3

HYPOTHESIS-STATEMENT AND FORMULATION

'It isa good morning exercise for a research
apet hypothesis scientist to discard
everyday before breakfast. It keeps him young.'
-Lorenz Konrad

Objectives
After reading this chapter, the reader shouid understand:
Definitions and meaningof Hypoth.esis.
Purpose of the Hypothesis in relation to different types of variables.
Difference between Hypothesis and Theory.
Components between Hypothesis and Theory.
Components of a Hypothesis and approaches to hypothesis
formulation.
Different types of hypothesis and Characteristics of a good Hypothesis.
Fundamental concepts related to Hypothesis Testing.

Hypothesis formulation is mosty related to causal research or

empjricál research. Usually, exploratory researches and its like are
exceptions, where a hypothesis is not a necessity. But, in most researches,
the researcher has an intelligent guess' about the outcome of his research.
This guess' paves way to a formal hypothesis. The term hypothesis has been
derived from the ancient Greek, hypotithenai, meaning "to put under" or "to
Suppose". A hypothesis is a particular kind of conjecture that dlearty
formlates a suggestion about the solution to a certain probiem. In common
 Hypothesis-Statement and Formulation 47
parlance, a hypothesis is an educated guess about the phenomencn being
studied. "Educated" meaning that it goes beyond mere speculation. That is.
the hypothesis is a translation of iníormation gathered from theories, other
research and casual observation. Fig. 3.1 illustrates the process of
Hypothesis generation.

Initia! ldeas
(Undeveloped)

Initial Cbservation Literature Review

Problem Statement

Operational Definitions of
Constructs

Research Hypothesis

Fig. 3.1 Generation of a Research Hypothesis

(Source: Kroel (2002)
Ahypothesis is a "supposition made as a starting point for further
investigation from known facts Pearsall & Trumble, 1996). Since a
hypothesis is aimed at generating new scientific knowledge, it shoud be
novel and contain a testable prediction.
However, hypotheses are not unique to research. Hypotheses are
constantly generated in the human mind as we work to understand day-to
day phenomena. By formulating a series of reasonable guesses of cause and
etfect, we are able to understand and explore the events in our surounding
environment (Leedy and Ormrod, 2001).
Hypothesis fomulation has been an integral part of philosophy since
the earty days. Wikepidea reports : In early usage, scholars often referred to
a clever idea or to aconvenient mathematical approach that simplified
cumbersome calculations as a hypothesis; when used this way, the word aia
n
48
Fundamentals of Research Methodology
not necessarily have any specific
fan:ous example of the older sense meaning.
of the
Cardinal Bcllarmine gave a
word in the warning issued to
Galileo in the early 17th century : that he must
Earth as a reality, but merely as a hypothesis. not treat the notion of ihe
In common usage in the 2lst
century, a hypothesIs refers to a
provisional idea whose merit needs evaluation.
framer of a hypothesis needs to define specificsForin proper evaluation, the
hypothesis requires more work by the researcher in order operational terms. A
or disprove it. In due course, a to either confirm
confimed hypothesis may become part of a
theory or occasionaly may grow to become a thecry itselt. Normally,
scientific hypotheses have the form of a mathematical model. Sometimes,
but not always, onc can also fomulate
thera as existential siatcments,
stating that some particular instance of the
some characteristic and causa! explanations,phenomenon being studied has
whichh have the general form of
universal statements, stating that every instance of the
particular characteristic. phenoimenon has a
Any usefç hypothesis will enable predictions, by reasoning
deductive reasoning). it might predict the outcome of an experiment (including
laboratory setting or the cbservation of a phernomenon in nature. inThea
predicion may also invoke statistics and only talk about probabilities. Karl
Popper, following others, has argued that a hypothesis must be falsifiable, and
that a proposition or theory cannot be called scientific if it does not admit the
possibility of being shown false. By this additional criterion, it must, at least in
principle, be possible to make an observation that would disprove the
proposition as false, even if one has not actually (yet) made that observation.
Afalsifiable hypothesis can greatly simplify the process of testing to determine
whether the hypothesis has instances in which it is false.
It is essential that the outcome be currently unknown or reasonably
under continuing investigation. COnly in this case does the experiment, test
or study potentially increase the probability of showing the truth of a
hypothesis. If the researcher already knows the outcome, it is called a
Consequence-and the researcher should have already considered this while
formulating the hypothesis. If the predictions are not assessable by
observation or by experience, the hypothesis is not yet useful, and must wait
for others who might come afterward to make possible the needed
observations. For example, a new technology or theory might make the
necessary experiments feasible.
PÚRPOSE OF THE HYPOTHESIS
One of the major purposes for which à hypothesis is formulated is
defining the relationshËp between variables. Since hypothesis states before
hand what the possible outcome of the research would be, hypotheses are
 49
Hypothesis Staterment and Formulation
and its corresponding changes
in one variable the
expressed in terms of changes key components of the hypothesis are that
Hence, the procedure
inanother variable.means that for a researcher, the research other words,
variables. This also truthfulness of this relationship. In
something else
follows is about checking the changes to one iten cause
experiment so that
they design an
way.
to vary inapredictable quantities are called variables. In an experiment,
1These changing independent, dependent, and controlled.
usually there are three kinds : changes, while
variables are those which the researcher
observed with
Independent variables that are
are those
dependent variables variables. The dependent variable
corresponding changes in the dependent
the scientist makes to the
independent
change
changes in response to thethe dependent variable is caused by and depends
variable. Thenew value of variable. Controlled Variables are the
the independent
on the alue of stay constant.
variables that the researcher wants to
example, a researcher studying the effect of dividend payment as
For
Net Asset Value (NAV) of a mutual fund, formulates a hypothesis constant
onthe NAV of a mutual fund at
Dividend paynent adversely affects the between
sensex level. Here, he would like the to seek out a relationship
and the NAV of mutual furnd. To test whether his
dividend payment collect data of NAVs of different
hypothesis is true or not, he will have to
and after the payment of dividends. The amount of
Mutual Funds before
(which will vary from one mutual fund to another) paid ill
dividend NAV which he will observe after
constitute the independent variable and the variable.
dependent
the bonus payment will form the
Sometirnes, we want to understand the process or variables through
This brings us to the idea of
which one variable affects another variable.
intervening variables (also called mediatorbetween or mediating variables).
two other variables.
Intervening variables are variables that occur
example, tissue damage is an intervening variable in the smoking and
For (which mean causes or effects)
lungcancer relationship. We can use arrows variable like this:
and draw the relationship that includes an intervening
Smoking >Tissue Damage ->Lung Cancer.
everyone;
Sometimes, a relationship does not generalize to how the
show
therefore, researchers often use moderator variables to For example,
relationship changes across the levels of arn additional variable.
and cognitive therapy
perhaps behavioural therapy works better for malesmoderator variable. The
works better for females. In this case, gender is the
cognitive) and
relationship by type of therapy (behavioural versus
psychological relief is moderated by gender.
Classification of Variables
Variables are classified according to two different schemes:
()using the range set, and
50
Fundamentals of Research Methodology
(i) using the level of
Statistic, Luu Pub, UK 2005, p. 24) measurement (Hale Robert, Investigating
According to the Range Set, there are two different levels of
measurement, using the range set method of classification. The first is called
discrete or categorical. Adiscrete variable is finite (you can count it), and has
aone-to-one
correspondence with real
allowed). An example of a discrete variable whole numbers (no decimals are
is the number of cars in the
parking lot, or in the case of our earlier example, whether the patient was
using the Drug A(1) or the Drug B (2).
The other level of measurement using the range set
continuoUS. A continuous variable has an infinite number method is called
of values
associated with it. As an example, think of measuring peoples weights. The
first person might weigh 135 pounds on the bathroom scale.
more accurate scale, their weight might be 135.5 pounds. However, on a
values, (say 135.0 and 135.5 pounds), there are an infiniteBetween ofany two
possibilities (i.e., 135.125 pounds). Variables that have this number other
known as continuous variables. p1operty are
Note that it is the variable that has this
property, and
instrument that you use to measure the variable. The instrument maynot not
the
be
able to produce all the possible values that the variable may
have.
According to the Level of Measurement : When we classify variables
according to the level of measurement, they are said to be measured at the
nominal; ordinal, interval, or ratio levels of measurement.
Nominal: When the numbers that are assigned to a variable only
name that variable, the variable is said to be measured at the nominal level.
Examples of nominal variables would be serial numbers on equipment parts.
Another example would be the 1for Drug Aand the 2 for Drug B. When
numbers are given for variable values that would more appropriately take
text values, the variable is usually a nominal variable. Obviously, very little
math can be done with these numbers. The only operators that are
reasonable to use are equals (-) and not equals (=). It would be silly to add
two parts numbers together (0042 + 0056 = 0098) and ask for the sum of
their parts. People make very few eors applying math operators with
nominal level variables. The equals and not equals signs are appropriate
because part 0042 is obviously not equal to part 0056.
Ordinal: Ordinal level variables have a unique characteristic that
the measurement has order to it. A3 is higher than a 2 which is higher than
a 1. Examples of ordinal level variables indude: positions in a race,
percentle ranks, grade equivalency scores, etc. Obviously equals and not
equals (=, are stll appropriate to use with these variables. Also
appropriate are greater than (>) and less than (k)operators.
 Hypothesis- Statement and Formulation
51
mentioned
Ordinal variables have gotten us into much difficulty. As scores. A
ordinal level
above, qrade level scores on tests are simply
scoring higher than a
youngster achieving at the 6th grade level in rèading isdifference between a
youngster at the 4th grade level!. However, the difference)
youngster at the 6th grade level and the 4th grade level (2 years grade level
is not the same as the difference between a youngster at the 4this an ordinal
and the 2nd grade level (again 2 years difference): Because thisa horse race.
levelscore, the difference (2 places) is not the same. Think of
The difference between 6th and 4th places in a race is not necessarily the
same as between 4th place and 2nd. The second place horse ulay iiave
been well ahead of the 4th place finisher while the 4th, 5th and 6th place
finishers may have crossed the finish ine so closa to one another that a
photo was required to teliwho was ahead of whom.
Interval: When a variable is measured at the interval level, the
numbers are in order, and the distance between the numbers are equal.
Thus, the number 4 is as far ahead of 3 as 3 is ahead of 2 etc. Examples of
interval level variables are temperature in degrees Fahrenheit and
centigrade, and most of the dependent variables we are interested in the
socii sciences. Appropriate operators include =, , >, <, and addition (+)
and subtraction (). It is worth mentioning here, pararnetric statistics assume
that the dependent variable has been measured at the interval level or
higher. We willneed to add our numbers together, and subtract one number
from another and for that result to make sense. It makes sense to say that
the difference between 40 degrees and 20 degrees (20 units) is the same as
the difference between 100 degrees and 80 degrees.
Ratio : Ratio levei variables have order, equal intervals, and the
added property that the zero in the scale is real. When a ratio level variable
reads zero, none of what it measures exists. Examples include density, and
as long as we confine ourselves to the time and space defined on earth,
height and weight. Another ratio variable is degrees Kelin. If something is 0
degrees Kelvin, it cant get any colder (all molecular motion is lost).
However, note, that 0 degrees fahrenheight is not as cold as it can get
(Many of us have all been in Fahrenheit temperatures below zero). For ratio
level scores, appropriate operators indude =,,>,<, +, ", and muliply (x)
and divide ). While ratio level dependent variables are not required for
parametric statistics, when they are fourd, parametric statistics are used.
Independent variables are almost always measured at the nominal level,
while in the social sciences, dependent variables are often measured at the
interval level.

DIFFERENCE BETWEEN HYPOTHESIS AND THEORY

Atheory is awell-established principle that has been developed to
explain some aspects of the natural word. A theory arises from repeated
52 Fundamentals of Research Meihodology
observation and testing and incorporates facts, laws, predictions, and tested
hypotheses that are widely accepted. Ahypothesis is a specific, testable
predicion about what you expect to happen in your study. For example, a
study designed to look at the relationship betuween leadership style and
motivation might have a hypothesis that states, "This study is designcd to
assess the hypothesis that managers with better leadership ability will be
more motivated." Unless the study is exploratory in nature, the typothesis
should always explain what you expect to happen durin1g the course of your
experiment or research. Pearsall and Trunble, 1996)
While the terms are, sometimes, used interchangeably in general
practice, the difference between a theory and a hypothesis is important
when stuciying experimental iesign. Sone inportant istincticns io note
indlude:
" A theory predicts events in general tems, while a hypothesis
makes a specific prediction about a specified set of circumstances.
" A theory is a causal chain, or statements of a causal chain, which
have been proven logically/mathematically to be true. In other
words, a statement is made about a causai relationship ("X
because Y or "X, therefore, Y") and that statement is tested.
Evidence is gathered and presented regarding the staternent, and
evaluated; this might include the resuits of observation or
experiment, but in all cases evidence must be grounded in the
objective, external world.
A hypothesis is an unproven theory, e.g., it is a statement of a
causal chain or relationship, perhaps (one hopes) inciuding an
explanation of why it is likely to be true. It has not, however, been
subject to definitive or rigorous testing
LeOMPONENTS OF AHYPOTHESIS
Ideally, a hypothesis should have four major components.
However, not every hypothesis can be fit into this framework, but
knowledge of these four components is very helpful if there is an
incompletely formed hypothesis.
The first component is the subject group. In other
should answer the question who are you interested in words, it
component describes the group that the researcher wouldstudying ? This
study. Subjects
Could be consumers, workers, or managers, e.g., if a
know that increasing the bonus will result in an researcher wants to
employees. The hypothesis formulated can be thereoverall performance of the
of bonus on the perfomance of the is a positive influence
employees at a given incustry bonus
 Hypothesis-Statement and Formulation 53
the effect of
rate', cleariy points that the researcher is interested in studying group would
bonus on the pertormance of the employees. Here, the subject
be the employees.
The second component is the treatment or exposure. in other
words, what is being done to part or allof your subject group. A treatment
implies an action on your part, such as providing information or applying a
new therapy. An exposure, on the other hand, implies some action that you
do not control, such as the bonus being paid in the overall industry, or the
infiuence of employee union on the workers. Continuing with our example
of bonus and employee performance, the hypothesis makes it clear that the
researcher will measure the employee perfomance at different levels of
bonus rates assuming thal the industry bonus rate remains constant. iere,
the treatment is the different levels of bonus that the employees will be
getting assuming everything else remains constant.
The third component is the outcome measure. In other words,
how or in what manner is the treatmernt or exposure going to be assessed ?
MaM N

Iis very important that the outcome measure be defined precisely and
unambiguously. For example, if your outcome is performance, then what is
meant by performance has to be defined. In our performance, it can be
measured in terms of 'number of units produced', or 'number of defective
items produced' or 'nurnber of accidents taking place'.
The fourth component is the control group. In other words, who
are you comparing to ? It is important for the control group to be as similar
as possible to those who receive a treatment or exposure. For example, an
advertiser wishes to test a new advertisement campaign. He can choose a
group of consumers and after exposing them to the new advertisement,
Compare their purchase intentions to another group who continue to be
exposed to the old advertisement.
As mentioned earlier, not every research hypotheses will have
all four components. For example, a cross-over design involves applying
both anew treatment and a standard treatment using the same subjects. For
this study, the hypothesis would not involve a separate control group. In our
example, the researcher can compare the performance output of the same
employee group at different time intervals. Corelation studies look at
relationships within asingle group, such as a study of the factors that
improve performance. This type of study would not have a treatment/
exposure. Some examples of Research Question along with their
hypotheses and variables are given in Table 3.1. (Wagner, 2006)
54 Fundamentals of Research Methodology
Table : 3.1 Some Very Simple Examples of Variables
Question Hypothesis Independent Dependert Controlled Comments
Variable Variables Variables
How much jH,: More the faucet Water faucet Volume of Water Abetter measure
water flows tap is opened nore opening water flouing pressure (how of the
through a is the flow of water,(closed, i/2 ineasured in much the independent
faucet ? at constant water |open, fully |itres per water is variable uwould be
presure. open) minute "pushing") to find area of the
opening in the
|pipe in square
centimetres.
How fast |H: The candle bums|Time Height of Use same In this case, time
does a at the rate of lcm measured in |candle type of is what causes the
candle 0per minute in an minutes. measured in candle for dependent
bum enciosed room. jcentimetres. every test variable to
Wind change. The
make sure scientist simply
there is starts the process,
none then observes and
records data at
|regular inrtervals.
Does |H,:The growth of a Amount of Growth of Same plants
fertilizer plant depends upon fertilizer the plant Same soil
make a the amount of measured in measured by Same size
plant grow fertilizer that it has grams its height pot
tigger? been given. Growth of Same
the plant amount of
measured by water and
the number light
of leaves Make
measure
ments of
growth at
the same
time
Does an |H,:Speed of Voltage of the Sped of Same motor
electric rotation of an electric electricity rotation for every
motor tum motor depends on supplied to the measured in test
faster if you the voitage given for motor |RPMS Same load
increase the constant load on the measured in on the
voltage ? motor, volts/ motor

NPORMULATING A HYPOTHESIS
Formulation of the hypothesis basically varies with the kind of
research conducted. Causal Studies tend to study the relationship that exists
within the variable, while qualitative studies tend to generalize upon the
subject. Qualitative studies are characterized by :
Use of words- what, how
Use of non-directional wording in the question.
The questions are under continual review and reformulation-will
evolve and change during study.
Hypothesis-Statement and Formulation 55

Use of a single focus.

research is characterized by :
On the other harnd, quantitative
fiom Theory.
The testable proposition to be deduced
separated and
Independent and dependent variables to be
measured separately.
method' is more useful in
For Qualitative studies deductive
method tries to establish a
formulation of a hypothesis. The deductive hypothesis (which gets
pattern after observation and formuiate a tentative
eventually leading to theory.
modified as the research progresses) and of water, when an
Archimedes observation that there is a rise in the level
hypothesis that there is a
obiect is immersed in water, let him to formulate a in water. Further
change in the level of water when a body is inuersed
in the level of
observation made hin1 refine the hypothesis from change hypothesis was
water to 'volume of water dispersed'. And finally after this established as a
found to be true for different objects, the hypothesis was
theory.
used for hypothesis
In quantitative studies, an 'inductive method' is starts
formulation. The inductive method contrary to the deductive methods
prior to
from an established theory and the hypothesis is formulated
observation and the hypothesis is, consequently, confirmed, related e.g., the
Demand theory states that the demand for a product or service is to
demand holds
its price. A researcher is interested in knowing if the theory of dass
true for a market segment of quality conscious upper-middle hypothesis,.
consumers. He will use the inductive method in formulating the
Fig. 3.2 graphically represents the two approaches to hypothesis
formulation. (Michael, 1985)

Theory Theory
Tentative
Hypothesis
Hypothesis
Pattern Observation

Observaticn Confirmaion

Fig. 3.2 Approaches to Hypothesis Formulation

56 Fundamentals of Research Methodology
FYPES OF HYPOTHESES
As has been stated in the preceding text, the key component of any
hypothesis is its testability. This irnplies that after ahypothesis has been
formulated, it is tested for its truthfulne_s, this gives rise to two scenarios :
E) the formulated hypothesis is correct and the assumptions
postulated are accepted,
(i) the fomulated hypothesis fails to assert itself and, hence, the
assumptions postulated are rejected.
As the hypothesis is formulated prior to the research, there is a 50%
chance that the hypothesis will fail the test for its truthfulness. Therefore,
researchers forulate two hypotheses, each complerment of the other. The
hypotheses, thus, formmulated are null hypothesis and alternate hypothesis.
Null Hypothesis
The null hypoth.esis is generally that which is presumed to be true
initially. Hernce, we reject only when we are quite sure that it is false, often
90, 95, or 99% confident that the data do not support it. The null
hypothesis is represented as H or H,. The null hypothesis represents a
thecry that has been put forward, either because it has been believed to be
true or because it has to be used as a basis for argument, but has not yet
been proved to be true. It has serious outcome if incorrect decision has been
made.

Alternate Hypothesis
Alternate hypothesis is complementary to null hypothesis and is
represented as H, or H, The alternative hypothesis is a statement of
what ahypothesis test is set up to establish. Alternate hypothesis is opposite
of Null. It holds true only if H, is rejected. Frequently, "alternative" is actual
desired conclusion of the researcher.
We give special consideration to the null Hypothesis. This is due to
the fact that the null hypothesis relates to thestatement being tested,
whereas the altemative hypothesis relates to the statement to be accepted
iff/when the null is rejected. The final conclusion, once the test has been
carried out, is always given in terms of the nul hypothesis. We either 'reject
H, in favour of H; or 'do not reject H,'; we never conclude 'reject
H, or even' accept H, If we conclude 'do not reject Ho, this does not
necessarly mean that the nullbypothesis is true, it only suggests that there is
not sufficient evidence against H, in favour of H; rejecting the null
hypothesis, then, suggests that the alternative hypothesis may be true.
Hypothesis-Statement and Formulation 57

null hypothesis
For example, in a clirical trial of a new drug, the
than the current drug.
might be that the newdrug is no better, on average, two drugs on
the
We would write H, : there is no difference between
new drug has a
average. The alternative hypothesis miqht be that: the
We would
different effect, on average, compared to that of the current drug.
the rew
write HI : the two drugs have different effects, on average. Or, if
write H1: the
drug is better, on average, than the current drug. We would
new drug is better than the current drug, on average.
Limitations of Null Hypothesis
Formulation of a null hypothesis simpliftes the research process, but
it alsc faces certain limitations. Anull hypothesis is only useful if it is possible
to calculate the probability of observing adata set with particular parameters
from it. in general, it is much harder to be precise about how probable the
data would be if the alternative hypothesis is true.
If experimental observations contradict the prediction of the null
hypothesis, it means that either the nul hypothes1s is false, or we have
observed an event with very low probability. This gives us high confidence in
the falsehood of the null hypothesis, which can be improved by increasing
the number of trials. However, accepting the alternative hypothesis only
commits us to a difference in observed parameters; it does not prove that
the theory or principles that predicted such a difference is true, since it is
always possible that the difference could be due to additional factors not
recognized by the theory.
For example, rejection of a null hypothesis (that, say, rates of
syiptom relief in a sample of patients who received a placebo and a sample
who received a medicinal drug will be equai) alows us to make a non-null
statement (that the rates differed); it does not prove that the drug relieved
the symptoms, though it gives us more confidenice in that hypothesis.
The fomulation, testing and rejection of null hypotheses is
methodologically consistent with the falsficationist model of scientific
discovery fomulated by Karl Popper and widely believed to appl to most
kinds of empirical research. However,, concerns regarding the high power of
statistical tests to detect differences in large samples have led to suggestions
for re-defining the null hypothesis, for example, as a hypothesis that an
effect falls within a range considered negligible. This is an attempt to address
the confsion among non-statisticians between significant and substantial,
since large enough samples are likely to be able to indicate differences,
however, minor.
The theory underlying the idea of a nuil hypothesis is closely
associated with the frequentist theory of probability, in which probabilistic
staterments can only be made about the relative frequencies of events in
5
 58
Fundamentals of Research Methodology
arbitrarily large samples. Afailure to reject the null hypothesis is
only in relation to an arbitrarily large popuiation from meaningful
sanple is supposed to be drawn. which the observed
Simple Hypothesis and Composite Hypothesis
A simple hypothesis is a hypothesis which speciies the
distribution completely, while a composite population
hypothesis is a hypothesis which
does not specify the population distribution.
more common because they allow us to Composite hypotheses are
decide whether a sample comes
from any distribution of a specific type. In this
situation, the form of the
Gisuiouion is of interest, regardless of the values of the parameters.
Unfortunately, composite hypotheses are more difficult to work with
because the critica values are often hard to compute.
CHARACTERISTICS OF AGOOD HYPOTHESIS
Agood hypothesis shouid have the following
characteristics :
S Based on information
contained in the review of
1literature : A good hypothesis is well-rooted in existing theories and facts.
Hypothesis should not be based on isolated observations and popular myth.
This also asserts the importance of literature
review prior to hypothesis
formulation.
ü) Includes the independent and dependent variables : The
main purpose of hypothesis fomulation is testing. The hypothesis should
clearly speak about the relationship that exists between variables. It should
also be formulated such that distinction between
variable can be made. Hypotheses are always inindependent
declarative
and dependent
sentence form,
and they relate, either generally or specifically, variables to variables.
(iii) Conceptually. Clear :The hypothesis to be framed must be
conceptually clear. As far as possible, care must be taken to define the terms
used in the formulation in precise language. Moreove, the
be those already in use. One method of obtaining clarity is by definitions must
writing down
the list of concepts used in the research outline, then try and define
an order. It is always better to state the hypothesis as simply as them in
the hypothesis is not dear and precise, the inferences drawn onpossible. If
its basis
cannot be taken as reliable.
iv) Should point towards the line of action or th research
design : A research to be practical should relate the hypothesis to the
research design. This equires the researcher to have a sound knowledge of
techniques to test the hypothesis and, thus, to formulate practical questions.
Ifa researcher fomulates such a hypothesis that does not have a
design, the hypothesis cannot be tested. If one is not able to research
identify
 Hypothesis Statement and Formuation 59
correctly the techniques used to test the proposition, it would be a good idea
io initiate research in that direction.
w{u) Empirical Referents : If a hypothiesis is to be useful, it should
contain only such propusitions which are verifiabie or suitable for empiical
analysis. lt should be remembered that no usable hypothesis can embody
moral judgments. In other words, while a hypothesis may study value
judgments, it must be separated trom a moral preachment or a plea for
acceptance of one's values.
(vi) Consistent with the data : If there is a mismatch between the
unit of analysis of our hypothesis and the unit of analysis in our data, this
can lead to problems. An example to this is Ecologica! Fallacy. The mistake
of assurririg tiat wheire ieiaionships are found among aggregate data, these
relationships will also be found among individuals or households. This is a
common mistake that marketing researchers commit, whereby they would
predict the buying behaviour of individual consumers based on data gathered
at a collective level.

HYPOTHESIS TESTING FUNDAMENTALS

After the hypothesis has been formulated, the hypothesis needs to
be tested for its truthfulness. Various statistical methods are available for
hypothesis testing which will be dealt in detail in the ensuing chapters. Here,
we will discuss some basic concepts of hypothesis testing. Hypothesis testing
is a four step procedure :
(a) It starts with stating the hypothesis (Null or Alternative)
(b) Setting the criteria for decision
(c) Collecting the data
(d) Evaluating the Nul Hypothesis
There are two categories of tests that are applied to hypothesis testing.
One-tailed Tests : If your prediction specifies a direction, and the
null, therefore, is the no difference prediction and the prediction of the
opposite direction, we call this a one-tailed hypothesis. Taking the previous
example of testing the effect of a new drug in comparison to a placebo,
there can be three oossible scenarios :
() there is no effect of the drug,
(i) there is a positive effect of the drug,
(ti) there is a negative effect of the drug.
It we guess that there is a positive effect of the drug, we will use the
one-tailed test for hypothesis, as we already have a specific direction for the
fest. The null and alternate hypotheses might be stated
something like this :
The null hypothesis for this study is : H: there is no difference
between the two drugs on average.
60 Fundamentals of Research Methodology
Which is tested against the altenative hypothesis :
H,:Hl: the new drug is better than ihe current drug, on an average.
The null must account for the other two possible conditions : no
difference or a negative effect. Fig. 3.3 shows a hypothetical distribution of
patients showing inprovement and their corresponding T-Test values. The
shaded region shows the 1ejection region (RR). We can see that the tem
"one-tailed" refers to the tail of the distribution on the outcome variable.

Group B
X Bai

Frequency

.05

1.65
tValues
Fig. 3.3 One-Tailed Test

Group B
X Bar

Frequency

.025

-2 or 1.96 -1 +1 +2 or 1.96
t Values

Fig. 3.4 Two-Tailed Test

Two-tailed Tests : When the 'prediction (hypothesis) does not
specify direction, then we use the two-tailed test. Continuing with the
Hypothesis-Statement and Formlation 61
a change in
previous example, if the researcher is sure that there has been
null and alternate
the health of the patients (either positive or negative), his
hypothesis may be stated as folilows :
no
The nuli hypothesis for this study will be : Ho: there is
difference between the two drugs on an average.
which
i ks Note that the rnull hypothesis remains the same in both cases,
isetested against the aiternative hypothesis : H: the two drugs have
different effects, on an average.
The term "two-tailed" reters to the tails of the distribution for your
outcome variable.

Errors in Hypothesis Testing : The hypothesis testing procedure

has the chance of making two types of errors. In a hypothesis test, a Type I
error occurs when the nul hypothesis is rejected when it is, in fact, true;
that is,H, is wrongly rejected. For example, in a cinica! trial of a new drg.
the null hypothesis might be that the new drug is no better, cn an average,
than the current drug; i.e., Ho: there is no difference between the two
drugs cn an average.
A type I error would occur if we concuded that the two drugs
produced different effects when, in fact, there was no difference between
them.

Table 3.2 gives a summary of possible results of any hypothesis test.

Table 3.2 Errors in Hypothesis Testing
Decision
Reject Ho Don't reject H,
Truth Ho Type IError Right decision
H, Right decision Type II Error
Atype I error is often considered to be more serious and, therefore,
more important to avoid, than a type lI error. The hypothesis test procedure
is, therefore, adjusted so that there is a guaranteed low probability of
rejecting the nul hypothesis wrongly; this probability is never 0: This
probability ofatype Ierror can be precisely computed as
P(type Ierror) =significance level= a
The exact probability of a type Il error is generally unknown. If we
do not reject the nul hypothesis, it may still be false (a type II error) as the
sample may rot be big enough to identify the falseness of the null
hypothesis (especially if the truth is very close to hypothesis).
 62 Fundamentals of Research Methodology
For any given set of data, type Iand type Il errors are inversely
related; the smaller the risk of one, the higher the risk of the other. A type l
error can also be reterred tO as an error of the first kind.
Type II Error
In a hypothesis test, atype ll eror occurs when the null
hypothesis
HÍ, is not rejected when it is, in fact, false. For example, in a clinical trial of
anew drug, the nul hypothesis might be that the new drug is no
an average, than the current drug, i.e.,
beiter, on
H: there is no difference between the two drugs on an
average.
Atype II error would occur if it was concluded that the two drugs
produced the same effect, i.e., there is no difference between the two drugs
on an average, when, in fact, they produced different ones. A type II error is
frequentiy due to sample sizes being too small.
The probability of atype II error is generally unknown, but is
symbolized by b and written as :
P(type l! error) =
A type lerror can also be referred to as an error of the second kind.
Test Statistic
A test statistic is a quantity calculated from our sample of data. Its
value is, used to decide whether or not the null hypothesis should be rejected
in our hypothesis test. The choice of a test statistic will depend on the
tssumed probability model and the hypotheses under question.
Critical Value(s)
The critical value(s) for a hypothesis test is a threshoid to which the
value of the test statistic in a sample is compared to determine whether or
not the nul hypothesis is rejected. The critical value for any hypothesis test
depends on the significance level at which the test is carried out, and
whether the test is one-sided or two-sided.
Critical Region
The critical region CR, or 1ejection region RR, is a set of values of
the test statistic for which the null hypothesis is rejected in a hypothesis test.
That is, the sample space for the test statistic is partitioned into two regions
one region (the critical region) will lead us lo reject the null hypothesis H,
the other will not. So, if the observed value of the test statistic is a member
of the critical region, we conclude "Reject H; if it is not a member of the
critical region, then we conclude "Do not reject H,".
 Hypothesis- Statement and Formulation 63
Significance Level
The significance level of a statistical hypothesis test is a fixed
probability of wrongly rejecting the null hypothesis H, if it is, in fact, true. It
is the probability of a type I error and is set by the investigator in relation to
the consequences cf such an error. That is, we want to make the
significance level as small as possible in order to protect the null hypothesis
and to'prevent, as far as possible, the investigator from inadvertently making
false claims.
The sionificance level is usually denoted by 'a'
Significance Level =P(type Ieror) =a
Üsualiy, the significance level is chosen to be 0.05 (or equivalently, 5%).
P-Value
The probability value (p-value) of a statistical hypothesis test is the
probability of getting a value of the test statistic as extreme as or more
extreme than that observed by chance alone, if the null hypothesis Ho, is
true. It is the probability of wrongly rejecting the null hypothesis if it is, in
fact, true. It is equal to the significance level of the test for which we would
only just reject the null hypothesis. The p-value is compared with the actual
significance level of our test and, if it is smaller, the result is significant. That
is, if the null hypothesis were to be rejected at the 5% significance level, this
would be reported as "p<0.05".
Smallp-values suggest that the null hypothesis is unlikely to be true.
The smaller it is, the more convincing is the rejection of the null hypothesis.
it indicates the strength of evidence for, say, rejecing the null hypothesis H,
rather than sinply conluding "Reject H," r "Do not reject H".
Power
The power of a statistical hypothesis test measures the test's ability to
reject the nul hypothesis when it is actually false that is, to make a correct
decision. n other words, the power of a hypothesis test is the probability of
not committing a type II error. It is calculated by subtracting the probability
of a type ll error from I, usually expressed as:
Power =1-P(type II error) =(- B)
The maximum power a test can have is , the minimum is 0. Ideally,
we wantatest to have high power, close to I.

HYPOTHESIS AS ABASIS FOR SCIENTIFIC ENQUIRY

Hypothesis provides a tentative explanation of phenomena and
facilitates the extension of knowledge in an area. It is based on existing
 64 Fundamentals of Research Methodology
theories and purports io extend the limits of existing knowledge. If a
hypothesis is accepted, it willeither reinforce the theories already prevalent,
or will add to a new dimension as in the case of the
hypothesis not being
accepted. This is also the reason that hypothesis is formulated
null and alternative hypothesis. In either of the two being in pairs of
an overall gain of accepted, there is
knowledge.
It provides the investigator with a relational statement that is
testable in a research study. This makes the overall directy
scientific deduction and analysis more objectiye andcumbersome process of
foiows is to test whether the hypothesis made is simplified. Whatever
providing direction to research. acceptable or not, thus,

Hypothesis provides a framework for reporting conclusions of the

study. It could be considered as the working
Hypctheses can be deduced from theory and from other instrument of theory.
hypotheses.
However, it is not always accepted that hypothesis is necessary for
conducting a study. It is said that in conducting studies to
hypothesis needs to be framed. Generally, the studies find out facts, no
government or autononous organisations would fit into initiated by the
this type. For
instance, inquiry ordered into firing, misappropriation of funds,
come under this category. The inquiries in these conditions are tokillings
only
etc.
facts. They are not supposed to hypothesise that the incident might or find
not have happened in a particular manner. might
Similarly, many studies
ordered by various independent organisations with specific terms areof
reference. In such cases,their job is to conduct inquiry on the lines that have
been suggested. For instance, a committee appointed by RBÍ to go
into the
prospects of Rural credit services in India goes on to review the situation and
report the conditions prevailing. The committee is not required to
hypothesise whether ihere is or there is no scope for rural credit in India.
However, researchers may start with certain implied or working
hypothesis and proceed to conduct an inquiry. As such, all scientific inquiries
are conducted only after fomulation of a few hypotheses and the fact
finding inquiries may not find a hypothesis necessary.
SUMMARY
Hypothesis formulation is mostly related to causal
empirical research. The term hypothesis has been derived from research
or
the ancient
Greek, hypotithenai, meaning "to put under" or "to suppose". Ahypothesis
is a "supposition made as a starting point for further
known facts". it provides the investigator with a relationalinvestigation from
statement that is
directly testable in a research study. One of the major purposes for which a
hypothesis is formulated is defining therelationship between variables.
 65
Hypothesis Statement and Forrrulation
two diffeient schemes :
Variables are classified according to
() using the range set, and
(ii) using the level of neasurement.
measurement, variables are either nominal, ordinal,
According to
ratio or interval.
has four components; the subject group, the treatment
iA hypothesis the control group. Two approaches are
and
group, the outcome measure the inductive method or the deductive
hypothesis,
used in forn1ulating a in pairs that are mutually complimentary
Hypothesis is formulated
method.
These are called null and alternate hypothesis. After the
of each other. the hypothesis needs to be tested for its
hypothesis has been formulate,
truthfulness.
It starts with stating the
Hypothesis testing is a four step procedure. the criteria for decision.
hypothesis (Null or Alternative), fllowed by setting
place, foliowed by evaluating the Null
After which collecting the data takesprocedure
hypothesis testing has the chance of making two
Hypothesis. The
Type Ierror and Type Il error
types of errors. In a hypothesis test, a
The significance level and the power of a test signify the ability of the
occur.
hypothesis.
test to accurately assess the truthfulness of a
REFERENCES
of Analysis, Worid Scientific
1. Chiang Chin Long, 'Statistical Methods
Publishing Company, Singapore, p. 154, (2003).
Introduction to
2. Donald Ary, Lucy Chese Jacobs & Asghar Razavieh, (1994).
Research in Education, Holt, Rinehart and Winston, Inc.
3. Earl R. Babbie, Survey Research Methods, Wadsworth publishing
Company, Inc. (1973).
4. Hale Robert, Investigating Statistic, Lulu Pub, UK, p. 24, (2005).
5. Leedy, Paul. D, and Ormrod, Jeanne E, Practical Research: Planning
and Design. (7th Ed.), PHI, NJ, (2001).
6. Michael VP, Research Methodology in Management, Himalaya
Publishing House,Bornbay, p. 110,(1985).
7. Net TOM, Intoduction to Research Methods,
http://cbdd.wsu. edu/edev/_NetTOM_ToT/Resources/Other/TOM614/
page45.htm,p. 45, (2006).
8. Pearsall, J. & Trumble, B. (eds), Oxford English Reference Dictionary,
Oxford: Oxford University Press, (1996).