
(2) It should be easy to understand
(3) It should be easy to calculate
Telegram: AgroMind Website: [Link]
 (4) It should be based on all observation
(5) It should be suitable for algebraic treatment
(6) It should not be affected by fluctuations of sampling
(7) It should not be affected by extreme values of the given data
There are two types of measure of dispersion
1. Absolute measure of dispersion
2. Relative measure of dispersion

Absolute measure of dispersion

A measure of dispersion is expressed in the same unit in which the original data
are given called absolute measure of dispersion. By using this measure of
dispersion the variability of two or more distributions with same units can be
compared.
Types of absolute measure of dispersion.
1. Range
2. Quartile deviation
3. Mean deviation
4. Standard deviation
Relative measure of dispersion
An absolute measure of dispersion is expressed as a percentage of measure of
central tendency (Mean, Median, Mode etc.,) gives relative measure of
dispersion. It is unit free measurement. By using this measure of dispersion the
variability of two or more distribution with different units of measurements can
be compared.
Type of relative measure of dispersion
1. Co – efficient of variation
2. Co – efficient of Quartile deviation
3. Co – efficient of Mean deviation
4. Co – efficient of Range. Absolute Measure of Dispersion
Range:
Telegram: AgroMind Website: [Link]
 It is the difference between the largest and smallest values of given data.
Merits:
i) It is simple to understand and easy to calculate.
Demerits:
i) It is affected by extreme values of the given data.
ii)It is not based on all observations.
Quartile deviation:
It is the half of the difference between upper and lower quartile values. It
is also called semi-inter quartile range. It is not based on all [Link] is not
a suitable measure of dispersion to know the variation of the given data.
Mean deviation:
It is mean of the absolute deviations of observation from its mean. It is
based on all observations, but neglecting signs is unjustified and not suitable for
algebraic treatment.
Standard deviation:
It is defined as the positive square root of the mean of the squared
deviations from mean. The square of S.D is called variance. It posses almost
allthe characteristics of good measure of dispersion.
Relative measure of dispersion Case 1 : Raw data
S.D = Sqrt ( (∑(x – x )2) / n - 1 )
n-1 = degree of freedom

S.D = Sqrt ( (∑x2 –( ( ∑x )2)/n ) / n - 1 )

n = total number of observation
This is known as variable square method.

Case 2 : Discrete data

S.D = Sqrt ( ∑fx2 –( ∑fx )2) / n - 1 )
n = total frequency

Telegram: AgroMind Website: [Link]

 Case 3 : Continuous data
S.D = Sqrt ( (∑fd2 –( ( ∑fd )2)/n ) / n - 1 ) x i
d= ( m – A ) / i
i= common class interval
m=midpoint
A= Assumed value.

[Link] of variation:
It is the most important relative measure of dispersion. It is the ratio of
S.D to mean expressed as a percentage. It is used to compare the variability of
different sets of data having different units of measurement. It is used to know
the reliability of data. For example, in an experiment the C.V. % of yield data
should be within 5% and 15% .If it is not in this range the data will not be reliable
to make conclusion of the experiment.
Note: Various formulas to calculate the above measure of dispersion can be
referred with the practical record note.

Co-efficient of Variation:
CV (%) = ((SD) / Mean) X 100
Solved Problems
Measures of Dispersion Standard
Deviation:
The standard deviation is defined as the positive square root of the mean of
the squared deviations of individual values from their mean.

Case 1 : Raw data

S.D = Sqrt ((∑(x – x )2) / n - 1 )
n-1 = degree of freedom

Telegram: AgroMind Website: [Link]

 S.D = Sqrt (( ∑x2 –( ( ∑x )2)/n ) / n - 1 )
n = total number of observation
This is known as variable square method.

Case 2 : Discrete data

S.D = Sqrt ((∑fx2 – (∑fx) 2) /n) / n - 1)
n = total frequency

Case 3 : Continuous data

S.D = Sqrt ((∑fd2 – ((∑fd) 2)/n) / n - 1) x i
d= (m – A) / i
i= common class interval
m=midpoint
A= Assumed mean.
Co-efficient of Variation:
CV (%) = (SD / Mean) X 100
1. Calculate standard deviation and co-efficient of variation from the data given
below.
9 2 4 5 8 9 12

9 7 2 4
2 7 -5 25
4 7 -3 9
5 7 -2 4
8 7 1 1
9 7 2 4
12 7 3 25
Total = 49 49 -2 72

Telegram: AgroMind Website: [Link]

 2

SD = (x − x ) =3.46
n −1
SD 3.46
CV = *100 = x *100
x 7

=
= 49.4%
Result:
Co-efficient of variance is less (49.4%) and the variability is less. So the data is
reliable data.
2. Calculate standard deviation and co-efficient of variation from the data given
below.
X 0 1 2 3 4
f 9 7 21 3 1

x f fx x2 fx2
0 9 0 0 0
1 7 7 1 7
2 21 42 4 84
3 3 9 9 27
4 1 4 16 16
Total 41 62 30 134

Telegram: AgroMind Website: [Link]

 2 fx
2
1.0075 =1.0037
fx −
n SD 1.003
SD =
n −1 = , =1.512,CV = *100 = x100, C.V. =66.4%
2

134 − ( )
62 f x 1.512
41 Result:
= 40 CV is more (66.4%) and the variability is more. So the
3844 reliability of data is less.
134 −
41 3. Calculate standard deviation and co-efficient of variation
= 40
from the data given below.
40.3
= 40 , =
fx 62 Classes Frequency
x=
41
0-5 2
5-10 5
10-15 7
15-20 13
20-25 21
25-30 16
30-35 8
35-40 3

Classes f m fm d=m-A d2 fd2 fd

A=22.5
0-5 2 2.5 5 -20 400 800 -40
5-10 5 7.5 37.5 -15 225 1125 -75
10-15 7 12.5 87.5 -10 100 700 -70
15-20 13 17.5 227.5 -5 25 325 -65
20-25 21 22.5 472.5 0 0 0 0
25-30 16 27.5 440 5 25 400 80
30-35 8 32.5 260 10 100 800 80
35-40 3 37.5 112.5 15 225 675 45
Total 75 160 1642.5 -20 1100 4825 -45

Telegram: AgroMind Website: [Link]

 2

fd 2 − ( )
fd
SD
SD = n
CV = *100
( −)
n 1
x
4825 − 27
= 74 = *100 =36.76%
= 64.8 =8.05 Result:
fm The CV value is less (36.76%) and the variability is less.
x=
f So the data is reliable one.

= = 21.9 4). The following table gives the frequency

distribution of Age of wheat plants. Calculate the average of wheat plant and
coefficient of variation.
Age(yr) No. of Plants
3–5 2
5–7 16
7–9 25
9 – 11 43
11 – 13 51
13 – 15 20

x f m fm d=m-A, d2 fd2 fd
A=12
3-5 2 4 8 -8 64 128 -16
5-7 16 6 96 -6 36 576 -96
7-9 25 8 200 -4 16 400 -100
9-11 43 10 430 -2 4 172 -86
11-13 51 12 612 0 0 0 0
13-15 20 14 280 2 4 80 40
Total 157 54 1626 124 1356

Telegram: AgroMind Website: [Link]

 2
fd 2 − ( )
fd
SD
S= n
CV = *100
n −1
x
66564
1356 −
157
= 156 C.V.==10.35, C.V.
=x100, 23.61%
1356 − 423.9
= 156 Result:
932.1 The CV value is less (23.61) ,the variability of data
= 156 , = 5.975 , = 2.44 is less. So the data will be reliable one.
fm
mean =
f

= =10.35,
x x x−x (x − x)2
0 3 -3 9
1 3 -2 4
2 3 -1 1
3 3 0 0
4 3 1 1
5 3 2 4
6 3 3 9
21 28
x=
x,
n

= ,=3

Telegram: AgroMind Website: [Link]

 (n −1)=3−1 = 2
2
(x − x ) = 28 =14 Result:
n −1 2 The CV value is more (124.66%) and the
2
(x − x ) 28 variability is more. So the data will be less
S= ,= , 14 , = 3.74
n −1 2 = reliable.
SD
CV = * 100
x
3.
C.V . = x100, =124.66%

Telegram: AgroMind Website: [Link]

 x 150 200 250 300 350 400 450 500 550
f 2 22 19 14 3 4 6 1 1

x f fx x2 fx2
150 2 300 22500 45000
200 22 4400 40000 880000
250 19 4750 62500 1187500
300 14 4200 90000 1260000
350 3 1050 122500 367500
400 4 1600 160000 640000
450 6 2700 202500 1215000
500 1 500 250000 250000
550 1 550 302500 302500
Total 72 20050 1252500 6147500
2

−( )
2 fx
fx
S= n
n −1
402002500
6147500 −
72
= 71
6147500 − 5583368.05
= 71 = 71
= 7945.520 , SD = 89.1376
fx 20050
mean = ,= , 278.47
f 72 =
SD
CV = *100
x 564131.95

C.V. = x100, = 32.007%

Result:
The CV Value is less (32.007%) and the variability is less. So the data will be
reliable one.

Telegram: AgroMind Website: [Link]

 Exercise
Part – I Explain
/ Define
i. Dispersion
ii. Absolute measures or dispersion iii. Types of
measures or dispersion
iv. Relative measures of dispersion
v. Range
vi. Mean Deviation vii. Quartile deviation viii. Standard deviation
ix. Co-efficient of variation
x.
Part – II
1. Characteristics of Good Measures of Dispersion

Part – III [Link] “Types of Absolute

and Relative Measures of dispersion” with its merits and demerits

Telegram: AgroMind Website: [Link]

 Skewness and Kurtosis

Skewness means “lack of symmetry (asymmetrical distribution)”. If the

frequency distribution uniformly distributed around the mean value then it is said
to be symmetrical distribution. If the frequency distribution not uniformly
distributed around the mean value is said to be “asymmetrical or skewed one”.
There are three types of skewness i.e. positively skewed, negatively skewed and
No skewness (or) symmetrical one. It can be determined by using coefficient of
skewness. If was given by Karl Pearson.

−Mode
coefficientof skewness= Mean
Stadard Deviation

Drawback of this measure is no limits and also determination of mode

position is not possible.
−Median)
3(Mean

coefficientof skewness=
S tandard Deviation

Limit of this measure lies between -3 and +3.

Other measure uses the β (read ‘beta’) coefficient which is given by β1=
32

where µ2 and µ 3 are the second and third central moments. The second
3

2 central moment u2 is nothing but variance. The sample estimate of

this coefficient is

b1 =mm3223

Where m2 and m3 are the sample central moments given by

(x− x) 2

m2 =variance= (or) f (x− x) 2 andm3 = (x− x) 3 (or) f (x− x) 3

Telegram: AgroMind Website: [Link]

 n−1 n−1 n−1 n−1
For a symmetrical distribution b1=0. Skewness is positive or regulative
depending upon whether m3 is positive (or) negative.

Kurtosis
Kurtosis is the measure of peaked ness (or) convexity of the curve. Based
on peaked ness, kurtosis divided into three types (i) lepto kurtic (ii) Meso kurti
and (iii) platy kurtic.

It can be determine based on β2 values β2 is nothing but coefficient of kurtosis.

β2= 4
2

m4
Sample estimate of this coefficient is b2 = 2

m2
Where m4 is the fourth central moment given by m4

= (x− x) 4 (or) f (x− x) 4 n−1

n−1

If b2 = 3 then it is said to be meso kurtic, if b2more than 3 then the

distribution is called as lepto kurtic. It is less than 3 the distribution is said to be
platy kurtic.
It also determine based on r2value r2 = β2 -3.
If
r2 = 0 ; meso kurtic
r2 > 0 ; lepto kurtic r2 < 0 ;
platy kurtic Problems
1. Verify whether the given frequency distribution x : 1 2 3
4 5ƒ :5 9 12 9 5 is symmetrical one (or)
not.
2. Find out the skewness and kurtosis for the following data and interpret
the result.
Telegram: AgroMind Website: [Link]
 x : 3 4 5 6 7 8 9 10
ƒ : 4 6 10 26 24 15 10 5
3. Calculate skewness and kurtosis for the following data and interpret it.
x :6 7 8 9 10 11 12 ƒ : 3 6 9
13 8 5 4

Types of Kurtosis

Mean

Types of Skewness

Mean

Symmetrical Negative skewness

Telegram: AgroMind Website: [Link]

 Positively skewned

UNIT-V CORRELATION AND REGRESSION

Correlation
If the change of one variable affects the change of another variable then the
variables are said to be correlated. In other words, the systematic
interrelationship between the variables is termed as correlation Eg. i) Price
and demand ii) Population and unemployment
iii) Yield and pest incidence Types
of Correlation
1. Simple correlation
2. Simple linear correlation
3. Simple Non-linear correlation
4. Multiple correlation
5. Multiple linear correlation
6. Multiple Non-linear correlation
7. Positive or Direct Correlation
8. Negative or Indirect Correlation
9. No Correlation or Independent variables
10. Pseudo Correlation
Simple correlation

Telegram: AgroMind Website: [Link]

 If the relationship between only two variables then the correlation is said
to be Simple correlation. Simple linear correlation
If the relationship between two variables are straight line then the
correlation is said to be Simple linear correlation.
Simple Non-linear correlation
If the relationship between two variables are not straight line then the
correlation is said to be Simple Non-linear correlation
Multiple correlation
If the relationship between more than two variables then the correlation is
said to be Multiple correlation.
Multiple linear correlation
If the relationship between more than two variables are in straight line
then the correlation is said to be Multiple linear correlation.
Multiple Non-linear correlation
If the relationship between more than two variables are not in straight line
then the correlation is said to be Multiple Non-linear correlation.
Positive correlation
An increase in one variable may cause an increase in other variable, or a
decrease in one variable may cause a decrease in the other variable. In other
words if the movement of the variables are in the same direction then they are
said to be positively correlated. It is also called direct correlation. Eg. i) Yield
and fertilizer application ii) Changes in the temperature and changes in length
of a iron bar iii) Increase in the radius of the circle and increase in the
circumference.
iv) Incidence of pests and weather factors
Negative correlation
It the movements of the variable are in opposite direction then the variables are
said to be negatively correlated. It is also called indirect correlation
i) Yield and pest incidence

Telegram: AgroMind Website: [Link]

 No Correlation or Independent Variables
If the variables are having no relationship then the variables are said to be
independent variables or no correlated one.
SCATTER DIAGRAM
In correlation problems, first we have to investigate whether there is any relation
between the variables say, X and Y. For this purpose we use scatter diagram.

Let (X1, Y1), (X2, Y2), (X3, Y3) . . . . . . . . . . . . (Xn, Yn) be n pairs of obsevations.
If the values of variables X and Y are plotted along the X axis and Y axis
respectively in the XY plane of graph sheet, the resultant diagram of dots is
known as scatter diagram. From the scatter diagram we can say whether there is
any correlation between X and Y, correlation is positive or negative and the
correlation is linear or curve linear.
Advantages:
[Link] diagram method is easy to understand and simple to follow.
[Link] does not involve too much of mathematics.
3. It helps to get a preliminary idea of correlation in a data.
Disadvantages:
1. It gives only the direction but not magnitude of correlation between the two
variables.
Correlation coefficient
The scatter diagram will give only a vague idea about the presence or absence
of correlation and the nature of the correlation. It will not indicate about the
strength or degree or relationship between two variables. The index of the degree
of relationship between two variables is known as correlation coefficient. It can
be determined by the following formula

r= SPxy
[Link]

Telegram: AgroMind Website: [Link]

 where , SPxy = [ Σ xy – (Σx Σy) / n ]
SSx = [ Σ x2 – ( Σ x )2/ n) ]
SSy = [ Σ y2 – ( Σ y )2/ n) ]

The two variables X & Y are said to be positively, negatively correlated

according as r is positive or [Link] r is equal to zero then the two variables
are said to be uncorrelated.

Properties of correlation co-efficient

i) ‘r’ will always lie between – 1 & +1 ii) The value of ‘r’ is unaltered if
X & Y are interchanged. This property is called symmetric property.
iii) Suppose the values of data on two variables are altered uniformly then
the value of ‘r’ will not be altered. This property is called
invariance property of ‘r’.

Limitation of correlation co-efficient

1. If the variable under study should be continuous and normally distributed

one.
2. Variables are linear in nature.
3. Each pair of observation unconnected with other pairs of observations.
REGRESSION
The concept of regression was given by Francis [Link] two variables under
study one may represent the cause and the other may represent the effect. The
variable representing the cause is known as independent variable and it is
denoted by X. The variable X is sometimes called predictor variable or regressor.
The variable representing the effect is known as dependent variable and is
denoted by Y. The variable Y is sometimes called predicted variable. The
relationship between the independent and dependent variables may be expressed

Telegram: AgroMind Website: [Link]

 as a function. Such functional relationship between two variables is termed as
regression.
When only two variables are involved the functional relationship is known as
simple regression. If the relationship between the two variables a straight line,
it is known as simple linear regression. Otherwise it is called as simple non
linear regression. When there are more than two variables and one of them is
assumed to be dependent upon the others, the functional relationship between
the variables is known as multiple regressions.
Curve fitting
In general when two variables are studied simultaneously the interest will be to
know the functional relationship between them. The scatter diagram is used to
identify the type of relationship between the two variables for example, when n
pairs of observations (X1, Y1), (X12 Y2),. . . . . . . . . . . . . . . .
(Xn, Yn) are plotted on a graph sheet.
The scatter diagram for a given data may result in non – linear forms.
The functional relationships may be given by algebraic expressions like
Y = a + bX
The above algebraic expression is known as curve. The method of
finding such relationship is known as curve fitting. A best fitting line is one for
which the sum of the squares of the residuals, (or errors) is minimum. For this
purpose the principle ,known as method of least squares is used.
Method of least square
A best fitting line is one for which sum of the squares of the residuals (errors) is
minimum. The principle which minimize the error in known as method of least
squares.
General form of simple linear equation is

Y=a+bX

then estimated values of y is

Telegram: AgroMind Website: [Link]

 ŷ = a + bx

Let
y-ŷ = ei

Application of methods of least squares principle

ei 2 = minimum

(y – ŷ)2 = 0
[y –(a+bx)]2 = 0
(y –a-bx)2 = 0 s =
(y-a-bx)2 = 0

For Minimization

8s

8a = 2 (y-a-bx) (-1)
8s
=0
8

-2 (y-a-bx) = 0
(y-a-bx) = 0
y- a- bx=0
y-na-b x=0  1 Note (c) = nc
ax = a x

Similarily

8s

8b = 2 ( y – a - bx ) (-x)
8s
=r
8b

-2 (y-a-bx) (x)=0

Telegram: AgroMind Website: [Link]

 (y –a – bx) (x)=0
(xy-ax-bx2) =0

a b
xy- x- x2 = 0  2

Equation and are normal equations 1 2

From normal equation 1

y – na- b x=0
.
n
.

y− na − b x =0

n n n

y −a−b x =0 n
n

a= y −b xn
n

a = y−bx

From Normal Equation (2)

xy - a x -b x2 = 0

Substitute a = y −b x then n
n

xy− y −b x x−b x2 =0

Telegram: AgroMind Website: [Link]

 n n

xy− x y −b( x)2 −b x2 =0

n n

xy− x y +b( x)2 −b x2 =0

n n

xy− x y =b x2 − ( nx)2

x− y

xy−
n
b= 2

2 ( x)

x− n

Substitute the value of a and b to the general form of simple linear

regression equation, then that equation y = a+bx is said to be fitted simple linear
regression equation y on x.

Regression coefficient
In the equation Y = a +bX, ‘b’ is the slope of the line, also called Regression
coefficient and ‘a’ is the intercept of the line with the Y-axis.
bYX = regression coefficient Y on X
bXY = regression coefficient X on Y
Propertices of regression coefficients

Telegram: AgroMind Website: [Link]

 i) Correlation coefficient is the geometric mean between the regression
coefficient
ii) If one of the regression Coefficients is greater than unity, the other must
be less than unity. iii) Arithmetic mean of regression coefficients is greater
than the correlation coefficient.
iv) Regression coefficient are independent of the origin but not the scale

Exercise
Part – I
Define / Explain
1. Correlation
2. Scatter diagram
3. Correlation coefficient
4. Properties of correlation co efficient
5. Regression
6. Curve fitting
7. Regression co-efficient
8. Properties of regression co-efficient
Part – II
Explain
1. Types of correlation
2. Scatter diagram with its advantages and disadvantages
Part – III
1. Briefly explain about correlation coefficient and its properties
2. Write the procedure to fit the regression equation Y on X and give
properties or regression co-efficient

Solved Problems
Correlation

1. The following data relates to yield per plant(gms) and plant height (cm) of
cotton.
Find the correlation between yield and height.
Yield ( gms )( Y) 3.0 3.5 4.5 5.0 6.0 6.5 7.0 7.0
Height (cms) (X) 3.5 5.5 7.5 6.0 8.5 10.5 7.5 11.5

Telegram: AgroMind Website: [Link]

 Yield ( gms )( Y) 8.5 8.5 8.5 8.5 9.5 9.5 10.0 11.0
Height (cms) (X) 10.0 8.5 10.5 12.5 14.5 13.5 14.5 13.0

Yield (y) Height (x) xy x2 y2

3 3.5 10.5 12.25 9
3.5 5.5 19.25 30.25 12.25
4.5 7.5 33.75 56.25 20.25
5 6 30 36 25
6 8.5 51 72.25 36
6.5 10.5 68.25 110.25 42.25
7 7.5 52.5 56.25 49
7 11.5 80.5 132.25 49
8.5 10 85 100 72.25
8.5 8.5 72.25 72.25 72.25
8.5 10.5 89.25 110.25 72.25
8.5 12.5 106.25 156.25 72.25
9.5 14.5 137.75 210.25 90.25
9.5 13.5 128.25 182.25 90.25
10 14.5 145 210.25 100
11 13 143 169 121
2 2
Σy=116.5 Σx=157.5 Σxy=1252.5 Σx =1716.25 Σy =933.25
xy− x y
n
2 2

x 2
−
( x) y 2
−
( y) ,
n n
r= ,

1252.5− (
116 .5)(157.5)
16
=, =
(157.5)2 (116.5)2
(1716.25)− )(933.25− 105.71 )
16 118
16 . 72
105.71 105.71
= ,= ,= , r = 0.890
(165.86)(84.99) 14096.44

Result:

Telegram: AgroMind Website: [Link]

 The value of r is 0.890 .It indicates the plant yield will be positively correlated
with plant height.

Telegram: AgroMind Website: [Link]

 2. Following data give the length and breadth of 30 leaves. Work out the
correlation coefficient and interpret the result.
Length (cm) Breadth (cm) Length Breadth (cm)
(x) (y) (cm) (y)
(x)
1.78 1.53 1.73 1.50
1.79 1.54 1.71 1.46
1.70 1.48 1.76 1.47
1.77 1.49 1.77 1.52
1.71 1.47 1.71 1.52
1.76 1.52 1.76 1.53
1.72 1.48 1.78 1.50
1.77 1.49 1.70 1.45
1.79 1.53 1.75 1.47
1.71 1.77 1.70 1.47
1.75 1.51 1.74 1.50
1.76 1.52 1.70 1.45
1.70 1.48 1.75 1.47
1.74 1.49 1.77 1.47
1.70 1.47 1.77 1.50

Length (x) Breadth (y) xy x2 y2

1.78 1.53 2.7234 3.1684 2.3409
1.79 1.54 2.7566 3.2041 2.3716
1.70 1.48 2.516 2.89 2.1904
1.77 1.49 2.6373 3.1329 2.2201
1.71 1.47 2.5137 2.92461 2.1609
1.76 1.52 2.6752 3.0976 2.3104
1.72 1.48 2.5456 2.9584 2.1904
1.77 1.49 2.6373 3.1329 2.2201
1.79 1.53 2.7387 3.2041 2.3409
1.71 1.77 2.0267 2.9241 3.1329
1.75 1.51 2.6425 3.0625 2.2801
1.76 1.52 2.6752 3.0976 2.3104
1.70 1.48 2.516 2.89 2.1904
1.74 1.49 2.5926 3.0276 2.2201
1.70 1.47 2.499 2.89 2.1609
1.73 1.50 2.565 2.9929 2.25

Telegram: AgroMind Website: [Link]

 1.71 1.46 2.4966 2.9241 2.1316
1.76 1.47 2.5872 3.0976 2.1609
1.77 1.52 2.6904 3.1329 2.3104
1.71 1.52 2.5992 2.9241 2.3104
1.76 1.53 2.6928 3.0976 2.3409
1.78 1.50 2.67 3.1684 2.25
1.70 1.45 2.465 2.89 2.1025
1.75 1.47 2.5725 3.0625 2.1609
1.70 1.47 2.499 2.89 2.1609
1.74 1.50 2.61 3.0276 2.25
1.70 1.45 2.465 2.89 2.1025
1.75 1.47 2.5725 3.0625 2.1609
1.77 1.47 2.6019 3.1329 2.1609
1.77 1.50 2.655 3.1329 2.25
Σx=52.25 Σy=45.05 Σxy=78.4679 Σx2=91.0303 Σy2=67.74

xy− x y
n
2 2

x 2
−
( x) − y 2
−
( y)
n n
r= ,

78.4679 − (
52.25) − (45.05)
30
= 2 2

91.0303 − (
52.25)
67.74 − (
45.05)
30 30
78.4672 − 78.4620
= ,
(91.0303 −91.002)(67.74 − 67.65)
0.0058 0.0058
= ,= , 0.1139
0.0509 =
(0.0283)(0.0932)
Result:
The value of r is [Link] indicates the length of the leaves will be positively
correlated with breadth of the leaves.

3. The following is the data of size of crop (hundreds of fruits) and percentage
of wormy fruits on 12 apple trees. Work out the correlation coefficient and
test its significance.
Size of crop 15 14 12 26 18 12 8 38 26 19 29 22

Telegram: AgroMind Website: [Link]

 Wormy 51 46 38 37 33 37 34 25 22 22 20 14
fruits

Size of crop (x) Wormy fruits xy x2 y2

(y)
15 51 765 225 2601
14 46 644 196 2116
12 38 456 144 1444
26 37 962 676 1369
18 33 594 324 1089
12 37 444 144 1369
8 34 272 64 1156
38 25 950 1444 625
26 22 572 676 484
19 22 418 361 484
29 20 580 841 400
22 14 308 484 196
239 379 6965 5579 13333
xy− x y
n
2 2

x2 − ( ) y2 −( )
x y
n − n
r= ,

6965 − (
239 x379)
12
= 2 2

5579 − ( ) 13333 − ( )
239 379
12 12
6965 − 7548.4
= ,
(5579 − 4760)(13333 −11970 )
− 583.4 , − 583.4
= = 1056.54 = −0.552
( )( )
819 1363

Result:
The value of r is -[Link] indicates the size of crop will be negatively correlated
with wormy fruits.

Telegram: AgroMind Website: [Link]

 Regression
1. The yield and height of the variety is given in the following table. Fit the
linear regression line.
Yield ( gms ) Y 3.0 3.5 4.5 5.0 6.0 6.5 7.0 7.0
Height (cms) X 3.5 5.5 7.5 6.0 8.5 10.5 7.5 11.5
Yield ( gms ) Y 8.5 8.5 8.5 8.5 9.5 9.5 10.0 11.0
Height (cms) X 10.0 8.5 10.5 12.5 14.5 13.5 14.5 13.0

Yield (y) Height xy x2

(x)
3.0 3.5 10.5 12.25
3.5 5.5 19.25 30.25
4.5 7.5 33.75 56.25
5.0 6.0 30.0 36.00
6.0 8.5 51.0 72.25
6.5 10.5 68.25 110.25
7.0 7.5 52.5 56.25
7.0 11.5 80.5 132.25
8.5 10.0 85.0 100.00
8.5 8.5 72.25 72.25
8.5 10.5 89.25 110.25
8.5 12.5 106.25 156.25
9.5 14.5 137.75 210.25
9.5 13.5 128.25 182.25
10.0 14.5 145 210.25
11.0 13.0 143 169
2
Σy=116.5 Σx=157.5 Σxy=1252.5 Σx =1716.25
xy− x y

b= n

x2 − ( x)2
n

1252.5 −(
116 .5)(157.5)
16
= 2

(157.5)
1716.25 −
16
Telegram: AgroMind Website: [Link]
 =

=
b = 0.63 a

= y−b xn
n

= −(0.63)
= 7.28 − 6.20
a =1.08
Result:
The fitted simple linear regression equation is y=1.08+0.63x. From the fitted
equation we will conclude 1 unit changes of x value, the y will change 0.63unit.

2. The yield of grain(Y) and weight of ear head(X) of sorghum is given in the
following table. Fit the linear regression line.
59.1 105.3 82.8 67.5 74.5 60.5 74.2 63.2 76.6 102.7
Y
X 39.4 67.3 46.2 37.1 43.4 34.1 29.7 33.0 37.2 76.3
Y 68.1 104.3 80.8 62.2 59.5 59.1 59.9 89.4 100.4 77.2
X 28.3 66.3 44.2 32.0 30.1 60.3 31.7 59.5 63.5 51.6

Grain yield Earhead xy x2

(y) (x)

59.1 39.4 2328.54 1552.36

105.3 67.3 7086.69 4529.29
82.8 46.2 3825.36 2134.44
67.5 37.1 2704.25 1376.41
74.5 43.4 3233.30 1883.56
60.5 34.1 2063.05 1162.81
74.2 29.7 2203.74 882.09
63.2 33.0 2085.6 1089.00
76.6 37.2 2849.52 1383.84
102.7 76.3 7836.01 5821.69
68.1 26.3 1927.23 800.89
Telegram: AgroMind Website: [Link]
 104.3 66.3 6915.09 4395.69
80.8 44.2 3571.36 1953.64
62.2 32.0 1990.4 1024.00
59.5 30.1 1790.95 906.01
59.1 60.3 3563.73 3636.09
59.9 31.7 1886.15 1004.89
89.4 59.5 5319.3 3540.25
100.4 63.5 6375.4 4032.25
77.2 51.6 3983.52 2662.56
Σy=1527.3 Σx=909.2 Σxy=73215.67 Σx2=45662.56
xy− x y

b= n

x2 − ( x)2
n

73215.67 − (
909.2 )(1527.3)
20
= 2

(909.2)
45662.56 −
20

=
b = 0.8739

a= y−b x
n n

= −(0.8739)
= 76.36 − (0.8739)(45.56)
= 76.36 −39.72
a = 36.64

Result: The fitted simple linear regression equation is Y = 36.64 + 0.87 X from
this, we will conclude 1 unit changes of x value and the y will change 0.87 units.

Telegram: AgroMind Website: [Link]

 Exercise
Part – I
Define / Explain
9. Correlation
10. Scatter diagram
11. Correlation coefficient
12. Properties of correlation co efficient
13. Regression
14. Curve fitting
15. Regression co-efficient
16. Properties of regression co-efficient
Part – II
Explain
3. Types of correlation
4. Scatter diagram with its advantages and disadvantages
Part – III
3. Briefly explain about correlation coefficient and its properties
4. Write the procedure to fit the regression equation Y on X and give
properties or regression co-efficient

Telegram: AgroMind Website: [Link]

 UNIT-VI PROBABILITY AND ITS DISTRIBUTION
Introduction
The theory of probability is a branch of applied mathematics dealing with the
effects of chance. The idea of probability was first introduced in the field of
games of chance about 1650. It was developed by Pascal, Fermat, Laplace,
Guass and James [Link] theory of statistics is entirely based on the theory
of probability.
Definitions
Experiment
An experiment is a systematic procedure that generates a set of data.
Example. Rolling of a dice or tossing of an unbiased coin.
Event
Event is a definite result got in an experiment conducted with certain conditions
Example . Consider tossing of a coin. In this experiment we get either a head or
a tail facing upwards. These are the two definite results (or) events. Trial
Trial is an attempt to produce an event.
Eg. Tossing a coin once is a trial.
Equally likely events
Events are said to be equally likely if one cannot be expected in preference to
the other.
Eg. In throwing of a die the six events 1,2,3,4,5,6 are equally likely.
Mutually exclusive events
Events are said to be mutually exclusive if the occurrence of one excludes the
occurrence of others.
Eg. Consider the tossing of a coin. If head occurs then tail cannot occur.
Similarly if tail occurs then head cannot occur. Getting of a head or a tail is said
to be mutually exclusive event.

Telegram: AgroMind Website: [Link]

 Exhaustive events
Events are said to be exhaustive if at least one of them must necessarily
happen when all the cases are considered.
Eg. Consider the tossing of a con. We have two events
i) getting a head ; ii) getting a tail
If the coin is unbiased one then we are sure to get any one of these two events.
Therefore these two events are said to be exhaustive.
Compound events
Events are said to be compound if they occur simultaneously.
Eg. Consider throwing of two dice. Here the simultaneous occurrence of faces
of two dice are called compound events.

The compound events are of two kinds.

i) Independent events ii)
Dependent events
Independent events
Events are said to be independent if the occurrence of one is not affected by the
occurrence or non occurrence of the other.
Eg. In throwing of two dice at a time the occurrence of any number of the first
die is not affected by the occurrence of a number of the second die. So these two
events are independent.
Dependent Events
Events are said to be dependent If the occurrence of one depends upon the
occurrence of other.
Eg. If a bag contains 6 white and 6 black balls and if 2 balls are drawn one by
one and if the first ball is not returned back then the occurrence of the second
one depends upon the first. Probability

Telegram: AgroMind Website: [Link]

 If there are ‘n’ mutually exclusive, equally likely and exhaustive cases, among
which ‘m’ are favorable to the event ‘E’ then the probability of that event is given
by the ratio m/n
Favorable number of cases
P(E) = -----------------------------------------------------
Total number of cases
Theorems on probability
i) Addition theorem ii) Compound theorem

Addition theorem
Statement
If the events E1, E2, E3 . . . . . . . . . . . . . . . . . En are mutually exclusive then the
probability of their sum is the sum of their probabilities..
P (E1+ E2+ E3+ . . . . . . . . . . . . . . . + En) = P (E1) + P (E2) + P (E3) + . . . . . . .
+ P(En)
Proof : Let ‘n’ be the total number of cases.
Let m1, m2, m3, . . . . . . . . . . . . . . . mn are favourable number of cases among ‘n’
events to the events E1, E2, E3 . . . . . . . . . . . . . . . . . En respectively. Since these
events are mutually exclusive all the cases m1, m2, m3, . . . . . . . . . . . . . . . mn are
entirely different and disjoint.
The number of cases favourable to either E1 or E2 or E3 . . . . .or En is m1 + m2
+ m 3+ .. +mn
By definition of probability, we write
P(E or E or E or …….or E)= (m1 + m2 + m 3+ . . . . . . . . . .+ mn)
------------------------------------------------------------
n
P (E1+ E2+ E3+ . . . . . . . + En) = m1/n + m2/n + m3/n + . . . . . . . . . mn/n
=P (E1) + P (E2) + P (E3) + . . . + P(En)
Hence the theorem

Telegram: AgroMind Website: [Link]

 Compound theorem
Statement
The probability of the simultaneous occurrence of two events E and E is the
product of the unconditional probability of E and the conditional probability of
E assuming that E has already occurred.
Proof
Let ‘n’ be the total number of cases among which ‘m’ are favourable to
the event E1. Let m1 be the favourable number of cases for both the events E 1
and E2.
By definition we write the probability for the simultaneous occurrence of E 1
and E2 as P(E1 E2) = m1/n The same can also be written as
P(E1 E2) = (m/n) x (m1/m)
The ratio m/n stands for the probability of the occureance of E1.
P(E1) = m/n
Since the event E1 has already happened there are only ‘n-m’ cases let for the
event E2. Out of these ‘n-m’ possible cases m1 favourable to the event E2. Hence
the ratio m1/m stands for the conditional probability of E2 while E1 has already
occurred.
Ie., P(E2/ E1) = m1/m
P(E1 E2) = (m/n) x (m1/m)
P(E1 E2) = P(E1) P(E2/E1)
Hence the theorem proved.
Random variable
The variable whose values are determined by the outcome of a random
experiment is called random variable. i) Discrete random variable
A random variable which assumes only a finite or countable number of distinct
values is called discrete random variable.
ii) Continuous variable

Telegram: AgroMind Website: [Link]

 A random variable which takes any numerical value on a continuous scale is
called continuous random variable.
Binomial distribution
The simplest probability distribution is binomial distribution. It is
used to represent the probability distribution of discrete random variables.
An experiment which has two possible outcomes is called a Bernoulli trail.
The two outcomes are usually called success or failure.
An experiment consisting of a repeated number of Bernoulli trails is called
a binomial experiment.
Definition
Let X be the number of success in ‘n’ independent Bernoulli trails,
with probability ‘p’ for success in each trail. Then X is called the
Binomial random variable with parameters ‘n’ and ‘p’
The binomial distribution is given by the probability mass function,
n
P(X=x) = px. q n - x
x
for all possible values of x. In the formula,

n = number of trials x =
number of successes in a trial
n – x = number of failures in a trial
p = probability of success = x/n
q = 1 – p = probability of failure

n
= the possible number of ways in which x successes can occur.
x

Example1: Consider the population of seeds of a crop.

If the seeds are allowed to germinate, then the observation will be such that
each seed has only any one of the two results viz., ‘germination’ or ‘no
germination’.The individuals or objects in the population are seeds. Therefore

Telegram: AgroMind Website: [Link]

 distribution of population of seeds is a binomial distribution as far as the variable
‘germination’ is concerned.
Example 2: Suppose a pathologist wants to know the proportion of sesamum
plants which will be affected by phyllody disease. The individual or object will
be a sesamum plant in the study. The possible outcomes of individual(or plant)
are any one of the two namely ‘diseased’ or ‘not diseased’.Therefore the
distribution of sesamum plants will be a binomial distribution as far as the
variable ‘incidence of disease’ is concerned.
Mean
The means of the binomial distribnution with parameters ‘n’ and ‘p’
Mean = np
Varience = npq

Poisson Distribution
The Poison distribution is also used to represent the probability
distribution of a discrete random variable. It is employed in describing random
events that occur rarely.

The Poison misdistribution is given by the probability mass function,

e-m . mx
P(X=x) = ------------
x!
In the formula,

m= np = mean number of times an event occurs,

x = the number of times the event occurs.

Poisson distribution is a limiting case of Binomial Distribution

i) When the number of trails is very large(n→ ) ii) When

Telegram: AgroMind Website: [Link]

 the Probability of success is very small(p→0) iii) The product

np always remains a constant(np→ )

Then binomial distribution can be approximate to a poison distribution.

Example of Poisson distribution
i) The number of typographical errors per page in typed material.
ii) The number of deaths per day due to a specific disease in a certain town.
iii) The number of born blind per year in a town.

Normal distribution
The most important and widely used probability distribution is normal
distribution. It is also known as Gaussian distribution.
The normal distribution is defined as to represent the probability
distribution of a continuous random variable. Its probability density function is
expressed by the relation,
2
1 x-
f (x) = -----. e -½ --------
2

In the above formula,

= a constant equality 22/7 e =

Naperian base equality 2.7183
= population mean
= population standard deviation
x = a given value of the random variable in the range -- x

The distribution of the standard normal variable is known as standard

normal distribution.
For standard normal distribution, = 0 and = 1.
It is given by

z2
Telegram: AgroMind Website: [Link]
 1 e- ---, in the range - z f
(z) = 2 2

= (z) ( is read as ‘phi’).

Properties of Normal distribution

1. The normal curve is perfectly symmetrical about the mean. Also the curve is
bell shaped
2. Mean = Median = Mode
3. It has only one mode. It is uni modal.
4. The ordinate at the mean of the distribution divides the total area under the
normal curve in to two equal parts.
5. Skewness is zero.
6. i) 2/3 of the individuals in the population will lie between ii) 95% of
the individuals in the population will lie between iii) 99% of the individuals
in the population will lie between This property is called area property of
normal distribution.

Solved Problems
BINOMIAL DISTRIBUTION
1. The probability that a coconut seedling good is ½. For n = 5, p=½ , q=1 – p =
½ , and x= 0, 1, 2, 3, 4 and 5, the probabilities can be worked out by using the
binomial distribution. Calculate the probability of getting atleast 3 good
seedlings?

n = 5, p =1/2, q=1/2.

P(X=x) = nCx px qn-x

P(X=3) = 5c3 (1/2)3(1/2)5-3

=(1/8)x(1/2)2
!

Telegram: AgroMind Website: [Link]

 = (1/8)x(1/4)
=10x(1/8)(1/ 4)
= 5/16
=
0.312
P(x=4) = 5C4(1/2)4(1/2)1
!
=x1/16x1/2

=5x(1/16)x(1/2)
= 5/32
= 0.15.

P(x=5)=5C5(1/2)5(1/2)5-5
5! 1
= x x1
1x5! 32
= 0.03

P(x = atleast3)= 0.312 + 0.15+ 0.03 =

0.492.
P(x = atleast3) = 0.492.

2. Suppose that a population of size 500 (=N) consists of 300 dominants and
200 recessives. For a sample of 10 (=n), calculate the probabilities, (i) exactly 2
individuals will be recessive, and (ii) at least 2 individuals will be recessive.
N=500
n=10
200- recessives
300 – dominance

Probability of recessive p=200/500 = 2/5 = 0.4.

q=1-p, = 1-0.4= 0.6
i.P(X = 2)=10C2 (0.4)2 (0.6)10−2
!
=x0.16x0.016

Telegram: AgroMind Website: [Link]

 = 45x0.16x0.016
= 0.11 ii. p + q
=1
Failure = p(x = 0) + p(x =1)

x(0.6)10
0

P(x = o)=10C0 (0.4) =

0.005 p(x =1)=10C1 (0.4)1
x(0.6)9
!
=x0.4x0.0096

=10x0.4x0.00096,= 0.038.
p =1− q, p =1− (0.005x0.038),=1− 0.04, p = 0.96

3. When 7 coins are tossed find the probability of getting (i) exactly 4 heads (ii)
6 or more heads.
p = 12,q = 12,n = 7

i. P(x = 4)= 7C4 12( ) x(12)4 7−4

x
73x6x5x4! 1 18
= x
16 x2x1x4!
= 35x 116x 18

= 0.27

( ) x(12)
ii. P(x = 6) = 7C6 12 6 1

71xx66!!1 x 12
= x 64

= 0.05

( )
p(x = 7)= 7C7 12 7

= 1128

Telegram: AgroMind Website: [Link]

 = 0.0078

p(x = 6) + p(x = 7) = 0.05+ 0.0078

= 0.0578 p(x = 6) + p(x = 7)
=0.0578

4. Comment on the following: The mean of a binomial distribution is 3 and

variance is 4.
np=3
npq=4 3q=4
q=4/3 =1.3
Comment:
The range of probability is 0 to [Link] in this problem probability of failure
is more than 1. So it is impossible to happen.

5. The probability that a bulb produced by a factory will fuse after 100 days
of use is 0.05. Find the probability that out of 5 such bulbs (i) none (ii) not
more than one (iii) more than one and (iv) at least one will fuse after 100
days of use.

p=0.05
q=1-0.05
q=0.95
n=5
i) None

P(x=0) = 5C0 (0.05)0 x (0.95)5

=1x (0.95)5
=0.77

ii) Not more than one

P(x=not more than one)

P (x=0) + p(x=1)
P(x=0) = 0.77
P(x=1) = 5C1(0.05)1 x (0.95)4
!
=x0.05x0.81

Telegram: AgroMind Website: [Link]

 = 0.20
p(x = 0) + P(x =1) = 0.77 + 0.20
= 0.97
iii. More than one

1-P(x=0) + p(x=1)
P(x=more than one) = 1-[p(x=0)+p(x=1)]
=1-0.97
=0.03

iv. At least one

P(x=at least one) = 1-p(x=0)
=1-0.77
=0.23

Answer: i.0.77, ii. 0.97, iii. 0.03, iv.0.23

Telegram: AgroMind Website: [Link]

 Poison Distribution:
1. In Field experiment, it was observed that the incidence of stem fly in black
gram was 6 percent. Suppose we examine 50 black gram plants in a field at
random. What is the probability that at the most 3 plants will be found to be
affected by stem fly?. n=50 p=0.06 λ=np
λ =50 * 0.06
λ=3
e− x

P(X=x) =
x!

e−3 −30
i) P(x=0) =
0!
= e = 1/e3
-3

= 0.04

e−3 x31
P(x=1) =
1!
1
= e3 *3,

=0.04*3,
= 0.12
e−3.32
p(x = 2) =
2!
0.

= 0.18

e−3.33
p(x = 3) =
3!
0.

= 0.018
= 0.04 + 0.12 + 0.18 + 0.018
Telegram: AgroMind Website: [Link]
 = 0.358

2. If X is a Poisson Variate such that P (X=1) =P (X=2), then find P (X=4).

p(x = 1) = p(x = 2) e− 1 e− . 2

1 = 2

1 2 =
=2
e−2 .24
p(x = 4) =
4!

= = 0.08
= Ans.0.08

3. Suppose that X has a Poisson distribution. If P (X=2) = 2/3 P (X=1), find (i)
P (X=0) (ii) P (X=3). p(x = 2) = 23 P(x =1)

2
e− . 2 = 3e− . 1

= 43 =1.33

e−1.33x(1.33)0
i.p(x = 0) =
1
= e−1.33

= 0.26

p(x = 3)
e−1.33x(1.33)3
=
3x2
= 0.10
Ans :0.10

Telegram: AgroMind Website: [Link]

 4. A car hire firm has two cars, which it hires out day by day. The number of
demands for a car on each day is distributed as a Poisson distribution with mean
1.5. Calculate the proportion of days on which some demand is refused.

=1.5
P(x = more than2) =1− P(x = 0) + P(x =1) + P(x = 2)
e−1.5 x(1.5)0
P(x = 0) =
0!

= 1e 1.5

= 0.22 p(x =1) =

e−1.5 x(1.5)
= 0.22x1.5
= 0.33
e−1.5 x(1.5)2
P(x = 2) = 2
0.
== 0.24

p(x = 0) + p(x =1) + p(x = 2) = 0.79

=1− 0.79
= 0.21
Ans : 0.21

Telegram: AgroMind Website: [Link]

 5. In a Poisson distribution the probability P (X=0) is twice the probability P
(X=1). Show that the mean of the distribution ½.
p(x = 0) = 2[p(x =1)]
e− . 0 −

= 2*e .
0!
2 =1

1
= 2
Result: It proves that the mean of the distribution is ½.

6. If a random variable X follows a Poisson distribution such that (P (X=1) = P

(x=2), find P (X=0).
p(x =1) = p(x = 2)
− e− .
e . =
2

2
1=

=2

p(x = 0) = e−2x20 , = e−2, = 1e2

0!

=
7.
= 0.135

Ex. No.7 Normal Distribution

1. The demand for a new product is assumed to be normally distributed with
mean
( ) =200 and standard deviation ( ) =[Link] x be the number of
units demanded, find the following probabilities.
Telegram: AgroMind Website: [Link]
 a) P (180<x<220)
b) P (225<x<250)
P(180 x 220)
X =180,220, u = 200, = 40 Z(x
−u
=130) = x

= =
=−0.5

Z(x = 220) = =
= 0.5
= P(−0.5 x0.5)
= P(0.19146 + 0.19146)
= 0.38292
Result

The number of unit demanded P(130<x<220)=0.38292

P(225<x<250)
= 200, = 40

− 200
Z(x = 225) = 22540 = 2540

= 0.625

Z(x = 250) = =
=1.25
P(0.625 x 1.25)
= 0.39435− 0.23565
= 0.1587
Result: The number of unit demanded P(225<x<250)=0.1587

2. Test conducted at a particular university marks are given to 700 students that
is found to be normally distributed. The mean score of students on test is 50 and
standard deviation is [Link] the number of students scored between 30 and 70.
Total students =700, μ=50, σ=27, x=30 and 70 X=30
x−

Telegram: AgroMind Website: [Link]

 Z(x=30) =

= =
= −0.740

Y=70

Z(x =10)= =
= 0.740
P(−0.740 x 0.740)
= 0.27035 + 0.27035
= 0.5407
= 700x0.5407
= 378.49
= 379
Result: The total number of students scored between 70 and 30=379 3. The
mean weight of 400 machines in a company is 250 pounds and the standard
deviation is [Link] the weights are normally distributed, find how many
machine weight between 190 and 265.
= 250, = 25, p(190 x 265)
x − 190 − 250 Z(x
−190) = =
25

=
=−2.4

Z(x = 265) = =
= 0.6
Pt2.4 x 0.6
P(0.49180 x 0.22575)
= 0.71755
[Link] mechines = 400
= 400x0.71755
= 287
Re sult :Themachine weightbetween190and 265is 287.
Telegram: AgroMind Website: [Link]
 4. If x is a normal variate with mean 80 and standard deviation is [Link]
the following probabilities by standardizing.
(i) P (65<x<100)
(ii) P (85<x<95).
= 80, = 10, x = 65 and 100 x
= 65
65 − 80 15
Z(x = 65) = =
10 10
= −1.5

x = 100
100 − 80 20
Z(x = 100) = =
10 10
=2

P(-1.5<x<2)
P(0.43319<x<0.47725)
=(0.43319+0.47725)
=0.91044
Result: The probability of standardizing P (65<x<100) =0.91044

Exercise
Part – I Explain
/ Define
1. Experiment
2. Trials
3. Event
4. Probability
5. Sample Space
6. Sample point
7. Random Variable
8. Discrete Random Variable
9. Continuous Random Variable
10. Binomial distribution
Telegram: AgroMind Website: [Link]
 11. Poison distribution
12. Normal distribution
Part-II
Explain
1. Types of Events
2. Addition Theorem and multiplication theorem of probability

Part – III
1. Define Normal distribution and explain properties of normal distribution.

Telegram: AgroMind Website: [Link]

 UNIT VII SAMPLING THEORY
Introduction
The basic input in statistical analysis is the data. The data can be collected either
by experimentation or by sampling methods. The sampling methods are used for
a variety of purposes.
For example,
i) A consumer wants to purchase a commodity, inspects a portion of
it ii) In industry the quality of a product is assessed through
sampling.
iii) In agriculture,
a) to determine the germination rate of a seed material
b) to estimate the area covered under a particular crop
c) to estimate incidence rates of pests and diseases etc.,
Population
The aggregate of all units is known as population
Examples. i) All the students of TNAU
ii) All paddy plants in a field
iii) All the fields under a specified crop
iv) All farms in a village
Sample
A portion of the population is known as sample
Example i) A handful or rice out of a bag of rice.
ii) A collection of plants from a field
Sampling
The process of selecting a sample from the population is known as sampling.
Sample survey
If the required information is collected through the sample then it is called
sample survey.
Census

Telegram: AgroMind Website: [Link]

 It is defined as the study of all objects in a population. It is a also known
as complete enumeration. Sample survey Vs Census
Sample survey Census
The information obtained Reliable
about the object is less reliable
Accuracy of information is not Possible
possible
Is economical in all aspects Consumes much time energy
and money Is the study of smaller part of Study of all
objects of population population.

Sampling error
The error arising in drawing inference about the population on the basis of few
observations (sample) is termed as sampling error.
Methods of Sampling
Sampling methods are classified into two types
1) Probability sampling; 2) Non – Probability sampling
Probability sampling
It is a sampling in which individuals if the population are selected with a known
amount of probability for the inclusion of the individuals in the sample.
Non – Probability sampling
It is a sampling in which the units are drawn using certain amount of judgement.
Types of probability sampling
1 .Simple random sampling
It is defined as every item of population is given an equal chance of being
include in the sample.
If the population consists of N units then the probability of selecting any unit is
1/N. Suppose we are sample of size ‘n’ from the population of size N then there
are NCn number of possible samples of size n. If all possible samples have an
equal probability 1/NCn of being drawn then the sampling is said to be random
sampling. There two methods available to select a random sample.
i) Lottery method ; ii) Random number table method

Telegram: AgroMind Website: [Link]

 I) Lottery method
In this method the individuals in the population are given different numbers or
names. The names or numbers are written on different bits of papers of equal
size and shape, The different paper bits are rolled. Care should be taken that the
paper rolls are identical. The named or numbered paper rolls are put into a box.
Select paper rolls equal to the required number of individuals in the sample.
Open the paper rolls and note the names or numbers.
The selected individuals will form the required random sample.
Advantage
It is simple to understand
Disadvantage
If the sample size and population size are large then this method becomes
laborious.
II. Random number table method
If the total number of units in the population is three digited or two digited or
one digited number then three or two or one digited random number table is taken
accordingly. In general if the total number of units in the population is a ‘n’
digited number, then n digited random number table will be taken.
First number is selected anywhere in the table and the second number is selected
from the downward direction after the first number. The drawing of random
numbers is stopped as soon as the required sample size is obtained. Eg. Suppose
it is required to take a random sample of five plants form a row o
50 plants form the 2 digited random number table the following number are
selected, 23, 34,
37, 19, 41. The required sample has 19th , 23rd , 34th, 37th, and 41st plants.
[Link] sampling
It is a sampling in which the first member is selected at random and
remaining members are selected systematically.

Telegram: AgroMind Website: [Link]

 Eg. To select 5 plants out of 5 0 plants in a row, select the first plant by random
number table method or lottery method from the first 10 plants Suppose 7 th
plant is selected then 12th, 17th, 22nd and 27th selected.
[Link] sampling
It is a sampling method in which the population is divided into different strata
(or classes) and samples are drawn form each stratum. Stratum is defined as a
group in which the individuals are more or less homogeneous.
Eg. To select trees from an orchard the trees of same age can be formed as groups.
Then each group will be a stratum because the trees with in the group are
more or less homogeneous. Then trees will be selected at random form each
stratum either by lottery or random number table method.
[Link] sampling
It is a sampling in which clusters are selected at random and individuals in the
selected clusters are studied without omitting even a single individual. Eg. In a
study of fruits of tree, the branches may be numbered and the required number
of branches will be selected at random. All the fruits in the selected branches
will be studied without omitting even a single fruit. In this sampling
different branches are clusters.

Types of non – probability sampling :

i) Purposive sampling It is a sampling in which sample is drawn with
definite purpose
Eg. To get information on turmeric growth is Tamilnadu, if turmeric growers in
Madurai, alone are taken as the sample then the sample is a purposive sampling.
Parameter
The distribution of observation in a population is described in terms of certain
measures like mean, S.D., etc… These measures are known as parameters.
2
Eg. In a normal population the parameters are mean and variance ( , )

Telegram: AgroMind Website: [Link]

 Statistic
In order to describe the distribution of sample observation, mean, SD, etc.. will
be calculated. These measures sample mean sample SD are termed as statistic
the are denoted by x and s.
The statistics is random variable it varies form sample to sample on the other
hand the parameter is a fixed constant. Generally the parameters are unknown
the have to be estimated by corresponding statistics.

Sampling distribution
From a population of size N we can drawn NC n number of different samples of
size ‘n’.
For each of sample drawn we can compute a statistics. These N C n values may
be grouped in to a frequency distribution.
The probability distribution of a statistic can be obtained from the frequency
distribution. This probability distribution is known as sampling distribution of
the statistic.
The important sampling distribution of
i) Student ‘t’ distribution ii)
Fisher’s distribution iii)
Chi – square distribution
Standard Error
The standard deviation of sampling distribution of a statistics is called
standard error of the statistic. The square of the standard deviation is known as
variance of the statistic.
Importance of standard error
It plays an important role in the inferential statistics. The reciprocal of the
standard error is taken as the measure of precision for statistics.

Telegram: AgroMind Website: [Link]

 σ
Standard error (S.E) = .
n

where
- population standard deviation.
n - sample size.
* Standard error used to determine the ‘limits within which the population
parameter may be expected to tie’.
* If also forms the basis of the testing of hypothesis.
Solved Problems

Exercise
Part – I
Explain / Define
1. Sample
2. Population
3. Sampling
4. Sample survey
5. Census
6. Sampling error
7. Sampling distribution
8. Standard error
9. Parameter
[Link]
Part – II
1. What is probability and non-probability sampling methods
2. List out the probability and non-probability sampling methods Part –
III
[Link] explain about “Types of probability and non-probability sampling.

UNIT-VIII TESTING OF HYPOTHESIS

Statistical hypothesis
“In any scientific work, a statement about a phenomenon occurring under certain
conditions is formulated as proposition called Scientific hypothesis”.

Telegram: AgroMind Website: [Link]

 A scientific hypothesis may be rejected or accepted. For this purpose the
scientific hypothesis is translated into statistical language.
Eg. The scientific hypothesis may by “ the yield of new paddy variety will be
3500 kg/hec”.
In statistical language it may be stated as a random variable (Yield of paddy) is
distributed normally with mean (3500 kg/hec). Such a statement or proposition
about the probability distribution of random variable is called “ a statistical
hypothesis”.
Simple hypothesis
A hypothesis specifies all parameters of a probability distribution is known as
simple hypothesis.
Eg. A statistical hypothesis is involved parameters of a normal distribution, mean
and standard deviation called simple hypothesis.
Composite hypothesis
If a hypothesis specifies only some of the parameters of the probability
distribution known as composite hypothesis.
Eg. A statistical hypothesis contain, the parameter mean or SD alone called
composite hypothesis.
Null hypothesis
The original hypothesis always is not clear and accurate.
Eg. Paddy a variety A gives increased yield per hectare than variety B. The
increased may be 1 quintal or 10 quintal or 20 quintal. Practically it is not
possible to verify all the alternatives, that is the original hypothesis can not be
tested directly some hypothesis may be put in the form Paddy variety A will give
same yield /hec as that if variety B or)there is no difference between the average
yield of paddy varieties A and B. These hypothesis are in definitive terms. They
can be tested directly. If this hypothesis accepted then the original hypothesis
will be rejected. On the other hand, if this hypothesis are rejected than original
hypothesis accepted.

Telegram: AgroMind Website: [Link]

 The hypothesis which reject are accept the original hypothesis is known as Null
hypothesis. The original hypothesis known as alternative hypothesis. Note : The
process of deciding whether to accept or reject the null hypothesis is known as
testing of hypothesis.

Type I and Type II Error

In testing of hypothesis, the decision can be made by accepting or rejecting the
null hypothesis. In this decision making, it is possible to commit one of two kind
of errors.
i) We may reject null hypothesis when it is correct ii) We
may accept null hypothesis when it is not correct These errors
are known as Type I and Type II errors respectively.
Level of Significance
The probability levels used for deciding whether to accept or to reject a null
hypothesis called levels of significance. The levels of significance is the
minimum probability of rejecting, a null hypothesis hypothesis which is in fact
true. That is probability of committing Type I error is known as level of
significance. Conventionally the levels of significance is taken as 5% or 1%
Critical region: The testing of hypothsis involues the choice of region on the
sampling distribution of [Link] the statistic falls within this region,the null
hypothesis is rejected otherwise it is accepted. This region is called the critical
region.
One tailed and two tailed test
The nature of the original hypothesis determines the nature of the alternative
[Link] the original hypothesis states that two groups will differ with
respect to means then the direction of the difference is not [Link] the
hypothesis states that the mean of one group is larger or smaller than the other
then the direction of the difference is [Link] case the direction of the
difference is known,the critical region will be on one end of the sampling

Telegram: AgroMind Website: [Link]

 [Link] the is known as one tailed [Link] the direction of difference is
not known,the critical region will on either end of the sampling
[Link] it is called two tailed tests.

Degrees of freedom
The number of degree of freedom is the number of observations that are free to
vary after certain restrictions have been placed on the data.
Eg. We are asked to write numbers with restriction imposed that their sum is 30.
Then we can choose, only 2 numbers and the third number is automatically
determined by relations X1+X2+X3=30 ; X3=30 – (X1+X2), here there are 2
degrees of freedom and the number of restriction is one (ie. The total should be
30)
In general if there are N observations and K restrictions then the degrees of
freedom will be N-K.
Steps in Testing of hypothesis
i) Formulation of Null and alternative hypothesis ii) Specification level of
significance iii) Selection of test statistic and its computation iv) Find out
critical value form tables using the level of significance, sampling
distribution and its degree of freedom.
v) Determination of the significance of test statistic vi) Decision about the
null hypothesis based on the significance of the test statistic.
vii) Writing the conclusions in such a way that will answer the question
on hand.
Test of Significance
It was used to test whether the sample mean equal to population mean (one
sample case)(or) whether to test the significant difference between the two
sample mean equal to population mean (or) not (two sample case).For testing the
above cases if the sample size is large(at least 30 ie.,≥30)apply “z”test (or) “t”
test when small sample size(less than 30 ie.,<30)as follows.

Telegram: AgroMind Website: [Link]

Telegram: AgroMind Website: [Link]
 Tests of significance for large samples (‘Z’ test)
One sample test case:
1.
Assumption
- Population is normal
- Sample is drawn at random Conditional
- known
- sample u large (> 30)

o
Test statistic z= x- N (0,1)
/ n

Case 2.
Conditional
- not known

z= x- o
s/ n

(xi-x )2
where s =
n-1

Or

s = ( xi)2 –
( xi)2 n-1

Two sample cases Case:1.

Assumption
Telegram: AgroMind Website: [Link]
 - Population is normal
- Samples are drawn at random.
Conditional:-
- is known
- large sample size.
Test statistic x1 -x2
z= N (0,1)
2
( 1/n1 + 1/ n2)

Case.2.
Conditional
- is not known
x1-x2
Test statistic z = N (0,1)
(s21)/(n1) 2
+ (s 2) /( n2)

Tests of significance for small samples (‘t’ test)

One sample test
Assumption
- Sample is drawn at random
- Samples are independent in nature Conditional
- sample is small
- not known

t= x- o
s/ n

(xi-x )2 t-distribution where

s=-
n-1 with(n-1)df

Or

Telegram: AgroMind Website: [Link]

 ( n-1
s = xi)2 – ( xi)2

Two sample cases(‘t’ test)

Case:1
Assumption
- Samples are drawn at random and independent

Conditional:-
- is not known with equal variances -
Small sample size.
x1 -x2
Test statistic t= t(n1+n2-2)df
s2 ( 1/n1 + 1/ n2)

Case:2
Assumption
- Samples are drawn at random and independent
Conditional:-
- is not known with unequal variances - Small sample size.
x1 -x2 t1/2(n1+n2)-1df
Test statistic t=
s2 ( 1/n1 + 1/ n2) if n1=n2

Telegram: AgroMind Website: [Link]

 Behrens-Fisher
x1 -x2 d-

Test statistic t= (S12/n1 + S22/ n2) distribution.

if n1≠n2

In this method the critical value of t (tw) is

calculated as follows

s12 s2
t1 n +t2 2

i.e., tw = 12 2 n2

s1 s2
+
n1 n2
Where, t1 is the critical value for n1 df. at a specified level of significance, t2 is
the critical value for n2-1df at the same level of significance.
Two sample cases (Paired ‘t’ test)
Case:1
Assumption
- Samples are drawn at random and dependent

Conditional:-
- is not known
- Small sample size.
d
t= s/ n
Where d di n-1 di
= xi-yi S2 =
( di)2 – ( di)2
n-1

Chi-square test
Telegram: AgroMind Website: [Link]
 It is used to test
i) the independence of attributes ii) the
goodness of fit iii) the homogeneity of
variances iv) the homogeneity of correlation
coefficients v) the linkage in genetic
problems
Contingency table: A two-way table with two attributes is known as
contingency table.
For example, an entomologist may be intersted to know the effectiveness of
different concentrations of a chemical in killing the insects. In this case the
concentrations of the chemical form one attribute. The state of insects ‘killed’
and ‘not killed’ forms another attribute. The results of this experiment can be
arranged in the form of a contingency table. The association between these two
2
attributes may be tested using statistic.
Goodness of Fit
In testing of hypothesis our objective may be to test whether a sample has come
from a population which has a specified theoretical distribution like normal,
binomial and Poisson. In order words, it may be ncessary to test whether an
obtained frequency distribution resembles a theoretical distribution. In plant
genetics, our interest may be to test whether the observed segregation ratios
deviate significantly from the Mendelian ratios. In such situations we want to
test agreement between the observed and theoretical frequenceis. Such test is
called as test of goodness of fit.
The null hypothesis in goodness of fit tests is that there is no disagreemtn
between the observed and theoretical distributions, or the observed distribution
fits to the theoretical distribution. The test statistic to test this null hypothesis is

(O-E)2
2
= ------------, where O = observed frequency,
E E = expected frequency
Telegram: AgroMind Website: [Link]
 2
In general, the degrees of freedom associated with this is k – 1, where k is the
number of classes or groups.
Solved Problems
One Sample‘t’ test

1. Based on field experiments, a new variety of green gram is expected to

give an yield of 12 quintals per hectare. The variety was tested on 10 randomly
selected farmers’ fields. The yields (quintal/hectare) were recorded as 14.3,
12.6, 13.7, 10.9, 13.7, 12.0, 11.4, 12.0, 12.6 and 13.1. Do the results confirm
the expectation?

μ=12, n=10
X X2
14.3 204.49
12.6 158.76
13.7 187.69
10.9 118.81
13.7 187.69
12.0 144
11.4 129.76
12.0 144
12.6 158.76
13.1 171.61
x=126.3 x2 =1605.77
Step: 1

H0 : = 0

H1 : ≠ 0

Step: 2
Level of significance = 1%, 5%, 10%

Telegram: AgroMind Website: [Link]

 x−
t cal =
S/ n
x 126.3
x= 12.63
n = 10 =
2

x −( )
2 x
S= n
n −1
(126.3) 2
1605.77 −
10
= 9

Step: 3 = 10.60
9
S =1.08
12.63−12

t=
1.08/ 10

=
=1.85

Step: 4 t distribution – critical

value
Degrees of freedom = (n-1)
= 10-1
=9
Level of significance
1%, 5%, 10% table values are 3.25, 2.26, 1.83respectively

Step: 5

Calculated value is less than tabular value, so the test is non significance.
Step: 6

Null hypothesis accepted.

Step: 7

The result confirm the expectation.

Telegram: AgroMind Website: [Link]

 2. Ten plants are chosen from a population at random whose heights in inches
are given below:-
48, 52, 55, 57, 61, 64, 65, 68, 70 and 71.
In the light of above data discuss the suggestion that the mean height of plants in
the population is 60 inches.

μ=12, n=10
X X2
48 2304
52 2704
55 3025
57 3249
61 3721
64 4096
65 4625
68 4900
70 5041
71
x =611 x2 = 37890
Step:1

H0 : 1= 60
H1 : 1 60

Step: 2
Level of significance = 1%, 5%, 10%

Step: 3
x−
t cal = S / n

x=
x
n

Telegram: AgroMind Website: [Link]

 = ,
= 61.1

37890 − ( )
611
S= 10
9
557.9
= 9
= 7.87

t=
1.1
= = 0.44
2.49

Step: 4 t-distribution
Degree of freedom = (n-1)
= 10-1 =9
Level of significance
1%, 5%, 10% values are 3.25, 2.26, 1.83 respectively

Step: 5
Calculated value is less than tabular value. So the test is non significant.

Step: 6
Accept null hypothesis

Step: 7
Heights of plants is equal to the Mean height of plants.

[Link] following data give the reduction in microbial population(10 6 g/l) for
spraying observed in maize:2,12,8,4,1,14,8,[Link] there any evidence in the data
for reduction in microbial population for spraying in maize?
X X2
2 4
12 144
8 64
4 16

Telegram: AgroMind Website: [Link]

 1 1
14 196
8 64
3 9
x = 52 x2 = 498

Telegram: AgroMind Website: [Link]

 Step:1

H0 : = 0

H1 : 0

Step: 2
Level of significance 1% or 5% or 10%

Step: 3
x−
t cal =
S/ n
x 52
, = , = 6.5
n 8
2

498 − ( )
52
8
8 −1
2704
498 −
8
7
498 − 338
7
160
= , 22.857 , S = 4.78
7 =
6.5 − 0
t cal =
4.78 / 8

= 4.
x=

S=

Telegram: AgroMind Website: [Link]

 6.5
= , = 3.834
1.695

Step: 4

t-distribution
Degree of freedom = n-1 = 8-1 =7
Level of significance
1%=3.449
5% =2.365
10%=1.895

Step: 5
Calculated value is more than tabular value. So the test is significant.

Step: 6
Reject null hypothesis.

Step: 7
There is reduction in microbial population for spraying in maize.

Telegram: AgroMind Website: [Link]

 F-test for equality of two variances

1. The following data present the yield in quintals of corn on ten subdivisions of
equal area of two agricultural plots. Test whether two samples taken from two
random populations have the same variances?
Plot - 1 6.2 5.7 6.5 6.0 6.3 5.8 5.7 6.0 6.0 5.8
Plot – 5.6 5.9 5.6 5.7 5.8 5.7 6.0 5.5 5.7 5.5
2

X1 X2 X21 X22
6.2 6.6 38.44 43.56
5.7 5.9 32.49 34.81
6.5 5.6 42.25 31.36
6.6 5.7 36 32.49
6.3 5.8 39.69 33.64
5.8 5.9 33.64 32.49
5.7 6.0 32.49 36
6.0 5.5 36 30.25
6.0 5.7 36 32.49
5.8 5.5 33.64 30.25
x1 =60.6 x2 =58 x21 = 368.2 x22 = 339.34

Step:1 H0

: 12 = 22

2 2

H0 : 1 2

Step: 2
Level of significance = 1% (or) 5%

Step: 3

2
x − n
( x1 )2
1
1

S 12 = n1 −1

368.2 −
Telegram: AgroMind Website: [Link]
 =
9
0.

= 0.1071
(58)2
337.34 −

s22 = 9 10

0.

= 0.104 s122

= 0.1071

s 2 0.104
=1.0298

Step: 4
F-distribution
Degree of Freedom = (n2-1), (n1-1)
= (10-1) (10-1)
= 9, 9
Level of significance = 1%, 5%
= 5.35. 3.18
Step: 5
Calculated value is less than tabular value. So the test is non significance.

Step: 6
Null hypothesis is accepted.

Step: 7
The samples are having same variance.

Telegram: AgroMind Website: [Link]

 2. Two random samples were drawn from two normal population and their values
are

A 66 67 75 76 82 84 88 90 92
B 64 66 74 78 82 85 87 92 93 95 97

Test whether the two populations have the same variance?

X1 X2 X21 X22
66 64 4356 4096
67 66 4489 4356
75 74 5265 5476
76 78 5776 6084
82 82 6724 6724
84 85 7056 7225
88 87 7744 7569
90 92 8100 8464
92 93 8464 8649
97 9025
9409
x1 =720 x2 =818 x21 = 58334 x2 = 68052
Step: 1

H0 : 12 = 222

H1 : 1 2

Step: 2
Level of significance = 1%, 5%

Step: 3
2
x −
( x1)2
1
n1

s 12 = n1 −1

Telegram: AgroMind Website: [Link]

 (720)2
58334 −
9
=
8

= = 91.75
2

68052 − (818) s 22

9 10
=

= =126.62

s 2 126.62 F = 22 =
s 1 91.75
=1.38

Step:4

F distribution
Degree of freedom = (n2-1)(n1-1)
= (11-1)(9-1)
=10, 8
Level of significance = 1% , 5%
= 5.82, 3.34

Step:5
Calculated value is less than tabular value so non significance.

Step: 6
Null hypothesis accepted.

Step: 7
Two sample having equal variance.

Telegram: AgroMind Website: [Link]

 3. Two samples are drawn from normal populations. From the following data test
whether two samples have the same variances or not?

Sample 60 65 71 74 76 82 85 87
1
Sample 61 66 67 85 78 63 85 86 88 91
2

X1 X2 X21 X22
60 61 3600 3721
65 66 4225 4356
71 67 5041 4489
74 85 5476 7725
76 78 5776 6084
82 63 6724 3969
85 85 7225 7225
87 86 7569 7396
88 9944
91 8281
x1 =600 770 x21 = 45636 x2 = 60490

Step:1

H0 : 12 = 222

H1 : 1 2

Step: 2
Level of significance 1% (or) 5%.

Telegram: AgroMind Website: [Link]

 2
x −
( x1)2
1
n1

S 12 = n1 −1

(600)2
45636 −
8
S 12 = 7

= = 90.85
2
x −
( x2)2
2
n2

s 22 = n2 −1

(770)2
60990 −
= 10
9

= = 133.3

Step.3 F = s222 = 133.3 s1

90.85
= 1.467
Step: 4
F distribution
Degree of freedom = (n2-1) (n1-1)
= (10-1)(8-1)
=9,7
Level of significance = 1%, 5%
= 6.71, 3.68
Step:5
Calculated value is less than tabular value. So the test is non significant

Step: 6
Accept null hypothesis.
Step: 7
Both the samples having equal variance.

Telegram: AgroMind Website: [Link]

 From the following table showing the number of plant having certain characters

that the flower colour is independent of the leaf shape.

Flower Color Shape of the leaf Totals

Flat Curled
White 99(a) 36(b) 135(r1)
Red 20(c) 5(d) 25(r2)
Total 119(c1) 41(c2) 160(N)
Step:
1
H0 = Two attributes are independent
H1 = Two attributes are dependent
Step:
2
Level of significance: 1% 5%
Sample size is large

Step: 3

2Cal = (ad −bc).N

R1,R2,C1,C2

= (99 5) − (36 20)2.160 = (−225)2.160

135 25 119 41 16466625

= = 0.4919
2
cal = 0.4919
Step:
4
Degrees of freedom = (r-1) (-1)
= (2-1) (2-1)
=1
= 6% 5%
Telegram: AgroMind Website: [Link]
 = 6.64, 3.84

Step: 5
Accept null hypothesis
Step:
6
Calculated value < tabular value so it is non significance

Step: 7
The flower colour is independent of the leaf shape.
********

2. In a survey a random sample of 198 farms were classified into three classes
according to tenure status as owned, rented and mixed. They were also
classified according to the level of the soil fertility as highly fertile, moderately
fertile and low fertile farms. The results are
Classification of farms by tenure status and fertility level.
Tenure status Total No. (%)
Soil fertility
Owned No. Rented No. Mixed No.
(%) (%) (%)
High 40 (64.5) 12 (19.4) 10 (16.1) 62 (100.0)
Moderate 22 (47.9) 10 (21.7) 14(30.4) 46 (100.0)
Low 22 (24.4) 26 (28.9) 42 (46.7) 90 (100.0
Total 84 (42.4) 48 (24.3) 66 (33.3) 198 (100.0)

Does tenure status depend on soil fertility?

O; E; O;-E; (O; - E;)2 (O;−E;)2

E;
40 26.3 13.7 187.69 7.13
12 15.03 -3.03 9.1809 0.61
10 20.6 -10.6 112.36 5.45
22 19.5 2.5 6.25 0.32

Telegram: AgroMind Website: [Link]

 10 11.15 -1.15 1.8225 0.118
14 15.33 -1.33 1.7689 0.115
22 38.8 -16.18 261.7924 6.85
26 21.81 4.19 17.5561 0.804
42 30.00 12.0 144.00 26.194

E11 = R1C1 = 62 84 = 5208 = 26.3

N 198 198

E12 = R1C2 = 62 48 = 2976 =15.03

N 198 198

E13 = R1C3 = 62 66 = 3864 = 20.6

N 198 198

E21 = R2C1 = 46 84 = 4092 =19.5

N 198 198

E22 = R2C2 = 46 48 = 2208 =11.15

N 198 198

E23 = R2C3 = 46 66 = 3036 =15.33

N 198 198
E31 = R3C1 = 90 84 = 7560 = 38.18
N 198 198

E32 = R3C2 = 90 48 = 4320 = 21.81

N 198 198

E33 = R3C3 = 90 66 = 5940 = 30.0

N 198 198

Step: 1
H0 = Two attribute are independent
H1 = Two attribute are dependent

Step: 2
Telegram: AgroMind Website: [Link]
 Level of significance = 1% 5%
Sample size large

Step: 3
2
cal = 26.197

Step: 4
Degree of freedom = (r-1) (-1) = (3-1) (3-1) = (2,2) = 1% 5%
= 9.21, 5.99

Step: 5
Calculated value = tabular value so it is significance

Step: 6
Reject the null hypothesis

Step: 7
Tuner status dependent on fertility level
********

3. In order to determine the possible effect of a chemical treatment on the rate

of germination of cottonseeds a pot culture experiment was conducted. 140
chemically treated seeds and 160 untreated seeds were sown. The results are
given in table.

Germinated Not Germinated Total

No. (%) No. (%) No. (%)
Chemically treated 118 (84.3) 22 (15.7) 140 (100)
Untreated 120 (75.0) 40 (25.0) 160 (100)
Total 238 (79.3) 62 (20.7) 300 (100)
Test whether the attributes are independent or not?
Step: 1
H0 = Two attribute are independent
H1 = Two attribute are dependent

Telegram: AgroMind Website: [Link]

 Step: 2
Level of significance = 1% 5%
Sample size large

Step: 3

2Cal = (ad −bc).N

R1,R2,C1,C2

=
2
cal = 3.926
Step: 4
Degree of freedom = (r-1) (-1) = (2-1) (2-1) = 1
= 1% 5%
= 6.64, 3.84

Step: 5
Calculated values < tabular value so it is non significance

Step: 6
Accept null hypothesis

Step: 7
Effect of a chemical treatment on rate of germination of cotton seeds a
pot calculus experiment was conducted is independent. ******

4. In an experiment on the effect of a growth regulator on fruit setting in

muskmelon the following results were obtained.
The effect of growth regulator on fruit setting in muskmelon.
Fruit set Fruit not set Total
Treated 16 9 25
Control 5 20 25
Total 21 29 50
Test whether two attributes are independent or not?
Telegram: AgroMind Website: [Link]
 Step: 1
H0 = Two attribute are independent
H1 = Two attribute are dependent

Step: 2
Level of significance = 1% 5%

Step: 3

2Cal = (ad −bc).N

R1,R2,C1,C2

= = 9.9343
2
cal = 9.9343

Step: 4
Degree of freedom = (2-1) (2-1) = (1, 1)
= 1% 5%
= 6.64, 3.84
Step: 5
Calculated > tabular value so it is significance

Step: 6
Reject the null hypothesis

Step: 7
The affect of a growth regulation on fruit setting in muskmelon is dependent
with fruit not set.

[Link] a cross between parents of the genetic constitution AAbb and aaBB, the
phenotypes in F2 sample is classified as follows:
AB Ab aB ab Total
87 29 32 12 160
Telegram: AgroMind Website: [Link]
 They are expected to occur in a 9:3:3:1 ratio. Does the segregation ratio agree
with the theoretical ratio?

Oi Ei Oi-Ei (Oi - Ei)2 (Oi − Ei)2

Ei
87 90 -3 9 0.1
29 30 -1 1 0.033
32 30 2 4 0.13
12 10 2 4 0.4

9
E11 = 160 = 90
16

3
E12 = 160 = 30
16

3
E2 = 160 = 30
16

1
E22 = 160 =10
16

Step: 1
H0 = Two attribute are independent
H1 = Two attribute are dependent

Step: 2
Level of significance = 1% 5%

Step: 3
2
cal = 0.66

Telegram: AgroMind Website: [Link]

 Step: 4
Degree of freedom = K – 1 = 4 – 1 = 3

Step: 5
Calculated value < tabular value so it is non significance

Step: 6
The segregation ratio agree with the theoretical ratio is independent.

*********

Telegram: AgroMind Website: [Link]

 PAIRED T TEST
[Link] an experiment on sorghum the plots were divided into two equal parts
each. One half of each plot received soil treatment. A and the other half
received soil treatment B. Each plot was planted with sorghum. Test whether
the sorghum yield is same in both the treatments. The results are given below.
Sorghum yield (kg/plot) under two soil treatments.
Plot No. Soil treatment
A B
1. 26.3 28.0
2. 24.2 25.7
3. 23.1 22.9
4. 21.6 22.4
5. 24.5 25.2
6. 23.3 25.4
7. 21.8 23.0
8. 22.9 23.4
9. 25.7 25.5
10. 24.1 24.8
11. 25.8 25.6
12. 26.1 25.7
Solution
Plot No. Soil Treatment d = A-B d2
A B
1 26.3 28.0 -1.7 2.89
2 24.2 25.7 -1.5 2.25
3 23.1 22.9 0.2 0.04
4 21.6 22.4 -0.8 0.64
5 24.5 25.2 -0.7 0.49
6 23.3 25.4 -2.1 4.41
7 21.8 23.0 -1.2 1.41
8 22.9 23.4 -0.5 0.25
9 25.7 25.5 0.2 0.04
10 24.1 24.8 -0.7 0.49
11 23.8 25.6 0.2 0.04

Telegram: AgroMind Website: [Link]

 12 26.1 25.7 0.4 0.16
d =−8.2 d2 =13.14

Step: 1
H0: 1 = 2

H1: 1 2

Step:
2
Level of significance is 1%, 5%, 10% Step: 3
−
d
Paired tcal =
sd
n
d −8.2
0.683
n = 12 = −

2 ( d )2
d −
n
n −1
(−8.2) 2
13.14 −
= 12
11
67.24
13.14 −
= 12
11
13.14 − 5.603
=
11
13.14 − 5.603
=
11
7.537
= 0.829
11 =
−

d=

sd =

Telegram: AgroMind Website: [Link]

 0.683
Paired tcal =
0.8291
12
0.683
=

=
tcal = 2.857
Step: 4 t
distribution
Degree of freedom = (n-1) = (12-1=11)
Level of significance = 1% 5% 10%
Step: 5
Calculated value is more than tabular value so it is significance.

Step: 6
Reject the null hypothesis

Step: 7
Sorghum yield under soil treatment. A is not equal to sorghum yield under
soil treatment B.

********

Telegram: AgroMind Website: [Link]

 2. Two laboratories carry out independent estimates of fat content for ice cream
made by a certain firm. A sample is taken from each batch, halved, and the
separate halves sent to the two laboratories. They obtain the following results.
Percentage of fat content in ice-cream.
Batch No. → 1 2 3 4 5 6 7 8 9 10
Lab. A 7 8 7 3 8 6 9 4 7 8
Lab. B 9 8 8 4 7 7 9 6 6 6

Is the testing reliable?

Batch no Lab A Lab B d =A– B d2

1 7 9 -2 4
2 8 8 0 0
3 7 8 0 1
4 3 4 -1 1
5 8 7 1 1
6 6 7 -1 1
7 4 4 0 0
8 4 6 -2 4
9 7 6 1 1
10 8 6 2 4
d =−3 d2 =17

Step: 1
H0: 1 = 2

H1: 1 2

Step: 2
Level of significance is 1%, 5%, 10%

Step: 3

Telegram: AgroMind Website: [Link]

 Paired tcal = −
d

sd
n
d −3
0.3
n = 10 = −
( d )2
d2 −
n
n −1
(3) 2
17 −
= 10
9
17 − 0.9
=
9
16.1
=
9
= 1.78 =1.33
10 = 3.16
−

d=

sd =

0.3
Paired tcal =

Telegram: AgroMind Website: [Link]

 =
tcal = 0.714

Step: 4 t
distribution
Degree of freedom = (n-1) = (10-1= 9)
Level of significance = 1% 5% 10%
= 3.250, 2.262, 1.833
Step: 5
Calculated value is less than tabular value so it is non-significance.

Step: 6
Accept the null hypothesis

Step: 7
Both laboratories are equal that content for ice-cream.

3. The following data relate to number of seeds set per pod in Lucerne.
Top flowers 4.0 5.2 5.7 4.2 4.8 3.9 4.1 3.0 4.6 6.8
Bottom 4.4 3.7 4.7 2.8 4.2 4.3 3.5 3.7 3.1 1.9
flowers

Test whether mean seeds set at top flowers is different from that of bottom
flowers.

Top flower Bottom flower di di2

4.0 4.4 -0.4 0.16
5.2 3.7 1.5 2.25
5.7 4.7 1.0 1
4.2 2.8 1.4 1.96
4.8 4.2 0.6 0.36
3.9 4.3 -0.4 0.16

Telegram: AgroMind Website: [Link]

 4.1 3.5 0.6 0.36
3.0 3.7 0.7 0.49
4.6 3.1 4.5 2.25
6.8 1.9 4.9 24.01
di =10 di2 = 33

d− =

di = 10 =1
n 10
Step: 1
H0: 1 = 2

H1: 1 2

Step: 2
Level of significance is 1%, 5%, 10%

Step: 3
−
d

sd
Paired tcal = n

2 ( di ) 2
di −
sd = n
n −1
100
33.44 −
= 10
9
33.44 −10
=
9
23.44
=
9
2.504 =1.593
Telegram: AgroMind Website: [Link]
 =

1
Paired tcal =
59
1.
10
1
= 1.59
3.16

= =1.99
0.

Step: 4 t
distribution
Degree of freedom = (n-1) = (10-1= 9)
Level of significance = 1% 5% 10%
= 3.250, 2.262, 1.833
Step: 5
Calculated value is less than tabular value so it is non significance

Step: 6
Accept the null hypothesis

Step: 7
Mean seed set at top flowers is same that of bottom flowers.

Telegram: AgroMind Website: [Link]

 INDEPENDENT T TEST
1. In order to compare the effectiveness of two sources of nitrogen, namely
ammonium chloride and urea, on grain yield of paddy, an experiment was
conducted. The results on the grain yield of paddy (kg/plot) under the two
treatment are given below.
Ammonium chloride : 13.4, 10.9, 11.2, 11.8, 14.0, 15.3, 14.2, 12.6, 17.0, 16.2,
16.5, 15.7.
Urea: 12.0, 11.7, 10.7, 11.2, 14.8, 14.4, 13.9, 13.7, 16.9, 16.0, 15.6, 16.0 Assess
which source of nitrogen is better for paddy.

X1 X2 X12 X22
13.4 12.0 179.56 144
10.9 11.7 118.81 136.89
11.2 10.7 125.44 114.49
11.8 11.2 139.24 125.44
14.0 14.8 196.00 216.04
15.3 14.4 234.09 267.56
14.2 139 201.64 193.24
12.6 13.7 158.76 187.69
17.0 16.9 289 285.61
16.2 16.0 262.44 256
16.5 15.6 275.25 243.36
15.7 16.0 247.49 256.00

8
X−1 = x1 = 168. =14.0
n1 12

9
X−2 = x2 = 166. =13.9
n2 12

Telegram: AgroMind Website: [Link]

 Step: 1
H0: 1 = 2

H1: 1 2

Step: 2
Level of significance is 1%, 5%, 10%

F= S222
S1

2 ( x1)2
x−
1
S12 = n
n −1
(168.8)2
= 2423.72 -
12
2423.72− 28493.44
= 12
11
= 4.479
S12 = 4.479

2 ( x2 )2
x−
2
S22 = n
n −1

= 2369.09 - (166.9)2

S12 = 4.344 f= = 0.969

Degree of freedom = (n2-1) (n1-1) (12-1) (12-1)

= 11, 11
Telegram: AgroMind Website: [Link]
 Step: 3
Level of significance = 1% 5%
= 2.82, 4.46

Step: 4
Calculated value is less than tabular value, so it is non significance and
accept null hypothesis

− −

X1− X 2
tcal =
1 1
S2 +
n1 n 2
2
S2 = (n1 −1) S1 − (n2 −1) 4.344
12 +12 − 2

=
S2 = 0.067

t cal = 14 −13.9 0.1

=
0.067( 1 + 1 ) 0.067(0.167)
2 2
0.1
= =10
0.01

*********

2. The summary of the results of an yield trial on onion with two methods of
propagation is given below. Determine whether the methods differ with regard
to onion yield. The onion yield is given in kg/plot.

Method I Method II
n1 = 12 n2 = 12

Telegram: AgroMind Website: [Link]

 x1 = 25.25 x2 = 28.83
SS1 = 186.22 SS2 = 737.6667
S11 = 16.9318 S22 = 67.0606

Step: 1
H0: 1 = 2

H1: 1 2

Step: 2
Level of significance 1%, 5%, 10%
− −

X1− X 2
t cal =
F = 3.96

Step: 3
Level of significance 1% 5%
Calculated value is more than tabular value so it is significance
F distribution = (n2-1) (n1-1)
= 2.82, 4.46
= 5% 1%

1
2 1 1
S n1 + n2

F = SS1222 = 1667..93180606 = 3.96

Telegram: AgroMind Website: [Link]
 S2 = SS1 +
SS2 = 186.22+ 737.6667 n1 +n2
12+12− 2

= = 41.99

t cal = 25.25− 28.83

= 41.99 = 6.741
1 1
12 12)

− 3.58
= −1.35
6.744
41.99−( +

t cal =

t distribution = n1 + n2 – 2
= 12 + 12 – 2
= 22

Step: 4
Determination of significance = 1% 5% 10%
2.819 2.074 1.717

Step: 5
Calculated value is less than tabular value, so it is non significance

Step: 6
Accept null hypothesis

Step: 7
The mean of the onion yield of its sample equal to the onion yield of 2 nd
sample
[Link] kernels of mature iodent corn were tested for crushing resistance.
Measured in pounds the
resistanceswere:50,36,34,45,56,42,53,25,65,33,40,42,39,43,[Link] batch of
15 kernels was tested after being harvested in the dough

Telegram: AgroMind Website: [Link]

 stage:43,44,51,40,29,49,39,59,43,48,67,44,46,54,64..Test the significance of the
difference between two means.

First batch Second batch X12 X22

50 43 2500 1849
36 44 1296 1936
34 51 1156 2601
45 40 2025 1600
56 29 3136 841
42 49 2809 1521
53 39 625 3481
25 59 4225 1849
65 43 1089 2304
33 48 1600 4483
40 48 1464 2404
42 67 1521 2116
39 41 1704 1936
43 46 1849 2916
42 64 1764 4096
x1 =645 x2 =720 x21 = 35128 x22 = 29970

Step: 1
H0: 1 = 2

H1: 1 2

Step: 2
Level of significance 1%, 5%, 10%
2 ( x1)2
x−
1
S12 = n
n −1

Telegram: AgroMind Website: [Link]

 29123
−
=
14

= = 99.14
2 ( x2 )2
x−
2
S22 = n
n −1

35936
−
=
14

= = 98.28
F - Test
S12 99.14
2 = =1.008
S2 98.28

Degree of freedom = n1+n2 – 2

= 15 + 15 - 2
= 28
Level of significance = 1% = 2.48, 5% = 3.70

Step: 3
Calculated value is less than tabular value, so it is non significance and
accept null hypothesis.
− −

t cal = X1− X 2
2
1 1
S n1 + n2

Telegram: AgroMind Website: [Link]

 43− 48
=
98.71 1+1
15 15

=
= 0.422

Step: 4
t – distribution
Degree of freedom = n1 + n2 - 2
= 15 + 15 – 2 = 28
level of significance = 1% = 1.7%, 5% = 2.048, 10% = 2.763

Step: 5
Calculated value is less than tabular value, so it is non significance

Step: 6
Accept null hypothesis

Step: 7
Crushing resistance in mature indent stage is equal to the dough stage
********

4. A new variety of cotton was evolved by a breeder. In order to compare

its yielding ability with that on a ruling variety, an experiment was conducted in
completely randomized design. The kapas yield (kg/plot) was observed . The
summary of the results are given below. Test whether the new variety of cotton
gives higher yield than the ruling variety.

New variety Ruling variety

n1 = 9 n2 = 11

x1 = 28.2 x2 = 25.9

Telegram: AgroMind Website: [Link]

 S21 = 186.22 S22 = 1.2822

Step: 1
H0: 1 = 2

H1: 1 2

Step: 2
Level of significance 1%, 5%, 10%
F - Test

S122 = 186.22 =145.23

S2 1.2822

Degree of freedom = (n2 – 1), (n1 – 1)

= (11-1) (9 -1)
= 10, 8
Level of significance = 1% = 3.34, 5% = 5.82

Calculated value is greatest than tabular value so it is significant.

− −

t cal = X 1− X 2 = 28.2 − 25.9

S1
2
S2
2
186.22 1.2822
n1 − n2 9 − 11

2.3 2.3
= =
20.69 − 0.11 20.58

=
= 0.5

(S12 ) (S22 ) t
+t
1 2
n1 n2
2 2
t tab = S1 S2
+
n1 n2

Telegram: AgroMind Website: [Link]

 =

= =
= 2.26

Step: 4
T – distribution
Degree of freedom = n1, n2 -1
= t, n – 1
= 9, 10
level of significance = 1% = 3.169, 5% = 2.228, 105 = 1.812

Step: 5
Calculated value is less than tabular value, so non significance

Step: 6
Accept null hypothesis

Step: 7
The new variety and ruling variety of cotton having same yield.

Exercise
Part-I Explain
/ Define
1. Statistical hypotheses
2. Scientific hypotheses
3. Simple hypotheses
4. Composites hypotheses
5. Null hypotheses
6. Original hypotheses alternate
7. Type I and Type II Error
8. Level of significances
9. Degrees of freedom
10. One tailed and two tailed test
11. Test of significances

Telegram: AgroMind Website: [Link]

 Part – II
1. Lift out the steps involved in testing of hypotheses 2.
Write about standard error and its importance
Part – III
1. Explain large and small sample test for comparision ofmeans(for both
one and two sample cases)

Telegram: AgroMind Website: [Link]

 UNIT IX DESIGN OF EXPERIMENTS

Design of experimental designs means the proper allotment of treatments to

available experimental units. Any object of comparison under the same or
identical condition is called a treatment. In agriculture the different varieties and
fertilizer tried in an experiment are called treatments. The treatments can be
individual factors or combination of factors. The experimental units are the
subjects of experimentation. The experimental unit depends on the type of the
experiment. The experimental unit may be a plant or a group of plants. The type
of experimental material to be used for the experiment will suggest what shall be
the experimental unit. The effect of treatment on the experimental units is called
the response due to the treatment. The experiment should clearly specify the
object of the experiment, the character under study, the type of experimental
material available, the treatments under comparison and the required precision
before designing the experiment. A good design of the experiment will provide
a proper method of statistical analysis of the data to get maximum information
with known confidence.

Basic principles of an experimental design

The basic principles of any experimental design are (i) Randomization; (ii)
Replication; (iii) Local control.
(i) Randomization
Randomization is defined as the allotment of treatments at random to the
experimental units. That means every unit of the treatment will get equal chance
of getting selected or allotted to an experimental plot.
It is process employed to avoid bias in the allotment of treatments to the
experimental units. Since the experimental units have inherent variations it is
necessary that no bias is brought in the allotment of these units to the different
treatments. The method of randomization is the only scientifically accepted

Telegram: AgroMind Website: [Link]

 method to avoid personal bias in the allotment and this objective approach is
necessary for the validity of the use of the standard errors of estimates in the
statistical analysis. The randomization can be done with the proper use of the
table of random numbers. The process of simple randomization consists of
providing equal chance to all experimental units to be put under any particular
treatment. The randomization procedure will be different for different types of
designs. It ensures an unbiased estimate in an experiment and enhances the
validity of the experiment.

(ii) Replication
Replication means, repetition of treatments to different experimental units. The
replication in an experimental design enhances the precision of that experiment.
That is it reduces the experimental error which is inversely proportional to the
replication, thereby the number of replications increase the efficiency of the
experimental design. If all the treatments are carried out an equal number of
experimental units resulting in equal number of observations under each
treatment, then all the treatments have equal number of replications. We can also
have different number of replications for different treatments, depending upon
the nature of the experiments.

(iii) Local Control

The local control also known as blocking, is used to mean the logical grouping
of the experimental units to control the variability among the experimental units.
The blocking is done with a view to group the available experimental units to a
few groups (blocks) consisting of more homogenous units than the whole set of
units. This may be done only when it is possible for the experimenter to group
the units according to an assignable source of variation which influences the
character under study.

Telegram: AgroMind Website: [Link]

 LAYOUT AND RANDOMIZATION

Layout of experimental design

Rules for Layout: First decide the number of plots. Number of plots will be
equal to the number of treatments(t) multiplied by number of replications (r).
The total number of experimental units will be ‘rt’.
Secondly divide the plots into number of rows and columns as many as the
number of treatments. Thirdly, number the plots serially from left hand side to
right hand side. After numbering the first row in the second row, the number is
given from right hand side to left hand side. For third row again the number is
given from left hand to right and for the fourth row from right to left and so on.
This process is continued till all the plots are numbered, where the number of
plots are equal to the number of experimental units (rt). ie. If all treatments are
alike, there is no need for any statistical analysis. A variable is one which
changes from unit to unit or individual to individual in any population. When
the treatment are not alike, different values will be obtained and it is possible for
statistical analysis.

Example: Layout an experiment which is having 6 treatments and 5 replication

In this case, they are 6 treatments and 5 replications and hence total number plots
is equal to 6 x 5 = 30. This can be divided into 5 rows and 6 columns. Each plot
can be obtained as below.
Layout

1 2 3 4 5 6
12 11 10 9 8 7
13 14 15 16 17 18
24 23 22 21 20 19
25 26 27 28 29 30

Telegram: AgroMind Website: [Link]

 COMPLETELY RANDOMIZED DESIGN

Completely randomized Design is the simplest 5 among all the

experimental designs. It is commonly used in highly homogeneous
experimental material of field. Generally for laboratory experiments this
design is most suitable.
Randomization
Procedure: The total number of plots (i.e.) treatment multiplied by replication
should be same as that of total number of random numbers. Selection of 3 digit
random numbers from statistical table is better compared to two digits or single
digit random number, since the repetition of random number might have been
more number of times, hence we select only 3 digit random numbers. Choose a
coloumn in the random number table at random and fix a number randomly.
Once the first number is fixed the other random numbers will automatically
follow. This process will be continued till the numbers selected is equal to the
total number of plots. If some random number is repeated in the selection, delete
that random number and select the succeeding one. After selecting the required
random numbers, rank them in ascending or descending order. The ranking of
these random numbers will be the plot numbers to be allotted to the various
treatments. After ranking divide the ranking into different groups according to
the number of replications. Here, the total number of groupings will be equal to
number of treatments. For example, if there are 6 treatments and 5 replications,
group the first 5 ranks and allot to the first treatment, a group another 5 rank to
the second treatments and so on. Here ranking means only the plots. (e.g.)
Randomise the 6 treatments and 5 replications.
Number of plots required = 6 x 5 = 30 plots.
S. No. Random numbers Ranks Treatment
1. 904 1
2. 416 16
3. 609 11 T1

Telegram: AgroMind Website: [Link]

 4. 762 6
5. 050 27
6. 734 7
7. 095 25
8. 247 19 T2
9. 795 5
10. 732 8
11. 047 28
12. 428 14
13. 061 26 T3
14. 034 29
15. 404 17
16. 702 09
17. 202 20
18. 843 04 T4
19. 164 22
20. 640 10
21. 852 03
22. 550 12
23. 424 15 T5
24. 299 18
25. 195 21
26. 438 13
27. 114 23
28. 010 30 T6
29. 111 24
30. 883 02

Layout
T1 T6 T5 T4 T2 T1
(1) (2) (3) (4) (5) (6)
T5 T1 (11) T4 T4 T2 T2
(12) (10) (9) (8) (7)
T6 (13) T3 T5 T1 T3 T5
(14) (15) (16) (17) (18)
T6 (24) T6 T4 T5 T4 T2 (19)
(23) (22) (21) (20)
T2 (25) T3 T1 T3 T3 T6 (30)
(26) (27) (28) (29)
Telegram: AgroMind Website: [Link]
 Eg. If there are 4 treatments and r replications the following will be the data table.
Treatments
Replication
I II III IV
1 X11 X21 X31 X41
2 X12 X22 X32 X42
3 X13 X23 X34 X43
… … … …
… … … …
… … … …
R X1r X2r X3r X4r

Total T1 T2 T3 T4

Since the effect of treatment is the only source of assignable cause of variation,
the total variation in the data under C.R.D. can be analysed as follows (one way
classification).

Analysis of Variance Table

Source of Degree of Sum of
Mean Square F-Ratio
Variation Freedom Squares
(M.S.) Calculated
(S.V.) (D.F.) (S.S)
Treatment Treatment Tr.M.S. =
(between) t-1 S.S Tr.S.S.
Treatment (Tr. S.S) t-1 F=
(within) E.M.S. = Tr.M.S./EMS
n-t Error S.S
(Error) E.S.S.
E.S.S. n-t
Total n-1 T.S.S.

t = No. of treatments
n = Total no. of observations

To calculate the sum of squares for sources of variations.

Steps:
1. Calculate the treatment totals and the grand total.

Telegram: AgroMind Website: [Link]

 2. Calculate the correction factor by making use of the formula. C.F. =
(Grand Total)2 / Total number of observations
3. Calculate the total sum of squares, treatment sum of squares and error
sum of squares.
Total sum of squares (T.S.S.) = Sum of squares of all observations
– C.F.
Treatment S.S. (Tr.S.S.) = {(T12 + T22 + …….. + T12) / r} – C.F
Error S.S. (E.S.S.) = T.S.S. – Tr. S.S.
Result: F value calculated has got (t-1), (n-t) degrees of freedom. If this value
of F exceeds the theoretical value at 5% level of significance then it is said to be
statistically significant. If the calculated value exceeds the tabular value at 1%
level of significance, F is declared to be highly significant.
If F is significant we conclude that the treatment effects are significantly
different. In that case, for further information we have to calculate the standard
error of difference of treatment means by the formula:
S.E. of differences of any two treatments ([Link]) = √2 x E.M.S. / r

The least significance difference (L.S.D.) or the critical difference (C.D) at 5%

and 1% are calculated by
1. Multiplying the standard error with value of t for (n-t) d.f. at 5% level.
2. Multiplying the standard error with values of t for (n-t) d.f. at 1% level
respectively.
The treatment means are calculated and also the difference between these
means are taken in pairs. Then differences of treatment means are compared
with C.D. and can be tested for statistical significance. When the actual
difference between any two treatments means exceeds the C.D. value then the
treatments are said to be statistically different in their performance. If that
difference is less than or equal to C.D, those treatments can be placed on par.
The Bar Chart can be prepared accordingly. Calculation of S.E. of difference of

Telegram: AgroMind Website: [Link]

 treatment means and the C.D. will have to be done only in case if F is found to
be significant. If F is not significant S.E. and C.D. need not be calculated.

Advantages of C.R.D.
It is easy to design and complete flexibility is allowed. Any number of
treatments and replications may be used. The number of replications can be
varied from treatment to treatment. All the available experimental materials can
be utilized.
The statistical analysis is simple and straight forward. Even if the number of
replications are not the same for all treatments it would not complicate the
statistical analysis.
The method of statistical analysis remain simple when the data from some units
or from whole treatments are missing or rejected. More over the relative loss of
information due to missing data is smaller than with any other design.
This design is particularly useful for comparison of growth promoting ability of
certain feeds on day old chicks or laboratory animals like mice, gainea pigs, etc.
In agricultural experiments if all the plots in the field are comparably equal
fertility, this design can be adopted to different treatments. (e.g.) Pot culture
experiments this design is more suitable for laboratory experiment.

Disadvantages
This design requires a homogenous set of experimental units. If the
homogeneity of the experimental units cannot be ascertained beforehand the
heterogeneous units make the experimental error large and this in turn makes the
treatment comparisons less efficient.
Therefore C.R.D. may be appropriate.
1. When the experimental material is homogenous,
2. When an appreciable fraction of units is likely to be destroyed or fail
to respond, and

Telegram: AgroMind Website: [Link]

 3. In small experiments where increased accuracy from alternative
designs does not outweigh the loss of error degree of freedom.
Example-1:
An experiment is conducted to find out which of the four available rations is
more efficient in promoting grain in weight of chicks at the end of a fixed period
of experimentation. There are 20 days old chicks from a hatch of one breed
available to carry out the experiment. How will you randomly allocate the 20
chicks to 4 feeds as per C.R.D?
Carry out the statistical analysis of variance and interpret the results of the data
obtained at the end of the experiment.

Randomization
The 20 chicks are wing handed with 1 to 20 numbers. The initial hatch weights
are recorded. Making use of table of random numbers, the experimental units
are allotted at random to r treatments.
Since the total number of experimental units (chicks) are 20, a 2 digit random
number table is chosen. Consider all the numbers upto 99 including 00 so that
all the chicks are given 5 chances each. Starting randomly from any 2 digit
random number, take the first random number, divide by 20 and the reminder is
the serial number of chick to be allotted to the treatment. Similarly second
random number is taken, divided by 20 and the reminder is the serial number of
chick selected for the treatment. But once a number has been taken for treatment,
that number will be omitted; if the same number comes subsequently because it
is not possible to use the same chick under the different treatments at the same
time. The first 5 serial number of series represent a batch of 5 chicks to be
allotted to the first treatment, the second set of 5 serial numbers represents the
chicks to be allotted to the second treatment; the third set of 5 to third treatment
and the remaining 5 chicks to the fourth treatment. After randomization, the
four groups of chicks are maintained separately feeding with one ration to each
group. At the end of the period of experimentation the chicks are weighted and
the data on weight gain is obtained. The data thus obtained is given below.
Gain in weight (Gms)
T1 I T2 II T3 III T4 IV
feed feed feed feed
11 8 12 6
Telegram: AgroMind Website: [Link]
 12 9 10 5
13 7 9 4
15 15 14 7
14 6 11 8

Total 65 45 56 30
Mean 13.0 9.0 11.2 6.0

Statistical analysis
Null hypothesis: Treatments means are not different Treatment
totals:
T1 = 65: T2 = 45: T3 = 56: and T4 = 30
Grand Total = 196
Correction factor = (Grand Total)2 / Total number of observations
C.F. = (196)2 20 = 38416 20 = 1920.8
Total sum of squares
(T.S.S.) = sum of squares of each observation - correction factor.
= (112 + 122 + 132 + …. +82) – CF
= 2142.0 – 1920.8 = 221.2
Treatment S.S.(Tr.S.S.) = {(T12+T22+T32+T42) / No. of observations under each
treatment (r)} – C.F.
= {(652 + 452 + 562 + 302) 5} – 1920.8
= 2057 – 1920.8 = 136.4
Error of S.S. (E.S.S.) = T.S.S. – Treatment S.S. = 221.2 – 136.4 = 84.8
Analysis of variance table
Degrees
Sum of Mean
Source of of
squares square F-Value
Variation (S.V.) freedom
(S.S.) (M.S.)
(d.f.)
Between 3 136.4 45.47 F = 45.47 5.30 =
treatments 8.58

Telegram: AgroMind Website: [Link]

 Error (within tr.) 16 84.8 5.30
Total 19 221.2

Table values of F at 3, 16 d.f. are 3.24 and 5.29 at 5% and 1% level respectively.
Since the calculated value of F is greater than the table values of F at 1% level,
the calculated F value is highly significant. Null hypothesis is rejected and
concluded that the treatments are significantly different. That is the effect of
feeds are different.
For further comparisons of different treatment means critical difference (C.D.)
between any two treatment means is calculated.
S.E. of difference of treatment means [Link]. = √ 2 EM.S. r

= √2 (E.M.S.) r = √ 2 x 5.3 5 = 2.12 = 1.45 gms.

C.D. at 5% = ‘t’ value at error d.f. at 5% x S.E. of treatment means
= 2.12 x 1.45 = 3.074 gms.
C.D. at 1% level = ‘t’ value at error d.f. at 1% x [Link] = 2.921 x 1.45 = 4.234 gms.

Comparisons of Difference Remarks

treatments

T1 – T2 4.0 Significant
T1 – T3 1.8 Not Significant
T1 – T4 7.0 Highly Significant
T2 – T3 2.2 Not Significant
T2 – T4 3.0 Not Significant
T3 – T4 5.2 Highly Significant

Bar Chart:

Treatments : T1 T3 T2 T4
Means: 13.0 11.2 9.0 6.0
Telegram: AgroMind Website: [Link]
 Problem-2
An experiment was conducted to find out the yielding abilities of 4 varieties of
a crop (v1, V2, V3 and V4). The experiment was conducted in a CRD. Each
treatment was replication 5 times. The layout planned with grain yield in kg/plot
of size 10 x 5m is given below. Analyse the data and draw your conclusions.
4 3 2 1
V2 41 V3 69 V4 18 V2 35
5 6 7 8
V3 53 V1 16 V4 16 V1 29
12 11 10 9
V4 17 V1 20 V1 20 V1 33
13 14 15 16

V2 33 V4 16 V3 76 V2 22
20 19 18 17
V3 77 V2 61 V4 18 V3 73

Telegram: AgroMind Website: [Link]

 Variety / Rep. V1 V2 V3 V4
R1 16 41 69 18
R2 29 35 53 16
R3 22 33 76 17
R4 20 22 77 16
R5 33 61 73 18
Total 120 192 348 85

Grand Total = 743

Step – 1 : Correction factor calculation

C.F. = (Grand Total)2 n = (GT)2 n = (745)2 20 = 555025 20 =
27751.25
Step – 2 : Total sum of squares
TSS = X12 – C.F. = 37323 – 27751.25 = 9571.75
Step – 3 : Varieties sum of squares
VSS = {(V12 + V22 + V32 + V42) 5} – C.F. = (179593 5) –
27751.25
= 35918.6 – 27751.25 = 8167.35
Step – 4 : Error sum of squares
ESS = TSS – VSS = 9571.75 – 8167.35 = 1404.40
After this, the values are to be tabulated in Analysis of variance table.

Analysis of Variance table

Sum of
Degrees of
Source squares Mean square (MS) F-calculated
freedom (d.f.)
(SS)
Between (t-1) = 4-1 =3 VSS=8167.3 VSS df=VMS F-VMS EMS
varieties 5 8167.35 3=2722.4 =2722.45 87.775
5 =31.016
Telegram: AgroMind Website: [Link]
 within (n-1)=20-1=19 TSS=9571.7 ESS df = EMS
varieties 5 =1404.40 16=87.7
75
Total (n-1) =20-1=19 TSS=9571.7
5

Fc=31.016 Ft=3.24

Telegram: AgroMind Website: [Link]

 Rules for Interpretation
i) If calculated ‘F’ value (Fc) is greater than the tabular value (Ft), then
we interpret that there exists a significant difference between the
treatment means.
ii) If Fc Ft it can be concluded that there is no significant difference
between treatment means.

Conclusion:- Fc > Ft (ie.31.06>3.24)

Hence, there exists a significant difference between the treatments means.

Therefore the C.D. is to be calculated for further information.

[Link] = √ 2EMS r; CD=SEd x ‘t’ 5% in error df

C.D = [Link] x ‘t’ value at 5% level with error degrees of freedom.

[Link] = √2 x 87.775 5 = √2 x 4.18 = 5.93

CD = 2.12 x 5.93 = 12.57

Summary of Results :-

Sl.
Varieties Mean S.E. C.D.P = (0.05)
No.
1. V1 24.0 4.18 12.57
2. V2 38.4
3. V3 69.6
4. V4 17.0

Conclusions:
V3 is superior to the remaining 3 varieties, V 2, V1 and V4. V2 is superior to V1
and V4. V1 and V4 are said to be on par with each other.

Bar chart: V3 V2 V1 V4
69.3 38.4 24.0 17.0

Telegram: AgroMind Website: [Link]

 Problem 3: The following data refers to the yield from an experiment layout in
completely
Randomized design with 5 treatments and 5 replications
D E 17 D 21 B E 17
20 15
A D A D B
9 17 9 23 15
A C E 18 C A
12 19 18 15
D A C E 16 B
16 8 18 19
B C E 16 B C
13 16 14 19

Solution
Step-1: One way classification table

Replication
/ R1 R2 R3 R4 R5 Total Mean
Treatment
A 9 12 8 9 15 53 10.6
B 12 15 14 15 19 76 15.2
C 19 16 19 18 18 90 18.0
D 26 16 17 21 23 97 19.4
E 13 17 17 16 16 84 16.8
Total 79 76 75 79 91 400

Correction factor C.F. = (G.T.)2 n = (400)2 25 = 6400

Step-2: Calculation of sum of squares

1. TSS = ∑ Xi2 – C.F. = 6822 – 6400 = 422

2. TrSS = {[(∑A)2 + (∑B)2 + (∑C)2 + (∑D)2 + (∑E)2] 5} – C.F. =

6432.86400 = 328
3. E.S.S. = TSS – TrSS = 422 – 328 = 94

Telegram: AgroMind Website: [Link]

 Step – 3: ANOVA TABLE
Source df S.S M.S Fc Ftab
Between (t-1) = 4 328 82.0 F= 2.87 at
treatment TMS EMS 5%
Within treatments (r-t) = 20 94 4.7 = 17.44 4.43 at
(Error 1%
)
Total 19 221.2

Since the calculated F value exceeds the table value, it is significant at 1% level
Step – 4:
C.D = √(2x4.7)

5
Bar Chart

D C E B A

19.4 18.0 16.8 15.2 10.6

Conclusion

(i) The treatments D and C; and E and B are statistically significant from
A, since the observed difference for D, C, E and B is more than CD
(ii) D and C as well as E and B are on par
(iii) A is significantly inferior to the treatments D, C, E and B. Solved
Problems
1. The following data gives the random observations on the yield of tomato in
CRD. To know the effect of tricontinol(growth regulator) at 50ppm, 100ppm,
150ppm, 200ppm and 250ppm concentrations. The figures given in the
following random layout plan refer fruit yield (kg/plot) for the treatment.
Analyse the data and draw the conclusions.

Telegram: AgroMind Website: [Link]

 T4 T2 T3 T4 T1
9.1 11.2 13.6 11.7 15.5

T1 T5 T1 T2 T3
16.0 10.8 15.8 11.7 14.1

T5 T2 T3 T1 T2
11.2 12.6 14.7 16.4 11.0

T4 T5 T5 T4 T3
11.9 11.2 11.9 16.2 14.1

Step: 1
Mathematical model for CRD
yij = +ti+eij
yij = jth observation on ith treatment effect
= grand mean (or) overall mean ti
= ith treatmental effect
eij = Error term

Solution
Step: 2
(grandtotal)2
Correction factors (CF): CF =
[Link]

= (260.7)2
20
CF = 3398.2
Step:
3
Total sum of squares (Tss)= yij2 −CF

Telegram: AgroMind Website: [Link]

 yij yij2
9.1 82.81
11.2 125.44
13.6 184.96
11.7 136.89
15.6 240.25
16.0 256.00
10.8 116.64
15.8 249.64
11.7 136.89
14.1 198.81
11.2 125.44
12.6 158.76
14.7 216.09
16.4 268.96
11.0 121.00
11.9 141.61
11.2 125.44
11.9 141.61
16.2 262.44
14.1 198.81
yij = 3488.49

= yij2 −CF = 3488.49 – 3398.2

TSS = 90.29
Step: 4
ti2

Treatmental sum of squares (Trss) = −CF

r
ti ti2
Telegram: AgroMind Website: [Link]
 T1 = 63.70 4057.69
T2 = 46.50 2162.25
T3 = 56.50 3192.25
T4 = 48.90 2391.21
T5 = 45.10 2034.01
Sti2 = 13837.41

= ti2 −CF
r

= − 3398.2
= 3459.35 – 3398.2
Trss = 61.15

Step: 5
Error sum of squares (ESS) = TSS – Trss
= 90.29 – 61.15
ESS = 29.14
Step:
6
Construction of ANOVA table:

Trss
TrMss =
trdf

= =15.28

EMss= Erss = 29.14 =1.94 Erdf

15
Source of Degress of SS Mean SS F
variation freedom
Treatmental T -1 = 5 – 4 Trss =
variation TrMss= =15.28
=4 61.15

Telegram: AgroMind Website: [Link]

 Error tr – t (H) = ESS = TrMss
Fcal =
variation EMss= =1.94
5(4) – 5 =15 29.14 =7.87 EMss
Total tr – 1 = 5 * 4 TSS =
-1 = 19 90.29

F = TrMSS = 15.28 = 7.87

ErMSS 1.94 Fcal = 7.87
Ftab = (4,15) at 5% LOS
= 3.06 at 5% LOS
= 4.89 at1% LOS
Step: 7
For finding out the best treatment
CD = SE (Difference) x t tab value of error df at 5% LOS

2EMSS
SE (Diff) =
r

= 2 1.94
4
= 0.98
CD = 0.98 x 2.131
= 2.09

T−1 =T1 = 63.7

=15.9 r 4
−

T2 = 11.62
−

T3 = 14.1
−

T4 = 12.2
−

T5 = 11.2
Arrange into descending order
15.9, 14.1, 12.2, 11.62, 11.2
− − − − −

T1 , T3 , T4 , T2 T5

Telegram: AgroMind Website: [Link]

 Difference:
− −

T1−T3 = 15.9 – 14.1

= 1.8 < C.D
So T1 and T3 are non-significant
Step:
8
Result
− −

T1 and T3 have equally performance

−

T1 is significantly different from T4, T2 and T5.

T3 and T4 have equally performance T3 is
significantly different from T2 and T5 T4,
T2 and T5 have equally performance.
2. The following data give yield (g) per plant of sesame varieties in a pot
culture [Link] the data and draw your conclusion.
Variety Replication
1 2 3 4 5 6 7
CO 1 4.7 5.2 6.4 7.1 6.5 3.5 4.7
TMV 3 3.5 5.3 5.6 6.3 6.3 4.2 4.7
TMV 6 3.2 3.2 4.5 5.0 5.1 3.0 3.5

Step: 1
Correction factor = (Grand Total)2 total no. of observation.

= =490.58

Step: 2
Total sum of squares = yij2 −CF
= 620.8917 – 490.58
= 30.3117

Step: 3
ti2

Telegram: AgroMind Website: [Link]

 Treatmental sum of squares = −CF
r

= − 490.58
= 499.524 – 490.58
= 8.944
Step: 4
Total variance = Known variance + unknown variance
Error = TSS – TrSS
Error = 30.31117-8.944
= 21.3677

Step: 5
Construction of Anova table

Source of Degrees of Sum of Mean sum of F value

variation freedom squares squares
variety (t-1) =2 8.944 4.472
Error (r-1) t(3*6) = 21.3677 1.1187 TRSS
18 = 3.767
ESS
(tr-1)=20 30.3117 1.575

T tabular value
5% = 3.55; 1% = 61
Calculated value < tabular value, It is non significant.
Null Hypothesis
− − −

T1 = T2 = T3

Alternative Hypothesis
T
So accept the null hypothesis
All the varieties have equal performance of yield.

Telegram: AgroMind Website: [Link]

 RANDOMISED BLOCK DESIGN (RBD)

When the experimental units are heterogeneous, CRD is not suitable since CRD
warrants homogeneity in the experimental units. Therefore it is necessary to
make the field into homogeneous by some other design to reduce the
experimental error. RBD is an experimental design in which the heterogeneity
of the experimental field is divided into blocks, such that the plots adjacent in a
block are homogeneous. Here, the randomization is made independently from
block to block. In RBD, the given field is divided into blocks perpendicular to
the fertility gradient. The number of blocks should be equal to the number of
replications of each treatment. Divide each blocks should be equal to the number
of replications of each treatment. Divide each block into a number of plots of
uniform size equal to the number of treatments.

Randomisation
First the experimental units are grouped according to a character which has a
direct influence on the character under study. In each group the number of
experimental units should be equal to or multiples of the number of treatments
under study. Then in each group (block) the number of experimental units are
randomly allotted to treatments using random number tables.
In RBD when each treatment is applied to one experimental unit within each
block then the number of replication of each treatment is the same as the number
of the blocks used. For an RBD with ‘b’ blocks and ‘t’ treatments, the data is
recorded as follows:
Treatment / Block
T1 T2 T3 … T1
Blocks Totals
B1 X11 X12 X13 .. X1t B1
B2 X21 X22 X23 .. X2t B2
.. .. .. .. .. .. ..
.. .. .. .. .. .. ..
Bb Xb1 Xb2 Xb3 .. Xbt Bb

Telegram: AgroMind Website: [Link]

 Treatment T1 T2 T3 .. T1 GT
Total

Statistical Analysis
(i) Grant Total = G.T.
(ii) Treatment totals = T1 + T2 + T3 + ………. + Tt
(iii) Block Totals = B1 + B2 + B3 + …….. + Bb

Correction factor = (G.T)2 Total no. of observations = (G.T)2 b.t

Total sum of squares

(i) T.S.S. = (Sum of squares of all values of X) – C.F.

(ii) Block SS = {( B12+ B22+ …….. + Bb2) / No. of treatments in each
block (t)} – C.F.
(iii) Treatment SS = {( T12 + T22 + T32 + ………. + Tt2) / No. of replications under
each
treatment (b)} –
C.F.
(iv) Error S.S = T.S.S. – (Treatments S.S. + Blocks S.S.)

Analysis of variance table

Source D.F. S.S. M.S. F
Treatments (t-1) Treatment S.S TrS.
Tr.S.S. Tr.M.S.= ------- Tr.M.S.
(t-1) F1 = -------
--

E.M.S.
Blocks (b-1) Block S.S. B.S.S.
B.S.S. B.M.S. = -------- B.M.S.
(b-1) F2 = --------

E.M.S.

Error (b-1) (t-1) Error S.S. E.S.S.

E.S.S. E.M.S. = -------
(t-1)
(b1)
Total bt-1 T.S.S.
Telegram: AgroMind Website: [Link]
 If the F value is found to be significant we will have to calculate the S.E.
of difference of 2 treatment means by using the formula
[Link] = √(2 x M.S.(Error)
No. of blocks
Then C.D. at 5% and 1% level calculated
C.D. at 5% level = The value at error d.f. at 5% x [Link].
C.D. at 1% level = The value at error d.f. at 1% x [Link].
with the values of C.D. the difference of treatment means are compared for their
significance as in C.R.D and the bar Chart can be prepared.

Advantages of R.B.D.
1. This is a simple design with one local control for more efficient
utilization of the available experimental units. R.B.D. takes into
account and eliminates one assignable source of variation among the
experimental units by means of grouping / blocking the more
homogeneous units together. This reduces the experimental error
considerably and the test of significance becomes more efficient in the
sense that even smaller differences among treatments become
detectable. Though the number of degrees of freedom of the Error
mean square in R.B.D is less than that of the Error mean square in
C.R.D., due to removal of d.f. for the blocks the effective reduction in
M.S. (Error) due to blocking is more.
2. Any number of treatments and any number of replications may be
included but each treatment should have the same number of
replications.

Disadvantages
When data from some individual units are lacking, missing plot technique has
to be used. If the missing observations are more, this design is less convenient
than C.R.D. for any statistical analysis.
Telegram: AgroMind Website: [Link]
 Example – 1
Five diets D1, D2, D3, D4 and D5 have to be compared for their efficiencies
measured by the gain in weight when fixed quantities of those different feeds
were fed. The experimental animals available were 20 piglets from four different
breeds in equal numbers. A randomized block design was adopted and the five
piglets of each breed were randomly allotted to the five different feeds. At the
end of the experimental period, the following data were obtained Analyse the
data and draw your conclusions.
Gain in weight by individual pigs (Kgs)
Breed / D1 D2 D3 D4 D5 Total
Diets
B1 7 10 11 12 9 49
B2 9 11 8 15 10 53
B3 6 9 7 16 9 47
B4 7 16 8 13 11 55
Total 29 46 34 56 39 204
Total number of observations = b x t = 4 x 5 = 20
Treatment totals : T1 = 29; T2 = 46; T3 = 34; T4 = 56; T5 = 39
Blocks totals : B1 = 49; B2 = 53; B3 = 47; B4 =55
Grand Total = G.T. = 204
C.F. = (G.T)2 (b.t) = (204)2 (4 x 5) = 41616 20 = 2080.8

T.S.S. = (sum of squares of all the 20 observations) – C.F.

=
(72+92+62+72+102+112+92+162+112+82+72+82+122+152+152+132+02+10
2
+92+112)– C.F. = 2248.0 – 2080.8 = 167.2
Treatments S.S. = [(T12+T22+T32+T42+T52) 4] – C.F.
= [(292+462+342+562+392) 4] – C.F.
= (8770 4) – C.F. = 2192.5 – 2080.8 = 111.7
Block S.S. = [(Br2 + B22 + B32 + B42) 5] – C.F.

Telegram: AgroMind Website: [Link]

 = (10444 5) – C.F. = 2088.8 – 2080.8 = 8.0
Error S.S = T.S.S. – (Tr.S.S. + Block S.S.) = 167.2 – (111.7+8.0)
= 167.9 – 119.7 = 47.5

Analysis of Variance
Source of
D.F. S.S. M.S. F
variation
Treatments 4 111.7 27.93 F1=27.93
3.96 =
7.053**
Blocks 5 3.0 2.67 F2 = 0.67 NS
Error 12 47.5 3.96
Total 19 167.2
** significant at 1% level of significance
NS – not significant
Table value of F at (4,12) d.f. at 5% level = 3.26
Table value of F at (4,12) d.f at 1% level = 5.41
Here the calculated value of F is significant at 1% level, since Fc>Ft. Therefore
the treatment means are significantly different.

To test the difference of treatment means the C.D is to be calculated as follows.

S.E. of differences of any two treatments = √ 2 x E.M.S.

b

= √(2 x 3.96) 4 = √ 1.98 = 1.4

C.D. at 5% level = ‘t’ value at d.f. x [Link]. at 5% = 2.179 x 1.4 = 3.05
C.D. at 1% level = ‘t’ value at 12 d.f. x [Link]. at 1% = 3.044 x 1.4 = 4.3
Treatment means: T1 = 7.25; T2 = 11.50; T3=8.5; T4 = 14.00; T5 = 9.75

Treatments Difference in treatment Significance

Comparisons
means

Telegram: AgroMind Website: [Link]

 T1 – T2 7.25 – 11.50 = -4.25 Significant at 5%
T1 – T3 7.25 – 8.50 = -1.25 Not Significant
T1 – T4 7.25 – 14.00 = -6.25 Significant at 1%
T2 – T3 11.50 – 8.50 = +3.00 Not Significant
T2 – T4 11.50 – 14.00 = -2.50 Not Significant
T2 – T5 11.50 – 9.75 = +1.75 Significance at 1%
T3 – T4 8.50 – 14.00 = -5.50 Significance at 5%
T3 – T5 14.00 – 9.75 = +4.25 Not significant
T3 – T5 8.50 – 9.75 = -1.25 Not significant

Bar Chart:

Treatments: T4 T2 T5 T3 T1

Means: 14 11.50 9.75 8.50

7.25

Conclusion
The feeds under consideration exhibited some gain in weights of the animal. T4
is superior followed by T2. T1 is the inferior feed when compared with other
feeds.

Example-2
A varietal experiment was conducted on black gram on RBD with 6 varieties V1,
V2, V3, V4, V5 and V6 in 4 randomised blocks. The following are the grain yield
in Kg/plot of 20m2 area. Analyse the data and draw your conclusion.
V6 V4 V3 V1 V5 V2 Total
B1
1.62 1.31 3.34 1.65 2.02 2.52 12.46
V2 V4 V3 V5 V1 V6
B2 10.74
2.21 0.91 2.99 1.51 1.51 1.61
V4 V2 V6 V3 V3 V1
B3 9.16
0.81 2.42 0.82 2.19 2.19 1.12

Telegram: AgroMind Website: [Link]

 V1 V4 V3 V5 V5 V6
B4 8.19
1.04 1.12 2.08 1.12 1.12 0.92

Step-1 (G.T)2 (40.55)2 1644.3025

Correction factor = ---------- = ------------ ---------------
n 24 24
C.F = 68.51

Step-2
Find the Treatment and Block totals

68.51=10.83
To estimate the known source of variations namely block sum of square and
variety sum of square, form a 2 way table as follows.
Variety /
V1 V2 V3 V4 V5 V6 Total
Blocks
B1 1.65 2.52 3.34 1.31 2.02 1.62 12.46
B2 1.51 2.21 2.99 0.91 1.51 1.61 10.74
B3 1.12 2.42 2.19 0.81 1.80 0.82 9.16
B4 1.04 1.91 2.08 1.12 1.12 0.92 8.19
Total 5.32 9.06 10.60 4.15 6.45 4.97 40.55

Block sum of square = {(b12+B22+B32+B42) ÷ 6} – CF

= {(155.25 +115.35+83.91+67.08) ÷ 6} – CF
= (306.26 ÷ 4) – 68.51
= 76.565 – 68.51 = 8.055
Step-5
Error sum of square = T.S.S. - (V.S.S. + B.S.S.)
E.S.S. = 10.84 – (1.755 + 8.055) = 1.03
Analysis of variance table
Source df S.S. M.S. F
Block (t-1) = 4-1 = 3 1.755 1.755 ÷ 3 =
0.585
Variety (r-1) = 6-1 = 5 8.055 8.055 ÷5 = 1.61

Telegram: AgroMind Website: [Link]

 Error (t-1) (r-1) 3 1.03 1.03 ÷15 = 1.61÷0.068
x 5 = 15 0.068 =23.66 (5, 15)
Total (rt-1) = 24-1 = 10.84 ---
23
Fc = 23.06 with df(5,15) and Ft at 5%=2.29 & at 1% = 4.56
Fc is highly significant. Since the ‘F’ value is significant it is necessary to
calculate the C.D. for further comparison.

[Link]. =√2 x (EMS÷r) = √ 2 x (0.136 ÷ 4) = 0.18

C.D. = [Link] x t 5% at error df = 0.18 x 2.13 = 0.3834

Summary of Results
Sl. No. Varieties Mean S.E. C.D
1 V1 1.33
2 V2 2.27 0.13 0.38
3 V3 2.65
4 V4 1.04
5 V5 1.79
6 V6 1.24
Conclusion
V3 V2 V5 V1 V6 V4
2.65 2.27 1.79 1.33 1.24 1.04
V3 V2 V5 V1 V6 V4
V3 is significantly superior to V5, V1, V6, V4 but is on par with V2. V2 is
significantly superior to V5, V2, V6, V4, V5 is significantly superior to V1, V6,
V4.

Example-3
Block I Block II Block III Block IV Block V
D17 C12 B12 C15 E23 A30 A28 F64 F75 C14
F70 B6 E26 A26 C16 D20 B9 D23 D20 B07
A20 E28 D10 F56 F56 B10 E33 C14 E30 A23
Analyse the data and interpret your results.

Analysis
Telegram: AgroMind Website: [Link]
 i) Tabulation of the data

The fist step in the analysis of data is to tabulate yield figures according to block
and treatments in the following manner.
Varieties Blocks Treatment
Treatment
Total
I II III IV V means
A 20 26 30 28 23 127 (T1)
B 6 12 10 9 7 44(T2)
C 12 15 16 14 14 71(T3)
D 17 10 20 23 20 90(T4)
E 28 26 23 33 30 140 (T5)
F 70 62 56 64 75 327 (T6)
TOTAL 153 151 155 171 169 799(GT)
(B1) (B2) (B3) (B4) (B5)
ii) Sum of squares of different sources
a) Correction factor (CF) = GT2 / (b x t), where GT is the grand total, ‘b’ is
No. of blocks; ‘t’ is treatments
= 7992/(5x6) = 21280.03

b) Total sum of squares : TSS = sum of squares of all the observations –

C.F. = 202 + 60 + ……….. + 752-C.F = 32013 – 21280.03
= 10732.97

c) Block sum of squares = [(B12+B22+……+B52) ÷ t] – C.F BSS =

[(1662+1512+……+1692) ÷ 6] – C.F
= 21339.5 – 21280.03 = 59.47

d) Treatments S.S = [T12+T22+….+T62) ÷6] – C.F

= [(1272+472+….+3272)] ÷ C.F
Tr.S.S = 31547 – 21280.03 = 10266.97

e) Error S.S = Total S.S. – (BSS – [Link])

ESS = 10732.97 – (10266.97 + 59.47) = 406.53

iii) Analysis of Variance Table

Telegram: AgroMind Website: [Link]

 Now these values will be set down in a table of analysis of variance as given
below.

Analysis of Variance table

Source of D.F. S.S. M.S. Variance F at 5%
variation
Ratio ‘F’
Blocks 4 59.47 14.87
Treatments 5 10266.97 2053.39 101.02* 2.71
Error 20 406.53 20.33
Total 29 10732.97
* Significant at 5% level of significance

It is clear from the table that this observed value of ‘F’ is significant at 5% level
of significance which proves that there are significant differences between the
treatment means. Now, we have to test the significance of the difference between
the individual treatments, and this will be done with the help of C.D. as usual.

iv) Critical difference

S.E. of the difference between any two treatment means is

[Link] = √(2 x EMS) ÷ r = √(2 x 20.91) ÷ 5 = 2.85

Critical difference (C.D.) = [Link]. x t 5% = 2.85 x 2.086 = 5.95
v) Conclusions represented symbolically
The treatment have been compared by setting them in the descending order of
their mean yields in the following manner.
Varieties: F E A D C
B
Mean yields: 65.4 28.0 26.4 18.0 14.2
8.8

Telegram: AgroMind Website: [Link]

 The treatments which do not differ significantly have been underlined by a
bar. The Variety ‘F’ has been found to be best and B happened to be inferior:
Solved Problems
1. An experiment was conducted in RBD to study the comparative performance
of fodder sorghum under rain fed conditions. The rearranged data are given in
the table.

Variety Replication
I II III IV
African Tall 22.9 25.9 39.1 33.9
Co-11 29.5 30.4 35.3 29.6
FS-1 28.8 24.4 32.1 28.6
K-7 47.0 40.9 42.8 32.1
C0-24 28.9 20.4 21.1 31.8

Solution
Mathematical model yijk
= +ti+ rj + eijk
yijk = kth observation on ith treatment effect and jth replication

= Grand mean
ti = ith treatmental effect rj
= jth replication effect
eijk = Error block

Step: 1
(Grandtotal)2
Correction factors (CF): CF =
[Link]
(625.5)2
=

Telegram: AgroMind Website: [Link]

 =
CF = 19562.51

Step: 2
Total sum of squares
TSS = yij2 −CF
= 20514.95 – 19562.51
= 952.44

Step: 3
Treatmental sum of squares

ti2
Trss = −CF
r
ti2 = 80332.17

= −19562.51
Trss = 520.5

Step: 4
Replication sum of squares

rj 2
Repss = −CF
t

rj2 = 98216.57

= −19562.51
Repss = 80.804
Step:
5
Error sum of squares
ESS = TSS – RSS – TrSS
= 952.44 – 80.804 – 520.5
= 351.14
Telegram: AgroMind Website: [Link]
 Step: 6
Construct the Anova Table
Source of Df SS MS F
variation
Treatments t–1 = 5 – 1 = 520.5 130.1
4
Replication r-1 = 4 -1 = 3 80.804 26.9 TrMSS
= 4.4
ErMSS
Error (r-1) (t-1) = 351.14 29.26
12
Total rt – 1 = 19

Fcal = 4.4
Ftab = 3.26 at 5% LOS
TrMSS
TrMSS =
TMdf

= =130.1

RMSS = Re pSS
RMdf

= = 26.9
ESS
EMSS =
EMdf

= = 29.26
Fcal > Ftab
Its is significant. So we reject the null hypothesis

Step: 7
Calculate CD:
CD = Standard Error Difference * t tab value for error df 5% LOS
t tab = 2.179

Telegram: AgroMind Website: [Link]

 SD = 2 EMSS
r
= 2 29.56
4
= 3.82
CD = 3.82 * 2.179
= 8.33

T1
T −1 = = 30.45
r
−

T2 = 31.2
−

T3 = 28.4
−

T4 = 40.7
−

T5 = 25.5

Arrange in descending order

− − − − −

T4 = 40.7; T2 = 31.2; T1 = 30.45; T3 = 28.4; T5 = 25.5

Difference
− −

T4−T2 = 40.7 – 31.2 = 9.5

− −

T4 −T1 = 40.7 – 30.45 = 10.25

− −

T4−T3 = 40.7 – 28.4 = 12.3

− −

T2 −T1 = 31.2 – 30.45 = 0.75

T2 – T3 = 31.2 – 28.4 = 2.8
T2 – T5 = 31.2 – 25.5 = 5.7
− −

T1−T3 = 30.45 – 28.4 = 2.05

− −

T1−T5 = 30.45 – 25.5 = 4.95

− −

T3−T5 = 28.4 – 25.5 = 2.9

Telegram: AgroMind Website: [Link]

 T4 – T5 = 40.7 – 25.5 = 15.2

Step: 8
Interpretation:
1. T4is significantly different from T2, T1, T3, T5
2. T2 , T1, T3, T5 are equally performance.

Result:
The best variety of fodder sorghum is K7.
**********

2. An experiment was conducted to test for the significance of yield differences

of niger varieties at RRS, Raichur by adopting [Link] the data and
draw your conclusions.

Yield(kg/ha)
Variety Replication
1 2 3
N-23 261 201 290
N-35 291 292 265
PCR-232 307 300 285
PCR-140 334 360 330
PCR-290 352 351 348
Phule-1 276 277 285
Phule-4 244 240 264

Step: 1
(Grandtotal)2
Correction factors (CF): CF =
[Link]

= (6153)2
Telegram: AgroMind Website: [Link]
 =
CF = 1802829
Step: 2
Total sum of squares
TSS = yij −CF
= 1837493 – 1802829
= 34664

Step: 3
Treatmental sum of squares

ti2
Trss = −CF
r

= −1802829
= 1831712.3 - 1802829
Step:
4
Replication sum of squares

rj 2
Repss = −CF
t

= −1802829
= 1803022.143 – 1802829
= 193.14286
Step:
5
Error sum of squares
ESS = TSS – RSS – TrSS
= 34664 – 193.14 - 28903
= 5587.26
Step:
6
Telegram: AgroMind Website: [Link]
 Construct the Anova Table
Source of Df SS MS F
variation
Treatment t–1 = 6 28803 4876.5
Block of r-1 = 3 – 1 = 193.14 32.19 TrMSS
replication 2 =
ErMSS
10.33
Error (r-1) (t-1) = 5587.86 465.65
12
Total rt – 1 = 20 34664 1733.2

F value 10.30
Level of significance = 5% 3.00 1% 4.82
Calculated value > Label value The test
is highly significant.

Null hypothesis
H0:T1= T2 = T3 = T4= T5= T6 = T7

Alternative hypothesis
H1:T1≠ T2≠ T3≠ T4≠ T5 ≠ T6≠ T7
Step:
7

2 MSS
Critical difference = t
r

= 2 463.96 2.179
3
CD = 38.31
−

T1 =107.43
−

T2 = 121.14
−

T3 = 127.42
−

Telegram: AgroMind Website: [Link]

 T4 = 146.28
−

T5 = 150.14

Arranging in descending order

− − − − − − −

T5 , T4 , T3 , T2 , T6 , T1 , T7
− −

T5−T4 = 3.86 < CD

− −

T4−T3 =18.86 < CD

− −

T3−T2 = 6.28 < CD

− −

T2−T6 = 1.44 < CD

− −

T6−T1 = 12.27 < CD

− −

T1−T7 = 0.57 < CD

It is non significant to each other so all have equal performance.

Conclusion
All the varieties have equal performance.

LATIN SQUARE DESIGN (LSD)

When the available experimental units are known to be subjected to two major
sources of variation, the experimental units are grouped according to these two
sources of variation so as to have a two-way elimination of variability in the
experimental units. That when the fertility gradient of the experiment field runs
in two directions perpendicularly to each other. The double groupings (double
blocking) are called rows and columns. In each row and each column every
treatment is applied once. This leads ot an arrangement of t-treatments in a
square of t-rows and t-columns such that every treatment is allotted once in every
row and every column. Such design is called Latin square design.
Telegram: AgroMind Website: [Link]
 The number of experimental units for a LSD (of t x t = t2) with 5 treatments, the
number of experimental units is 25; with 6 it is 36; with 10 it is 100 and so on.
When more experimental units are required for the experiment, then these units
are likely to be very heterogeneous and also the allotment of the treatment is
more troublesome. For treatments of less than 5, the error df will be less. So, in
general, LSD is adopted for treatments from 5 to 12.

Layout
The allotment of the different treatments to the different units without repetition
of any treatment in any row and in any column should be done. This is done as
follows:
1. Get a random Latin square of the required size t x t, where, t = No. of
treatments from the statistical table. The choice of the random Latin
Square is decided by the random number which is less than the total
number of given square in the table.
2. Number the rows and get a random arrangement of these rows and change
the position of the rows of the selected square according to the random
order got for the rows.
3. Keeping the first column unchanged, change the order of the other
columns with the random numbers 1 to (t-1)

The latin square obtained after the above mentioned operations is a Latin
square in which, the treatments are allotted randomly to the experimental units.
These treatments are applied on the experimental plots which have already been
grouped according to 2-way variations among the units.
The treatments are denoted by the letters A, B, C, D, E. etc. In 5 x 5 Latin
Square the following table of data may be obtained.
Rows Columns
I II III IV V

Telegram: AgroMind Website: [Link]

 I IC B A B D
II IIB C D E A
III IIIE A C D B
IV IVD B E A C
V VA D B C E

Statistical Analysis:
Step-1
a) Calculation of treatment total: TA, TB, TC, TD, TE
b) Calculation of column totals : C1, C2, C3, C4, C5
c) Calculation of Row Totals : R1, R2, R3, R4, R5
d) Calculation of Grand Total (G.T): Sum of treatment of column or row
totals.
e) Calculation of C.F. = (G.T.)2 ÷ Total number of treatment = (G.T)2 ÷ t
xt
Step-2
Total S.S. = (Sum of squares of all ‘t x t’ observations – C.F.)
Tr. S.S. = [(T12+T22+T32+T42+t52) ÷ 5] – C.F
Column SS = [(C12+C22+C32+C42+C52) ÷ 5] – C.F. Row
SS = [(R12+R22+R32+R42+R52) ÷ 5] – C.F.
Error S.S. = T.S.S. - (Treatment S.S + Row S.S. + Column S.S.)

Analysis of variance table

Source D.F. S.S. M.S. F

Treatments (t-1) Tr.S.S Tr.S.S ÷ (t-1) Tr.M.S. ÷ E.M.S.
Columns (t-1) Col.S.S Col.S.S. ÷ (t-1)
Row (t-1) Row S.S Row S.S ÷ (t-1)

Telegram: AgroMind Website: [Link]

 Error (t-1) (t-2) E.S.S. E.S.S. ÷ (t-1) (t-
2)
Total (t2-1) T.S.S

Compare the calculated value of F with table values. If it is not significant,

the mean differences are not different from each other.

If the F value is significant, to test the differences of treatment means C.D is

to be calculated at 5% and 1% levels.

S.E. of difference of any two treatment means ([Link].) = √(2 x E.M.S.)

t
CD = [Link]. x t value with error df at 5% level.

With these [Link] the differences between the means of treatments are tested for
significance.

Telegram: AgroMind Website: [Link]

 Advantages of LS.D.
This design through it double blocking of the experimental units in capable of
reducing the experimental error when the experimental units are subject to two
major sources of variation. This reduction in experimental error compensates
the loss in degree of freedom if the double blocking is effective.

Disadvantages
It requires many experimental units as the square of the number of treatments
and these experimental units should be such that they are classifiable equally to
the row blocking. As the number of treatments increases the design of the
experiment becomes more difficult.
Statistical analysis of the data under L.S.D. will become difficult if the data on
some experimental units under the experiment are not available.

Example – 1.
An experiment to compare 5 feeds on kinds was conducted on 25 kinds which
were first groped into 5 different age groups and 5 different weight groups. The
design adopted was latin square. The following data on gain in weight during
the experimental period were recorded. Analyse the data and draw the
conclusions.
Gain in Weight (Kgs)
Age group Weight in group
1 2 3 4 5 Total
1 5(F2) 3 (F1) 6(F5) 3(F4) 2(F3) 19
2 5(F5) 3 (F3) 6(F4) 3(F2) 2(F1) 16
3 5(F1) 3 (F5) 6(F2) 3(F3) 2(F4) 19
4 5(F4) 3 (F2) 6(F3) 3(F1) 2(F5) 17
5 5(F3) 3 (F4) 6(F1) 3(F5) 2(F2) 15
Total 18 21 16 15 16 86

Telegram: AgroMind Website: [Link]

 Analysis:
(i) Grand Total: GT = Sums of all treatments = 86
(ii) C.F. = 862 ÷ 25 = 295.84
(iii) TSS = (52+32+….+52+42) – CF = 368.0 – 295.84=72.16
(iv) Row S.S (Age group) = [(192+162+192+172=152) ÷ 5] – C.F.
= 368.0 – 295.84 = 2.56
(v) Col.S.S. (weight) = [(182+212+162+152+162) ÷ 5] – C.F.
= 300.4 – 295.8 = 4.56 (vi)
Treatment totals
F1 = 11; F2 = 24; F3 = 10; F4=12; F5 = 29
(vii) Tr.S.S. = [(112+242+102+122+292) ÷ 5 ] – C.F.
= 356.4 – 295.8 = 60.56
(viii) E.S.S. = 72.2 – (2.56 +4.56 +60.56) = 4.48 Analysis of
variance
Source D.F. S.S. M.S. Fc
Age group 4 2.56 0.64 F1 = 1.71
Weight group 4 4.56 1.14 F2 = 3.05
Feeds 4 60.56 15.14 F3 = 40.55
Error 12 4.48 0.37
Total 24 72.16
Table value of F at 4, 12, d.f. = 3.26 at 5% level
Table value of Fat 4,12 d.f. = 5.41 at 1% level
There is no significant differences among the age groups as well as the weight
groups. There exists highly significant variation among the different treatments
(feeds). For further comparison, the critical difference is to be calculated as
follows.

[Link]. = √(2x0.37) ÷ 5 = 0.3847

C.D = t value at 12. d.f x [Link]. = 2.179 x 0.3847 = 0.84
Treatment means in ascending order

F3 F1 F4 F2 F5
Telegram: AgroMind Website: [Link]
 2.0 2.2 2.4 4.8 5.8

Conclusion
F1, F3, F4 have equal effects and F2 and F5 are different. F5 is superior than all
the other followed by F2.
LATIN SQUARE DESIGN
Example II.
The following table gives the yield data of 6 varieties conducted in LSD.
Analysis the data and give the conclusion.
Column
Row Total
C1 C2 C3 C4 C5 C6
r1 D3 D2 D1 D6 D5 D4
68 48 20 56 92 98 382
r2 D4 D3 D2 D1 D6 D5
102 72 52 24 58 58 396
r3 D5 D4 D3 D2 D1 D6
84 114 66 46 26 60 396
r4 D6 D5 D4 D3 D2 D1
56 86 108 67 46 28 399
r5 D1 D6 D5 D4 D3 D2
24 54 87 112 69 48 394
r6 D2 D1 D6 D5 D4 D3
46 26 50 88 100 70 380
Total 380 400 383 393 399 392 2347

Analysis
(i) Correction factor = [(G.T.)2] ÷ 36 = [(2347)2] ÷ 36 = 5580409÷ 36 =
153011.36
(ii) T.S.S. = ∑X2-C.F = 178299 – 153011.36 = 25287.64
(iii) R.S.S. = [(r12 + r22 + r32 + r42 + r52 + r62) ÷6] – C.F
= [(3822+3962+3962+3992+3942+3802) ÷6] – 153011.36
= (918393 ÷6) – 153011.36 = 153065.5 – 153011.36 = 54.14
(iv) C.S.S. = [(C12+C22+C32+C42+C52+C62) ÷ 6] – C.F

Telegram: AgroMind Website: [Link]

 = [(3802+4002+3932+3992+3922) ÷ 6] – C.F
= (91804.3÷6) – 153011.36 = 153067.16 – 153011.36 = 55.80
Treatment Totals
D1 = 148; D2=294; D3=412; D4 = 634; D5=525; D6 = 334

(v) Tr.S.S. = [(D12+D22+D32+D42+D52+D62) ÷6] – C.F

= (1067221 ÷ 6) – 153011.36 = 24858.80
C.D = 4.81
The treatment are significantly different from each other.
Er.S.S. = T.S.S. – (R.S.S. + C.S.S + Tr.S.S)
= 25287.64 – (54.14 +55.8 + 2485.8)
= 318.9 Analysis of variance table
Degree Total sum F-table
Sources of of Mean sum of
of F-ratio 5%
varieties freedom squares
squares 1%
Rows 6-1 = 5 55.14 = 54.14÷5= F1=S1/S2 = 2.71 4.1
10.828 10.828/15.945
= 0.679
Columns 6-1 = 5 55.80 = 55.80 ÷ 5 F2 = 11.16 /
= 11.6 15.945 =
0.699
Treatment 6-1 = 5 24858.80 = 24858.80 F3 = 311.80
÷5
= 4971.76
Error 35-15 = 318.9 = 318.9 ÷ 20
20 = 15.945
Total 36-1 = 25287.64 - -
35

Results and Conclusion:

The calculated F3 value was greater than the table F value at 1% level with
(5,20) degrees of freedom. Therefore the F ratio is highly significant. For further
comparison the critical difference is to be calculated.
[Link]. = √(2 x E.M.S) ÷ r =√(2 x 15.995) ÷ 6 = 2.30
C.D. = [Link]. x t 5% = 2.30 x 2.086 = 4.809.
Telegram: AgroMind Website: [Link]
 Conclusion
The means of treatments are presented in the ascending order of magnitude.
Treatments: D1 D2 D6 D3 D5
D4
Means: 24.66 49.00 55.66 68.66 87.50
105.66
Bar Chart:
Treatments: D4 D5 D3 D6 D2 D1
Means: 105.7 87.5 68.7 55.7 49.0
24.7
C.D: 4.809
The treatments are different. That is the yields are different. Variety D4 is
superior followed by D5.
Solved Problems
1. The plant and yields in kg/plot of a 5x5 Latin square experiment on the
wheat crop observations are given below. Analyze the data and draw
the conclusions.

B 15 A E 17 D C
8 29 17
A D C E 16 B 13
9 21 19
C B 12 D A E 17
18 23 8
E 18 C A B 15 D
16 10 23
D E 15 B 13 C A
22 18 10
Solution

Telegram: AgroMind Website: [Link]

 Mathematical mode yijkl
= +ti+ rj + ck + eijkl
yijkl = ith observation on ith treatment effect and jth row effect and kth column
effect
= Grand mean
ti = ith treatmental effect
rj = jth row effect ck =
kth column effect
eijkl = Error term

Step: 1
(Grandtotal)2
Correction factors (CF): CF =
[Link]

= (86 + 78+82 + 78+ 78)2

25

= (402)2

=
CF = 6464.16

Step: 2
TSS = yijkl2 −CF
= 7102 – 6464.161
= 637.84

Step: 3
ti2
Trss =

−CF r

Telegram: AgroMind Website: [Link]

 = − 6464.16
= 7041.2 – 6464.16
Trss = 577.04

Step: 4
rj 2
Row SS = −CF
t

= − 6464.16
= 6474.6 – 6464.16
Row SS = 10.24

Step: 5
ck 2
Col SS = −CF
t

= − 6464.16
= 6485.6 – 6464.16
Col SS = 21.44
Step:
6
ESS = TSS – TrSS – RSS – Col SS
= 637.84 – 577.04 – 10.24 – 21.44
ESS = 29.12

Step: 7
Construct the Anova Table
Source of Df SS MSS F
variation
Treatmental t–1 = 5 – 1 = 4 577.04 144.26
variation
Telegram: AgroMind Website: [Link]
 Row variation t-1 = 5 -1 = 4 10.24 2.56
Column t-1 = 5 -1 = 4 21.44 5.36 59.61
variation
Error variation (t-1) (t-2) 29.12 2.42
(5-1)(5-2)
4*3=12
Total t2-1=25-1=24

F cal value = 59.61

F tab value = 3.26 F cal > F tab =
so, it is significant We reject the
null hypothesis.
Step:
8
CD = SE(df) x t tab value of 5 % LOS
2 EMSS
SE =
t
2 2.42
= 2.179
5
= 0.98 x 2.179
CD = 2.14

Mean:
−

A == 9
−

B ==13.6
−

C ==17.6
−

D == 23.6
−

E ==16.6

Arranging in descending order

− − − − −

Telegram: AgroMind Website: [Link]

 D= 23.6, C=17.6 , E=16.6 , B=13.6, A= 9

Interpretation
(i) C and E have equal performance
(ii) Plant C is significantly different from the plants E,B,A.
************

[Link] varieties of rice IR

20(V1),ADT36(V2),ASD18(V3),ADT39(V4)andCO47(V5) were
tested in a5X5 [Link] following layout gives average yield of 5 plants
selected at random in a [Link] the data and draw conclusions.

9.3(V2) 10.0(V4) 8.8(V5) 14.4(V3) 12.0(V1)

9.2(V4) 9.5(V2) 14.9(V3) 9.2(V1) 9.3(V5)
7.7(V1) 9.2(V5) 9.7(V4) 9.4(V2) 11.5(V3)
8.3(V3) 13.3(V1) 8.0(V2) 9.5(V5) 8.4(V4)
9.6(V5) 9.5(V3) 10.2(V1) 9.2(V4) 6.8(V2)

Step: 1
(Grandtotal)2
Correction factors (CF): CF =
[Link]

= (246.9)2

=
CF = 2438.38
Step: 2
To find the to total sum of squares (TSS)
TSS = yijkl2 −CF
= 91.49
Step: 3

Telegram: AgroMind Website: [Link]

 To find the treatmental sum of squares (TrSS)

ti2
Trss = −CF
r
Trss = 30.4
Step: 4
To find the column sum of squares (CSS)

cj 2
CSS = −CF
t
CSS = 8.92

Step: 5
To find the row sum of squares (RSS)

rk 2
RSS = −CF
t
= 11.47
Step:
6
Error sum of squares
ESS = TSS – TrSS – CSS – RSS
= 91.49 – 30.4 – 8.92 – 11.47
ESS = 40.7

Step: 7
Construction of the Anova table
Sources of DF SS Mean sum of F value
variation squares
Treatment (t-1) = 4 30.4 7.6
Column (t-1) = 4 8.92 2.23
Row (t-1) = 4 11.47 2.86 2.24
Error (t-1)(t-2) = 12 40.7 3.39
Total (t2-1) = 24 91.49 3.81

Telegram: AgroMind Website: [Link]

 Calculated value = 2.24
Tabular value
5% - 3.26; 1% - 5.4
Calculated value < tabular value
It is non significant
Accept the null hypothesis

Conclusion
All the varieties have equal performance.

FACTORIAL EXPERIMENTS
Frequently it can be seen that several factors may affect the treatments and
we wish to estimate the effect of each of the factors and how the effects of one
factor varies over the levels of the other factors. This situation can be handled
with factorial experiments.
The treatments will be the combination of two or more factors and this
combination should be considered as treatment. This will facilitates the
estimation of the different effects (i.e) main and interaction effects.
Example:
Suppose if we have 2 varieties V1 and V2 and three levels of nitrogen N1,
N2 and N3, then the treatments will be 6 combination (V1 N1, V1 N2, V1 N3, V2
N1, V2 N2, V2 N3). These 6 treatments will give 5 degrees of freedom that d.f.
again is to be broken and divided in to one d.f. due to main effect of varieties and
2 d.f. for the main effect on N levels and 2 d.f. for V x N interaction effect.
Therefore in the factorial experiments the possibilities of estimating main effects
and interaction effect will be there.
The package of practices for different crops are worked out only through
factorial experiments. In the factorial experiments with reduced cost, labour and
area. We will be able to get the main and interaction effect. In the factorial

Telegram: AgroMind Website: [Link]

 experiment, the concept of hidden replications will operate and the precision also
will be more.
Block 1 Block 2 Block 3 Block 4
V1F0 52 V1F1 51 V2F2 53 V2F0 45
V2F0 49 V2F2 47 V2F1 47 V1F2 49
V1F1 52 V2F1 48 V1F1 51 V2F2 43
V1F0 48 V2F0 49 V2F0 48 V1F0 48
V2F2 48 V1F0 51 V1F2 47 V1F1 47
V2F1 54 V1F2 54 V1F0 51 V2F1 43

Analyse the data and compare the various effects.

Step-1:
Rewrite the above data systematically in a two way table.
‘Bloc x Treatment’ Table

Treatments Blocks Total

1 2 3 4
V1F0 52 51 51 48 202
V1F1 52 51 51 47 201
V1F2 54 54 47 49 204
V2F0 49 49 48 45 191
V2F1 48 48 47 43 186
V2F2 48 47 53 43 191
Total 303 300 297 275 1175(GT)

Step-2:
Rewrite the treatment combination totals in two way table as given below
F0 F1 F2 Total
V1 202 201 204 607

Telegram: AgroMind Website: [Link]

 V2 191 186 191 568
393 (F0) 387 (F1) 395 (F2) 1175
Total
GT

Step-3:
Find the row totals and column totals of the above table. Denote the row
totals as V1 and V2, denote the column totals as F0, F1 and F2. Check up that
sum of column totals is equal to sum of row totals. Check also that these are
equal to grand total (GT).

Step-4:
Correlation factor (CF) = GT2 ÷ (r x f x v); where r is no. of blocks, f is [Link]
fertilizer
level and v is [Link] varieties.
= 11752 F ÷ (4 x 3 x 2) = 1380625 ÷ 24 = 57526.04
VSS = [(V12 + V22) ÷ (r x f)] – CF
= [(6072 + 5682) ÷ (4 x 3)] – 57526.04 = 63.38
VSS denotes variety SS, which is the SS for the main effect of variety.
Note that 12 is the division because V1, V2 are sum of 12 observations.

Step-5:
FSS = [(f02 + f12 +f22) ÷ (r x v)] – CF
= [(3932 + 3872 + 3952) ÷ (4 x 2)] – 57526.04 = 4.33
FSS denotes fertilizer SS, Which is the SS of the main effect of fertilizer.
(Note that 8 is the divisor because f0, f1 and f2 are sum of 8 observations)

Step-6:
Tr.S.S = {[(202)2 + …….. + (191)2] ÷ 4 } – 57526.04
= 57651.75 – 57526.04 = 68.71
(V x F) SS = T.S.S – (V.S.S + F.S.S) = 68.71 – (63.38 + 4.33) = 0.99
B.S.S = [(3032 + 3002 + …… + 2752) ÷ 6] – CF = 81.13

Telegram: AgroMind Website: [Link]

 T.S.S = (522 + 512 + …. + 432) – C.F. = 208.96

Step-7:
Step up the A.V. Table and carryout ‘F’ test to test main effects and interaction
effects.
Source Df SS MS Fc Ft
5% 1%
Blocks 4-1=3 81.13 27.04
Treatment 6-1=5 68.71 13.74 3.49* 2.9 4.56
V 2-1=1 63.38 63.38 16.09** 4.54 8.68
F 3-1=2 4.33 2.17 0.55 NS 3.68 6.36
VxF 5-(2+1) = 2 6.56 0.50
Error 4-1 x 6-1 = 15 59.12 3.94
Total 24- 1 = 23 208.96

* Significant at 5% level of significance

** Significant at 1% level of significance.
NS – Not Significant

Step-8: Bar Chart

i) The ‘F’ test for ‘r’ has indicated significant differences among varieties. The
C.D value is calculated by the formula.

C.D = [√(2 x EMS) ÷ (r x f) ] x t12 (5%) = 0.154 x 2.131 = 1.76

Bar chart:
V1 V2
51.08 46.83
ii) The ‘F’ test for fertilizer has indicted significant difference among the
fertilizer level. The C.D value is calculated by the formula. CD = = [√(2 x
EMS) ÷ (r x v) ] x t15 (5%) = 2.28

Telegram: AgroMind Website: [Link]

 Bar Chart: F2 F0 F1

iii) The ‘F’ test for the interaction effect reveals the existence of interaction of
fertilizer with varieties. Construct the bar chart as detailed below:

CD = = [√(2 x EMS) ÷ r ] x t15 (5%) = 0.266 x 2.131 = 0.57

Mean table
V1 V2
F0 50.50 47.75
F1 50.25 46.50
F2 52.50 47.75

Bar chart : Comparison of fertilizer at each variety

V1 : F2 F0 F1
V2 : F0 F1 F2
Comparison of variety at each fertilizer
F0 : V1 V2
F1 : V1 V2
F2 : V1 V2

Example-2:
With the objective of studying the effect of N application on two varieties of
sugarcane in 2 x 2 FRBD was conducted. The following are details of
experimental design in FRBD. Treatment s all the possible combinations of
varieties V1 and V2 and N level n1 and n2. Number of block is 6. Gross plot size
7 x 6m2 and net plot size 6 x 5 m2. Yield is (Sugarcane) given in kg / plot.
Analyse the data and draw the conclusion

Block /
R1 R2 R3 R4 R5 R6
Treat

Telegram: AgroMind Website: [Link]

 V1n1 186 205 225 240 208 238
V2n1 275 311 308 273 301 301
V1n2 491 362 413 558 502 532
V2n2 420 338 389 351 393 423

Solution:
Step-1 : Two way table for treatment and Block
Block /
R1 R2 R3 R4 R5 R6 Total
Treat
V1n1 186 205 225 240 208 238 1302
V2n1 275 311 308 273 301 301 1769
V1n2 491 362 413 558 502 532 2858
Vn 420 338 389 351 393 423 2314
Total 1372 1216 1335 1422 1404 1494 8243

Step-2: Two way table for V x N (V x N Table)

N/ V n1 n2 Total V mean
V1 1302 2858 4160 346.67
V2 1769 2314 4083 340.25
Total 3071 5172 8243

Mean 3071 ÷ 12 5172 ÷ 12

= 255.92 = 431
Step-3:
i) Correction factor CF = (8243)2 ÷ (4 x 6) = 2831127 ii) TSS = ∑
X2 – CF = (1862 + 2752 + … + 4232) – CF = 264197.96 iii) BSS =
(11368601 ÷ 4) – 2831127 = 11023.21 iv) VSS = (3396489 ÷ 12) –
2831127 = 247
v) NSS = (36180625÷ 12) – 2831127 = 183925.08 vi) In V x
N = 286760.5 – (247 + 183925) = 42588.42 vii) ErSS =
Telegram: AgroMind Website: [Link]
 264197.96 – (11023.21 + 226760.5) = 26411.32 viii) V x N
table SS = (18347325 ÷ 6) – 2831127 = 226760.5
Step-4: AV Table
Source df SS MS Fc
Block 5 11023.21 2204.64
Treatment 3 226760.50 75586.83 42.9**
Variety 1 247.00 247.00 0.1403NS
Nitrogen 1 183925.08 183925.08 104.46**
NxV 1 42588.42 42588.42 24.19**
Er SS 15 26414.25 26414.25
Total 23 264197.96

Step-5: Construction Bar chart

CD(N) = [√(2 x 1760.95) ÷ 12 ] x 2.131 = 36.50
CD (V x N) = [√(2 x 1760.95) ÷ 6] x 2.131 = 51.63 Mean
Table:
Interaction effect:
VxN V1n1 V1n1 V2n2 V2n2
Particulars
Total 1302 2858 1769 2314
Mean 217 476.33 294.83 38.67

1. Comparison of Nitrogen effect:

n2 = 431 C.D = 36.50
n1 = 255.92 C.D = 175.08
Bar Chart:
n2, n1

2. Interaction effect : CD = 5163

476.33 385.67 294.83 217

Telegram: AgroMind Website: [Link]

 V1n1 V2n2 V2n1 V1n1
Conclusion:
(i) Two varieties are equal in effect. They are statistically non significant (ii)
In two doses of Nitrogen, n2 is superior than n1.
(iii) In interaction, V1n1 is superior than others V2n2 is superior than V2n1 and
V1n1.

SPLIT – PLOT DESIGN

It is defined as the design in which (1) the blocks will be divided into larger plots
to which whole plot treatments will be randomized and (2) each larger plot will
be further divided into smaller plots equal to the number of subplots to which
sub-plots treatments will be allotted at random such that (3) the randomization
of sub-plot treatments are independent of the randomization of main-plot
treatments.
In general the treatments require larger areas will be allotted to the main plots
and other set of treatments will be allotted to the sub-plots. Treatments requiring
larger areas like irrigation will be known as main plot treatments and the other
factors like variety will be called sub-plot treatments. For example, let there be
four irrigation levels, I1, I2, I3 and Ir. Let there be two varieties V 1 and V2.
Therefore I1, I2, I3 and I4 will be the main plot treatments and V1 The
randomization of main plot will be as follows:
Block I Block II Block III Block IV Block V
I2 I2 I3 I3 I3
I3 I4 I4 I1 I1
I1 I1 I2 I2 I2
I4 I3 I4 I4 I4

With regard to the randomization of sub-plot treatments, divide each larger plot
in a block into smaller plots equals to the number of sub-plot treatments.
Division of the larger plots equal into smaller plots will be perpendicular to the

Telegram: AgroMind Website: [Link]

 division of blocks into larger plots. Here each block was divided horizontally
into 4 larger plots. Each larger plot will be divided vertically into two smaller
plots. The sub-plots are randomized as follows:

Block 1 Block 2 Block 3 Block 4 Block 5

V2I2 V1I2
V1I3 V2I3
V2I1 V1I1
V1I4 V2I4
Here the accuracy, precision and validity will be many times more for the sub-
plot treatment. Because, the sub-plot treatments are allotted to smaller plots in
comparison to plot size of main plot treatments. Small plots will be more
homogenous than bigger plots in comparison to plot size of main plot treatments.
Small plots will be more homogenous than bigger plots; therefore sub-plot
treatments are estimated more accurately than main plot treatments. The plot
size of sub-plot treatment is same as the plot size of the combination of main plot
treatments with sub-plot treatments. Hence validity, precision and accuracy of
estimates of sub-plot treatments and interaction effects are same. If situation
warrants, the split plot design could be extended to a further stage by splitting
the sub-plots known as second order sub-plots to be assigned at random to a
further set of treatments. The analysis is almost similar to the split-plot with
additional estimation of the error mean square for the second order of sub-plots.
The sub-plot effects would be estimated with the greatest precision as a result of
most effective local control, since the second order subplots would be compared
under more homogenous condition than the first order of sub-plots and the main
plots.
Analysis of Variance Table
Source D.F S.S M.S F-value
Replication r-1

Telegram: AgroMind Website: [Link]

 Main plot t-1
Error I (r-1) (t-1)
Sub plot s-1
treatment
Main Plot x sub (t-1) (s-1)
Error II T(r-1) (s-
1)
Total trs – 1

Example-1:
Given below the plan and yields of a cultural trial on sugarcane in which a split-
plot lay-out was adopted. The treatments consisted of 4 dates of planting, namely
planting in October (S1), November (s2), February (S3) and March (S4) and 3
methods of planting, namely, planting in trenches and turning into ridges (M 1),
planting on flat and turning into ridges (M2) and planting on flat and leaving as
such (M3). The 4 levels of the first factor were allotted to 4 main plots of each
block at random and the 3 levels of the second factor were randomly assigned to
the 3 sub-plots into which each main plot was divided.
Table – Plan and yields of a split – plot experiment on Sugarcane
Replication S2 M3 7.7 S4 M1 1.3 S3 M1 0.3 S1 M1 5.5
I
M1 5.2 M2 1.0 M2 2.1 M3 2.6
M2 2.6 M3 1.3 M3 0.4 M2 2.1
Replication S3 M2 4.2 S1 M3 4.4 S4 M1 6.8 S2 M2 4.4
II
M1 1.3 M2 4.0 M2 6.9 M1 3.5
M3 1.4 M1 3.7 M3 5.3 M3 1.2
Replication S2 M2 8.3 S1 M3 3.1 S3 M3 7.3 S4 M2 0.
III
M1 5.4 M2 2.9 M1 7.8 M1 0.9
M3 8.3 M1 4.8 M2 6.4 M3 1.4
Replication S2 M3 4.2 S4 M3 5.3 S1 M1 16.0 S3 M1 4.9
IV
Telegram: AgroMind Website: [Link]
 M1 7.3 M1 6.3 M3 10.0 M2 4.0
M2 7.5 M2 4.3 M2 9.0 M3 5.1
Replication V S4 M2 2.6 S2 M3 1.5 S3 M1 8.5 S1 M2 4.9

M3 4.3 M1 2.0 M2 6.1 M3 2.2

M1 4.2 M2 2.1 M3 5.9 M1 3.3
Replication S1 M3 22.1 S3 M2 2.5 S4 M1 6.3 S2 M3 1.3
VI
M1 19.0 M1 2.0 M3 3.8 M2 0.9
M2 14.9 M3 2.9 M2 1.6 M1 2.3

Telegram: AgroMind Website: [Link]

 Main Plot Treatments Sub-Plot Treatments
Dates of Planting Methods of Planting
S1 October M1 – Planting in trenches and turning into ridges
S2 November M2 – Planting in flat and turning into ridges
S3 February M3 – Planting in flat and leaving as such
S4 March
(Yield in 50 kg units per sub-plot; sub-plot size 36 x 2.5 m)

Two-way table of blocks and dates of planting in split –plot design:

(Each figure is a total of 3 sub-Plots)
Blocks S1 S2 S3 S4 Total
I 10.2 15.5 3.4 3.6 32.7
II 12.1 9.1 6.9 19.0 47.1
III 10.8 22.0 21.5 2.6 56.9
IV 35.9 19.0 14.0 15.9 84.8
V 10.4 5.6 20.5 11.1 47.6
VI 56.0 4.5 7.4 11.7 79.6
Total 135.4 75.7 73.7 63.9 348.7

C.F. = (348.7)2 ÷ (6x4x3) = 159121591.69 ÷ 72 = 1688.77

Total S.S. (23 df) = [(10.22 + …..+ 11.72) ÷ 3] – 1688.77 = 1046.49
S.S. for planting dates ( 3 df) = [(135.42 +…….+ 63.92) ÷ 18] – 1688.77 =
176.71
S.S. for Blocks (5df) = [(32.72+47.12+…..+79.62) ÷ 12] – 1688.77 = 171.09
S.S for Main plot Error, Error {a} (23-5-3=15 df) = 1046.49 – 176.71 – 171.09
= 698.69
Analysis of Variance of Main plot yields
Source of D.F. S.S. M.S. F
Variation
Blocks 5 171.09 34.22 0.73 NS
Sowing dates 3 176.71 58.90 1.26 NS
Error (a) 15 698.69 46.58
Total 23 1046.49

Telegram: AgroMind Website: [Link]

 Table – Two – way table of treatment totals for split – plot experiment
Methods of Dates of planting Total
planting S1 S1 S1 S1
M1 56.3 25.7 25.4 25.8 133.2
M2 41.9 25.8 25.3 16.7 109.7
M3 37.2 24.2 23.0 21.4 105.8
Total 135.4 75.7 73.7 63.9 348.7

We have
S.S for methods of planting (2df) = [(133.22 +…+ 105.82) ÷ 24] – CF =
18.31 and
S.S for interaction SM (6 d.f.) = Table III SS – (MPSS +SPSS)
Also from the entire table of all the 72 plot yields,
Total SS (71 d.f.) = 2873.75 – CF
Tab II SS = [(11437.65 ÷ 6) – CF] = 1906.28 – 1688.77 = 217.51
• The sum of squares for sub-plot error is obtained by subtracting from the
total sum of squares the sums of squares for blocks and all the treatment
effects as also the sum of squares for main plot error.
• Sum of square for sub-plot error, or Error (b) = T.S.S. – B.S.S. – Tr. S.S.
– Main lot error
= 1184.98 – 338.54 – 395.15 – 19.72 – 19.51 – 321.04 =
97.62
We have then the following table of complete Analysis of Variance.

Source of variation D.F S.S M.S F-value

Blocks (r-1) = 5 171.09 34.22 0.7376
Planting dates (S) (t-1) = 3 176.71 58.90 1.264
Error (I) (r-1) (t-1) = 15 698.69 46.58 -
Methods of Planting (s-1) = 2 8.31 9.16 3.75
(M)
Interaction: Planting (t-1) (s-1) = 6 22.49 3.74 1.53
dates x methods (SM)

Telegram: AgroMind Website: [Link]

 Error II T (r-1) (s-1) = 40 97.69 2.44

Total (rst – 1) = 71 1184.98

The mean square between planting dates is compared against mean square for
the main plot error (Error (a) for significance and those for the main response to
methods of planting and the interaction, planting dates x methods of planting,
being effects belonging to sub-plots, are tested against the mean square for the
sub-plot error (Error (b)). It is seen that both the main effects are significant, the
differences between planting dates being significant at 1 per cent level. We
notice too from two way table that the differences between methods of planting
are of a much smaller order than those between dates of planting.
Comparison
Calculation of SE and CD
(i) Standard error for the difference between two main plot treatments.
[Link].

= √2.378 = 1.542
CD = [Link]. x t 5% = 1.542 x 2.131 = 3.286
(ii) Standard error for the difference between two sub-plot treatments

[Link]. = √2EMSII ÷ (r x t) = √(2x2.44) ÷(6 x 4) = √4.88 ÷ 24 = 0.4509

C.D. = [Link]. x t value at 5% = 0.4509 x 2.0221 = 0.9113 Example-2:
In an experiment to study the yield of a crop, the split plot design was adopted.
I1, I2, I3 and I4 are irrigation levels allotted to main plots. V1 and V2 are varieties
allotted to sub plots. The data as per the random layout of the experiment is
given below.
Block I Block II Block III Block IV Block V
I2V2 I2V1 I2V2 I2V2 I3V1 I3V2 I3V1 I3V2 I3V2 I2V1
38.4 41.2 43.5 39.3 32.5 30.0 31.8 25.6 28.5 31.5
I3V1 I3V2 I4V1 I4V2 I1V2 I1V1 I1V2 I1V1 I1V1 I1V2
30.6 26.9 38.8 31.4 31.5 32.2 30.3 34.4 34.5 30.6
I1V2 I1V1 I1V1 I1V2 I2V2 I2V2 I2V2 I2V1 I2V2 I2V1
31.3 35.2 32.3 29.4 42.4 40.6 39.6 44.6 39.5 42.9
Telegram: AgroMind Website: [Link]
 I4V1 I4V2 I3V2 I3V1 I4V1 I4V2 I4V1 I4V2 I4V1 I4V2
39.5 30.2 28.5 31.2 37.9 29.9 38.8 30.4 38.9 30.5
Analyse the data, compare the main-plot treatment, sub-plot treatments and the
interaction effects.
Step-1:
Rearrange the observations in a systematic table. There will three-way
classification table in this case. This table is called the master table.
Block 1 Block 2 Block 3 Block 4 Block 5 Total
I1 V1 35.2 32.3 32.2 34.4 34.5 168.6
V2 31.3 29.4 31.5 30.3 30.6 153.1

I2 V1 41.2 43.5 42.4 44.6 42.9 214.6

V2 38.4 39.3 40.6 39.6 39.5 197.4

I3 V1 30.6 31.2 32.5 31.8 31.5 157.6

V2 26.9 28.5 30.0 28.6 28.5 142.5

I4 V1 39.5 38.8 37.9 38.8 38.9 193.9

V2 30.2 31.4 29.9 30.4 30.5 152.4

Note:
The plot size of whole plot treatments is not equal to the plot size of subplot
treatments. Therefore the interaction of whole plot treatments with blocks
(which is called experimental error) will constitute an error variation. The sum
of (1) interaction of sub-plots with block and (2) interaction of treatment
combination with blocks will constitute another error variation. Hence there will
be two error effects in split plot design. The yield of a plot is the sum of (1)
effect of blocks, (2) effect of whole plot treatments, (3) error effect of whole
plots, (4) effect of sub-plot treatments, (5) effect of interaction of sub-plot and
main plot treatments and (6) effect of sub-plot. Step-2: To calculate total SS
(denoted as TSS)
C.F. = [(GT)2 ÷ (r x 1 x V)]; {where r – no. of blocks; I-levels of irrigation;
-no. of varieties}

Telegram: AgroMind Website: [Link]

 = [(1380.1)2 ÷ (5 x 4 x 2)] = 47616.9 TSS
= SS of all observation – C.F.
= (35.22 + 32.32+…+30.52) – 47616.9 = 48599.5 – 47616.9 =
982.25
Step-3:
From a two – way table with blocks as one way and main plots is another way.
Find the row totals and column totals and denote it as table (1).
Table-1.
Main plot Blocks
treatment Total
I II III IV V
I1 66.5 61.7 63.7 64.7 65.1 321.7
I2 79.6 82.8 83.0 84.2 82.4 412.0
I3 57.5 59.7 62.5 60.4 60.0 300.1
I4 69.7 70.2 67.8 69.2 69.4 346.3
Total 273.3 274.4 277.0 278.5 276.9 1380.1

The value of first cell in the above table is obtained by adding the yield of
all treatment combinations which involve I1 in block 1 and so on.
Each cell value is sum of two observations or plot values. Row totals are
the totals of whole plot treatments and it is the sum of 10 observations or plot
values. Column total are the block totals and it is the sum of 8 observations or
plot values.

Step-4: Find SS based on the cell values of Table (1) and denote it as table (1)
SS.
Table (1) SS = [(166.52 +….+ 69.42) ÷ 2] – CF = 48342.03 – CF = 725.13
Step-5: Find block SS denoted as BSS
BSS = {(273.32+…+276.92) ÷ 8} – CF = 47619.14 – CF = 2.24 Step-6:

Telegram: AgroMind Website: [Link]

 Main plot treatments are irrigation levels in the present example. Therefore SS
for main plot treatment will be denoted as ISS. It is calculated as follows:
ISS = [(321.72 +…+ 346.32) ÷ 10] – CF = 48321.86 – CF = 704.96 Step-7:
The table (1) SS consists of variation due to blocks, whole plot treatments and
interaction of whole plot treatments with blocks. Therefore if block SS and
whole plot treatment SS are subtracted from Table (1) SS, interaction SS of
whole plot treatments will be obtained. The interaction of whole plot treatments
and block is the effect of experimental error of whole plot treatments. So the SS
for interaction of whole plot treatments and blocks will be denoted as error (1)
SS. Therefore,
Error (1) SS = Table (1) SS – BSS – ISS = 725.12 – 2.24 – 704.96 =
17.92.
Step-8:
Now, form a two way table with whole plot treatments as one way and sub –
plot treatments as another way. Denote the table as table(2). Find row totals and
column totals of table (2).
Table-2
Main plot Sub-plot treatments
Total
treatment V1 V2
I1 168.6 153.1 321.7
I2 214.6 197.4 412.0
I3 157.6 142.5 300.1
I4 193.9 152.4 346.3
Total 734.7 645.4 1380.1

The value of the first cell in the above table is obtained by adding the yield of
all plots receiving the treatment combination I1, V1, i.e. by adding the plot values
receiving the I1, V1 combination in all the blocks. In the master table, it can be
verified that it is equal to the row totals. Therefore, the row totals of mater table
can be found out and rewritten in table (2) according to treatment combination
Telegram: AgroMind Website: [Link]
 of each cell. The row total of master table is sum of 5 observations or plot values.
Therefore each cell value in table-2 is the sum of five observations. Row totals
in table (2) are the totals of whole plot treatments. This may be checked up with
the whole plot treatments of table (1) for correctness of the formation of the table.
Column totals in table (2) are subplot treatment totals and it is the sum of 20
observation or plot values.

Step-9: Find SS based on the cell values of table (2) and denote it as table (2)
SS.
Table (2) SS = [(168.52 + …. +152.42) ÷ 5] – CF = 48570.49 – CF =
953.59
Step-10:
Sub-plot treatments are the varieties in the present example. Therefore
SS for sub-plot treatment will be denoted as VSS. It is calculated as follows,
VSS = [(734.72 + 645.42) ÷ 20] – CF = 47816.26 – CF = 199.36.
Step-11
Table (2) SS consists of variation due to whole treatments sub plot treatments
and interaction or whole plot treatments with subplot treatments. Therefore, if
SS for whole plot treatments and SS for subplot treatments are subtracted from
Table (2) SS, interaction SS of whole plot treatments and subplot treatments will
be obtained.
(I x V) SS = Table (2) SS – ISS – VSS
= 953.59 – 704.95 – 199.36 = 49.28
Step-12: Find error (2) SS as follows,
Error (2) SS = TSS – BSS – ISS – Error (1) SS – (I x V) SS
= 982.25 – 2.24 – 704.96 – 17.93 – 199.36 – 49.2 = 8.49
Step-13:
Set up the various SS calculated above in analysis of variance table. Calculate
the means square for each source of variation. Divide main plot mean square by
Error (1) mean square divide subplot mean square by error (2) mean square.
Telegram: AgroMind Website: [Link]
 Divide means square of interaction of main plot treatment and subplot treatment
by Error (2) mean square.
Source D.F S.S M.S Fc
Blocks 5-1 = 4 2.24 0.56
Main plot 4-1 = 3 704.96 234.98 234.98 ÷ 1.49 =
treatments (1) 157.29
Error (I) 4 x 3 = 12 17.93 1.49
Subplot treatments 2-1 = 1 199.36 199.36 199.36 ÷ 0.53 =
(V) 375.71
IxV 3x1=3 49.27 16.42 16.42 ÷ 0.53 =
30.95
Error (2) 39 – (sum of 8.49 0.53
above df) 39-
23 = 16
Total 40-1 = 39 982.25
Result and conclusion
The variation of whole plot treatments is found to be significant since the
calculated value of F is greater than table value of F with (3, 12) df.
The variation of sub-plot treatments is found to be significant since the
calculated value of F5 is greater than table value of F with (1, 16) df. The
variation of interaction of whole plot and sup-plot treatments is also found to be
significant since calculated value of MS is greater than table value of F with (3,
6)df.

Step-14: Comparisons
Bar Chart
a) CD to construct bar chart when whole plot treatments are significant.
CD = {√2error (1) MS ÷ 10} x t (5%) error (1) df
= {√(2x1.494) ÷ 10} x 2.179 = 1.191
b) CD to construct bar chart when subplot treatments are significant.
CD = {√(2 x error (2) MS) ÷ 20} x t (5%) error (2) df
= {√(2x0.531) ÷ 20} x 2.12 = 0.488

Telegram: AgroMind Website: [Link]

 c) Bar chart for interaction effects when variation of interaction effect is
significant, a single CD cannot be used to construct bar chart as two cases
arise.
Case I: The variation among whole plot treatments at each subplot
treatments.
Case II: The variation among subplot treatments at each main plot
treatments. The calculation of CD for case I and II are different.
Let t (1) is the ‘t’ value for error (1) df, t(2) is the tabulated ‘t’ value for error
(2) df, E(1) is the error (1) Mean square and E(2) is the Error (2) mean square.

Case I : C.D. = {√[2(2-1) E(2) + E(1)] ÷ (2 x 5) } x tw (1) where,

tw(1) = [(2-1) E (2)t(2) + E(1) t (1)} ÷ {[(2-1) E(2)] + E(1)}
Note:
(2-1) is the subplot treatment df. 5 is the number of replications. In
general, if ‘b’ is number of subplot treatments and ‘r’ is the number of
replications, then
CD = {√[2(b-1) E (2) + E (1) ÷ (b x r) } x tw (1) where
tx(1) = {[(b-1)E(2) + [E(1) t (1)]} ÷ {[(b-1) E (2) + E(1)} In
the present case

CD = {√[2[(1x0.53) + 1.494] ÷ (2 x 5) }x 1.6462 = 1.047 since tw (1) = [(1.0531

x 2.12) + (1.494 x 2.179)] ÷ [(1 x 0.531) + 1.494] =
1.6462
Case II:

CD = {√2E(2) ÷ 5 } x t(2) = 0.977

Note : In general

CD = {√2E(2) ÷ r} x t(2)
The bar charts in the present case are
Bar chart for main plot treatments
Telegram: AgroMind Website: [Link]
 I2 I4 I1 I3

Bar chart for subplot treatments

V1 V2
Bar chart for interaction – Sub at main
I1 : V1 V2
I2 : V1 V2
I3 : V1 V2
I4 : V1 V2
Main at Sub
V1 : I2 I4 I1 I3
V2 : I2 I1 I4 I3

STRIP – PLOT DESIGN

It is a design some what analogous to the split-plot arrangement, in which
two different sets of treatments can be tried in large plots with one set of plots
superposed over the other set at right angles. Such an arrangement may be
convenient in a cultural experiment involving factors i.e. spacing, ploughing,
where the use of small plots by splitting larger plot is not feasible.

It is also defined as a design in which a block is divided into (a) horizontal strips
(plots) equal to the level of one factor (b) vertical strips (plots) equal to the levels
of another factor (c) the levels of first factor are randomized to the horizontal
strips (d) the levels of second factor are randominsed to the vertical strips and
(e) the above steps (a) to (d) repeated to the other block.

Layout
This is also a variation of the split plot design. In every replicate the levels of
one factor will be applied in horizontal strips and the levels of another factor will
be applied in vertical strips. The design will appear as follows;

Telegram: AgroMind Website: [Link]

 Rep.1 Rep. 2 Rep.3
a1 a2 a3 a4 a3 a1 a0 a2 a1 a2 a0 a3
b1
b0 b b
b2 1 1
b b
0 0
b b
2 2
Strip plot design will be helpful in such experiments where the lot size should be
larger for both the factors used. It will be an ideal design for experiments on
sugarcane and cotton. This design has a drawback that it tests the interaction A
x B more sensitively than the main effects A and B. The partitioning of the
variation will be as follows. If r replications are used with A and B factors at s
and t level respectively.
The Anova Table
Source d.f
Blocks r-1
A s-1
Blocks x A (Error a) (r-1) (s-1)
B t-1
Blocks x B (Error b) (r-1) (t-1)
Ax B (s-1) (t-1)
Error (c) (t-1) (r-1) (s-1)
Total rst - 1

The above partitioning of the degrees of freedom clearly indicate how the sums
of squares are to be computed for each component. Significance of main effect
A will be tested against error (a), significance of main effect B will be tested
against error (b) and significance of interaction A x B will be tested against error
(c).

Telegram: AgroMind Website: [Link]

 Example-1:
An experiment was conducted with 4 varieties of rice and 3 levels of N (0 kg,
60 kg 120kg N/ha) with 3 replications. Analyse the data obtained from strip plot
design draw your conclusion.

Telegram: AgroMind Website: [Link]

Telegram: AgroMind Website: [Link]
R1 R2 R3
V2 V3 V1 V4 V1 V2 V4 V3 V3 V1 V2 V4
N0 40 44 33 26 N1 64 73 53 28 N2 6 86 63 32
N2 70 68 73 16 N0 39 58 33 37 N1 6 49 72 44
N1 56 55 41 39 N2 68 83 51 27 N0 4 44 50 33

Solution Tabulate the data as follows

R1 R2 R3 Total
V1N0 23 39 44
106
V1N1 41 64 49
154
V2N2 73 68 86
227
V2N0 40 58 50
148
V2N1 66 73 72
211
V3N2 30 83 63
176
V3N0 44 37 46
127
V3N1 55 28 60
143
V4N2 68 27 61
156
V4N0 26 33 33
92
V4N1 29 53 44
126
V4N2 16 51 32
99
Total 511 614 640 1805

CF = (Grand Total)2 ÷ No. of observations = 18052 ÷ 36 = 90500.69

TSS = [(1232 + …. + 322) ÷ 1] – CF = 10788.3060
R x A Table (Variety)
R1 R2 R3 Total
V1 137 171 179 487
V2 166 214 185 565
V3 167 92 167 426

Telegram: AgroMind Website: [Link]

 V4 81 137 109 327
Total 551 614 640 1805

RSS = [(5512 + 6142 + 6402) ÷ 12] – CF = 349.056

ASS = [(4872 + 5652 + 4262 + 3272) ÷ 9] – CF = 3365.8616
R x A Stable SS = [(1372 + … + 1092) ÷ 3] – CF = 5869.6393
Error (a) = R x A table SS – RSS – ASS
= 5859.6393 – 349.0
– 3365.81 = 2144.72
R x B Table
R1 R2 R3 Total
N0 133 167 173 473
N1 191 218 225 634
N2 227 229 242 398
Total 551 614 644 1805

RSS = 349.056
BSS = [(4732 + 6342 + 6982) ÷ 12] – CF = 2240.056
R x B Table SS = [(1332 + … + 2422)] – CF = 2667.056
Error (b) = R x B Table SS – RSS – BSS
= 2667.056 – 349.056 – 2240.056 = 77.9440

A x B interaction Table
V1 V2 V3 V4 Total
N0 106 148 127 92 473
N1 154 201 143 136 634
N2 227 216 156 99 698
Total 487 565 426 327 1805

ASS = 3365.8616
Telegram: AgroMind Website: [Link]
 BSS = 2240.056
A x B Table SS = [(1062 + … + 992) ÷ 3] – CF = 7204.9727 Error
(c) = A x B Table SS – ASS – BSS
= 7204.9729 – 3365 – 2240.056 = 1011.6116

Anova table
Source df SS M.S Fc Ft5% 1%
Replication (r-1) = 2 349.056 174.5
Variety (A) (a-1) = 3 3365.36 1121.9 3.13 4.76 5.76
Error (a) (r-1) (a-1) = 6 2144.72 357.45
Nitrogen (b-1) = 2 2240.0 1120.0 57.47 6.74 7.90
(B)
Error (b) (r-1) (b-1) = 4 77.94 19.68
Ax B (a-1) (b-1) = 6 1599.0 266.5 3.16 3.00 4.00
Error (c) (r-1) (a-1) (b-1) = 1011.6 84.3
12
Total (rab-1) 10788.3

Calculation of CD

(1) For two ‘A’ means, SED = √x\2Err(a) ÷ rb = √(2 x 357.4) ÷ (3 x 3) =

8.91

CD = 8.91 x 2.776 = 21.809

CD = 5.00

(3) For two ‘A’ means at a fixed levels of ‘B’

√2(b-1) (Err a + Err c) ÷ (r x b) = √2 x (3-1) (357.0 + 84) ÷ (3 x 3)

Telegram: AgroMind Website: [Link]

 T = 2.179 SED = 13.32 CD = 29.04

(4) For two ‘B’ means at a fixed levels of ‘A’

√2(a-1) (Err b + Err c) ÷ (r x b) = √2 x (4-1) (19.45 + 84.3) ÷ (4 x

3)

t = 2.179 SED = 4.878 CD = 10.628

Conclusion
(1) For comparing four A means CD = 21.89
V2 V1 V3 V4
62.77 54.11 47.33 36.33
V1, V2, V3 and V4 are all on par

(2) For comparing three B means CD = 5.0026

N2 N1 N0
58.16 52.83 39.41
(a) N2 is significantly superior to N1 and N0
(b) N1 is significantly superior to N0

(3) For A at fixed levels of B, CD = 29.04

At N0 : V2 V3 V1 V4
49.33 42.33 35.33 30.67
At N0 : V2 V1 V3 V4
67.0 51.3 47.7 45.3
At N0 : V1 V2 V3 V4

Telegram: AgroMind Website: [Link]

 75.66 72 52 33
(4) For B at fixed levels of A, CD = 10.63 At
V1 75.7 51.3 35
V2 72 67 49
V3 52 67 42
V4 45 33 30
At V1 : N2 is superior to N1 and N0; N1 is superior to N0
At V2 : N1 and N2 are on par but superior than N0
At V3 : All are on par
At V4 : N1 is superior than N2; N0 on par

Exercise
Part – I
Explain / Define
1. Experiment
2. Experimental unit
3. Treatment
4. Experimental material
5. Experimental error
6. Replication
7. Randomization
8. Local control
9. Analysis of variance
10. CRD
11. RBD
12. LSD
13. Split Plot Design
14. Strip Plot Design

Part-II
Explain
1. Advantages and disadvantages of CRD
2. Advantages and disadvantages of RBD
3. Advantages and disadvantages of LSD
4. Randomization of CRD with equal as unequal
5. Randomization of RBD with equal as unequal
6. Randomization of LSD with equal as unequal
Telegram: AgroMind Website: [Link]
 Part – III

1. Explain “Basic principles of experimental Designs”

2. Give the Analysis Procedure for CRD with equal and unequal
3. Give the Analysis Procedure for RBD with equal and unequal
4. Give the Analysis Procedure for LSD with equal and unequal
5. Give the Analysis Procedure for Split Plot Design
6. Give the Analysis Procedure for Strip Plot Design

Telegram: AgroMind Website: [Link]