1. DR. NALINI R
ASSOCIATE PROFESSOR OF COMMERCE
DEPARTMENT OF MBA
MAHARANI’S WOMEN’S COMMERCE AND
MANAGEMENT COLLEGE, MYSURU
2. HYPOTHESES : CONTENTS
FORMULATION OF HYPOTHESES
HYPOTHESIS : NECESSARY OR NOT?
TYPES
SOURCES
FUNCTIONS
CHARACTERISTICS
TESTING OF HYPTHESIS
TYPES OF ERROR
3. Formulation of Hypotheses
Meaning of Hypothesis: A tentative proposition
formulated for empirical testing
It’s a tentative answer to a research question
Good & Hatt: “ a proposition which can be put
to a test to determine its validity”
4. Hypotheses: Necessary or Not?
In mere fact finding investigation: no need
In exploratory studies: no need as it is not
possible & the very purpose is to help in
formulating hypotheses
In analytical & experimental studies: must be
set up to give clear direction to the study
6. Null Hypothesis: These are hypothetical statements
denying what are explicitly indicated in working hypo.
They do not exist in reality; nor intended to exist in
reality
They state that no difference exists between the
parameter and the statistic being compared to it
Ex: “ There is no relationship between families’ income
level and expenditure on recreation”
Null hypotheses are formulated for testing statistical
significance
As the test would nullify the null hypo they are called
so (utility of Null hypo: Objectivity)
7. Sources of Hypotheses
Theory : Goal of business (theory) Hypo: the rate of
return on CE is an index of business success; higher
the EPS more favorable is the financial leverage
Observation: Ex: price & demand for a product
Intuition & Personal experience
Findings of Studies
Continuity of research
8. Functions or Role of Hypotheses
Guides the direction of study
Gives an idea for setting order among facts
Specifies sources of data
Determines data needs
Suggests type of research
Determines the technique of analysis
Helps in development of theories
9. Characteristics of a good hypothesis
Conceptually clear
Specificity
Testability
Availability of techniques
Theoretical relevance
Consistency
Objectivity
Simplicity
10. Testing of Hypotheses
Embodies major part of research process
Consists of operationalisation of concepts,
construction of data gathering tools, collection of data
, statistical analysis of data and drawing inferences
from the results.
Tests of significance are applied
Facts may confirm the hypothesis or reject
11. Types of Error in testing of hypothesis
We may reject a hypothesis which is true and
should not be rejected (Type I Error);
We may accept a hypothesis which is false and
should be rejected (Type II Error)
12. Null Hypothesis
Develop a Null Hypo for statistically testing the hypo
The H0 is formulated for testing possible rejection or
nullification
It is accompanied by alternative hypo Ha
Our regular research hypo is the logical opposite of
null hypo
Ex: Implementation of Employment scheme has not
led to an increase in the average rural income level
Null hypo always states that there “ no change” or “ no
difference” or “no relationship”
13. Hypothesis testing
Both the null hypo and the research hypo are
expressed in terms of the population parameters not
in terms of the sample statistics
Null hypo is tested directly; the research hypo is
supported when the null hypo is rejected
As a sample is likely to vary somewhat from its
population, the sample result is subject to sampling
error & so it does not always reflect true population
value.
So sample results cannot be interpreted directly; a
decision rule is needed to enable us to accept or reject
a hypothesis on the basis of sample result
14. When does the sample result justify the rejection of null
hypothesis?
When it could not be a result of chance (when it is highly
improbable)
Statisticians consider only when if it is among the extreme
5% or 1% of possible outcomes
The 5% or 1% value defines the proportion of a sampling
distribution that contains the specified highly improbable
outcomes
This proportion of the distribution is called the rejection
region and the 5% or 1% are called levels of significance or
levels of confidence. These are levels of probability at
which the null hypo can be rejected with confidence.
The rejection of H0 supports the research hypo