Bas 103

1
LECTURE NOTES
Course No: STCA-101
STATISTICS
TIRUPATI
Karl Pearson R. A. Fisher
Karl Pearson (né Carl Pearson)
Sir Ronald Aylmer Fisher (1890-1962)
Born: 27 March 1857
Islington, London, England
Born: 17 February 1890
East Finchley, London , England
Prepared By
Dr. G. MOHAN NAIDU M.Sc., Ph.D.,
Assistant Professor & Head
Dept. of Statistics & Mathematics
S.V. Agricultural College
TIRUPATI
ACHARYA N.G. RANGA AGRICULTURAL UNIVERSITY

2
LECTURE OUTLINE
Course No. STCA-101 Credits: 2 (1+1)
Course Title: STATISTICS
THEORY
S. No. Topic/Lesson
1 Introduction to Statistics, Definition, Advantages and Limitations.
2 Frequency distribution: Construction of Frequency Distribution table.
3 Measures of Central Tendency: Definition, Characteristics of Satisfactory
average.
4 Arithmetic Mean, Median, Mode for grouped and ungrouped data – Merits and
Demerits of Arithmetic Mean.
5 Measures of Dispersion: Definition, standard deviation, variance and
coefficient of variation.
6 Normal Distribution and its properties. Introduction to Sampling: Random
sampling, concept of standard error of Mean.
7 Tests of Significance: Introduction, Types of errors, Null hypothesis, level of
significance and degrees of freedom, steps in testing of hypothesis.
8 Large sample tests: Test for Means – Z-test, One sample and Two samples
with population S.D. known and Unknown.
9 Small sample tests: Test for Means – One sample t – test, Two samples t-test
and Paired t-test.
10 Chi-Square test in 2x2 contingency table with Yate‟s correction, F-test.
11 Correlation: Definition, types, properties, Scatter diagram, calculation and
testing.
12 Regression: Definition, Fitting of two lines Y on X and X on Y, Properties,
inter relation between correlation and regression.
13 Introduction to Experimental Designs, Basic Principles, ANOVA its
assumptions.
14 Completely Randomized Design: Layout, Analysis with equal and unequal
replications.
15 Randomized Block Design: Layout and Analysis.
16 Latin Square Design: Layout and Analysis.

3
PRACTICALS
S .No. Topic
1 Construction of Frequency Distribution tables
2 Computation of Arithmetic Mean for Grouped and Un-grouped data
3 Computation of Median for Grouped and Un-grouped data
4 Computation of Mode for Grouped and Un-grouped data
5 Computation of Standard Deviation and variance for grouped and
ungrouped data
6 Computation of coefficient of variation for grouped and ungrouped data
7 SND (Z) test for single sample, Population SD known and Unknown
8 SND (Z) test for two samples, Population SD known and Unknown
9 Student‟s t-test for single and two samples
10 Paired t-test and F-test
11 Chi-square test – 2x2 contingency table with Yate‟s correction
12 Computation of correlation coefficient and its testing
13 Fitting of simple regression equations Y on X and X on Y
14 Completely Randomized Design: Analysis with equal and unequal
replications
15 Randomized Block Design: Analysis
16 Latin Square Design: Analysis

1
S T A T I S T I C S
Statistics has been defined differently by different authors from time to time. One
can find more than hundred definitions in the literature of statistics.
Statistics can be used either as plural or singular. When it is used as plural, it is a
systematic presentation of facts and figures. It is in this context that majority of people
use the word statistics. They only meant mere facts and figures. These figures may be
with regard to production of food grains in different years, area under cereal crops in
different years, per capita income in a particular state at different times etc., and these are
generally published in trade journals, economics and statistics bulletins, news papers, etc.,
When statistics is used as singular, it is a science which deals with collection,
classification, tabulation, analysis and interpretation of data.
The following are some important definition of statistics.
Statistics is the branch of science which deals with the collection, classification and
tabulation of numerical facts as the basis for explanations, description and comparison of
phenomenon - Lovitt
The science which deals with the collection, analysis and interpretation of numerical data
- Corxton & Cowden
The science of statistics is the method of judging collective, natural or social phenomenon
from the results obtained from the analysis or enumeration or collection of estimates -
King
Statistics may be called the science of counting or science of averages or statistics is the
science of the measurement of social organism, regarded as whole in all its manifestations
- Bowley
Statistics is a science of estimates and probabilities -Boddington
Statistics is a branch of science, which provides tools (techniques) for decision making in
the face of uncertainty (probability) - Wallis and Roberts
This is the modern definition of statistics which covers the entire body of statistics
All definitions clearly point out the four aspects of statistics collection of data,
analysis of data, presentation of data and interpretation of data.
Importance: Statistics plays an important role in our daily life, it is useful in almost all
sciences – social as well as physical – such as biology, psychology, education, economics,
business management, agricultural sciences etc., . The statistical methods can be and are

2
being followed by both educated and uneducated people. In many instances we use sample
data to make inferences about the entire population.
1. Planning is indispensable for better use of nation‟s resources. Statistics are
indispensable in planning and in taking decisions regarding export, import, and
production etc., Statistics serves as foundation of the super structure of
planning.
2. Statistics helps the business man in the formulation of polices with regard to
business. Statistical methods are applied in market and production research,
quality control of manufactured products
3. Statistics is indispensable in economics. Any branch of economics that require
comparison, correlation requires statistical data for salvation of problems
4. State. Statistics is helpful in administration in fact statistics are regarded as
eyes of administration. In collecting the information about population, military
strength etc., Administration is largely depends on facts and figures thud it
needs statistics
5. Bankers, stock exchange brokers, insurance companies all make extensive use
of statistical data. Insurance companies make use of statistics of mortality and
life premium rates etc., for bankers, statistics help in deciding the amount
required to meet day to day demands.
6. Problems relating to poverty, unemployment, food storage, deaths due to
diseases, due to shortage of food etc., cannot be fully weighted without the
statistical balance. Thus statistics is helpful in promoting human welfare
7. Statistics are a very important part of political campaigns as they lead up to
elections. Every time a scientific poll is taken, statistics are used to calculate
and illustrate the results in percentages and to calculate the margin for error.
In agricultural research, Statistical tools have played a significant role in the
analysis and interpretation of data.
1. In making data about dry and wet lands, lands under tanks, lands under
irrigation projects, rainfed areas etc.,
2. In determining and estimating the irrigation required by a crop per day, per
base period.
3. In determining the required doses of fertilizer for a particular crop and crop
land.

3
4. In soil chemistry also statistics helps classifying the soils basing on their
analysis results, which are analyzed with statistical methods.
5. In estimating the losses incurred by particular pest and the yield losses due to
insect, bird, or rodent pests statistics is used in entomology.
6. Agricultural economists use forecasting procedures to determine the future
demand and supply of food and also use regression analysis in the empirical
estimation of function relationship between quantitative variables.
7. Animal scientists use statistical procedures to aid in analyzing data for decision
purposes.
8. Agricultural engineers use statistical procedures in several areas, such as for
irrigation research, modes of cultivation and design of harvesting and
cultivating machinery and equipment.
Limitations of Statistics:
1. Statistics does not study qualitative phenomenon
2. Statistics does not study individuals
3. Statistics laws are not exact laws
4. Statistics does not reveal the entire information
5. Statistics is liable to be misused
6. Statistical conclusions are valid only on average base
Types of data: The data are of two types i) Primary Data and ii) Secondary Data
i) Primary Data: It is the data collected by the primary source of information i.e
by the investigator himself.
ii) Secondary Data: It is the data collected from secondary sources of information,
like news papers, trade journals and statistical bulletins, etc.,
Variables and Attributes:
Variability is a common characteristic in biological sciences. A quantitative or qualitative
characteristic that varies from observation to observation in the same group is called a
variable. In case of quantitative variables, observations are made using interval scales
whereas in case of qualitative variables nominal scales are used. Conventionally, the
quantitative variables are termed as variables and qualitative variables are termed as
attributes. Thus, yield of a crop, available nitrogen in soil, daily temperature, number of
leaves per plant and number of eggs laid by insects are all variables. The crop varieties,
soil types, shape of seeds, seasons and sex of insects are attributes.

4
The variable itself can be classified as continuous variable and discrete variable.
The variables for which fractional measurements are possible, at least conceptually, are
called continuous variables. For example, in the range of 7 kg to 10 yield of a crop, yield
might be 7.15 or 7.024kg. Hence, yield is a continuous variable. The variables for which
such factional measurements are not possible are called discrete or discontinuous
variables. For example, the number of grains per panicle of paddy can be counted in full
numbers like 79, 80, 81 etc. Thus, number of grains per panicle is a discrete variable. The
variables, discrete or continuous are dented by capital letters like X and Y.
Construction of Frequency Distribution Table:
In statistics, a frequency distribution is a tabulation of the values that one or more
variables take in a sample. Each entry in the table contains the frequency or count of the
occurrences of values within a particular group or interval, and in this way the table
summarizes the distribution of values in the sample.
The following steps are used for construction of frequency table
Step-1: The number of classes are to be decided
The appropriate number of classes may be decided by Yule‟s formula, which is as follows:
Number of classes = 2.5 x n1/4
. where „n‟ is the total number of observations
Step-2: The class interval is to be determined. It is obtain by using the relationship
Maximum value in the given data – Minimum value in the given data
C.I = -------------------------------------------------------------------------------------
Number of classes
Step-3: The frequencies are counted by using Tally marks
Step-4: The frequency table can be made by two methods
a) Exclusive method
b) Inclusive method
a) Exclusive method: In this method, the upper limit of any class interval is kept the
same as the lower limit of the just higher class or there is no gap between upper limit of
one class and lower limit of another class. It is continuous distribution

5
Ex:
C.I. Tally marks Frequency (f)
0-10
10-20
20-30
b) Inclusive method: There will be a gap between the upper limit of any class and the
lower limit of the just higher class. It is discontinuous distribution
Ex:
0-9
10-19
20-29
To convert discontinuous distribution to continuous distribution by subtracting 0.5 from
lower limit and by adding 0.5 to upper limit
Note: The arrangement of data into groups such that each group will have some numbers.
These groups are called class and number of observations against these groups are called
frequencies.
Each class interval has two limits 1. Lower limit and 2. Upper limit
The difference between upper limit and lower limit is called length of class interval.
Length of class interval should be same for all the classes. The average of these two limits
is called mid value of the class.
Example: Construct a frequency distribution table for the following data
25, 32, 45, 8, 24, 42, 22, 12, 9, 15, 26, 35, 23, 41, 47, 18, 44, 37, 27, 46, 38, 24, 43,
46, 10, 21, 36, 45, 22, 18.
Solution: Number of observations (n) = 30
Number of classes = 2.5 x n1/4
= 2.5 x 301/4
= 2.5 x 2.3
= 5.8  6.0

6
Class interval =
classesofNo
valueMinvalueMax
..
.. 
=
6
846 
=
6
38
= 6.3  6.0
Inclusive method:
8-14 | | | | 4
15-21 | | | | 4
22-28 | | | | | | | 8
29-35 | | 2
36-42 | | | | 5
42-49 | | | | | | 7
Total 30
Exclusive method:
7.5-14.5 | | | | 4
14.5-21.5 | | | | 4
21.5-28.5 | | | | | | | 8
28.5-35.5 | | 2
35.5-42.5 | | | | 5
42.5-49.5 | | | | | | 7
Total 30
MEASURES OF CENTRAL TENDENCY
One of the most important aspects of describing a distribution is the central value
around which the observations are distributed. Any mathematical measure which is
intended to represent the center or central value of a set of observations is known as
measure of central tendency (or )
The single value, which represents the group of values, is termed as a „measure of central
tendency‟ or a measure of location or an average.

7
Characteristics of a Satisfactory Average:
1. It should be rigidly defined
2. It should be easy to understand and easy to calculate
3. It should be based on all the observations
4. It should be least affected by fluctuations in sampling
5. It should be capable of further algebraic treatment
6. It should not be affected much by the extreme values
7. It should be located easily
Types of average: 1. Arithmetic Mean
2. Median
3. Mode
4. Geometric Mean
5. Harmonic Mean
Arithmetic Mean (A.M): It is defined as the sum of the given observations divided by the
number of observations. A.M. is measured with the same units as that of the observations.
Ungrouped data:
Direct Method: Let x1, x2, ………,xn be „n‟ observations then the A.M is computed from
the formula:
A.M. =
n
x
n
xxx
n
i
i
n


 121 ........
where 
n
i
ix
1
= sum of the given observations
n = Number of observations
Linear Transformation Method or Deviation Method: When the variable constitutes large
values of observations, computation of arithmetic mean involves more calculations. To
overcome this difficulty, Linear Transformation Method is used. The value xi is
transformed to di.
and A.M. = A+
n
d
n
i
i1
where A = Assumed mean which is generally taken as class mid point of middle class
or the class where frequency is large.

8
di = Axi  = deviations of the ith
value of the variable taken from an assumed
mean and n = number of observations
Grouped Data:
Let f1, f2, ………., fn be „n‟ frequencies corresponding to the mid values of the class
intervals x1, x2, ………xn then
A.M. =
N
xf
f
xf
fff
xfxfxf
n
i
ii
n
i
i
n
i
ii
n
nn


 




 1
1
1
21
2211
..........
.............
(direct method)
and
A.M. =A+ C
N
df
AC
fff
dfdfdf
n
i
ii
n
nn



















 1
21
2211
..........
.............
(indirect method)
where di = deviation =
c
Axi 
; f = frequency; C = class interval;
x = mid values of classes.
Arithmetic mean, when computed for the data of entire population, is represented by the
symbol „  ‟. Where as when it is computed on the basis of sample data, it is represented
as X , which is the estimate of  .
Properties of A.M.:
i) The algebraic sum of the deviations taken from arithmetic mean is zero
i.e. (x-A.M.) = 0
ii) Let 1x be the mean of n1 observations, 2x be the mean of the n2
observations ……… kx be the mean of nk observations then the mean x
of n = (n1+n2+…..nk) observations is given by
x =
k
kk
nnn
xnxnxn


....
......
21
2211
=




k
i
i
k
i
ii
n
xn
1
1
Merits:
1. It is well defined formula defined
2. It is easy to understand and easy to calculate
3. It is based upon all the observations

9
4. It is amenable to further algebraic treatments, provided the sample is randomly
obtained.
5. Of all averages, arithmetic mean is affected least by fluctuations of sampling
Demerits:
1. Cannot be determined by inspection nor it can be located graphically
2. Arithmetic mean cannot be obtained if a single observation is missing or lost
3. Arithmetic mean is affected very much by extreme values
4. Arithmetic mean may lead to wrong conclusions if the details of the data from which
it is computed are not given
5. In extremely asymmetrical (skewed) distribution, usually arithmetic mean is not a
suitable measure of location
Examples:
i) Ungrouped data:
If the weights of 7 ear-heads of sorghum are 89, 94, 102, 107, 108, 115 and 126 g. Find
arithmetic mean by direct and deviation methods
Solution:
i) Direct Method: A.M. =
n
x
n
i
i1
ii) Deviation Method: A.M. =
n
d
A
n
i
i
 1
Where n = number of observations; xi = given values;
A = arbitrary mean (assumed value); d = deviation =xi -A
xi di = xi –A
89 89-102= -13
94 94-102= -8
102 102-102= 0
107 107-102= 5
108 108-102= 6
115 115-102= 13
126 126-102= 24
 x 741  d 29
here A = assumed value = 102
i) A.M. =
n
x
n
i
i1
(ii) A.M. =
n
d
A
n
i
i
 1

10
=
7
741
= 






7
27
102
= 105.86 g = 105.86 g.
ii) Grouped Data:
The following are the 405 soybean plant heights collected from a particular plot. Find the
arithmetic mean of the plants by direct and indirect method:
Plant height
(Cms)
8-
12
13-
17
18-
22
23-
27
28-
32
33-
37
38-
42
43-
47
48-
52
53-
57
No. of
plants( if )
6 17 25 86 125 77 55 9 4 1
Solution:
a) Direct Method:
A.M. =
N
xf
n
i
ii
 1
Where xi = mid values of the corresponding classes
N = Total frequency = 
n
i
if
1
fi = frequency
b) Deviation Method:
A.M. = C
N
df
A
ii










Where di = deviation ( i.e.
C
Ax
d i
i

 )
Length of class interval (C ) = 5; Assumed value (A) = 30
C.I fi xi fixi C
Ax
d i
i


ii df
8-12 6 10 60 -4 -24
13-17 17 15 255 -3 -51
18-22 25 20 500 -2 -50
23-27 86 25 2150 -1 -86
28-32 125 30 3750 0 0
33-37 77 35 2695 1 77
38-42 55 40 2200 2 110
43-47 9 45 405 3 27
48-52 4 50 200 4 16
53-57 1 55 55 5 5
Total N = 405 Σfi xi = 12270 Σfidi = 24

11
a) Direct Method:
A.M. = 
405
12270
30.30 cms.
b) Deviation Method:
A.M. = 5
405
24
30 






= 30.30 cms.
MEDIAN
The median is the middle most item that divides the distribution into two equal
parts when the items are arranged in ascending order.
Ungrouped data: If the number of observations is odd then median is the middle value
after the values have been arranged in ascending or descending order of magnitude. In
case of even number of observations, there are two middle terms and median is obtained
by taking the arithmetic mean of the middle terms.
In case of discrete frequency distribution median is obtained by considering the
cumulative frequencies. The steps for calculating median are given below:
i) Arrange the data in ascending or descending order of magnitude
ii) Find out cumulative frequencies
iii) Apply formula: Median = Size of
2
1N
, where N=  f
iv) Now look at the cumulative frequency column and find, that total which is either equal
to
2
1N
or next higher to that and determine the value of the variable corresponding to it,
which gives the value of median.
Continuous frequency distribution:
If the data are given with class intervals then the following procedure is adopted for the
calculation of median.
i) find
2
1N
, where N = f
ii) see the (less than) cumulative frequency just greater than
2
1N
iii) the corresponding value of x is median

12
In the case of continuous frequency distribution, the class corresponding to the cumulative
frequency just greater than
2
1N
is called the median class and the value of median is
obtained by the following formula:
Median = l +














f
m
N
2
1
C
Where l is the lower limit of median class
f is the frequency of the median class
m is the cumulative frequency of the class preceding the median class
C is the class length of the median class
N = total frequency
Examples:
Case-i) when the number of observations (n) is odd:
The number of runs scored by 11 players of a cricket team of a school are
5, 19, 42, 11, 50, 30, 21, 0, 52, 36, 27
To compute the median for the given data, we proceed as follows:
In case of ungrouped data, if the number of observations is odd then median is the middle
value after the values have been arranged in ascending or descending order of magnitude.
Let us arrange the values in ascending order:
0, 5, 11, 19, 21, 27, 30, 36, 42, 50, 52
Median =
th
n





 
2
1
value =
th





 
2
111
value
= 6th
value
Now the 6th
value in the data is 27.
 Median = 27 runs
Case-ii) when the number of observations (n) is even:
Find the median of the following heights of plants in Cms:
6, 10, 4, 3, 9, 11, 22, 18

13
In case of even number of observations, there are two middle terms and median is
obtained by taking the arithmetic mean of the middle terms.
Let us arrange the given items in ascending order
3, 4, 6, 9, 10, 11, 18, 22
In this data the number of items n = 8, which is even.
Median = average of 





2
n
th and 





1
2
n
th terms.
= average of 9 and 10
Median = 9.5 Cms.
Grouped Data:
Find out the median for the following frequency distribution of 180 sorghum ear-heads.
Weight of ear-heads (in g) 40-
60
60-
80
80-
100
100-
120
120-
140
140-
160
160-
180
180-
200
No.of ear-heads 6 28 35 45 30 15 12 9
Solution: Median = C
f
m
N
l














 2
1
------------------ (1)
Where l is the lower limit of the median class
f is the frequency of the median class
m is the cumulative frequency of the class preceding the median class
C is the class interval of the median class
and N =  f = Total number of observations

14
Weight of ear-
heads (in g)
No. of ear-heads Cumulative
Frequency (CF)
40-60 6 6
60-80 28 34
80-100 35 69 - m
100-120 45 – f 114
(Median class)
120-140 30 144
140-160 15 159
160-180 12 171
180-200 9 180
N =  f =180
Here
2
1N
= 5.90
2
181
 . Cumulative frequency just greater than 90.5 is 69 and the
corresponding class is 100-120. The median class is 100-120.
N = 121; L =40; f = 24; m = 39 and C = 10
Substituting the above values in equation (1), we get
Median = 20
45
695.90
100 




 

= 109.56 g
Merits and Demerits of Median:
Merits:
1. It is rigidly defined.
2. It is easily understood and is easy to calculate. In some cases it can be located
merely by inspection.
3. It is not at all affected by extreme values.
4. It can be calculated for distributions with open-end classes
Demerits:
1. In case of even number of observations median cannot be determined exactly. We
merely estimate it by taking the mean of two middle terms.

15
2. It is not amenable to algebraic treatment
3. As compared with mean, it is affected much by fluctuations of sampling.
MODE
Mode is the value which occurs most frequently in a set of observations or mode is the
value of the variable which is predominant in the series.
In case of discrete frequency distribution mode is the value of x corresponding to
maximum frequency
In case of continuous frequency distribution, mode is obtained from the formula:
Mode = l + 







21
1
2
)(
fff
ff
C
Where l is the lower limit of modal class
C is class interval of the modal class
f the frequency of the modal class
f1 and f2 are the frequencies of the classes preceding and succeeding
the modal class respectively
If the distribution is moderately asymmetrical, the mean, median and mode obey the
empirical relationship:
Mode = 3 Median – 2 Mean
Example: Find the mode value for the following data:
27, 28, 30, 33, 31, 35, 34, 33, 40, 41, 55, 46, 31, 33, 36, 33, 41, 33.
Solution: As seen from the above data, the item 33 occurred maximum number of times
i.e. 5 times. Hence 33 is considered to be the modal value of the given data.
Grouped Data:
Example: The following table gives the marks obtained by 89 students in Statistics. Find
the mode.
Marks 10-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49
No. of
students
4 6 10 16 21 18 9 5
Solution: Mode =
21
1
2
)(
fff
ff
l


 xC
Where l = the lower limit of the modal class ; C = length of the modal class

16
f = the frequency of the modal class
1f = the frequency of the class preceding modal class
2f = the frequency of the class succeeding modal class
Sometimes it so happened that the above formula fails to give the mode. In this case, the
modal value lies in a class other than the one containing maximum frequency. In such
cases we take the help of the following formula;
Mode =
21
2
ff
f
l

 xC
Where f , c, 1f and 2f have usual meanings.
Marks No. of students (f)
10-14 4
15-19 6
20-24 10
25-29 16f1
30-34 21f
35-39 18f2
40-44 9
45-49 5
From the above table it is clear that the maximum frequency is 21 and it lies in the class
30-34.
Thus the modal class is 29.5-34.5
Here L = 29.5 , c = 5, f = 21, 1f =16, 2f = 18
Mode = 5*
181621*2
1621
30 






= 30 + 3.13
= 33.63 m

17
Merits and Demerits of Mode:
Merits:
1. Mode is readily comprehensible and easy to calculate.
2. Mode is not at all affected by extreme values.
3. Mode can be conveniently located even if the frequency distribution has class-
intervals of unequal magnitude provided the modal class and the classes preceding
and succeeding it are of the same magnitude. Open-end classes also do not pose
any problem in the location of mode
Demerits:
1. Mode is ill defined. It is not always possible to find a clearly defined mode. In
some cases, we may come across distributions with two modes. Such distributions
are called bi-modal. If a distribution has more than two modes, it is said to be
multimodal.
2. It is not based upon all the observations.
3. It is not capable of further mathematical treatment.
4. As compared with mean, mode is affected to a greater extent by fluctuations of
sampling.
Dispersion
Dispersion means scattering of the observations among themselves or from a
central value (Mean/ Median/ Mode) of data. We study the dispersion to have an idea
about the variation.
Suppose that we have the distribution of the yields (kg per plot) of two Ground nut
varieties from 5 plots each. The distribution may be as follows:
Variety 1: 46 48 50 52 54
Variety 2: 30 40 50 60 70
It can be seen that the mean yield for both varieties is 50 k.g. But we can not say that the
performances of the two varieties are same. There is greater uniformity of yields in the
first variety where as there is more variability in the yields of the second variety. The first
variety may be preferred since it is more consistent in yield performance.

18
Measures of Dispersion:
These measures give us an idea about the amount of dispersion in a set of
observations. They give the answers in the same units as the units of the original
observations. When the observations are in kilograms, the absolute measure is also in
kilograms. If we have two sets of observations, we cannot always use the absolute
measures to compare their dispersion. The absolute measures which are commonly used
are:
1. The Range
2. The Quartile Deviation
3. The Mean Deviation
4. The Standard Deviation and Variance
5. Coefficient of Variation
6. Standard Error
Standard Deviation: It is defined as the positive square root of the arithmetic mean of the
squares of the deviations of the given values from arithmetic mean. The square of the
standard deviation is called variance.
Ungrouped data:
Let x1, x2, …….,xn be n observations then the standard deviation is given by the
formula
S.D. =
n
MAx
n
i
i
2
1
.).( 
where A.M. =
n
x
n
i
i1
,
where n = no. of observations.
Simplifying the above formula, we have
or S.D. =






















 

2
1
2
1
1
n
x
x
n
n
i
in
i
i
by linear transformation method, we have

19
 =






















 

2
1
2
1
1
n
d
d
n
n
i
in
i
i
where di = Axi  ; A= Assumed value; xi = Given values
Continuous frequency distribution: (grouped data):
Let f1, f2, ……,fn be the „n‟ frequencies corresponding to the mid values of the
classes x1, x2, ……,xn respectively, then the standard deviation is given by
S.D. =
N
MAxf
n
i
i

1
2
.).(
where Nf
n
i
i 1
Simplifying the above formula, we have
S.D. =




























 

2
1
1
21
N
xf
xf
N
n
i
iin
i
ii
by linear transformation method, we have
S.D. = C










N
df
df
N
ii
ii
2
2 )(1
where di =
C
Axi 
; A = assumed value; and C = class interval
S.D. for population data is represented by the symbol „‟
Ungrouped data:
Example:
Calculate S.D. for the kapas yields (in kg per plot) of a cotton variety recorded from seven
plots 5, 6, 7, 7, 9, 4, 5
i) Direct method:
S.D. =






















 

2
1
2
1
1
n
x
x
n
n
i
in
i
i

20
ii) Deviation Method:
S.D. =






















 

2
1
2
1
1
n
d
d
n
n
i
in
i
i
Where xi = given values
Assumed value (A) = 7
di = deviation (i.e. di =xi-A)
ix
2
ix
Axd ii 
2
id
5 25 5-7=-2 4
6 36 6-7=-1 1
7 49 7-7=0 0
7 49 7-7=0 0
9 81 9-7=2 4
4 16 4-7=-3 9
5 25 5-7=-2 4
Σxi = 43 Σxi
2
= 281 Σdi = -6 Σ
2
id = 22
i) Direct method:
S.D. =
 







2
7
43
281
7
1
= 1.55 kg.
S.D. =
 





 

2
7
6
22
7
1
= 1.55 kg

21
Grouped Data:
Example:
The following are the 381soybean plant heights in Cms collected from a particular plot.
Find the Standard deviation of the plants by direct and deviation method:
Plant heights (Cms) No. of Plants (fi)
6.8 -7.2 9
7.3 -7.7 10
7.8-8.2 11
8.3-8.7 32
8.8-9.2 42
9.3-9.7 58
9.8-10.2 65
10.3-10.7 55
10.8-11.2 37
11.3-11.7 31
11.8-12.2 24
12.3-12.7 7
Solution:
i) Direct method:
A.M. =
N
xf
n
i
ii
 1
; where N = 
n
i
if
1
S.D. =






















 
 N
xf
xf
N
n
i
iin
i
ii
2
1
1
21
A.M. = C
N
df
A
n
i
ii













1

22
S.D. = C






















 
 N
df
df
N
n
i
iin
i
ii
2
1
1
21
C.I. fi xi fixi
2
ii xf C
Ax
d i
i


fidi
2
ii df
6.8-7.2 9 7.0 63 441 -5 -45 225
7.3-7.7 10 7.5 75 562.5 -4 -40 160
7.8-8.2 11 8.0 88 704 -3 -33 99
8.3-8.7 32 8.5 272 2312 -2 -64 128
8.8-9.2 42 9.0 378 3402 -1 -42 42
9.3-9.7 58 9.5 551 5234.5 0 0 0
9.8-10.2 65 10 650 6500 1 65 65
10.3-10.7 55 10.5 577.5 6063.75 2 110 220
10.8-11.2 37 11.0 407 4477 3 111 333
11.3-11.7 31 11.5 356.5 4099.75 4 124 496
11.8-12.2 24 12.0 288 3456 5 120 600
12.3-12.7 7 12.5 87.5 1093.75 6 42 252
N
=381
Σfixi
=3793.5
Σfixi
2
=38346.25
Σfidi
=
348
Σ
2
ii df =
2620
i) Direct method:
A.M. =
381
5.3793
= 9.96 Cms
S.D. =
 







2
381
5.3793
25.38346
381
1
=  71.3777025.38346
381
1

= 5106.1 = 1.23 Cms.
A.M. = 9.5+ 5.0
381
348






= 9.96 Cms

23
S.D. = 0.5
 







381
348
2620
381
1
2
= 0.5x47.98 = 1.23 Cms.
Measures of Relative Dispersion:
These measures are calculated for the comparison of dispersion in two or more
than two sets of observations. These measures are free of the units in which the original
data is measure. If the original is in dollar or kilometers, we do not use these units with
relative measure of dispersion. These are a sort of ratio and are called coefficients.
Suppose that the two distributions to be compared are expressed in the same units and
their means are equal or nearly equal. Then their variability can be compared directly by
using their standard deviations. However, if their means are widely different or if they are
expressed in different units of measurement. We can not use the standard deviations as
such for comparing their variability. We have to use the relative measures of dispersion in
such situations.
Coefficient of Variation (C.V.)
Coefficient of variation is the percentage ratio of standard deviation and the
arithmetic mean. It is usually expressed in percentage. The formula for C.V. is,
C.V. =
Mean
DS ..
x100
Where S.D. =






















 

2
1
2
1
1
n
x
x
n
n
i
in
i
i and
n
x
Mean
n
i
i
 1
The coefficient of variation will be small if the variation is small of the two groups,
the one with less C.V. said to be more consistent.
Note: 1. Standard deviation is absolute measure of dispersion
2. Coefficient of variation is relative measure of dispersion.

24
Example: Consider the distribution of the yields (per plot) of two ground nut varieties.
For the first variety, the mean and standard deviation are 82 kg and 16 kg respectively.
For the second variety, the mean and standard deviation are 55 kg and 8 kg respectively.
Then we have, for the first variety
C.V. = 100
82
16
x = 19.5%
For the second variety
C.V. = 100
55
8
x = 14.5%
It is apparent that the variability in second variety is less as compared to that in the first
variety. But in terms of standard deviation the interpretation could be reverse.
Example: Below are the scores of two cricketers in 10 innings. Find who is more
„consistent scorer‟ by Indirect method.
A 204 68 150 30 70 95 60 76 24 19
B 99 190 130 94 80 89 69 85 65 40
Solution: Let the player A = x
And the player B = y
Coefficient of variation of x= xVC .).( = 100X
x
x
Where x =










n
d
A
x
where Axd ix 
Standard deviation of x = x
 








 
2
21
n
dx
d
n
x
and coefficient of variation of y = yVC .).( = 100X
y
y
Standard deviation of y = y
 








 
2
21
n
dy
d
n
y
Where y =










n
d
B
y
where Byd iy 

25
Here A = 150 and B = 190
xi yi dx=xi –A dy=yi –B dxi
2
dyi
2
204 99 54 -91 2916 8281
68 190 -82 0 6724 0
150 130 0 -60 0 3600
30 94 -120 -96 14400 9216
70 80 -80 -110 6400 12100
95 89 -55 -101 3025 10201
60 69 -90 -121 8100 14641
76 85 -74 -105 5476 11025
24 65 -126 -125 15876 15625
19 40 -131 -150 17161 22500
Σ dx = -704 Σdy = -959 Σdx
2
= 80078 Σdy
2
=107189
x = 




 

10
704
150 y = 




 

10
959
190
= 79.6 runs = 94.1 runs
x
 





 

2
10
704
80078
10
1
y
 





 

10
959
107189
10
1
2
=  6.4956180078
10
1
 =  1.91968107189
10
1

= 55.24 runs = 39.01 runs
xVC .).( = 100
6.79
24.55
X yVC .).( = 100
1.94
01.39
x
= 69.4% = 41.46%
Coefficient of variation of A is greater than coefficient of variation of B and hence we
conclude that coefficient of player B is more consistent

26
NORMAL DISTRIBUTION
The Normal Distribution (N.D.) was first discovered by De-Moivre as the limiting
form of the binomial model in 1733, later independently worked Laplace and Gauss.
The Normal distribution is „probably‟ the most important distribution in statistics.
It is a probability distribution of a continuous random variable and is often used to model
the distribution of discrete random variable as well as the distribution of other continuous
random variables. The basic from of normal distribution is that of a bell, it has single
mode and is symmetric about its central values. The flexibility of using normal
distribution is due to the fact that the curve may be centered over any number on the real
line and it may be flat or peaked to correspond to the amount of dispersion in the values of
random variable.
Definition: A random variable X is said to follow a Normal Distribution with parameter 
and and 2
if its density function is given by the probability law
f(x) =
2
2
2
)(
2
1 




x
e - < x < ; - <  < ;  > 0
where  = a mathematical constant equality = 22/7
e = Naperian base equaling 2.7183
 = population mean
 = population standard deviation
x = a given value of the random variable in the range - < x <
Characteristics of Normal distribution and normal curve:
The normal probability curve with mean  and standard deviation  is given by the
equation
f(x) =
2
2
2
)(
2
1 




x
e ; - < x <
and has the following properties
i. The curve is bell shaped and symmetrical, about the mean 
ii. The height of normal curve is at its maximum at the mean. Hence the mean
and mode of normal distribution coincides. Also the number of observations
below the mean in a normal distribution is equal to the number of observations
about the mean. Hence mean and median of N.D. coincides. Thus, N.D. has
Mean = median = mode

27
iii. As „x‟ increases numerically, f(x) decreases rapidly, the maximum probability
occurring at the point x = , and given by
p[(x)]max =
 2
1
iv. Skewness 1 = 3
2
2
3


= 0
v. Kurtosis = 2 = 2
2
4


= 3 ( 432 ,,  andi are called central moments)
vi. All odd central moments are zero‟s
i.e. 0..............531  
vii. The first and third quartiles are equidistant from the median
viii. Linear combination of independent normal variates is also a normal variate
ix. The points of inflexion of the curve is given by






 2
1
2
1
)(, exfx


x. If 


1)( dxxf then
the area under the normal curve is distributed as follows
i) - < x < + covers 68.26% of area
ii) -2 < x < +2 covers 95.44% of area
iii) -3 < x < +3 coves 99.73% of area
Area under Normal curve

28
The Normal Curve: The graph of the normal distribution depends on two factors - the
mean and the standard deviation. The mean of the distribution determines the location of
the center of the graph, and the standard deviation determines the height and width of the
graph. When the standard deviation is large, the curve is short and wide; when the
standard deviation is small, the curve is tall and narrow. All normal distributions look like
a symmetric, bell-shaped curve, as shown below.
The curve on the left is shorter and wider than the curve on the right, because the curve on
the left has a bigger standard deviation.
Standard Normal Distribution: If „X‟ is a normal random variable with Mean  and
standard deviation  , then Z =

X
is a standard normal variate with zero mean and
standard deviation = 1.
The probability density function of standard normal variate „z‟ is
f(z) = 2
2
2
1
z
e


and 




dzzf )( =1
A graph representing the density function of the Normal probability distribution is also
known as a Normal Curve or a Bell Curve (see Figure below). To draw such a curve, one
needs to specify two parameters, the mean and the standard deviation. The graph below
has a mean of zero and a standard deviation of 1, i.e., (m =0, s =1). A Normal distribution
with a mean of zero and a standard deviation of 1 is also known as the Standard Normal
Distribution.
Standard Normal Distribution

29
Testing of Hypothesis
Introduction: The estimate based on sample values do not equal to the true value in the
population due to inherent variation in the population. The samples drawn will have
different estimates compared to the true value. It has to be verified that whether the
difference between the sample estimate and the population value is due to sampling
fluctuation or real difference. If the difference is due to sampling fluctuation only it can
be safely said that the sample belongs to the population under question and if the
difference is real we have every reason to believe that sample may not belong to the
population under question. The following are a few technical terms in this context.
Hypothesis: The assumption made about any unknown characteristics is called hypothesis.
It may or may be true.
Ex: 1.  = 2.3;  be the population mean
2.  = 2.1 ;  be the population standard deviation
Population follows Normal Distribution. There are two types of hypothesis, namely null
hypothesis and alternative hypothesis.
Null Hypothesis: Null hypothesis is the statement about the parameters. Such a
hypothesis, which is usually a hypothesis of no difference is called null hypothesis and is
usually denoted by H0.
or
any statistical hypothesis under test is called null hypothesis. It is denoted by H0.
1. H0:  = 0
2. H0: 1 = 2
Alternative Hypothesis: Any hypothesis, which is complementary to the null hypothesis,
is called an alternative hypothesis, usually denoted by H1.
Ex: 1. H1:  # 0
2. H1: 1 # 2
Parameter: A characteristics of population values is known as parameter. For example,
population mean () and population variance (2
).

30
In practice, if parameter values are not known and the estimates based on the sample
values are generally used.
Statistic: A Characteristics of sample values is called a statistic. For example, sample
mean ( x ), sample variance (s2
) where x =
n
xxx n ..........21
and s2
=






















n
i
n
i
i
i
n
x
x
n 1
1
2
21
Sampling distribution: The distribution of a statistic computed from all possible samples
is known as sampling distribution of that statistic.
Standard error: The standard deviation of the sampling distribution of a statistic is
known as its standard error, abbreviated as S.E.
S.E.( x ) =
n

; where  = population standard deviation and n = sample size
Sample: A finite subset of statistical objects in a population is called a sample and the
number of objects in a sample is called the sample size.
Population: In a statistical investigation the interest usually lies in the assessment of the
general magnitude and the study of variation with respect to one or more characteristics
relating to objects belonging to a group. This group of objects under study is called
population or universe.
Random sampling: If the sampling units in a population are drawn independently with
equal chance, to be included in the sample then the sampling will be called random
sampling. It is also referred as simple random sampling and denoted as SRS. Thus, if the
population consists of „N‟ units the chance of selecting any unit is 1/N. A theoretical
definition of SRS is as follows: Suppose we draw a sample of size „n‟ from a population
size N; then there are 





Nn
possible samples of size „n‟. If all possible samples have an
equal chance,






Nn
1
of being drawn, then the sampling is said to be simple random
sampling.
Simple Hypothesis: A hypothesis is said to be simple if it completely specifies the
distribution of the population. For instance, in case of normal population with mean 

31
and standard deviation  , a simple null hypothesis is of the form H0:  = 0 ,  is
known, knowledge about  would be enough to understand the entire distribution. For
such a test, the probability of committing the type-1 error is expressed as exactly  .
Composite Hypothesis: If the hypothesis does not specify the distribution of the
population completely, it is said to be a composite hypothesis. Following are some
examples:
H0 :  0 and  is known
H0 :  0 and  is known
All these are composite because none of them specifies the distribution completely.
Hence, for such a test the LOS is specified not as  but as „at most  ‟.
Types of Errors: In testing of statistical hypothesis there are four possible types of
decisions
1. Rejecting H0 when H0 is true
2. Rejecting H0 when H0 is false
3. Accepting H0 when H0 is true
4. Accepting H0 when H0 is false
1 and 4th
possibilities leads to error decisions. Statistician gives specific names to
these concepts namely Type-I error and Type-II error respectively.
the above decisions can be arranged in the following table
H0 is true H0 is false
Rejecting H0 Type-I error
(Wrong decision)
Correct
Accepting H0 Correct Type-II error
Type-I error: Rejecting H0 when H0 is true
Type-II error: Accepting H0 when H0 is false
The probabilities of type-I and type-II errors are denoted by  and  respectively.
Degrees of freedom: It is defined as the difference between the total number of items and
the total number of constraints.
If „n‟ is the total number of items and „k‟ the total number of constraints then the degrees
of freedom (d.f.) is given by d.f. = n-k
Level of significance(LOS): The maximum probability at which we would be willing to
risk a type-I error is known as level of significance or the size of Type-I error is level of
significance. The level of significance usually employed in testing of hypothesis are 5%

32
and 1%. The Level of significance is always fixed in advance before collecting the sample
information. LOS 5% means the results obtained will be true is 95% out of 100 cases and
the results may be wrong is 5 out of 100 cases.
Critical value: while testing for the difference between the means of two populations, our
concern is whether the observed difference is too large to believe that it has occurred just
by chance. But then the question is how much difference should be treated as too large?
Based on sampling distribution of the means, it is possible to define a cut-off or threshold
value such that if the difference exceeds this value, we say that it is not an occurrence by
chance and hence there is sufficient evidence to claim that the means are different. Such a
value is called the critical value and it is based on the level of significance.
Steps involved in test of hypothesis:
1. The null and alternative hypothesis will be formulated
2. Test statistic will be constructed
3. Level of significance will be fixed
4. The table (critical) values will be found out from the tables for a given level of
significance
5. The null hypothesis will be rejected at the given level of significance if the value of
test statistic is greater than or equal to the critical value. Otherwise null hypothesis
will be accepted.
6. In the case of rejection the variation in the estimates will be called „significant‟
variation. In the case of acceptance the variation in the estimates will be called „not-
significant‟.

33
STANDARD NORMAL DEVIATE TESTS
OR
LARGE SAMPLE TESTS
If the sample size n >30 then it is considered as large sample and if the sample size n< 30
then it is considered as small sample.
SND Test or One Sample (Z-test)
Case-I: Population standard deviation () is known
Assumptions:
1. Population is normally distributed
2. The sample is drawn at random
Conditions:
1. Population standard deviation  is known
2. Size of the sample is large (say n > 30)
Procedure: Let x1, x2, ………x,n be a random sample size of n from a normal population
with mean  and variance 2
 .
Let x be the sample mean of sample of size „n‟
Null hypothesis (H0): population mean (µ) is equal to a specified value 0
i.e. H0 :  = 0
Under H0, the test statistic is
Z =
n
x

0
~ N(0,1)
i.e the above statistic follows Normal Distribution with mean „0‟ and varaince‟1‟.
If the calculated value of Z < table value of Z at 5% level of significance, H0 is accepted
and hence we conclude that there is no significant difference between the population
mean and the one specified in H0 as 0.
Case-II: If  is not known
Assumptions:
1. Population is normally distributed
2. Sample is drawn from the population should be random
3. We should know the population mean

34
Conditions:
1. Population standard deviation  is not known
2. Size of the sample is large (say n > 30)
Null hypothesis (H0) :  = 0
under H0, the test statistic
Z =
n
s
x 0
~ N(0,1) where s =
 










2
21
n
x
x
n
i
i
x = Sample mean; n = sample size
and hence we conclude that there is no significant difference between the population mean
and the one specified in H0 otherwise we do not accept H0.
The table value of Z at 5% level of significance = 1.96 and table value of Z at 1% level of
significance = 2.58.
Two sample Z-Test or Test of significant for difference of means
Case-I: when  is known
Assumptions:
1. Populations are distributed normally
2. Samples are drawn independently and at random
Conditions:
1. Populations standard deviation  is known
2. Size of samples are large
Procedure: Let 1x be the mean of a random sample of size n1 from a population with mean
1 and variance
2
1
Let 2x be the mean of a random sample of size n2 from another population with mean 2
and variance
2
2
Null hypothesis H0 : 1 = 2
Alternative Hypothesis H1 : 1  2
i.e. The null hypothesis states that the population means of the two samples are identical.
Under the null hypothesis the test statistic becomes

35
Z =










2
2
2
1
2
1
21
nn
xx

~ N(0,1) )1(
i.e the above statistic follows Normal Distribution with mean „0‟ and varaince‟1‟.
If 1
2
= 2
2
= 2
(say) i.e both samples have the same standard deviation then the test
statistic becomes
Z =








21
21
11
nn
xx

~ N(0,1) )2(
otherwise rejected.
If H0 is accepted means, there is no significant difference between two population means
of the two samples are identical.
Example: The Average panicle length of 60 paddy plants in field No. 1 is 18.5. cms and
that of 70 paddy plants in field No. 2 is 20.3 cms. With common S.D. 1.15 cms. Test
whether there is significant difference between tow paddy fields w.r.t panicle length.
Solution:
Null hypothesis: H0: There is no significant difference between the means of two paddy
fields w.r.t. panicle length.
H0: 21  
Under H0, the test statistic becomes
Z =








21
21
11
nn
xx

~ N (0,1) ------------ (1)
Where 1x first field sample mean = 18.5 inches
2x second field sample mean = 20.3 inches
n1 = first sample size = 60
n2 = second sample size = 70
 = common S.D. = 1.15 inches
Substitute the given values in equation (1), we get

36
Z =
70
1
60
1
15.1
3.205.18


=
176.015.1
8.1
x
= 8.89
Calculated value of Z = 5.1
Cal. Value of Z > table value of Z at 5% LOS(1.96), H0 is rejected. This means, there is
highly significant difference between two paddy fields w.r.t. panicle length.
Case-II: when  is not known
Assumptions:
1. Populations are normally distributed
Conditions:
1. Population standard deviation  is not known
2. Size of samples are large
Null hypothesis H0 : 1 = 2
Under the null hypothesis the test statistic becomes
Z =










2
2
2
1
2
1
21
n
s
n
s
xx
~ N(0,1) )2(
Where 1x = 1st
sample mean , 2x = 2nd
sample mean
2
1s = 1st
sample variance,
2
2s = 2nd sample variance
n1 = 1st
sample size, n2 = 2nd
sample size
otherwise rejected.
Example:
A breeder claims that the number of filled grains per panicle is more in a new variety of
paddy ACM.5 compared to that of an old variety ADT.36. To verify his claim a random
sample of 50 plants of ACM.5 and 60 plants of ADT.36 were selected from the
experimental fields. The following results were obtained:
(ForACM.5) (For ADT.36)
1x 139.4-grains/panicle 2x 112.9 grains/panicle
s1 = 26.864 s2 = 20.1096
n1 = 50 n2 = 60
Test whether the claim of the breeder is correct.

37
Sol: Null hypothesis H0 : 21  
(i.e. the average number of filled grains per panicle is the same for both ACM.5 and
ADT.36)
Z =
2
2
2
1
2
1
21
n
s
n
s
xx


~ N (0,1) ------------ (1)
Where 1x first variety sample mean = 139.4 grains/panicle
2x Second variety sample mean = 112.9 grains/panicle
s1= first sample standard deviation = 26.864
s2= second sample standard deviation = 20.1096
n1 = first sample size = 50
and n2 = second sample size = 60
Substitute the given values in equation (1), we get
Z =
60
)1096.20(
50
)864.26(
9.1124.139
22


=
7399.64335.14
5.26

= 4.76
Calculated value of Z > Table value of Z at 5% LOS (1.96), H0 is rejected. We conclude
that the number of filled grains per panicle is significantly greater in ACM.5 than in
ADT.36

38
SMALL SAMPLE TESTS
The entire large sample theory was based on the application of “normal test”.
However, if the sample size „n‟ is small, the distribution of the various statistics, e.g., Z =
n
x

0
are far from normality and as such „normal test‟ cannot be applied if „n‟ is small.
In such cases exact sample tests, we use t-test pioneered by W.S. Gosset (1908) who wrote
under the pen name of student, and later on developed and extended by Prof. R.A. Fisher.
Student’s t-test: Let x1, x2,……….,xn be a random sample of size „n‟ has drawn from a
normal population with mean  and variance 2
then student‟s t – is defined by the
statistic
t =
n
s
x 0
where
n
x
x
n
i
i
 1
and s2
=


























n
i
n
i
i
i
n
x
x
n 1
2
12
1
1
this test statistic follows a t – distribution with (n-1) degrees of freedom (d.f.). To get the
critical value of t we have to refer the table for t-distribution against (n-1) d.f. and the
specific level of significance. Comparing the calculated value of t with critical value, we
can accept or reject the null hypothesis.
The Range of t – distribution is - to +.
One Sample t – test
One sample t-test is a statistical procedure that is used to know the population
mean and the known value of the population mean. In one sample t-test, we know the
population mean. We draw a random sample from the population and then compare the
sample mean with population mean and make a statistical decision as to whether or not the
sample mean is different from the population mean. In one sample t-test, sample size
should be less than 30.
Assumptions: 1. Population is normally distributed
2. Sample is drawn from the population and it should be random
3. We should know the population mean
Conditions: 1. Population S.D.  is not known
2. Size of the sample is small (<30).

39
Procedure: Let : Let x1, x2, …….,xn be a random sample of size „n‟ has drawn from a
normal population with mean  and variance 2
.
Null hypothesis (H0): population mean (µ) is equal to a specified value 0
i.e. H0:  = 0
t =
n
s
x 0
and follows student‟s t – distribution with (n-1) degrees of freedom.
where
n
x
x
n
i
i
 1
and s2
=
 












2
2
1
1
n
x
x
n
We now compare the calculated value of t with the tabulated value at certain level of
significance
If calculated value of t < table value of t at (n-1) d.f., the null hypothesis is accepted and
hence we conclude that there is no significant difference between the population mean and
the one specified in H0 as 0 .
Example: Based on field experiments, a new variety of greengram is expected to give an
yield of 13 quintals per hectare. The variety was tested on 12 randomly selected farmer
fields. The yields (quintal/hectare) were recorded as 14.3, 12.6, 13.7, 10.9,13.7, 12.0, 11.4,
12.0, 13.1, 12.6, 13.4 and 13.1. Do the results conform the expectation?
Solution:
Null Hypothesis: H0 :  = 0 = 13
i.e. the results conform the expectation
The test statistic becomes
t =
n
s
x 0
~t (n-1) d.f.
Where
n
x
x
 and s =











 n
x
x
n
i
i
2
2 )(
1
1
is an unbiased estimate of 

40
Let yield = xi (say)
xi xi
2
14.3 204.49
12.6 158.76
13.7 187.69
10.9 118.81
13.7 187.69
12 144
11.4 129.96
12 144
13.1 171.61
12.6 158.76
13.4 179.56
13.1 171.61
Σx = 152.8 Σx2
= 1956.94
12
8.152
x = 12.73
s = 






12
)8.152(
94.1956
11
1 2
= 1.01
t =
12
01.1
1373.12 
=
29.0
27.0
= 0.93 qa/h.
t-table value at (n-1) = 11d.f. at 5 percent level of significance is 2.20.
Calculated value of t < table value of t, H0 is accepted and we may conclude that the
results conform to the expectation.
t-test for Two Samples
Assumptions: 1. Populations are distributed normally
Conditions: 1. Standard deviations in the populations are same and not known
2. Size of the sample is small
Procedure: If two independent samples xi ( i = 1,2,….,n1) and yj ( j = 1,2, …..,n2) of sizes
n1and n2 have been drawn from two normal populations with means 1 and 2
respectively.
Null hypothesis H0 : 1 = 2

41
The null hypothesis states that the population means of the two groups are identical, so
their difference is zero.
Under H0, the test statistics is t =








21
11
nn
S
yx
Where S2
= =
   










 

2
2
2
1
2
2
21 2
1
n
y
y
n
x
x
nn
or
S2
= pooled variance =
2
)1()1(
21
2
22
2
11


nn
snsn
where s1
2
and s2
2
are the variances of the first and second samples respectively.
and
1
1
1
n
x
x
n
i
i
 and
2
1
2
n
y
y
n
j
j
 ; where x and y are the two sample means.
Which follows Student‟s t – distribution with (n1+n2-2) d.f.
If calculated value of t < table value of t with (n1+n2-2) d.f. at specified level of
significance, then the null hypothesis is accepted otherwise rejected.
Example: Two verities of potato plants (A and B) yielded tubes are shown in the
following table. Does the mean number of tubes of the variety „A‟ significantly differ
from that of variety B?
Tuber yield, kg/plant
Variety-A 2.2 2.5 1.9 2.6 2.6 2.3 1.8 2.0 2.1 2.4 2.3
Variety-B 2.8 2.5 2.7 3.0 3.1 2.3 2.4 3.2 2.5 2.9
Solution:
Hypothesis H0 : 1 = 2
i.e the mean number of tubes of the variety „A‟ significantly differ from the variety „B‟
Statistic t =








21
11
nn
S
yx
~ t fdnn .)2( 21 
n1 = 1st
sample size; n2 = 2nd
sample size
x = Mean of the first sample; y = mean of the second sample

42
x =
1n
x and
2n
y
y

Where 2
S =
   































2
2
2
1
2
2
21 2
1
n
y
y
n
x
x
nn
x y 2
x
2
y
2.20 2.80 4.84 7.84
2.50 2.50 6.25 6.25
1.90 2.70 3.61 7.29
2.60 3.00 6.76 9.00
2.60 3.10 6.76 9.61
2.30 2.30 5.29 5.29
1.80 2.40 3.24 5.76
2.00 3.20 4.00 10.24
2.10 2.50 4.41 6.25
2.40 2.90 5.76 8.41
2.30 Σy = 27.40 5.29 Σy2
= 75.94
Σx = 24.70 Σx2
= 56.21
x =
11
70.24
10
40.27
y
= 2.25 Kg = 2.74 Kg
Where 2
S = 



















 10
)40.27(
94.75
11
)70.24(
21.56
21011
1 22
=     07.7594.7546.5521.56
19
1

= 0.09 Kg2
S = 2
S = 0.3 Kg.
Test statistic t =








10
1
11
1
3.0
74.225.2
=
13.0
49.0
= 3.77
Calculated value of t = 3.77
Table value of t for 19 d.f. at 5 % level of significance is 2.09
Since the calculated value of t > table value of t, the null hypothesis is rejected and hence
we conclude that the mean number of tubes of the variety „A‟ significantly not differ from
the variety „B‟

43
Paired t – test
The paired t-test is generally used when measurements are taken from the same subject
before and after some manipulation such as injection of a drug. For example, you can use
a paired t test to determine the significance of a difference in blood pressure before and
after administration of an experimental pressor substance.
Assumptions: 1. Populations are distributed normally
Conditions: 1. Samples are related with each other
2. Sizes of the samples are small and equal
3. Standard deviations in the populations are equal and not known
Hypothesis H0: 0d
Under H0, the test statistic becomes,
t =
n
S
d
 t(n-1) d.f.
where
n
d
d
n
i
i
 1
and S2
=























 
 n
d
d
n
n
i
in
i
i
2
1
1
2
1
1
where S2
is variance of the deviations
n = sample size; where di = xi-yi ( i = 1,2,……,n)
If calculated value of t < table value of t for (n-1)d.f. at α% level of significance, then the
null hypothesis is accepted and hence we conclude that the two samples may belong to the
same population. Otherwise, the null hypothesis rejected.
Example: The average number of seeds set per pod in Lucerne were determined for top
flowers and bottom flowers in ten plants. The values observed were as follows:
Top
flowers
4.2 5.0 5.4 4.3 4.8 3.9 4.2 3.1 4.4 5.8
Bottom
flowers
4.6 3.5 4.8 3.0 4.1 4.4 3.6 3.8 3.2 2.2
Test whether there is any significant difference between the top and bottom flowers with
respect to average numbers of seeds set per pod.
Solution:
Null Hypothesis H0: d = 0

44
Under H0 becomes, the test statistic is
..~ )1( fdt
n
s
d
t n
Where
n
d
d
 and
 










 

2
22
1
1
n
d
d
n
s
x y d=x-y d2
4.2 4.6 -0.40 0.16
5.0 3.5 1.50 2.25
5.4 4.8 0.60 0.36
4.3 3.0 1.30 1.69
4.8 4.1 0.70 0.49
3.9 4.4 -0.50 0.25
4.2 3.6 0.60 0.36
3.1 3.8 -0.70 0.49
4.4 3.2 1.20 1.44
5.8 2.2 3.60 12.96
Σd = 7.90
Σd2
=
20.45
n
d
d
 =
10
90.7
= 0.79







10
)90.7(
45.20
9
1 2
2
s
= 2.27
s = 27.22
s
= 1.51
10
51.1
79.0
t
= 1.65
Calculated value of t = 1.65
Table value of t for 9 d.f. at 5% level of significance is 2.26
Calculated value of t < table value of t, the null hypothesis is accepted and we conclude
that there is no significant difference between the top and bottom flowers with respect to
average numbers of seeds set per pod.

45
F – Test
In agricultural experiments the performance of a treatment is assessed not only by
its mean but also by its variability. Hence, it is of interest to us to compare the variability
of two populations. In testing of hypothesis the equality of variances, the greater variance
is always placed in the Numerator and smaller variance is placed in the denominator.
F- test is used to test the equality of two population variances, equality of several
regression coefficients, ANOVA .
F- test was discovered by G.W. Snedecor. The range of F : 0 to ∞
Let 121 ......,,........., xnxx and 221 ,......., ynyy be the two independent random samples of sizes
n1 and n2 drawn from two normal populations ),(
2
11 N and ),(
2
22 N respectively.
S1
2
and S2
2
are the sample variances of the two samples.
Null hypothesis H0 :
2
2
2
1  
F = 2
2
2
1
S
S
where, S1
2
> S2
2
Which follows F-distribution with (n1-1, n2-1)d.f.
Where
 










 

1
2
2
1
2
1
1
1
n
x
x
n
S and S2
2
=
 












2
2
2
2 1
1
n
y
y
n
or F = 2
1
2
2
S
S
where S2
2
> S1
2
Which follows F-distribution with (n2-1, n1-1)d.f.
If calculated value of F < table value of F with (n2-1, n1-1)d.f at specified level of
significance, then the null hypothesis is accepted and hence we conclude that the variances
of the populations are homogeneous otherwise heterogeneous.
Example: The heights in meters of red gram plants with two types of irrigation in two
fields are as follows:
Tap water (x) 3.5 4.2 2.8 5.2 1.7 2.6 3.5 4.2 5.0 5.2
Saline water (y) 1.9 2.6 2.3 4.3 4.0 4.2 3.8 2.9 3.7
Test whether the variances of the two system of irrigation are homogeneous.
Solution:
H0: The variances of the two systems of irrigation are homogeneous.

46
i.e.
2
2
2
1  
F = 2
2
2
1
S
S
; (
2
2
2
1 SS  )
Where
2
1S = fist sample variance =
 










 
1
2
2
1 1
1
n
x
x
n
and
2
2S = second sample variance =
 










 
2
2
2
2 1
1
n
y
y
n
x y 2
x 2
y
3.5 1.9 12.25 3.61
4.2 2.6 17.64 6.76
2.8 2.3 7.84 5.29
5.2 4.3 27.04 18.49
1.7 4 2.89 16
2.6 4.2 6.76 17.64
3.5 3.8 12.25 14.44
4.2 2.9 17.64 8.41
5 3.7 25 13.69
5.2  y 29.7 27.04  2
y 104.33
x =37.9  2
x 156.35
2
1S = 






10
)9.37(
35.156
9
1 2
= 1.41 mt2
and
2
2S = 






9
)7.29(
33.104
8
1 2
= 0.79 mt2
F = 2
2
2
1
S
S
=
79.0
41.1
= 1.78
F calculated value = 1.78
Table value of F0.05 for (n1-1, n2-1) d.f. = 3.23
Calculated value of F < Table value of at 5% level of significance, H0 is accepted and
hence we conclude that the variances of the two systems of irrigation are homogeneous.
2
1
F
Fi
 or
2
1
1
F
F 

47
Chi-square (2
) test
The various tests of significance studied earlier such that as Z-test, t-test, F-test
were based on the assumption that the samples were drawn from normal population.
Under this assumption the various statistics were normally distributed. Since the
procedure of testing the significance requires the knowledge about the type of population
or parameters of population from which random samples have been drawn, these tests are
known as parametric tests.
But there are many practical situations in which the assumption of any kind about
the distribution of population or its parameter is not possible to make. The alternative
technique where no assumption about the distribution or about parameters of population is
made are known as non-parametric tests. Chi-square test is an example of the non
parametric test. Chi-square distribution is a distribution free test.
If Xi  N(0,1) then 
2
ix ~ n
2

Chi-square distribution was first discovered by Helmert in 1876 and later independently
by Karl Pearson in 1900. The range of chi-square distribution is 0 to ∞.
Measuremental data: the data obtained by actual measurement is called measuremental
data. For example, height, weight, age, income, area etc.,
Enumeration data: the data obtained by enumeration or counting is called enumeration
data. For example, number of blue flowers, number of intelligent boys, number of curled
leaves, etc.,
2
– test is used for enumeration data which generally relate to discrete variable where as
t-test and standard normal deviate tests are used for measure mental data which generally
relate to continuous variable.
2
–test can be used to know whether the given objects are segregating in a theoretical
ratio or whether the two attributes are independent in a contingency table.
The expression for 2
–test for goodness of fit
2
= 
n
i i
ii
E
EO
1
2
)(
where Oi = observed frequencies
Ei = expected frequencies
n = number of cells( or classes)
Which follows a chi-square distribution with (n-1) degrees of freedom

48
The null hypothesis H0 = the observed frequencies are in agreement with the expected
frequencies
If the calculated value of 2
< Table value of 2
with (n-1) d.f. at specified level of
significance (), we accept H0 otherwise we do not accept H0.
Conditions for the validity of 2
–test:
The validity of 2
-test of goodness of fit between theoretical and observed, the
following conditions must be satisfied.
i) The sample observations should be independent
ii) Constraints on the cell frequencies, if any, should be linear oi = ei
iii) N, the total frequency should be reasonably large, say greater than 50
iv) If any theoretical (expected) cell frequency is < 5, then for the application of chi-square
test it is pooled with the preceding or succeeding frequency so that the pooled frequency is
more than 5 and finally adjust for the d.f. lost in pooling.
Applications of Chi-square Test:
1. testing the independence of attributes
2. to test the goodness of fit
3. testing of linkage in genetic problems
4. comparison of sample variance with population variance
5. testing the homogeneity of variances
6. testing the homogeneity of correlation coefficient
Test for independence of two Attributes of (2x2) Contingency Table:
A characteristic which can not be measured but can only be classified to one of the
different levels of the character under consideration is called an attribute.
2x2 contingency table: When the individuals (objects) are classified into two categories
with respect to each of the two attributes then the table showing frequencies distributed
over 2x2 classes is called 2x2 contingency table.

49
Suppose the individuals are classified according to two attributes say intelligence (A) and
colour (B). The distribution of frequencies over cells is shown in the following table.
A
B
A1 A2 Row totals
B1 a B R1+(a+b)
B2 c D R2 = (c+d)
Column total C1 = (a+c) C2 = (b+d) N = (R1+R2) or (C1+C2)
Where R1 and R2 are the marginal totals of 1st
row and 2nd
row
C1 and C2 are the marginal totals of 1st
column and 2nd
column
N = grand total
The null hypothesis H0: the two attributes are independent ( if the colour is not dependent
on intelligent)
Based on above H0, the expected frequencies are calculated as follows.
E(a) =
N
xCR 11
; E(b) =
N
xCR 21
; E(c) =
N
xCR 12
; E(d) =
N
xCR 22
Where N = a+b+c+d
To test this hypothesis we use the test statistic
2
= 
n
i i
ii
E
EO
1
2
)(
the degrees of freedom for mxn contingency table is (m-1)x(n-1)
the degrees of freedom for 2x2 contingency table is (2-1)(2-1) = 1
This method is applied for all rxc contingency tables to get the expected frequencies.
The degrees of freedom for rxc contingency table is (r-1)x(c-1)
If the calculated value of 2
< table value of 2
at certain level of significance, then H0 is
accepted otherwise we do not accept H0
The alternative formula for calculating 2
in 2x2 contingency table is
2
=
2121
2
)(
xCxCxRR
Nbcad 
Example: Examine the following table showing the number of plants having certain
characters, test the hypothesis that the flower colour is independent of the shape of leaf.
Flower colour Shape of leaf Totals
Flat leaves Curled leaves
White flowers 99 (a) 36 (b) R1 = 135
Red flowers 20( c) 5 (d) R2 = 25
Totals C1 = 119 C2 = 41 N = 160

50
Solution:
Null hypothesis H0: attributes „flower colour‟ and „shape of leaf‟ are independent of each
other.
Under H0 the statistic is
i
n
i
ii
e
eo

 1
2
2
)(

where oi = observed frequency
and ei = expected frequency
Expected frequencies are calculated as follows.
E(a) =
160
119*135* 11

N
CR
= 100.40 where R1 and R2 = Row totals
E(b) =
160
41*135* 21

N
CR
=34.59 C1 and C2 = column totals
E( c ) =
160
119*25* 12

N
CR
=18.59 N= Grand totals
E(d) =
160
41*25* 22

N
CR
=6.406
io ie ii eo  (oi- ei)2
i
ii
e
)2e-(o
99 100.40 -1.4 1.96 0.02
36 34.59 1.41 1.99 0.06
20 18.59 1.41 1.99 0.11
5 6.41 -1.41 1.99 0.31
i
n
i
ii
e
eo

1
2
)(
=0.49
Calculated value of
i
n
i
ii
e
eo

 1
2
2
)(
 = 0.49
Direct Method:
Statistic: 2
 =
2121
2
)(
CCRR
bcadN 
here a = 99, b = 36, c = 20 and d = 5 and N = 160
2
 =
41*119*25*135
)20*365*99(160 2


51
=
25.164666
50625*160
=
16466625
18100000
= 0.49
Calculated value of 2
 = 0.40
Table value of 2
 for (2-1) (2-1) = 1 d.f. is 3.84
 < Table value of 2
 at 5% LOS for 1 d.f. , Null hypothesis is
accepted and hence we conclude that two characters, flower colour and shape of leaf are
independent of each other.
Yates correction for continuity in a 2x2 contingency table:
In a 2x2 contingency table, the number of d.f. is (2-1)(2-1) = 1. If any one of
Expected cell frequency is less than 5, then we use of pooling method for 2
–test results
with `0‟ d.f. (since 1 d.f. is lost in pooling) which is meaningless. In this case we apply a
correction due to Yates, which is usually known a Yates correction for continuity.
Yates correction consists of the following steps; (1) add 0.5 to the cell frequency which is
the least, (2) adjust the remaining cell frequencies in such a way that the row and column
totals are not changed. It can be shown that this correction will result in the formula.
2
(corrected) =
2121
2
2
CCRR
N
bcadN 




Example: The following data are observed for hybrids of Datura.
Flowers violet, fruits prickly =47
Flowers violet, fruits smooth = 12
Flowers white, fruits prickly = 21
Flowers white, fruits smooth = 3. Using chi-square test, find the association between
colour of flowers and character of fruits.
Sol: H0: The two attributes colour of flowers and fruits are independent.
We cannot use Yate‟s correction for continuity based on observed values. If only
expected frequency less than 5, we use Yates‟s correction for continuity.
The test statistic is

52
2
 (corrected) =
2
2121
2
CCRR
N
bcadN 






Flowers Violet Flowers white Total
Fruits Prickly 47(48.34) 21(19.66) 68
Fruits smooth 12(10.66) 3(4.34) 15
Total 59 24 83
The figures in the brackets are the expected frequencies
)(
2
corrected =
2
24*59*15*68
2
83
)12*21()3*47(83 






=
 2
24*59*15*68
5.4125214183 
=
1444320
75.400910
= 0.28
 = 0.28
Table value of 2
 for (2-1) (2-1) = 1 d.f. is 3.84
 < table value of 2
 , H0 is accepted and hence we conclude that
colour of flowers and character of fruits are not associated
CORRELATION
When there are two continuous variables which are concomitant their joint
distribution is known as bivariate normal distribution. If there are more than two such
variables their joint distribution is known as multivariate normal distributions. In case of
bivariate or multivariate normal distributions, we may be interested in discovering and
measuring the magnitude and direction of the relationship between two or more variables.
For this purpose we use the statistical tool known as correlation.
Definition: If the change in one variable affects a change in the other variable, the two
variables are said to be correlated and the degree of association ship (or extent of the
relationship) is known as correlation.
Types of correlation:
a). Positive correlation: If the two variables deviate in the same direction, i.e., if the
increase (or decrease) in one variable results in a corresponding increase (or decrease) in
the other variable, correlation is said to be direct or positive.

53
Ex: (i) Heights and weights
(ii) Household income and expenditure
(iii) Amount of rainfall and yield of crops
(iv) Prices and supply of commodities
(v) Feed and milk yield of an animal
(vi) Soluble nitrogen and total chlorophyll in the leaves of paddy.
b). Negative correlation: If the two variables constantly deviate in the opposite direction
i.e., if increase (or decrease) in one variable results in corresponding decrease (or
increase) in the other variable, correlation is said to be inverse or negative.
Ex: (i) Price and demand of a goods
(ii) Volume and pressure of perfect gas
(iii) Sales of woolen garments and the day temperature
(iv) Yield of crop and plant infestation
c) No or Zero Correlation: If there is no relationship between the two variables such that
the value of one variable change and the other variable remain constant is called no or zero
correlation.
Figures:
Methods of studying correlation: 1. Scatter Diagram 2. Karl Pearson‟s Coefficient of
Correlation 3. Spearman‟s Rank Correlation 4. Regression Lines
1. Scatter diagram: It is the simplest way of the diagrammatic representation of bivariate
data. Thus for the bivariate distribution (xi,yi); i = j = 1,2,…n, If the values of the
variables X and Y be plotted along the X-axis and Y-axis respectively in the xy-plane, the
diagram of dots so obtained is known as scatter diagram. From the scatter diagram, if the
points are very close to each other, we should expect a fairly good amount of correlation

54
between the variables and if the points are widely scattered, a poor correlation is expected.
This method, however, is not suitable if the number of observations is fairly large.
If the plotted points shows an upward trend of a straight line then we say that both the
variables are positively correlated
Positive Correlation
When the plotted points shows a downward trend of a straight line then we say that both
the variables are negatively correlated
Negative Correlation
If the plotted points spread on whole of the graph sheet, then we say that both the variables
are not correlated.
No Correlation

55
Karl Pearson’s Coefficient of Correlation: Prof. Karl Pearson, a British Biometrician
suggested a measure of correlation between two variables. It is known as Karl Pearson‟s
coefficient of correlation. It is useful for measuring the degree of linear relationship
between the two variables X and Y. It is usually denoted by rxy or „r‟.
i) Direct Method: rxy =
)()(
),cov(
yVxV
yx
=
2
1
2
1
1
)(
1
)(
1
))((
1






n
i
i
n
i
i
i
n
i
i
yy
n
xx
n
yyxx
n
After simplification =
 
    



22
)()
yyxx
yyxx
or
r =





 


}
)(
{}
)(
{
))((
2
2
2
2
n
y
y
n
x
x
n
yx
xy
ii) Deviation method:
r =





 


}
)(
{}
)(
{
))((
2
2
2
2
n
dy
dy
n
dx
dx
n
dydx
dxdy
Where x = S.D. of x and y = S.D. of Y
n = number of items; dx = x-A, dy = y-B
A = assumed value of and B = assumed value of y
The correlation coefficient never exceed unity. It always lies between –1 and +1 (i.e. –1
 r  1)
If r = +1 then we say that there is a perfect positive correlation between x and y
If r = -1 then we say that there is a perfect negative correlation between x and y
If r = 0 then the two variables x and y are called uncorrelated variables
No unit of measurement.

56
Test for significance of correlation coefficient
If „r‟ is the observed correlation coefficient in a sample of „n‟ pairs of observations from a
bivariate normal population, then Prof. Fisher proved that under the null hypothesis
H0 :  = 0
the variables x, y follow a bivariate normal distribution. If the population correlation
coefficient of x and y is denoted by , then it is often of interest to test whether  is zero or
different from zero, on the basis of observed correlation coefficient „r‟. Thus if „r‟ is the
sample correlation coefficient based on a sample of „n‟ observations, then the appropriate
test statistic for testing the null hypothesis H0 :  = 0 against the alternative hypothesis H1:
 # 0 is
t = )2(
1 2


n
r
r
t follows Student‟s t – distribution with (n-2) d.f.
If calculated value of t > table value of t with (n-2) d.f. at specified level of significance,
then the null hypothesis is rejected. That is, there may be significant correlation between
the two variables. Otherwise, the null hypothesis is accepted.
Example: From a paddy field, 12 plants were selected at random. The length of panicles
in cm (x) and the number of grains per panicle (y) of the selected plants were recorded.
The results are given in the following table. Calculate correlation coefficient and its
testing.
y 112 131 147 90 110 106 127 145 85 94 142 111
x 22.9 23.9 24.8 21.2 22.2 22.7 23.0 24.0 20.6 21.0 24.0 23.1
Solution:
a) Direct Method:
Correlation coefficient rxy = r =





 


}
)(
{}
)(
{
))((
2
2
2
2
n
y
y
n
x
x
n
yx
xy
Where n = number of observations
Testing the correlation coefficient:
Null hypothesis H0: Population correlation coefficient „‟ = 0

57
t =
2
1
2
r
nr


~ )2( nt d.f.
x y
2
x 2
y xy
112 22.9 12544 524.41 2564.8
131 23.9 17161 571.21 3130.9
147 24.8 21609 615.04 3645.6
90 21.2 8100 449.44 1908.0
110 22.2 12100 492.84 2442.0
106 22.7 11236 515.29 2406.2
127 23.0 16129 529.00 2921.0
145 24.0 21025 576.00 3480.0
85 20.6 7225 424.36 1751.0
94 21.0 8836 441.00 1974.0
142 24.0 20164 576.00 3408.0
111 23.1 12321 533.61 2564.1
 x 1400  y 273.4  2
x =168450  2
y 6248.20  xy 32195.6
r =














12
)4.273(
20.6248
12
)1400(
168450
12
)4.273)1400(
6.32195
22
x
=
2367.195308.71
9333.298
x
= 0.95
The value of coefficient of determination (r2
) = 0.90
t =
2
)95.0(1
21295.0


=
3122.0
0042.3
= 9.6
t critical (table) value for 10 d.f. at 5% LOS is 2.23
Since calculated value i.e. 9.6 is > t table value i.e. 2.23, it can be inferred that there exists
significant positive correlation between (x, y).

58
b) Indirect Method:
Here A = 127 and B = 24
x y Axdx  Bydy  2
xd
2
yd
yxdd
112 22.9 -15 -1.1 225 1.21 16.5
131 23.9 4 -0.1 16 0.01 -0.4
147 24.8 20 0.8 400 0.64 16
90 21.2 -37 -2.8 1369 7.84 103.6
110 22.2 -17 -1.8 289 3.24 30.6
106 22.7 -21 -1.3 441 1.69 27.3
127 23 0 -1 0 1 0
145 24 18 0 324 0 0
85 20.6 -42 -3.4 1764 11.56 142.8
94 21 -33 -3 1089 9 99
142 24 15 0 225 0 0
111 23.1 -16 -0.9 256 0.81 14.4
 xd =
-124
 yd =
-14.6

2
xd =
6398

2
dd =
37
 yxdd =
449.8
r =
  


 




n
d
d
n
d
d
n
dd
dd
y
y
x
x
yx
yx
2
2
2
2 )()(
=
12
)6.14(
37
12
)124(
6398
12
)6.14)(124(
8.449
22





=
2367.196667.5116
9333.298
x
=
3860.45309.71
9333.298
x
= 0.95

59
REGRESSION
The term `regression‟ literally means “stepping back towards the average”. It was first
used by a British Biometrician Sir Francis Galton.
The relationship between the independent and dependent variables may be expressed as a
function. Such functional relationship between two variables is termed as regression.
In regression analysis independent variable is also known as regressor or
predictor or explanatory variable while dependent variable is also known as regressed or
explained variable.
When only two variables are involved the functional relationship is known as
simple regression. If the relationship between two variables is a straight line, it is known
as simple linear regression, otherwise it is called as simple non-linear regression.
Direct Method:
The regression equation of Y on X is given as
Y= a+bX
Where Y= dependent variable; X = independent variable and a = intercept
b = the regression coefficient (or slope) of the line. a and b are also called as
Constants
the constants a and b can be estimated with by applying the `least squares method‟. This
involves minimizing    
n
i
i
i
bxaye
22
. This gives
byx = b =
)(
)(
Xv
XYCov
=



 


n
X
X
n
YX
XY
2
2
)(
))((
and a = XbY 
where b is called the estimate of regression coefficient of Y on X and it measures the
change in Y for a unit change in X.
Similarly, the regression equation of X on Y is given as
X = a1
+b1
Y
where X = dependent variable and Y = independent variable
bxy = b1
= =
)(
)(
Yv
XYCov
=



 


n
Y
Y
n
YX
XY
2
2
)(
))((

60
and a1
= X - b1
Y
where b1
is known as the estimate of regression coefficient of X on Y and `a‟ is intercept
Deviation Method:
The regression equation of Y on X is
(Y-Y ) = b (X- X )
 Y = )( XXbY 
The regression equation of X on Y
(X- X ) = b1
(Y-Y )
 X = X + b1
(Y-Y )
Properties of regression coefficient:
1. Correlation coefficient is the geometric mean of the two regression coefficients.
i.e r =  1
.bb
2. If one of the regression coefficient is greater than unity, the other must be less than
unity.
3. Arithmetic mean of the regression coefficients is greater than the correlation
coefficient `r‟.
4. Regression coefficients are independent of the change of origin but not of scale.
5. Units of „b‟ are same as that of the dependent variable.
6. Regression is only a one-way relationship between y (dependent variable) and x
(independent variable).
7. The range of b is from -∞ to ∞. - ∞ for negative b and +∞ for positive b.
Note:
1. Both the lines regression pass through the point ( YX, ). In other words, the
mean values ( YX, ) can be obtained as the point of intersection of the two regression lines
2. If r = 0, the two variables are uncorrelated, the lines of regression become
perpendicular to each other
3. If r =  1, in this case the two lines of regression either coincide or they are parallel to
each other
4.If the regression coefficients are positive, r is positive and if the regression
coefficients are negative, r is negative

61
Distinguish between correlation and regression:
Correlation Regression
1. Correlation is the relationship between two
or more variables. Where the change in one
variable affects a change in other variable
2. correlation coefficient measures extent of
relationship between two variables
3. correlation is a two way relationship
4. correlation coefficient is independent of
units of the variables.
5. correlation coefficient always lies
between –1 and +1
6. In correlation both the variables are
random
7. correlation coefficient calculated by
rxy =
)()(
),cov(
yVxV
yx
8. r = 1
bb
we can not predict
1. Regression is mathematical
measure of the average
relationship between two or more
variables. Where one variable is
dependent and other variable is
independent
2. regression coefficient estimates
the change in one variable for a
unit change in other related
variable
3. regression is a one way
relationship
4. regression coefficient is
expressed in the units of
dependent variable
5. regression coefficient lies
between
- and +
6. In regression one variable will
be in dependent and other will be
dependent
7. the regression coefficient of y on
x is b =



 


n
x
x
n
yx
xy
2
2
)(
))((
the regression coefficient of x on y is
b1
=



 


n
y
y
n
yx
xy
2
2
)(
))((
b = r
 
 



2
2
xx
yy
;we can predict

62
Example: The following data give the yield per plant (gm) (Y) and days to flowering (X)
of 11 pigeonpea plants. Fit regression equations of Y on X and X on Y. Also estimate Y
when the flowering period is 149.
Solution:
The regression equation of Y on X is given by
Y= a+bX
Where Y= dependent variable; X = independent variable and a = intercept
b = the regression coefficient of Y on X
The constants a and b can be estimated with the help of `least squares method‟
Where b =



 


n
X
X
n
YX
XY
2
2
)(
))((
; a = XbY 
Similarly, the regression equation of X on Y is given by
X = a1
+ b1
Y
Where X = dependent variable and Y = independent variable
Where b1
=



 


n
Y
Y
n
YX
XY
2
2
)(
))((
; and a1
= X - b1
Y
Y X Y2
X2
XY
24.72 91 611.08 8281 2249.52
20.25 76 410.06 5776 1539.00
38.56 98 1486.87 9604 3778.88
74.72 136 5583.08 18496 10161.92
72.75 128 5292.56 16384 9312.00
78.45 119 6154.40 14161 9335.55
69.8 142 4872.04 20164 9911.60
80.4 135 6464.16 18225 10854.00
160.2 147 25664.04 21609 23549.40
165.75 145 27473.06 21025 24033.75
77.56 122 6015.55 14884 9462.32
Y =863.16 X =1339
 2
Y =
90026.91
 2
X =
168609
XY =
114187.94
b=
11
)1339(
168609
11
)16.863)(1339(
94.114187
2



63
=
18.5616
83.9117
= 1.62
73.121
11
1339


n
X
X and
n
Y
Y
 =
11
16.863
= 78.47
and a = 78.47 – 1.62*121.73 = - 118.74
The regression equation of Y on X is Y = a+bX = Y = -118.74 + 1.62X
To estimate Y when X= 149
Yˆ = -118.74 + 1.62 * 149 = 122.64
b1
=
11
)16.863(
91.90026
11
)16.863)(1339(
94.114187
2


=
53.22295
83.9117
= 0.41
a1
= X - b1
Y = 121.73 – 0.41*78.47 = 89.56
The regression equation of X on Y is X = 89.56 + 0.41Y
Result:
1.The regression equation of Y on X is Y = a + bX = Y = -118.74 + 1.62X
2. The regression equation of X on Y is X = 89.56 + 0.41Y
3. Estimated Yield is 122.64

64
Analysis of Variance (ANOVA)
The ANOVA is a powerful statistical tool for tests of significance. The test of
significance based on t-distribution is an adequate procedure only for testing the
significance of the difference between two sample means. In a situation when we have
two or more samples to consider at a time, an alternative procedure is needed for testing
the hypothesis that all the samples have been drawn from the same population. For
example, if three fertilizers are to be compared to find their efficacy, this could be done by
a field experiment, in which each fertilizer is applied to 10 plots and then the 30 plots are
later harvested with the crop yield being calculated for each plot. Now we have 3 groups
of ten figures and we wish to know if there are any differences between these groups. The
answer to this problem is provided by the technique of ANOVA.
The term ANOVA was introduced by Prof. R.A. Fisher in 1920‟s to deal with
problem in the analysis of agronomical data. Variation is inherent in nature
The total variation in any set of numerical data is due to a number of causes which
may be classified as i) Assignable causes and ii) chance causes
The variation due to assignable causes can be detected and measured where as the
variation due to chance causes is beyond the control of humans and cannot be traced
separately
ANOVA: The ANOVA is a simple arithmetical process of sorting out the components of
variation in a given data
Types of ANOVA: There are two types i) One way classification and ii) Two way
classification
Assumptions of ANOVA:
1. The observations are independent
2. Parent population from which observations are taken is normal
3. Various treatment and environmental effects are additive in nature
4. The experimental errors are distributed normally with mean zero and variance 2
Experimental Designs
In order to verify a hypothesis pertaining to some scientific phenomena we have to
collect data. Such data are obtained by either observation or by experimentation. The
main topics connected with data collection are Theory of Sample Surveys and
Experimental Designs. In sample survey, a researcher makes observations on existing
population and records data without interfering with the process that is being observed. In

65
experimentation, on the other hand, the researcher controls or manipulates the
environment of the subjects that constitute the population. The experiments allow a
researcher to study the factors of his interest and show that these factors actually cause
certain effects. Hence, whenever the objective is to study the effects of variables rather
than simply to describe a population, we prefer the data collection through
experimentation.
Modern concepts of experimental design are due primarily to R.A. Fisher. He
developed them in the planning of agricultural field experiments. They are now used in
many fields of science.
Basic concepts:
Blocks: In agricultural experiments, most of the times we divide the whole experimental
unit (field) into relatively homogeneous sub-groups or strata. These strata, which are more
uniform amongst themselves than the field as a whole are known as blocks.
Treatments: the objects of comparison in an experiment are defined as treatments
For example: i) suppose an Agronomist wishes to know the effect of different
spacings on the yield of a crop, different spacings will be treatments. Each spacing
will be called a treatment.
ii) If different of fertilizer are tried in an experiment to test the responses of a crop
to the fertilizer doses, the different doses will be treatments and each dose will be a
treatment.
ii) A teacher practices different teaching methods on different groups in his class to
see which yields the best results.
iii) A doctor treats a patient with a skin condition with different creams to see
which is most effective.
Experimental unit: Experimental unit is the object to which treatment is applied to
record the observations.
For example i) In laboratory insects may be kept in groups of five or six. To each group,
different insecticides will be applied to know the efficacy of the insecticides. In this study
different groups of insects will be the experimental unit.
ii) If treatments are different varieties, then the objects to which treatments are applied to
make observations will be different plot of land. The plots will be called experimental
units.

66
Basic principles of experimental designs:
The purpose of designing an experiment is to increase the precision of the experiment. In
order to increase the precision, we try to reduce the experimental error. For reducing the
experimental error, we adopt some techniques. These techniques form the basic principles
of experimental designs. The basic principles of the experimental designs are replication,
randomization and local control.
1. Replication: Repetition of treatment to different experimental units is known as
Replication. In other words, the repetition of treatments under investigation is known as
replication. We have no means of knowing about the variations in the results of a
treatment. Only when we repeat the treatment several times we can estimate the
experimental error.
A replication is used (i) to secure more accurate estimate of the experimental error, a term
which represents the differences that would be observed if the same treatments were
applied several times to the same experimental units;
(ii) to reduce the experimental error and thereby to increase precision, which is a measure
of the variability of the experimental error.
The standard error of treatment mean is
r

. Where  is S.D. of treatment in the
population and `r‟ is the number of replications. As `r‟ increases, the standard error of
mean decreases. Also in the analysis of variance the replication of treatments provides
estimate of experimental error which is essential for the application of F-tes.
2. Randomization: when all the treatments have equal chances of being allocated to
different experimental units it is known as randomization
or
Random allocation of treatments to different experimental units known as randomization.
The purpose of randomization is to remove bias and other sources of extraneous variation
which are not controllable. Another advantage of randomization (accompanied by
replication) is that it forms the basis of any valid statistical test. Hence the treatments
must be assigned at random to the experimental units. Randomization is usually done by
using tables of random numbers.
3.Local control: It has been observed that all extraneous sources of variation are not
removed by randomization and replication. This necessitates a refinement in the
experimental technique. For this purpose, we make use of local control, a term referring
to the grouping of homogeneous experimental units.

67
The main purpose of the principle of local control is to increase the efficiency of an
experimental design by decreasing the experimental error.
Replication
II III
Random Local control
distribution
Validity of estimate Reducing the
of experimental error experimental error
Shape of blocks and plots: the shape and size of the blocks will usually depend up on the
shape and size of the plots. In order to control the experimental error it is desirable to
divide the whole experimental area into different subgroups (blocks) such that within each
block there is as much homogeneity as possible but between blocks there is maximum
variation. Further each block is to be divided into as many plots as the number of
treatments. For maximum precision the plots should be rectangular in shape with their
long sides parallel to the direction of the fertility gradient and the blocks should be
arranged one after the other along the fertility gradient as shown in the figure.
Fertility gradient
Block-I Block-II ……….. ………. Block-K
Novariationinthisdirection
Plots
Plots

68
COMPLETELY RANDOMIZED DESIGN (CRD)
The CRD is the simplest of all the designs. In this design, treatments are allocated
at random to the experimental units over the entire experimental material. In case of field
experiments, the whole field is divided into a required number of plots equal size and then
the treatments are randomized in these plots. Thus the randomization gives every
experimental unit an equal probability of receiving the treatment.
In field experiments there is generally large variation among experimental plots
due to soil heterogeneity. Hence, CRD is not preferred in field experiments. In laboratory
experiments and green house studies, it is easy to achieve homogeneity of experimental
materials. Therefore, CRD is most useful in such experiments.
Layout of CRD: The placement of the treatments on the experimental units along with the
arrangement of experimental units is known as the layout of an experiment.
For example, suppose that there are 5 treatments A,B,C,D and E. each with 4 replications,
we need 20 experimental units. Here, since the number of units is 20, a two digit random
number of table will be consulted and a series of 20 random numbers will be taken
excluding those which are greater than 20. suppose, the random numbers are
4,18,2,14,3,7,13,1,6,10,17,20,8,15,11,5,9,12,16,19. After this the plots will be serially
numbered and the treatment A will be allotted to the plots bearing the serial numbers 4,
18, 2, 14 and so on.
1
B
2
A
3
B
4
A
5
D
6
C
7
B
8
D
9
E
10
C
11
D
12
E
13
B
14
A
15
D
16
E
17
C
18
A
19
E
20
C

69
Statistical analysis:
Let us suppose that there are `k‟ treatments applied to `r‟ plots. These can be represented
by the symbols as follows:
Treatments 1….2…….. j………..n Totals means
t1
t2
.
.
.
ti
.
.
tk
y11 y12, … y1j… y1r
y21 y22, … y2j… y2r
. . . .
. . . .
. . . .
yi1 yi2, … yij…….. yir
. . . .
. . . .
yk1 yk2, … ykj… ykr
T1
T2
.
.
.
Ti
.
.
Tk
T 1
T 2
.
.
T i
.
.
T k
G.T.
Mathematical model:
yij =  + i + ij (i = 1,2,….k; j = 1,2,……r)
Where yij is the jth
replication of the ith
treatment
 = general mean effect
i = the effect due to ith
treatment =  iT
ij = error effect (ij  N(0,
2
))
Null hypothesis H0 : There is no significant difference between the treatment effects
 k  ....................21 =0
Where   ii , ( i = 1, 2,……..,k)
The null hypothesis can be verified by applying the ANOVA procedure. The steps
involved in this procedure are as follows:
1) Correction factor =
N
TG 2
.).(
2) Treatment Sum of Square (Tr.S.S.) =
 
n
TTT K
22
2
2
1 ............. 
- CF
= CF
n
T
k
i
i

1
2
3) Total Sum of Square (TSS) = { y11
2
+ y12
2
+ y13
2
+ …………+ ykn
2
} – CF
= yij
2
- CF
4) Error Sum of Square (ESS) = TSS-Tr.S.S.

70
ANOVA table
Sources of
variation
D.F. S.S. M.S. F-cal.value F-table value
Treatments
Error
k-1
N-k
Tr.S.S.
ESS
TMS =
1
...
k
SSTr
EMS=
kN
ESS

Ft =
EMS
TMS
-
F[k-1, N-k] at
 % LOS
Total N-1 TSS
If the calculated value of F < table value of F at certain level of significance, we accept H0
and hence we may conclude that there is no significant difference between the treatment
means
If the calculated value of F > table vale of F, H0 is rejected. Then the problem is to know
which of the treatment means are significantly different. For this, we calculate critical
difference (CD)
CD = SED x t – table value for error d.f. at 5% LOS.
where SED = Standard Error of Difference between the Treatments.
SED =
r
EMS2
(equal No. of replications), where r is number of replications
SED =









ji rr
EMS
11
, where i = 1,2,…..,k and j = 1,2,…..,k (unequal No. replications)
The treatment means are arranged first in descending order of magnitude. If the difference
between the two treatment means is less than CD value, it will he declared as non
significant otherwise significant
Advantages and disadvantages of CRD:
1. This design is most commonly used in laboratory experiments such as in Ag.
Chemistry, plant pathology, and animal experiments where the experimental
material is expected to be homogeneous.
2. This design is useful in pot cultural experiments where the same type of soil is
usually used. However, in greenhouse experiments care has to be taken with
regard to sunshade, accessibility of air along and across the bench before
conducting the experiment.

71
3. Any number of replications and treatments can be used. The number of
replications may vary from treatment to treatment.
4. The analysis remains simple even if information on some units are missing
5. This design provides maximum number of degrees of freedom for the estimation of
error than the other designs
6. The only draw back with this design is that when the experimental material is
heterogeneous, the experimental error would be inflated and consequently the
treatments are less precisely compared. The only way to keep the experimental
error under control is to increase the number of replications thereby increasing the
degrees of freedom for error.
Applications: 1. CRD is most useful in laboratory technique and methodological studies.
Ex: in physics, chemistry, in chemical and biological experiments, in some
greenhouse studies etc.
2. CRD is also recommended in situations where an appreciable fraction of
units is likely to be destroyed or fail to respond.
Case i) CRD with equal repetitions:
Example: In order to find out the yielding abilities of five varieties of sesamum an
experiment was conducted in the greenhouse using a completely randomized design
with four pots per variety. Analyze the data and state your conclusions
Seed yield of sesamum, g/pot
Varieties
1 2 3 4 5
8 10 18 12 8
8 12 17 10 11
6 13 13 15 9
10 9 16 11 8
Sol: H0: There is no significant difference between effect varities.
i.e. 321   = 4 = 5
Correction factor (CF) =
N
GT 2
)(
Total sum of squares (TSS) = {y11
2
+ y12
2
+ y13
2
+ …………+ y45
2
} – CF

72
Variety Sum of squares (VSS) =
r
vvvvv     2
5
2
4
2
3
2
2
2
1 )()()()()(
- CF
Error Sum of Square (ESS) = TSS- VSS
Varieties
1 2 3 4 5
8 10 18 12 8
8 12 17 10 11
6 13 13 15 9
10 9 16 11 8
Variety
totals
32 44 64 48 36 GT =
224
Means 8 11 16 12 9
CF =
20
)224( 2
= 2508.8
TSS = 2.2078.2508)8()11()16(..................)18()10()8{( 222222

VSS =
4
)36()48()64()44()32( 22222

- 2508.8 = 155.2
ESS= TSS-VSS = 207.2-155.2 = 52.0
ANOVA TABLE
Sources of
variation
d.f. S.S. M.S. F-cal value F- table value
Varieties 5-1 =4 155.20 38.80 11.19 F0.05(4,15) =3.06
Error 19-4=15 52.0 3.47
Total 20-1=19 207.20
Calculated value of F > Table value of F at 5%LOS, H0 is rejected and hence we conclude
that there is significant difference between variety means.
Critical difference (CD): SED x t0.05 for error d.f.
Where SED =
r
EMS2
t-table value for 15 d.f at 5%LOS is 2.13
CD =
4
47.3*2
x 2.13 = 2.80

73
The varieties means are arranging descending order of magnitude. If the difference
between the verities mean is less than CD value, it will be declared as non significant
otherwise significant
V3 V4 V2 V5 V1
16 12 11 9 8
________
________
________
The varieties which do not differ significantly have been underlined by a bar.
Coefficient of variation (CV) = Coefficient of variation =
X
EMS
x100
Where X = Grand mean
Coefficient of variation = 100
2.11
47.3
x = 16.6%
Conclusion: Lesser CV% indicates more consistency in the data
Case –ii) CRD with unequal repetitions:
Example: A Completely Randomized Design was conducted with the three treatments A
B and C where treatment A is replicated 6 times and B is replicated 4 times and C is
replicated 5 times. Analyze the data and state your conclusions.
A B C
16.5 15.0 18.2
17.0 13.8 24.3
16.0 14.0 25.0
12.0 17.9 18.9
18.0 - 21.0
14.0 - -
Sol: Null hypothesis H0: There is no significant difference between the effect treatments
i.e. 321  
First treatment A is replicated r1 = 6 times
Second treatment B is replicated r2 = 4 times
Third treatment C is replicated r3 = 5 times
Correction factor (CF) =
N
GT 2
)(
Total sum of squares (TSS) = {y11
2
+ y12
2
+ y13
2
+ …………+ y35
2
} – CF

74
Treatment of Sum of square (Tr.S.S.) =
3
2
2
2
1
2
)()()(
r
c
r
B
r
A   - CF
Error Sum of square(ESS) = TSS-Tr.S.S.
Standard error of difference = SED =









ji rr
EMS
11
Where i = 1, 2, ….., k and j = 1,2,….., r
Coefficient of variation =
X
EMS2
X100; where X = Grand mean
Treatments Totals Means
A 16.5 17.0 16.0 12.0 18.0 14.0 ΣA=93.5 15.58
B 15.0 13.8 14.0 17.9 - - ΣB=60.7 15.18
C 18.2 24.3 25.0 18.9 21.0 - ΣC=107.4 21.48
GT=261.6
CF =
15
)6.261( 2
= 4562.30
TSS = (16.5)2
+ (17.0)2
+ (16.0)2
+ (12.0)2
+ (18.0)2
+ (14.0)2
+ (15.0)2
+ (13.8)2
+ (14.0)2
+ (17.9)2
+ (18.2)2
+ (24.3)2
+ (25.0)2
+ (18.9)2
+ (21.0)2
– 4562.30
= 4758.04 – 4562.30
= 195.74
Tr.S.S. =
5
)4.107(
4
)7.60(
6
)5.93( 222
 - 4562.30
= 4685.11 – 4562.30
= 122.81
ESS = TSS-Tr.S.S.
= 195.74 – 122.81
= 72.93
AN0VA TABLE
Sources d.f S.S. M.S. F-cal. Value F- table Value
Treatments 3-1=2 122.81 61.405 10.10 F0.05(2,12) = 3.89
Error 12 72.93 6.0775
Total 15-1=14 195.74
Calculated value F(Tr) > Table value of F, H0 is rejected and hence we conclude that there
is significant difference between treatment means.

75
In CRD with unequal number of replications, we have to calculate CD for each pair of
treatment means. As there are 3 treatments, we have to calculate 3C2 = 3 CD values.
Treatment pair SED =









ji rr
EMS
11
CD = SEDxt0.05 d.f.
-----------------------------------------------------------------------------------------------------------
AB 






4
1
6
1
08.6 = 1.59 =1.59*2.18=3.47
AC 






5
1
6
1
08.6 = 1.49 =1.49*2.18=3.25
BC 






5
1
4
1
08.6 = 1.65 =1.65*2.18=3.61
Bar Notation:
The treatment means are arranged according to their ranks
T3 T1 T2
21.48 15.58 15.18
_________
i) Those pairs not scored are significant
ii) Those pairs under scored are non-significant
X
EMS
x100
where x = Grand mean
= 100
44.17
08.6
x =14.14%
Among three treatments, the third treatment i.e. T3 is found to be superior one
RANDOMIZED BLOCK DESIGN (RBD)
We have seen that in a completely randomized design no local control measure
was adopted excepting that the experimental units should be homogeneous. Usually,
when experiments require a large number of experimental units, completely randomized
designs cannot ensure precision of the estimates of treatment effects.
In agricultural field experiments, usually the experimental materials are not
homogeneous. In such situations the principle of local control is adopted and the
experimental material is grouped into homogeneous sub groups. The subgroup is

76
commonly termed as block. Since each block will consist the entire set of treatments a
block is equivalent to a replication.
The blocks are formed with units having common characteristics which may
influence the response under study. In agricultural field experiments the soil fertility is an
important character that influences the crop responses. The uniformity trial is used to
identify the soil fertility of a field. If the fertility gradient is found to run in one direction
(say from north to south) then the blocks are formed in the opposite direction (from east to
west).
If the number of experimental units within each group is same as the number of
treatments and if every treatment appears precisely once in each group, then such an
arrangement is called a randomized block design.
Layout: Let us consider 5 treatments A, B, C, D and E each replicated 4 times. We divide
the whole experimental area into 4 relatively homogeneous blocks and each block into 5
plots. Treatments are then allocated at random to the plots of a block, fresh randomization
being done for each block. A particular layout as follows.
Block-1 A E B D C
Block-2 E D C B A
Block-3 C B A E D
Block-4 A D E C B
Let us select one digit random numbers in the order of their occurrence in the table leaving
0 and greater than 5. suppose we get random numbers from 1 to 5 as: 1,3,5,4,2. So in the
First block we allocate treatment A to the 1st
plot and B to 3rd
plot and so on.
Statistical Analysis: The results from RBD can be arranged in two way table according to
the replications (blocks) and treatments; there will be `rk‟ observations in total. The data
can be arranged in the following table.

77
Treatments Blocks
b1…..b2 ..….bj……. ..br
Treatment
Totals
Means
t1
t2
.
.
.
ti
.
.
tk
y11 y12, … y1j………y1r
y21 y22, … y2j………y2r
. . . .
. . . .
. . . .
yi1 yi2, … yij…….. yir
. . . .
. . . .
yk1 yk2, … ykj……. ykr
T1
T2
.
.
.
Ti
.
.
Tk
T 1
T 2
.
.
T i
.
.
T k
Block
totals
B1 B2 ………Bj …. …Br G.T.
Means B 1 B 2 ……… B j …. . B r
Mathematical model:
Yij =  + ijji   (i = 1,2,….k; j = 1,2,……r)
Where yij is the response of the jth
block and ith
treatment
i = the effect due to ith
treatment
j = the effect due to jth
block
ij is the error effect (ij  N(0,
2
)
Null hypothesis: i) H01 : There is no significant difference between the treatment effects.
i.e. k  ....................21
ii) H02: There is no significant difference between the block effects
i.e. r  ....................21
The null hypothesis can be verified by applying the ANOVA procedure. The different
steps are in the analysis of data are:
1) Correction factor =
rk
TG 2
).(
2) Treatment Sum of Squares (Tr.S.S.) = CF
r
TTT K

 )(.............)()(
22
2
2
1
= CF
r
T
k
i
k

1
2

78
3) Block Sum of Squares (BSS) = CF
k
BBB r

 )(...........)()(
22
2
2
1
= CF
k
B
r
j
j

1
2
4) Total Sum of Squares (TSS) = {y11
2
+ y12
2
+ y13
2
+ ………+ykr
2
} – CF
=  

k
i
r
j
ij CFy
1 1
2
5) Error Sum of Square (ESS) = TSS -Tr.S.S .- SSB
ANOVA TABLE
Sources of
variation
D.F S.S M.S. F-Cal.Value F-table value
At 5% LOS
Treatments
Blocks
(Replications)
Error
k-1
r-1
(r-1)(k-1)
Tr.S.S.
BSS
ESS
TMS =
1
...
k
SSTr
BMS=
1r
BSS
EMS =
)1)(1(  kr
ESS
Ft =
EMS
TMS
Fb=
EMS
BMS
F[k-1,{(r-1)
(k-1)}]
F[r-1, {(r-1)
(k-1)}]
Total rk-1 TSS
If the calculated value of F (Treatments) < table value of F, we accept H0, and hence we
may conclude that there is no significant difference between the treatment means.
If calculated value of F (Treatments) > table value of F, we reject H0 and hence we may
conclude that there is significant difference between the treatment means.
If the treatments are significantly different, the comparison of the treatments is carried out
on the basis of Critical Difference (C.D.).

79
C.D. = SED(Tr) x t(r-1)(k-1) at  level of significance
Where SED =
r
xEMS2
i) F (Blocks) should be not significant, if the planning of experiment is
well manner.
ii) Desirable C.V. (%) in field experiment and lab experiment.
Advantages and disadvantages of RBD:
1. The principle advantage of RBD is that it increases the precision of the experiment.
This is due to the reduction of experimental error by adoption of local control.
2. The amount of information obtained in RBD is more as compared to CRD. Hence,
RBD is more efficient than CRD.
3. Flexibility is another advantage of RBD. Any number of replications can be
included in RBD. If large number of homogeneous units are available, large
number of treatments can be included in this design.
4. Since the layout of RBD involves equal replication of treatments, statistical
analysis is simple. Even when some observations are missing of certain
treatments, the data can be analysed by the use of missing plot technique.
5. When the number of treatments is increased, the block size will increase. If the
block size is large it may be difficult to maintain homogeneity within blocks.
Consequently, the experimental error will be increased. Hence, RBD may not be
suitable for large number of treatments. But for this disadvantage, the RBD is a
versatile design. It is the most frequently used design in agricultural experiments.
6. The optimum blocks size in field experiments is 21 plots. i.e. we can not compare
treatments which are > 21 in RBD to preserve homogeneity of plots, within a
block.
Example: The yields of 6 varieties of a crop in lbs., along with the plan of the experiment,
are given below. The number of blocks is 5, plot of size is 1/20 acre and the varieties have
been represented by A, B, C, D and E and analyze the data and state your conclusions
B-I B
12
E
26
D
10
C
15
A
26
F
62
B-II E
23
C
16
F
56
A
30
D
20
B
10
B-III A
28
B
9
E
35
F
64
D
23
C
14
B-IV F
75
D
20
E
30
C
14
B
7
A
23
B-V D
17
F
70
A
20
C
12
B
9
E
28

80
Solution:
Null hypothesis H01: There is no significant difference between variety means
4321   = 65  
H02: There is no significant difference between block means
54321  
Correction factor =
rk
TG 2
).(
Variety Sum of Squares due to varieties (VSS) = CF
r
vi


2
Block Sum of square(BSS) = CF
k
bj


2
Total sum of squares (TSS) = yij
2
- CF
Error Sum of Square (ESS) = TSS- VSS- BSS
First rearrange the given data
Blocks Varieties Block totals Means
A B C D E F
B1 26 12 15 10 26 62 ΣB1 = 151 25.17
B2 30 10 16 20 23 56 ΣB2 =155 25.83
B3 28 9 14 23 35 64 ΣB3 = 173 28.83
B4 23 7 14 20 30 75 ΣB4 = 169 28.17
B5 20 9 12 17 28 70 ΣB5 = 156 26.00
Variety
totals
ΣA =
127
ΣB =
47
ΣC =
71
ΣD =
90
ΣE =
142
ΣF =
327
GT = 804 -
Means 25.4 9.4 14.2 18 28.4 65.4 - -
CF =
30
)804( 2
= 21547.2
VSS = 2.21547
5
)327()142()90()71()47()127( 222222


= 31714.4 – 21547.2 = 10167.2
BSS = 2.21547
6
)156()169()173()155()151( 22222


= 21608.67 – 21547.2 = 61.47
TSS = (12)2
+ (26)2
+ ( 10)2
+ ( 15)2
+ + ………… + (12)2
+ (9)2
+ (28)2
- 21547.2

81
= 32194 – 21547.2
= 10646.8
ESS = TSS – BSS – Tr.S.S.
= 10646.8 – 61.47 – 10167.2
= 418.13
ANOVA TABLE
Sources of
variation
d.f S.S. M.S. F-cal. Value F- table Value
Blocks 5-1=4 61.47 15.37 0.74 F0.05 (4, 20) =2.87
Varieties 6-1=5 10167.2 2033.44 97.25 F0.05 (5, 20) = 2.71
Error 29-4-5= 20 418.13 20.91
Total 30-1-29 10646.8
Calculated value of F (Treatments) > Table value of F, H0 is rejected and hence we
conclude that there is highly significant difference between variety means.
Where SEm =
r
EMS
=
5
91.20
= 2.04
SED = 2 * SEm = 1.414 * 2.04 = 2.88
Critical difference = SED x t-table value for error d.f. at 5% LOS
 CD = 2.88 * 2.09
= 6.04
X
EMS
X100 =
8.26
91.20
X100 = 17%
Bar Notation:
F E A D C B
65.4 28.4 25.4 18.0 14.2 9.40
_________
_______
________
i) Those pairs not scored are significant
ii) Those pairs underscored are non-significant
Variety F gives significantly higher yield than all the other varieties; varieties D,C and B
are on par and gives significantly higher yield than variety A.

82
LATIN SQUARE DESIGN (LSD)
When the experimental material is divided into rows and columns and the
treatments are allocated such that each treatment occurs only once in a row and once in a
column, the design is known as latin square design. In this design eliminating fertility
variations consists in an experimental layout which will control variation in two
perpendicular directions
[Latin square designs are normally used in experiments where it is required to
remove the heterogeneity of experimental material in two directions. This design
requires that the number of replications (rows) equal the number of treatments]. In
LSD the number of rows and number of columns are equal. Hence the arrangement will
form a square.
Layout of LSD: In this design the number of rows is equal to the number of columns and
it is equal to the number of treatments. Thus in case of `m‟ treatments, there have to be
mxm = m2
experimental units (plots) arranged in a square so that each row as well as each
column contain `m‟ plots. The `m‟ treatments are then allocated at random to these rows
and columns in such a way that every treatment occurs once and only once in each row
and each column such a layout is known as mxm L.S.D and is extensively used in
agricultural experiments. The minimum and maximum number of treatments required for
layout of LSD is 5 to 12.
In LSD the treatments are usually denoted by alphabets like A,B,C…etc. For a
latin square with five treatments the arrangement may be as follows
Square –I Square – II……………………….. etc.,
A B C D E
B A D E C
C E A B D
D C E A B
E D B C A
Statistical analysis: the mathematical model for LSD is given by
Yijk =  kji  ijk ( i = j = k = 1,2,……....,m)
where Yijk denote the response from the unit (plot) in the ith
row, jth
column and
receiving the kth
treatment
A B C D E
B A E C D
C D A E B
D E B A C
E C D B A

83
i = ith
row effect
j = jth
column effect
k = kth
treatment effect;
ijk = error component
we know that total variation = variation due to rows + variation due to columns +
Variation due to treatments + variation due to error
Null hypothesis (H0) = There is no significant difference between Rows, Columns and
Treatment effects.
i.e. i) H01 : m  ....................21
ii) H02: m  ....................21 and
iii) H03: m  ....................21
The steps in the analysis of the data for verifying the null hypothesis are:
Different component variations can be calculated as follows:
1) C.F = 2
2
).(
m
TG
2).Row Sum of Squares (RSS) =
m
r
m
i
i1
2
-C.F
3) Column Sum of Squares (CSS) =
m
c
m
j
j1
2
- C.F
4) Treatment Sum of Squares (Tr..S.S). =
m
t
m
k
k1
2
- C.F
5) Total Sum of Squares (TSS) =   
m
i
m
j
m
k
ijky
1 1 1
2
- C.F
6) Error Sum of Squares (ESS) = TSS –RSS-CSS-Tr.S.S.

84
ANOVA TABLE
Sources D.F S.S M.S. F-cal. value F-table value at
5%LOS
Rows
Columns
Treatments
Error
m-1
m-1
m-1
(m-1)(m-2)
RSS
CSS
Tr.S.S.
ESS
RMS =
1m
RSS
CMS=
1m
CSS
TMS =
1
...
m
SSTr
EMS =
1m
ESS
FR =
EMS
RMS
FC =
EMS
CMS
FT =
EMS
TMS
FR[m-1, {(m-1)(m-2)}]


Total m2
-1 TSS -
If calculate value of F(Tr) < table value of Fat 5%LOS, H0 is accepted and hence we may
conclude that there is no significance difference between treatment effects.
If calculate value of F(Tr) > table value of F at 5%LOS, H0 is rejected and hence we may
conclude that there is significance difference between treatments effects.
If the treatments are significantly different, the comparison of the treatments is carried out
on the basis of Critical Difference (C.D.).
C.D. = SED (Tr) x t(r-1)(k-1) at  level of significance
Where SED =
m
xEMS2
, where m = number of rows
If F is significant, the significance of any treatment contrast can be tested by using the CD
value.
Advantages of Latin Square Design: 1) With two way grouping or stratification LSD
controls more of the variation than C.R.D. or R.B.D.
2) L.S.D. is an incomplete 3-way layout. Its advantage over complete 3-way layout is that
instead of m3
experimental units only m2
units are needed. Thus a 4x4 L.S.D. results in
saving of 64-16 = 48 observations over a complete 3-way layout.
3) The statistical analysis is simple though slightly complicated than for R.B.D. Even with
missing data the analysis remains relatively simple.
4) More than one factor can be investigated simultaneously.
5) The missing observations can be analysed by using missing plot technique.

85
Example: An experiment on cotton was conducted to study the effect of foliar application
of urea in combinations with insecticidal sprays in the cotton yield. Five treatments were
tried in a 6x6 Latin Square Design. The layout plan and yield is given below:
T2
4.9
T4
6.4
T5
3.3
T1
9.5
T3
11.8
T3
9.3
T1
4.0
T2
6.2
T5
5.1
T4
5.4
T4
7.0
T3
15.4
T1
6.5
T2
6.0
T5
4.6
T5
5.3
T2
7.6
T3
13.2
T4
8.6
T1
4.9
T1
9.3
T5
6.3
T4
11.8
T3
15.9
T2
7.6
Analyze the data and state your conclusions
Sol:
Null hypothesis: Rows, Columns and Treatments effects are equal
or
H01 : 54321   ;
H02 : 54321  
H03: 54321  
CF = 2
2
)(
m
GT
RSS = CF
m
R
m
i
i

1
2
CSS = CF
m
C
m
j
j

1
2
Tr.S.S = CF
m
T
m
k
k

1
2
TSS = CFy
m
i
m
j
m
k
ij   1 1 1
2
ESS = TSS-RSS-CSS-Tr.S.S.

86
ANOVA Table
Rows Columns Row
totals1 2 3 4 5
1 T2
4.9
T4
6.4
T5
3.3
T1
9.5
T3
11.8
R1 = 35.9
2 T3
9.3
T1
4.0
T2
6.2
T5
5.1
T4
5.4
R2 = 30.0
3 T4
7.0
T3
15.4
T1
6.5
T2
6.0
T5
4.6
R3 = 9.5
4 T5
5.3
T2
7.6
T3
13.2
T4
8.6
T1
4.9
R4 = 39.6
5 T1
9.3
T5
6.3
T4
11.8
T3
15.9
T2
7.6
R5= 50.9
Column
totals
C1 = 35.8 C2 = 39.7 C3 = 41.0 C4 = 45.1 C5 = 34.3 GT=195.9
Treatment Totals: ΣT1 = 34.2; ΣT2 = 32.3; ΣT3= 65.6; ΣT4 = 39.2; ΣT5 = 24.6
CF =
25
)9.195( 2
= 1535.07
RSS = CF
RRR


5
..........
2
5
2
2
2
1
= 07.1535
5
)9.50()6.39()5.39()0.30()9.35( 22222


= 46.54
CSS = CF
CCC


5
..........
2
5
2
2
2
1
= 07.1535
5
)3.34()1.45()0.41()7.39()8.35( 2222


=14.77
Tr.S.S =
5
)(...........)()( 2
5
2
2
2
1   TTT
- CF
= 07.1535
5
)6.24()2.39()6.65(()3.32()2.34( 22222


= 196.55

87
TSS = CFYYYY 
2
55
2
13
2
12
2
11 ..................
= (4.9)2
+ (6.4)2
+ (3.3)2
+ ……..+ (15.9)2
+ (7.6)2
– 1535.07
= 1821.07 – 1535.07 = 286.0
ESS = TSS – RSS – CSS – Tr.S.S
=286.0 - 46.54 - 14.77 - 196.55 = 28.14
ANVOA table
Sources d.f. S.S. M.S. F cal . value F table value
Rows 5-1 = 4 46.54 11.64 4.95 F0.05(4, 12) = 3.26
Columns 5-1 = 4 14.77 3.69 1.57 F0.05(4, 12) = 3.26
Treatments 5-1 = 4 196.55 49.14 20.91 F0.05(4, 12) = 3.26
Error 24 – 12 = 12 28.14 2.35
Total 25-1 = 24 286.0
Calculated value of F (treatments) > Table value of t at 5% LOS, H0 is rejected and hence
effect of foliar application of urea, have significant effect in the yield. To determine
which of the treatment pairs differ significantly we have to calculate the critical difference
(C.D.)
SEm =
m
EMS
=
5
35.2
= 0.69
SED = 2 * SEm = 1.414 * 0.69 = 0.97
CD = SED x t – table value for 12 d.f. at 5 % LOS
= 0.98 * 2.18
= 2.11
Coefficient of variation (CV) = 100*
X
EMS
=
84.7
35.2
X100 =19.55%

88
Bar Notation:
3T 4T 1T 2T 5T
13.12 7.84 6.84 6.46 4.92
______________
_______________
________________
i) The pairs not scored are significant
ii) The pairs under scored are non significant
From the bar chart it can be concluded that third treatment i.e. T3 significantly higher than
all the other treatments.
References Books:
1. Statistics for Agricultural Sciences, G. Nageswara Rao, Second Edition, BS
Publications, Hyderabad
2. A Text book of Agricultural Statistics, R. Rangaswamy, New Age International (P)
Limited, publishers
3. Statistical Methods, K.P. Dhamu and K. Ramamoorthy, AGROBIOS (INDIA)
4. Fundamentals of Mathematical Statistics, S.C. Gupta and V.K. Kapoor, Sultan
Chand & Sons Educational Publications
5. Fundamentals Applied Statistics, S.C. Gupta and V.K. Kapoor, Sultan Chand &
Sons Educational Publications
6. Design Resources Server: www.iasri.res.in

Bas 103

More Related Content

What's hot (20)

Similar to Bas 103 (20)

More from shivendra kumar (20)

Recently uploaded (20)

Bas 103