[Singh, 4(1): January, 2015] ISSN: 2277-9655

Scientific Journal Impact Factor: 3.449

(ISRA), Impact Factor: 2.114

IJESRT
INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH
TECHNOLOGY
HYPOTHESIS TESTING IN RESEARCH METHODOLOGY: A REVIEW
Balkeshwar Singh*, Asha Singh
*Mechanical Engineering Section, Salalah College of Technology, Sultanate of Oman
Department of Business Study, Mewar University, India
ABSTRACT
Hypothesis is usually considered as the principal instrument in research. It plays a major role in research. Its main
function is to suggest new experiments and observations. It occupies a very small space in the thesis. A researcher
cannot proceed in the research work without formulating one or more than one hypothesis. Hypothesis brings clarity,
specificity and focus to a research problem. There are two types of hypothesis which are called the Null Hypothesis
and the Alternative Hypothesis. There are four steps in hypothesis testing. After getting the results, the researcher tests
whether the collected facts support the hypothesis formulated or not. There exist a number of statistical tools like t-
test, F-test, chi-square test, D-W test, etc., to test the validity of hypothesis. Hypothesis testing shows whether the
hypothesis should be accepted or rejected. It has been rightly said that “You torture the data until they confess”. This
paper include the introduction, steps of hypothesis testing, function of hypothesis and testing of hypothesis.

KEYWORDS: Research, Hypothesis, null hypothesis, alternate hypothesis, population, significance.

INTRODUCTION (v) I can reject Hypothesis 5 because my electric
Hypothesis testing allows us to carry out inferences bill payment was up to date. By repairing or
about population parameters using data from a sample. replacing the cord, I can count on adequate
Suppose I came home after dark, open the front door light from the lamp in the near future. This
and reach inside for the switch that turns on a nearby way Hypothesis 4 solved my problem.
table lamp i switch the light but there is no light. That
time I begin to construct a series of reasonable STEPS OF HYPOTHESIS TESTING
guesses-hypothesis-to explain the lamp’s failure: In order to test a hypothesis in statistics, we must
(i) The bulb has burned out. perform the following steps:
(ii) The lamp has not plugged into the wall outlet. Step 1: State the hypotheses.
(iii) A recent weather event interrupted our Step 2: Set the criteria for a decision.
electrical service. Step 3: Compute the test statistic.
(iv) The wire from the lamp to the wall outlet has Step 4: Make a decision.
defective. Next, we will examine each one of these steps.
(v) I forgot to pay our electrical bill. Step 1: State the hypotheses: In order to undertake
Now I will go in search of information to determine hypothesis testing we need to express our research
which hypothesis is correct. hypothesis as a null and alternative hypothesis. If we
(i) I got a flashlight from my car, find a new are to compare method A with method B about its
bulb, and put the new bulb in the lamp. The superiority and if we proceed on the assumption that
lamp fails to light. (Hypothesis 1 is rejected). both methods are equally good, then this assumption
(ii) I plugged the lamp in the wall outlet. is termed as null hypothesis (H0) and if we think that
(Hypothesis 2 is rejected). the method A is superior or the method B is inferior,
(iii) I saw my neighbour’s house. Everyone has we are then stating what is termed as alternative
electric power. (Hypothesis 3 is rejected). hypothesis (Ha).
(iv) I lifted the cord that connects the lamp to the In other words we can say
wall outlet. Then lamp lighted briefly and H0: The Null Hypothesis
then goes out. I lifted the cord again. Again This is the hypothesis or claim that is initially assumed
the lamp lighted briefly. The connecting cord to be true.
has defective. (Hypothesis 4 is accepted). Ha: The Alternative Hypothesis
This is the hypothesis or claim which we initially

http: // www.ijesrt.com © International Journal of Engineering Sciences & Research Technology

[729]
[Singh, 4(1): January, 2015] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

assume to be false but which we may decide to accept the critical value obtained.
if there is sufficient evidence. 5. Make a decision.
Suppose we have to prove given hypothesis: If the test statistics falls in the critical region: Reject
H1: The automobile A is performing as well as H0 in favour of Ha.
automobile B. If the test statistics does not fall in the critical region:
Therefore, the null and alternative hypotheses are: Conclude that there in not enough evidence to reject
H0: The automobile A is performing much better than H0.
automobile B.
Ha: The automobile A is not performing much better Step 4: Make a decision: If P, is less than 0.05 or
than automobile B. 0.001, we should reject the null hypothesis in favor of
Types of errors in testing of hypothesis: When a the alternative. Alternatively, if P is greater than 0.05,
hypothesis is tested there are four possibilities: we should not reject the null.
1. The hypothesis is true but our test rejects
it.(Type I error) THE FUNCTIONS OF A HYPOTHESIS
2. The hypothesis is true but our test rejects A hypothesis serves the following functions:
it.(Type II error)  The formulation of a hypothesis provides a
3. The hypothesis is true but our test rejects study with focus. It tells us what specific
it.(correct decision) aspects of a research problem to investigate.
4. The hypothesis is true but our test rejects  A hypothesis tells us what data to collect and
it.(correct decision) what not to collect, thereby providing focus
Type I error means rejection of hypothesis which to the study.
should have been accepted. It is denoted by  (alpha)  The construction of a hypothesis enhances
known as  error, also called the level of significance objectivity in the study.
of test.  It enables us to conclude specifically what is
Type II error means accepting the hypothesis which true or what is false.
should have been rejected. It is denoted by  (beta) TESTING OF HYPOTHESIS
known as  error. Statisticians have developed several tests of
If Ha is of the type greater than (or of the type lesser hypothesis which are also known as the tests of
than), we use a one-tailed test, but when Ha is of the significance and can be classified as:
type “whether greater or smaller” then we use a two 1. Parametric Tests or Standard tests of
tailed test. hypothesis.
2. Non-parametric Tests or Distribution-free
Step 2: Set the criteria for a decision: The tests of hypothesis.
significance level for a given hypothesis are tested on Parametric Tests or Standard tests of hypothesis:
a pre-determined level of significance, generally, in These tests are based on the assumption of normality.
practice, either 5%level or 1%level is adopted. (i) Z-test
(ii) T-test
Step 3: Compute the test statistic: Hypothesis (iii) 2-test
establishes the relationships of concept with theory (iv) F-test
and specifies the test to be applied especially in the Non-parametric Tests or Distribution-free tests of
context of a meaningful judgment. To test a hypothesis hypothesis: Non-parametric tests do not make an
we need to go through a process that comprises five assumption about the parameters of the population and
phases: thus do not make use of the parameters of the
1. Identify the Null hypothesis H0 and the distributions. These are:
Alternate Hypothesis Ha . (i) Sign Tests
2. Determine a test size (significance level). (ii) Fisher-Irwin Test
3. Select the test statistic and determine its value (iii) McNemer Test
from the sample data. This value is called the (iv) Wilcoxon Matched-pairs Test
observed value of the test statistic.’t’ statistic (v) Rank Sum Tests
is usually appropriate for a small number of (vi) One sample Runs Test
samples; for large number of samples, a ‘z’ (vii) Spearman’s Rank Correlation
statistic can work well if data are normally
distributed. CONCLUSION
4. Compare the observed value of the statistic to The relationship between the two concepts is a
http: // www.ijesrt.com © International Journal of Engineering Sciences & Research Technology
[730]
[Singh, 4(1): January, 2015] ISSN: 2277-9655
Scientific Journal Impact Factor: 3.449
(ISRA), Impact Factor: 2.114

hypothesis. The testing of hypothesis reveals to the Kaur, sage publication, 2011, ISBN – 978-1-
researcher the truth. It determines the validity of the 84920-301-2.
assumption. In formulating a hypothesis it is 3. Research Methodology, Dr. Prasant Sarangi,
important to ensure that it is simple, specific and Taxmann Publication, 2010 ISBN – 978-81-
conceptually clear, able to be verified, rooted in an 7194-737-9.
existing body of knowledge and able to be 4. Practical Research, Tenth Edition, Paul D.
operationalised. The null hypothesis and alternate Leedy & Jeanne Ellis Ormrod, Pearson,
hypothesis are chosen before the sample has drawn. 2013, 978-0-13289950-5.
Generally, in hypothesis testing we proceed on the 5. Research Methodology, Second Edition,
basis of null hypothesis, keeping the alternative C.R. Kothari, New age international
hypothesis in view because on the assumption that null publishers, 2004, ISBN (13):978-81-224-
hypothesis is true, one can assign the probabilities to 2488-1.
different possible sample results, but this cannot be 6. Dawson, Catherine, Practical Research
done if we proceed with the alternative hypothesis. In Methods, New Delhi, UBS Publishers’
the process of research testing of formulated Distributors, 2002
hypothesis is very important step. 7. Shuttleworth,Martin(2008).
http://www.experiment
REFERENCES resources.com/definition-of-search.html.
1. Research Methods in Business Studies, Third 8. Creswell, J. W. (2008). Educational
Edition, Pervez Ghauri &Kjell Gronhaug, Research: Planning, conducting, and
Pearson Education, 2005, ISBN – 0-273- evaluating quantitative and qualitative
68156-7. research (3rd ed.). Upper Saddle River:
2. Research Methodology, Third Edition, Ranjit Pearson.

http: // www.ijesrt.com © International Journal of Engineering Sciences & Research Technology

[731]