The Research Process and the Hypothesis
Formulation
1. Introduction to Research
Research is
present in every walk of life. The
concept of research in different forms existed from time immemorial. It assumed significance after the formal
education and organizations came into existence. It does not mean that the formal education and
organizations gave birth to research.
Certainly the process of research was used by different civilizations to
know the past, understand the present and predict the future. We could define research as the process used
to move from unorganized way of thinking and decision making to the organized
way of thinking and decision making.
Formally, research could be defined as the process and tools used in decision
making that reduce the risk of wrong decisions. Research is a systematic
inquiry into a phenomenon to provide information to the decision maker in any
field. Research is undertaken in all the
subjects of inquiry. While scientific
research involves systematic and scientific experimentation which could be
replicated and verified, the replication and verifiability may not be central
to social, political, economic, commerce, commercial and business
research. Business research involving
commerce and management can be defined as the systematic inquiry into a
phenomenon to provide information to the managerial process of decision
making. A large part of the research in
accounting today is empirical and has focused on different issues related to
accounting concepts, principles, the treatment of specific items of income,
expenses, assets and liabilities etc.
Research has helped the human
kind in all walks of life. The
importance of research grew as organizations, both ‘for profit’ and ‘not for
profit’, grew in size and started to recognize that research can provide useful
inputs into the decision making.
Countries started to recognize the importance of research as the
economies became more complex and many variables started to influence the
outcome of a scheme/policy. Scientific
research has been of immense help in understanding the most complex things in
science and technology. Research in
medical field has solved many a problems of the human and animal life. Without the basic research, human life would
have been miserable today. It is because
research which is abstract in nature was undertaken by scientists, we have been
able to view the TV, travel from place to place, get electricity and light in
all the houses, save the human and animal life from deadly diseases, launch the
satellites into geostationary orbits and organise the working of business
organisations and find out the results of the operations for each year. While the basic or pure research may not look
at the immediate application of the research results in any field, the applied
research usually looks for solving the current or immediate future
problems. It is not uncommon for the
researchers in the educational institutions to be at the receiving end whenever
some problems crop up in a state/country.
The policy makers immediately criticize that the researchers have not
been able to solve their problems.
Unfortunately what they do not realize is that the problem which arises
may have to be adequately researched before the tentative solutions are
found. These types of research are
undertaken to address the current problems and find policy measures to solve
the problems.
Every research involves
systematic steps. Understanding of these
steps is essential to know the best ways of doing research. This paper has the limited objective of
discussing the research process and the hypothesis formulation. The paper is organised in five parts. Part 2 focuses on the research process, part
3 discusses the issues involved in hypothesis formulation and testing, part 4
discusses the definitions of variables used in the research and the last part
presents the conclusions.
2. The Research Process
Research can be
pure or basic research or applied research.
Empiricism is assuming a lot of importance in the recent times. Most of the research in commerce, economics
and management is empirical in nature.
Therefore, many researchers use the data to either describe a phenomenon
or prove or disprove a relationship.
When a survey research involves collection of qualitative and
quantitative data, researchers would first define the problem, then set the
objectives and then go on to collect the data relevant for their work. Generally, the research process involve the
following steps
2.1 Idea generation:
A research idea may be generated by general/specific problem that is
being faced currently, the problem that may be imagined by either the
researcher or others. It could also
emerge from a brain storming session, Delphi
method, researcher’s reading, researcher’s/others’ imagination. The problems currently being encountered
could be related to the economy, industry, company, society, a social
organization, an administration division of the government, world level organisations
etc. For example, India is facing the
problem of higher than expected food inflation today. Therefore, the relevant idea could be what
causes food inflation. We have seen the
fall of multinational financial institutions during the financial crisis that
originated in the US
2.2 Problem identification: Once the researcher generates research idea
based from different sources, the researcher needs to identify the problem in
concrete terms. An idea may be a hazy
one, but that needs to be refined while identifying the problem. The problem identification will define the
scope of the research and sets the boundary line for the research. If the problem is too broad and cannot be
approached in a single research, then we may not arrive at any
conclusions. Therefore, we need to
identify the problem that the researcher wants to address in his/her work.
2.3 Problem definition: At this stage the researcher not only knows
the problems that needs to be addressed but also defines it. The problem definition clearly identifies the
phenomenon that needs to be investigated, the variables needed to define the
problem, the relationship that needs to be established tentatively. The problem definition is done with the help
of reading of the related materials and works.
2.4 Review of related concepts, theories and the literature: Once the research problem is tentatively
defined, the researcher needs to review the related concepts, theories and the
empirical and theoretical works that have already been carried out. This stage is crucial to understand whether
or not to undertake the the proposed research work. If the review of related concepts, theories
and literature reveals that the problem under investigation has already been
sufficiently investigated and robust conclusions have been drawn, the problem
need not be researched again.
Alternatively, this stage may reveal that the researcher needs to modify
the problem and redefine in the light of the evidences that are already
available. The review also helps the researcher to understand the problems that
have already been investigated, the methodology that are followed, the
variables used and the phenomenon investigated and the conclusions that are
drawn. A critical review of the
literature would establish the methodological problems, the data collection problems
and the need for refining and reinvestigating the problems. The literature review can be compared to the
flow of a river. The river starts from
some place in a small way, collects water from small streams and tributaries
before becoming a big river and establishing its own identity and then flowing
in some definite direction. The
literature review would also reveal what are the different dimensions of the
problem that is being investigated and what are the ways in which different
researchers have defined and investigated the phenomenon. The review would also give an idea of
different variables that are need to be studied, the type of relationship
between different variables/attributes that are to be studied, and the
tentative relationships that could be formulated as a possible explanation of
the problem under investigation.
2.5 Formulation of hypotheses: The problem identification, definition and
review of related literature would reveal the different dimensions of the
problems, the variables used in the study, the methodologies followed and the
conclusions drawn by various researchers.
A critical review of the studies should also reveal the gaps in the
problems, the problems of drawing conclusions using the limited data and the questionable
methodologies used in the studies. It
would also establish the new variables that need to be considered, the refinements
or new methodologies to be used and the data collection tools to be applied
while taking up the researcher’s current research problem. This should be the basis on which the
researcher should formulate the hypotheses.
Each of the hypotheses should have a basis and be supported by either
the earlier established conclusions of different studies or established
practices etc. The hypothesis
formulation is crucial as the entire data collection work of the researcher
hinges on this. Therefore, the researcher
should exercise extreme care in formulating and stating the hypotheses.
2.6 Research Design including sampling Design: While research design encompasses the entire
research work starting from research idea generation and ending with the
drawing of conclusions, sampling design deals with the data and the data
collection techniques that are employed by the researcher. Sampling design is the process of defining
the population, sample, the sampling frame and the sampling techniques. This process helps make the data a
representative one for the entire population.
Since the entire population cannot be studied by the researcher for want
of time and cost of investigation, he/she uses sample to draw conclusion about
the entire population. Therefore, it is extremely
important that the researcher uses the correct sampling techniques to make the
sample representative of the population. The use of statistical techniques for
analysis of the data assumes that the sample is a random one and is
representative of the entire population. Therefore, if this condition is not
satisfied, the entire conclusions will be wrong.
2.7 Data collection:
The researcher uses the appropriate data collection instrument and
administers it to the sampling units chosen in the above step. The data
collection instrument should be carefully designed to minimize the sampling and
non-sampling errors. Further, hallow
effect should be taken care while finalizing the data collection instrument. The researcher may use pilot study to redraft
and revise the data collection instrument.
It has to be ensured that there are sufficient numbers of observations
from the data collected to draw conclusions.
Use of statistical techniques would also call for a minimum number of
observations. If this is not ensured in
data collection stage, the analysis of data poses problems.
2.8 Data checking and validation: If sufficient care is taken at the data
collection stage, the problem of checking and validation of the data is
easy. If the data collection is not done
properly, a lot of inconsistencies may be encountered while reading the
data. The first thing the researcher
should do after collecting the data is to put the data to “eye ball test”(EBT). EBT
involves observing the data carefully by the researcher for the possible
errors, omissions, inaccuracies and inconsistencies. A careful observation of the data may reveal
that there are certain problems with the data collected. Before, proceeding further, the researcher
should correct mistakes in the data and clean it. The process of ensuring the accuracy of data
leads to data validation. This stage
would avoid the possible blunders the researcher would commit by using the
wrong data and subjecting it to mountain of statistics. No sophisticated statistical tests can
rectify the wrong conclusions arising out of the inaccurate data. Therefore, it is necessary that the
researcher first cleanses the data before subjecting it to any statistical
tools.
2.9 Data analysis:
The cleaned data may be summarized and analysed subjecting it to
statistical tools. The statistical tools
may be broadly classified into two categories for the purpose of data
analysis. The first category of the
tools like measures of central tendency, dispersion, skewness and kurtosis may
be used to summarise the mass data that would help the researcher to get a
limited number of parameters that would describe the characteristics of the
data that has been collected. The second
category of the statistical tools like the measures of correlation, regression,
the various statistical tests can be used for establishing and testing the
relationship/hypotheses. These tools
would help to know whether or not the research hypothesis can be accepted or
rejected.
2.10
Interpretation: Most of the researchers stop by saying
whether or not the hypotheses formulated are either accepted or rejected. Mere statements of acceptance or rejection of
hypotheses is not the purpose of research.
The interpretation of the results of the data analysis and knowing
whether the hypotheses are accepted or rejected and then stating the results in
terms of the original problem is what is intended in the interpretation
stage. At this stage the researcher
should also give possible interpretations for either acceptance or rejection of
the hypotheses. This is also the stage
at which the researcher should compare his/her results with those of the works
studied in the literature review.
2.11
Conclusions: The researcher should draw the final
conclusions of the research in terms of the original problem being investigated. The researcher should also highlight the
problems of generalizations and the possible problems that could have been
investigated as part of his/her research but could not be done for various
reasons. This discussion would lead to
the problems that other researcher could undertake. The researcher should also discuss the
implications of the conclusions drawn.
Merely stating the conclusion is not sufficient.
3. Hypothesis Formulation
Most of the
research students in economics, commerce and management and in other fields
would be using empirical research. While
the empirical work in science involves experimentation in a laboratory; the
entire economy, industry, companies, social organizations etc are the
laboratories for students in economics, commerce and management and in other
fields in social sciences. When we use
the empirical research, the researchers are defining the problems in terms of
the variables/attributes and their tentative relationships. Therefore, when the relationships between
different variables/attributes are defined and formulated as a statement to
explain a phenomenon, it should emerge from some source. That source could be the observation,
literature, theory, an exploratory study undertaken specifically for this
purpose, current problems being faced, the problems that are likely to be
faced, the abstract formulation of the relationship that may or may not be
relevant in the current context, the phenomenon that was investigated for a long
time but conclusions could not be arrived at.
Therefore, the sources of hypotheses could be the previous literature,
the existing theory, law, an exploratory study, social/economic/demographic/scientific
phenomenon etc. It is not necessary that
every research study should have a hypothesis.
However, when a hypothesis is formulated, it should be done in a
scientific way. This is what is being
discussed here. Before, proceeding
further we need to define hypothesis.
3.1 What is a hypothesis?
A hypothesis is a tentative statement that establishes a
possible relationship between the variables, which in turn, explains some
phenomenon or relationship. It can be
defined as ‘a statement about the population’.
This statement could be stated in terms of the relationship between the
variables/attributes, cause and effect relationship, a phenomenon that explains
a problem etc. A hypothesis is a
statement which may include a prediction.
A fact cannot be a hypothesis. A
hypothesis should not be confused with facts.
For example, the population of India India
3.2. How to formulate a hypothesis?
A hypothesis cannot emerge from
vacuum. It has to have a strong
basis. The source of the hypothesis
could be the current problem, social and demographic phenomenon, economic problems,
country’s problems, scientific problems etc.
The hypothesis should have basis of either the previous literatures or
strong prevailing notions that are challenged by people, policies, practices,
law, social customs, scientific experiments, nature etc. The basis of research hypothesis is usually
the critical review of previous literature.
The critical review of the literature would reveal that there are
problems about the conclusions drawn and therefore, there is a further need for
investigation. This conclusion could
give rise to the new hypothesis. Therefore, a researcher should formulate each
hypothesis on the basis of the literature review he does and the discussion of
the reviews.
3.3. Examples of what are not Hypotheses
The following cannot be the
hypotheses.
- Commercial
Banks in
- The
purpose of Commercial Banks in
- The
customers of a company consist of male and female.
- A
large number of customers visit bus terminus every day.
- Money
supply influences the inflation.
- The
voters in
- The
age of the voters in
- The
government of
- Export
oriented companies earn foreign exchange for
- The
industrial units located in Special Economic Zones (SEZs) are eligible for
fiscal incentives.
- The
financial crisis originated in the
- WTO
ensures the free trade in goods and services across the member countries.
- The
women in
3.4. Examples of what are Hypotheses
- The
following can be the hypotheses.
- The
capital structure of a firm influences the value of the firm.
- The
value of dividend paying firm is more than that of firms not paying
dividends.
- There
is a positive relationship between wages and production
- Training
methods influence the productivity of firms.
- The
products produced by the MNCs are superior to that of Indian companies.
- The
TV viewing improves the IQ of children.
- The
use of chemical fertilizers may adversely affect the fertility of the
soil.
- The
capital structure of a firm influences the cost of capital
- Accounting
earnings influence the share prices.
- Publicly
available earnings information is absorbed by the stock market
instantaneously.
- There
is no significant difference between the Indian Accounting Standards and
International Accounting Standards.
- Accountants
involve in earnings manipulation when their pay packages are based on the
accounting earnings.
- Accountants
chose those accounting principles that maximize their personal wealth as
opposed to firm’s wealth.
The
following example shows how to formulate a Hypothesis. We know that there are different theories
that explain the capital structure of firm.
Many variables could influence the capital structure of a firm. A definite answer to the capital structure
puzzle is yet to be found. Therefore,
researcher could examine the factors that influence the capital structure of a
firm. A variable that is used to
measure the firm’s ability to generate capital internally as examined by Kester
(1986), Titman and Wessels (1988), Wald (1999) and Pandey et al (2000), among
others, is earnings before interest and taxes (EBIT) scaled by total
assets. Profitability
is found to be negatively correlated with leverage. Mallikarjunappa and
Goveas (2007) tested the factors influencing the capital structure of a firm
and concluded that profitability do not have significant relationship with the
debt ratio. However, a significant
negative relationship between profitability and debt ratio supports the pecking
order hypothesis that the firms with liquid assets and internal accruals would
use less debt. Accordingly, we can hypothesize
that firm with high profitability will use less debt capital in the capital
structure. Therefore, we can hypothesize
as follows.
Hypothesis: Profitability is inversely related to debt
ratios.
3.5. How do we test the hypothesis?
The following steps are followed
in hypothesis testing.
3.5.1. Formulation
of hypotheses
3.5.2. Sampling
design
3.5.3. Deciding
sampling distributions
3.5.4. Setting
the level of significance.
3.5.5. Deciding
the acceptance and rejection region.
3.5.6. Data
collection
3.5.7. Computation
of the values of test statistic based on the appropriate sampling distribution
and the data collected.
3.5.8. Deciding
whether the computed value of test statistic falls in the acceptance or rejection
region.
3.5.9. Deciding
whether to accept or reject the hypothesis.
3.5.10. Drawing
conclusions in terms of original problem.
3.5.11. Interpreting
the implications of conclusions.
4. Operational definitions of variables used in the research
It is very important that we provide operational
definitions of the variables used in the hypotheses. For example, if the hypothesis is: ‘morale
and production are positively related’; it important that we define morale and
the way it is measured. Similarly
production should be defined and measurement should be specified. If the survey work is involved, the data
collection instrument should contain questions related to the operational
definitions provided by the researcher.
This should be done before the data collection and not after the data
collection. Most often researchers do
not even know about this and will end up with problems after the data
collection. Therefore, every
variable/attribute in the hypotheses should be clearly defined. Further, measurements of these
variables/attributes should also be specified and incorporated into the data
collection instrument. It is always a
good idea to use the pilot survey and put the data collected to the statistical
tests to know and understand whether there are any problems with definitions
and measurements. It should also be ensured
that the variables are understood by all the respondents in the similar
way. There should not be any ambiguity
in the variable measurements. For
example, if a respondent in the production department is asked about morale, he
may understand morale in his own way and others may understand morale in their
own way. Therefore, it is preferable to
give the indicators of high or low morale so that the responses are correct and
comparable. This task is easy in case of variable like production as we can define
it in terms of number of units. Here
also it is important to specify that production is measured in terms of number
of units and even the units should be specified. However, it may not be so easy in case
variables like morale. It is here that
the researcher should search adequate literature and clearly define the
variables and their measurements. Consider
another hypothesis: education increases income of the people. This hypothsis specifies a positive
relationship between the concepts “education” and “income.” This abstract statement of hypothesis cannot
be tested without the operational definitions of the two variables or
concepts. First, it must be
operationalized or situated in the real world by rules of interpretation. To
test this hypothesis, the abstract meaning of education and income must be
derived or operationalized. The concepts should be capble of being understood
uniformly and measured. Education could be defined and measured by years of
schooling, professional eduction, general eduction, technical eduction, or
highest degree completed etc. Income could be measured by hourly rate of pay,
monthly income, yearly salary, etc. Similarly
if the hypothesis to be tested is ‘accounting earnings influences the stock
prices’, we need to define the two variables, accounting earnings and stock
prices. Accounting earnings may have
different meanings. These could be net
profit, gross profit, profit before taxes, profit before taxes and interest,
cash profit, operating profit, operating cash profit, profit before taxes,
interest and depreciation, etc.
Similarly stock prices may be the opening price, closing price, highest
price, lowest price, unadjusted price, adjusted price etc. We need to clearly define these two and
clarify that accounting earnings are measured as profit before taxes, interest
and depreciation and stock prices are the closing adjusted prices. This makes the entire hypothesis clear and
unambiguous. In the absence of operational
definitions the researchers and the readers may form their own opinion about
the variables and understand the conclusions drawn. To avoid differences in the understandings
the variables and concepts need to be defined after the formulation of the
hypotheses and before the data collection.
Without the operational definitions of the variables/concepts, the
researcher will end up with erroneous analysis, interpretation and conclusions. This can damage the foundations of a good
research work.
5. Conclusion
Research
is a long process. It involves commitment
of time, cost and dedication. Many a
time the researcher is prepared to spend time and money but do not have
dedication. It is lack of dedication
that leads to a lot of gaps in the research.
Even the dedicated researchers may go wrong if they do not have proper
direction. This paper is intended to
show the process involved in the research and the hypothesis formulation. There are many other issues related to
hypothesis formulation and testing. All
these cannot be discussed in short paper like this. The interested readers should read a good
book on research methodology. This paper
is intended to clarify the issues related to research process and hypothesis
formulation. The detailed discussion on
the testing process and the concepts involved in this is beyond the scope of
this paper. I hope that this paper will
help the researchers to clearly understand the process of research and
hypothesis formulation. I have examined
many PhD and MPhil theses and found that the researchers have erred in this
stage. Therefore, this short paper has
been written to help the researchers to improve their skills in research. Obviously this cannot be a final say on the
issue. Researchers can discover their
own innovative ways of formulating the hypothesis. The research process can be
improved by the collective wisdom of the researchers over a period of
time. I do hope that the researchers
will share this collective wisdom with the young researchers.
No comments:
Post a Comment