Академический Документы
Профессиональный Документы
Культура Документы
change in the other variable, it is called non- method all the values of the two variable are
linear correlation. plotted in a chart in the form of dots.
Degrees of correlation: Therefore, it is also known as dot chart. By
Correlation exists in various degrees observing the scatter of the various dots, we
1. Perfect positive correlation: If an can form an idea that whether the variables
increase in the value of one variable is are related or not. A scatter diagram
followed by the same proportion of increase indicates the direction of correlation and
in other related variable or if a decrease in tells us how closely the two variables under
the value of one variable is followed by the study are related. The greater the scatter of
same proportion of decrease in other related the dots, the lower is the relationship
variable, it is perfect positive correlation. Merits of Scatter Diagram method
2. Perfect Negative correlation: If an 1.it is easy to plot the points
increase in the value of one variable is 2. simple to understand
followed by the same proportion of decrease 3. abnormal values in the data can be easily
in other related variable or if a decrease in detected
the value of one variable is followed by the 4. extreme value do not affect it.
same proportion of increase in other related Demerits of Scatter diagram method
variably it is Perfect Negative Correlation. 1.the degree of correlation cannot be easily
3. Limited Degree of Positive correlation:: estimated
When an increase in the value of one 2. algebraic treatment is not possible
variable is followed by a non-proportional Correlation graph Method
increase in other related variable, or when a Under correlation graph method the
decrease in the value of one variable is individual values of the two variables are
followed by a non proportional decrease in plotted on a graph paper. Then dots relating
other related variable, it is called limited to these variables are joined separately so as
degree of positive correlation. to get two curves. By examining the direction
4. Limited degree of Negative correlation: and closeness of the two curves, we can infer
When an increase in the value of one whether the variables are related or not. If
variable is followed by a non-proportional both the curves are moving in the same
decrease in other related variable, or when a direction( either upward or downward)
decrease in the value of one variable is correlation is said to be positive. If the
followed by a non proportional increase in curves are moving in the opposite directions,
other related variable, it is called limited correlation is said to be negative.
degree of negative correlation. Correlation Coefficient:
5. Zero Correlation/ no correlation: If Correlation analysis is actually an attempt to
there is no correlation between variables it is find a numerical value to express the extent
called zero correlation. In other words, if the of relationship exists between two or more
values of one variable cannot be associated variables. The numerical measurement
with the values of the other variable, it is showing the degree of correlation between
zero correlation. two or more variables is called correlation
Methods of measuring correlation coefficient. Correlation coefficient ranges
1) Scatter Diagram between -1 and +1. Coefficient of correlation
2) Correlation graph can be computed by applying the methods
3) coefficient of correlation given below
Scatter Diagram 1) Karl Pearson’s Co-efficient of correlation
This is the simplest method for ascertaining 2) Spearman’s Rank correlation method
the correlation between variables. Under this 3) Concurrent deviation method
QT – (Al Jamia Arts and Science College, Poopalam)
1.Regression Equation of X on Y:- This is is a number lying between ‘0’ and ‘1’ ‘0’
used to describe the variations in the values denotes the even which cannot occur, and ‘1’
of X for given changes in Y. denotes the event which is certain to occur
here equation is X = a + by Terms use in Probability.
∑x = na +b ∑y Random Experiment: A random
∑xy = a∑y + b ∑y2 experiment is an experiment that has two or
more outcomes which vary in an
2. Regression Equation of Y on X :- This is unpredictable manner from trial to trail
used to describe the variations in the value when conducted under uniform conditions.
of Y for given changes in X. In a random experiment, all the possible
here equation is Y = a + bx outcomes are known in advance but none of
∑y = na +b ∑x the outcomes can be predicted with
∑xy = a∑x + b ∑x2 certainty. For example, tossing of a coin is a
Difference between correlation and random experiment because it has two
regression outcomes (head and tail), but we cannot
Correlation Regression predict any of them which certainty.
It studies degree of It studies the nature Sample Point: Every indecomposable
relationship of relationship outcome of a random experiment is called a
between variables between variables sample point. It is also called simple event or
It is not used for It is basically used elementary outcome. Eg. When a die is
prediction purposes for prediction thrown, getting ‘3’ is a sample point.
purposes Sample space: Sample space of a random
It is basically used It is basically used experiment is the set containing all the
as a tool for as a tool for sample points of that random experiment.
determining the studying cause and Eg:- When a coin is tossed, the sample space
degree of effect relationship is (Head, Tail)
relationship Event: An event is the result of a random
There may be There is no such experiment. It is a subset of the sample space
nonsense nonsense regression of a
correlation between random experiment.
two variables Sure Event (Certain Event): An event
There is no question There must be whose occurrence is inevitable is called sure
of dependent and dependent and even. Eg:- Getting a white ball from a box
independent independent containing all while balls.
variables variables Impossible Events: An event whose
THEORY OF PROBABILITY occurrence is impossible, is called impossible
INTRODUCTION event. Eg:- Getting a white ball from a box
Probability refers to the chance of happening containing all red balls.
or not happening of an event. Any problem Uncertain Events: An event whose
which contains uncertainty about the occurrence is neither sure nor impossible is
happening of the event is the problem of called uncertain event. Eg:- Getting a white
probability. ball from a box containing white balls and
Definition of Probability black balls.
The probability of given event may be Equally likely Events: Two events are said
defined as the numerical value given to the to be equally likely if anyone of them cannot
likely hood of the occurrence of that event. It. be expected to occur in preference to other.
For example, getting herd and getting tail
QT – (Al Jamia Arts and Science College, Poopalam)
when a coin is tossed are equally likely tendency of frequency ratio to stabilize, is
events. common in all random experiments. This
Mutually exclusive events: A set of events phenomenon is known as statistical
are said to be mutually exclusive of the regularity.
occurrence of one of them excludes the Relative Frequency Definition or
possibility of the occurrence of the others. Empirical Approach
Exhaustive Events:: A group of events is According to Relative Frequency definition,
said to be exhaustive when it includes all the probability of an event can be defined as
possible outcomes of the random experiment the relative frequency with which it occurs in
under consideration. an indefinitely large number of trials.
Dependent Events: Two or more events are Here, probability has between 0 and 1, i.e. 0
said to be dependent if the happening of one ≤ P(A) ≤ 1
of them affects the happening of the other. Axiomatic Approach (Modern Approach)
DIFFERENT SCHOOLS OF THOUGHT ON to Probability
PROBABILITY Let ‘S’ be the sample space of a random
Different Approaches/Definitions of experiment, and ‘A’ be an event of the
Probability random experiment, so that ‘A’ is the subset
There are 4 important schools of thought on of ‘S’. Then we can associate a real number to
probability :- the event ‘A’. This number will be called
1. Classical or Priori Approach Objective probability of ‘A’ if it satisfies the following
Probability three axioms or postulates :-
2. Relative frequency or Empirical Approach 1.P (A) is a real number such that P (A) ≥ 0
3. Modern or Axiomatic Approach for every subset of S
1. Classical or Priori Approach 2. P (S) = 1 where ‘S’ is the sample space
If out of ‘n’ exhaustive, mutually exclusive 3. P (A ⋃ B) = P(A) + P (B), where A and B
and equally likely outcomes of an are two non-intersecting subset of S
experiment; ‘m’ are favourable to the THEOREMS OF PROBABILITY
occurrence of an event ‘A’, then the There are two important theorems of
probability of ‘A’ is defined as to be P(A) = probability. They are :
1. Addition Theorem
Frequency ratio : consider a random
2. Multiplication Theorem
experiment. Let ‘A’ be an event associated
Addition Theorem
with the random experiment. Let us repeat
Here, there are 2 situations.
the experiment ‘n’ times. Let the event A
(a) Events are mutually exclusive
happen ‘f ‘ out of ‘n’ repetition of the
(b) Events are not mutually exclusive
experiment. Then f/n is called frequency
(a) Addition theorem (Mutually Exclusive
ratio.
Events): If two events, ‘A’ and ‘B’, are
Law of statistical regularity : the outcomes
mutually exclusive the probability of the
of a random experiment vary in an
occurrence of either ‘A’ or ‘B’ is the sum of
unpredictable manner. But if we consider
the individual probability of A and B.
sequence of frequency ratios, we can observe
P(A or B) = P(A) + P(B)
a regularity. We notice that these frequency
i.e., P(A⋃B) = P(A) + P(B)
ratios have a tendency to become more and
(b)Addition theorem (Not mutually
more stable as the number of repetitions of
exclusive events): If two events, A and B are
the experiment increases. The frequency
not mutually exclusive the probability of the
ratio appears to approach a fixed number as
occurrence of either A or B is the sum of
the number repetitions is increased. This
their individual probability minus
QT – (Al Jamia Arts and Science College, Poopalam)
types of test statistics. All these test statistics test statistic follows t-distribution. Uses of t-
can be classified into two groups. They are test are:-
a. Parametric Tests 1. To test the population mean when the
b. Non-Parametric Tests sample is small and the population s.D.is
PARAMETRIC TESTS unknown.
The statistical tests based on the assumption 2. To test the equality of two sample means
that population or population parameter is when the samples are small and population
normally distributed are called parametric S.D. is unknown.
tests. The important parametric tests are:- 3. To test the difference in values of two
1. z-test dependent samples.
2. t-test 4. To test the significance of correlation
3. f-test coefficients.
Z-test: The following are the important assumptions
Z-test is applied when the test statistic in t-test:-
follows normal distribution. It was 1. The population from which the sample
developed by Prof.R.A.Fisher. The following drawn is normal.
are the important uses of z-test:- 2. The sample observations are independent.
1. To test the population mean when the 3. The population S.D.is known.
sample is large or when the population 4. When the equality of two population
standard deviation is known. means is tested, the samples are assumed to
2. To test the equality of two sample means be independent and the population variance
when the samples are large or when the are assumed to be equal and unknown.
population standard deviation is known. F-test:
3. To test the population proportion. F-test is used to determine whether two
4. To test the equality of two sample independent estimates of population
proportions. variance significantly differ or to establish
5. To test the population standard deviation both have come from the same population.
when the sample is large. For carrying out the test of significance, we
6. To test the equality of two sample calculate a ration, called F-ratio. F-test is
standard deviations when the samples are named in honour of the great statistician
large or when population standard R.A.Fisher. It is also called Variance Ration
deviations are known. Test.
7. To test the equality of correlation Uses of F-distribution:-
coefficients. 1. To test the equality of variances of two
Z-test is used in testing of hypothesis on the populations.
basis of some assumptions. The important 2. To test the equality of means of three or
assumptions in z-test are:- more populations.
1. Sampling distribution of test statistic is 3. To test the linearity of regression
normal. Assumptions of F-distribution:-
2. Sample statistics are dose the population 1. The values in each group are normally
parameter and therefore, for finding distributed.
standard error, sample statistics are used in 2. The variance within each group should be
place where population parameters are to be equal for all groups.
used. 3. The error (Variation of each value around
T-test: its own group mean) should be independent
t-distribution was originated by W.S.Gosset for each value.
in the early 1900. t-test is applied when the
QT – (Al Jamia Arts and Science College, Poopalam)
Study
population or not, with the of the statistical
technique, called F – test.
Characteristics of Analysis of Variance:
1. It makes statistical analysis of variance of
two or more samples.
2. It tests whether the difference in the
well…
means of different sample is due to chance or
due to any significance cause.
3. It uses the statistical test called, F – Ratio.
Types of Variance Analysis:
There are two types of variance analysis.
They are:-
1. One way Analysis of Variance
2. Two way analysis of Variance
One way Analysis of Variance:
In one way analysis of variance, observations MUHAMMED RIYAS N
are classified into groups on the basis of a ASST. PROF.
single criterion. For example, yield of a crop AL JAMIA ARTS AND SCIENCE COLLEGE
is influenced by quality of soil, availability of POOPALAM
rainfall, quantity of seed, use of fertilizer, etc. PERINTHALMANNA
It we study the influence of one factor, It is PH : 9747799772
called one way analysis of variance. E-mail- riyasmuhammed89@gmail.com
If we want to study the effect of fertilizer of
yield of crop, we apply different kinds of
fertilizers on different paddy fields and try to
find out the difference in the effect of these
different kinds of fertilizers on yield.
two way analysis of variance
Two way analysis of variance is used to test
the effect of two factors simultaneously on a
particular variable.