Market research is the systematic design, 

collection, analysis and reporting of data and findings relevant

to a specific marketing situation facing the company.

CHARACTERISTICS OF GOOD
MARKETING RESEARCH

Following are the characteristics of good marketing research

1. Scientific method. Effective marketing research uses the principles of the

scientific method: careful observation, formulation of hypotheses, prediction, and
testing.
2. Research creativity. At its best, marketing research develops innovative ways
to solve a problem.
3. Multiple methods. Competent marketing researchers shy away from over-
reliance on any one method, preferring to adapt the method to the problem
rather than the other way around. They also recognize the desirability of
gathering information from multiple sources to give greater confidence.
4. Interdependence of models and data. Competent marketing researchers
recognize that the facts derive their meaning from models of the problem. These
models guide the type of information sought and therefore should be made as
explicit as possible.
5. Value and cost of information. Competent marketing researchers show
concern for estimating the value of information against its cost. Value/cost
evaluation helps the marketing research department determine which research
projects to conduct, which research designs to use, and whether to gather more
information after the initial results are in. Research costs are typically easy to
quantify, while the value is harder to anticipate. The value depends on the
reliability and validity of the research findings and management's willingness to
accept and act on its findings.
6. Healthy skepticism. Competent marketing researchers will show a healthy
skepticism toward assumptions made by managers about how the market works.
 7. Ethical marketing. Most marketing research benefits both the sponsoring
company and its consumers. Through marketing research, companies learn more
about consumers' needs, and are able to supply more satisfying products and
services. However, the misuse of marketing research can also harm or annoy
consumers. There are professional ethical standards guiding the proper conduct
of research.

objectives of market research

1. Market Attractiveness Evaluation:

o Market sizing - estimating the size of a total market, how much is
accessible by the business, and/or what market share ambitions would
be realistic.
o Competitive presence and customer preference.
o Customer spending patterns, budget cycles, and intent.
o Channel trends, preferences, and allegiances.

2. Customer Insight: Specific customer needs, aspirations, buying behaviors,

usage patterns, decision models, preferences, favourability, intentions, etc

3. Competitive Forces: Current and potential basis of competition in a

market.

4. Communications Planning: What information sources do prospective

customers pay attention to, how to reach them, opportunities / vehicles for
influencing target customers and which are most effective.

5. Product Testing: Evaluation of product improvements, alternatives,

packaging etc.

6. Concept Testing: Evaluation of potential products and solutions,

clarification of needs, wants, and preferences.
 7. Advert Testing: Evaluation of alternative brand promises, impact / cut
through ability, persuasiveness, strength of call to action, out-take versus
intent, etc.

8. Customer Satisfaction: Measurement of quality of customer experience,

perceptions, reaction, loyalty, intent, etc.

9. Pricing: Testing of price / feature / quality / packaging / positioning

combinations, price points, promotions, loyalty schemes, terms and
conditions, etc.

10. List Building: Compilation of information about prospective customers

for direct marketing purposes (NB may not be compatible with ethical
guidelines followed by many market research practitioners).

activities of market research

Market researchers have expanded their activities and techniques. The ten most common activities of
market research are –

1. Determination of marketing characteristics

2. Measurement of market potentials
3. Market share analysis
4. Sales analysis
5. Studies of business trends
6. Short range forecasting
7. Competitive product studies
8. Long range forecasting
9. Pricing studies, and
 10. Testing of existing products

Importance of marketing research

Marketing research can be defined as the process of gathering recording and analyzing the data related
to certain products and services. This need for market research is derived from the concept that only by
understanding the needs and wants of the target audience and by effectively meeting them, you will be
able to achieve the organizational goals and surpass the competition in the specific market. Thus, arises
the need to collect data about the customers, competitors, and other forces in the marketplace. This data
in turn is collected and analyzed to make relevant marketing decisions, be it in relation to setting up a
business, developing a product, creating a brand or coming up with an advertising campaign. 

There are two methods of marketing research, namely primary research or secondary market research.
While primary research seeks to understand customer motivations, opinions and needs through
quantitative and qualitative field research. In contrast to that, secondary market research uses already
existing sources of information to gather the data.
What is the Importance of Marketing Research?

To Make Marketing Decisions: Marketing research helps the marketers to make a decision about the
product or service. Sometimes a marketer might believe that the new product or service is useful for the
customers. However, research may show that customers do not need a product or are meeting their
needs with a certain competitor product and so on. Similarly good marketing research strives to provide
options for the successful introduction of new products and services. This makes the market entry of a
new product or service less risky.

Survive the Competition: Marketing research helps in ascertaining and understanding competitor

information such as their identity, marketing network, customer focus and scale of operations. This helps
in surviving and in certain cases, even leaving behind the competition. Moreover, with market research
you can also help understand the under-served consumer segments and consumer needs that have not
been met. 

Helps to Decide Target Markets: Research helps provide customer information in terms of their location,
age, buying behavior and gender. This helps the marketers zero in on the target markets and customers
for their products and services.

Maximize Profits: Apart from profit maximizing steps such as item optimization, customer profitability
analysis, and price elasticity, marketing research allows you to find out methods that can help you
maximize profits. For example, a product's price elasticity research can help you ascertain the impact of
an increased price on the sales and the profits of a product. This emphasis on profitability also helps the
company's focus to shift from maximizing sales to increasing the profits of a company. This helps the
company survive in the long run and maximize its profits.

Increasing the Sales: Increasing the sales of your products or services helps a company in maximizing
its profits. By understanding the customer's needs, wants and attitude towards the products and
determining whether your products fit the bill, marketers can increase their sales. This helps in not only
increasing the sales to the target customers and people already using the product but also converting the
non users into customers for the product.

Limitations of Marketing Research – 

1) Marketing research offers suggestions and not decisions:- Marketing research is not substitute for
decision making process. It only offers possible suggestion to marketing problem. It actually acts as a tool
which facilitates decision making process. It guides marketing managers in taking balanced, result oriented
and rational decisions. The suggestions offered by marketing researchers are usually possible solution but
not the exact solution. 

2) Marketing research cannot predict accurately: - In MR, efforts are being made to estimate or
predict the possible future situation. For this, certain research studies are undertaken. However, the results
conclusions arrived at may not be complete, perfect or accurate. They predict possible tendencies but not
certainties.

3) MR conclusions are not dependable:- there are many who are sceptics of MR. their criticism is that
MR conclusions are not dependable. There are examples where the research failed to deliver desired result
or a product failed even when the research had shown promising market demand and consumer support.
The classical example is that coke. Its MR should that 68% customers in US liked the taste of the new
formula developed by the coca-cola corporation. However, the new coke failed and in less than six months
of its launch, the coke management had to relaunch old coke under the brand name coke classic. However,
this failure of MR was mainly due to conventional approach.
4) Marketing research cannot study all marketing problem:- marketing research is rather very wide in
scope. However, it cannot study all marketing problem particularly where it is difficult to collect relevant
data. Similarly, research study is not useful for dealing with urgent marketing problem where quick
decisions and follow up actions are necessary. Thus, all marketing problem are not researchable and all
research problem are not answerable. MR fails to offer guidance to manager while dealing with specific
problem. It is also argued that very many times, marketing research faced by a business unit marketing
research tends to be fragmentary in its approach it become difficult to have an overall perspective in which a
marketing problem is to be viewed and studied.

5) Resistance by marketing executives: researchers study the marketing problems and offers
information and guidance to marketing executives in their decision-making process. However, some
executives are reluctant to use the solutions suggested by the researchers. They feel that such extensive
use will act as a threat to their personal status. Findings of the research work may bring them in difficulties
if the policy-decisions taken accordingly prove to be wrong.

6) Lengthy and time-consuming activity: MR is a lengthy and time-consuming activity. It involves

various steps which need to be completed in an orderly manner. It is not desirable to conduct research work
is an haphazard manner. Naturally, the research work take longer period for completion and findings when
available may prove to be old and outdated. Moreover, research work cannot be conducted overnight and
marketing decision cannot be postponed till the research work is completed. 

7) Marketing research is a costly affair: marketing research is a costly as research work require the
service of expert with knowledge, maturity and skill. Such experts include economists, management
scientists, statisticians, psychologists and computer experts. Advance training in economics, computer
technology, psychology, sociology, etc. is also necessary on the part of research staff. 

8) Non-availability of qualified staff: For scientific MR, professional marketing researcher with proper
qualification, training and experience are necessary. Research work is likely to be incomplete in the absence
of such expert staff.

9) Complexity of subject: marketing research fails to give complete and full-proof solution to
management. This is because marketing research itself is not an exact science. It is concerned with the
study of human beings and human behaviour is always difficult to predict.

10) Changing behaviour of consumers: Consumer is a focal point in the market research. However, his
buying motives are difficult to judge precisely and accurately. This brings some sort of uncertainty in the
conclusion drawn from the research activity. The findings of the research work may not prove to be
accurate.

Benefits of MIS
1)helps to recognise market trends: MIS helps managers to recognise market trends, in respect of
price, designs of products, fashions, etc. timely information of the market trend enables the firm to
follow the right course of action.

2)facilitates marketing planning and control: effective marketing planning is required in terms of
product planning, pricing, promotion and distribution. such planning will be possible only if the
company is possessing adequate and relevant information. this is also possible through MIS.

3)quick supply of information: today a firm has to take quick decisions. for this purpose it requires
fast flow of information which is facilitated by a properly designed MIS.

4)improves quality of decision making: a properly designed MIS supplies reliable and relevant
information. with the help of computers and other data processing equipments, the marketing
managers can make the right decisions at the rught time.

Research design process

The steps in the design process interact and often occur simultaneously. For example, the design of a
measurement instrument is influenced by the type of analysis that will be conducted. However, the type of
analysis is also influenced by the specific characteristics of the measurement instrument.

Step 1: Define the Research Problem

Problem definition is the most critical part of the research process. Research problem definition involves
specifying the information needed by management. Unless the problem is properly defined, the
information produced by the research process is unlikely to have any value. Coca-
Cola Companyresearchers utilized a very sound research design to collect information on taste
preferences. Unfortunately for Coca-Cola, taste preferences are only part of what drives the soft drink
purchase decision.

Research problem definition involves four interrelated steps: (1) management problem / opportunity

clarification, (2) situation analysis, (3) model development, and (4) specification of information
requirements.

The basis goal of problem clarification is to ensure that the decision maker’s initial description of the
management decision is accurate and reflects the appropriate area of concern for research. If the wrong
management problem is translated into a research problem, the probability of providing management
with useful information is low.
Situation Analysis
The situation analysis focuses on the variables that have produced the stated management problem or
opportunity. The factors that have led to the problem/opportunity manifestations and the factors that
have led to management’s concern should be isolated.

A situation analysis of the retail trade outflow problem revealed, among other things, that (1) the local
population had grown 25 percent over the previous five years, (2) buying power per capita appeared to be
growing at the national rate of 3 percent a year, and (3) local retail sales of nongrocery items had
increased approximately 20 percent over the past five years. Thus, the local retailers sales are clearly not
keeping pace with the potential in the area.

Step 2: Estimate the Value of the Information

A decision maker normally approaches a problem with some information. If the problem is, say, whether
a new product should be introduced, enough information will normally have been accumulated through
past experience with other decisions concerning the introduction of new products and from various
other sources to allow some preliminary judgments to be formed about the desirability of introducing
theproduct in question. There will rarely be sufficient confidence in these judgments that additional
information relevant to the decision would not be accepted if it were available without cost or delay. There
might be enough confidence, however, that there would be an unwillingness to pay very much or wait very
long for the added information.

Step 3: Select the Data Collection Approach

There are three basic data collection approaches in marketing research: (1) secondary data, (2) survey
data, and (3) experimental data. Secondary data were collected for some purpose other than helping to
solve the current problem, whereas primary data are collected expressly to help solve the problem at
hand.

Step 4: Select the Measurement Technique

There are four basic measurement techniques used in marketing research: (1) questionnaires, (2) attitude
scales, (3) observation, and (4) depth interviews and projective techniques.

Primary Measurement Techniques

I. Questionnaire – a formalized instrument for asking information directly from a respondent concerning
behavior, demographic characteristics, level of knowledge, and/or attitudes, beliefs, and feelings.

II. Attitude Scales – a formalized instrument for eliciting self-reports of beliefs and feelings concerning an
object(s).

A. Rating Scales – require the respondent to place the object being rated at some point along a
numerically valued continuum or in one of a numerically ordered series of categories.
B. Composite Scales – require the respondents to express a degree of belief concerning various attributes
of the object such that the attitude can be inferred from the pattern of responses.

C. Perceptual maps – derive the components or characteristics an individual uses in comparing similar
objects and provide a score for each object on each characteristic.

D. Conjoint analysis – derive the value an individual assigns to various attributes of a product.

I. Observation – the direct examination of behavior, the results of behavior, or physiological changes.

II. Projective Techniques and Depth Interview – designed to gather information that respondents are
either unable or unwilling to provide in response to direct questioning.

A. Projective Techniques – allow respondents to project or express their own feelings as a characteristic of
someone or something else.

B. Depth Interviews – allow individuals to express themselves without any fear of disapproval, dispute, or
advice from the interviewer.

Step 5: Select the Sample

Most marketing studies involve a sample or subgroup of the total population relevant to the problem,
rather than a census of the entire group.

Step 6: Select the Model of Analysis

It is imperative that the researcher select the analytic techniques prior to collecting the data. Once the
analytic techniques are selected, the researcher should generate fictional responses (dummy data) to the
measurement instrument. These dummy data are then analyzed by the analytic techniques selected to
ensure that the results of this analysis will provide the information required by the problem at hand.

Step 7: Evaluate the Ethics of the Research

It is essential that marketing researchers restrict their research activities to practices that are ethically
sound. Ethically sound research considers the interests of the general public, the respondents, the client
and the research profession as well as those of the researcher.

Step 8: Estimate Time and Financial Requirements

The program evaluation review technique (PERT) coupled with the critical path method (CPM) offers a
useful aid for estimating the resources needed for a project and clarifying the planning and control
process. PERT involves dividing the total research project into its smallest component activities,
determining the sequence in which these activities must be performed, and attaching a time estimate for
each activity. These activities and time estimates are presented in the form of a flow chart that allow a
visual inspection of the overall process. The time estimates allow one to determine the critical path
through the chart – that series of activities whose delay will hold up the completion of the project.

Step 9: Prepare the Research Proposal

The research design process provides the researcher with a blueprint, or guide, for conducting and
controlling the research project. The blueprint is written in the form of a research proposal. A
writtenresearch proposal should precede any research project.

 Qualitative marketing research

Qualitative research is a set of research techniques, used in marketing and the social sciences, in which data are obtained
from a relatively small group of respondents and not analyzed with statistical techniques. This differentiates it
from quantitative research in which a large group of respondents provides data that is statistically analyzed.

1 The role of qualitative research

Qualitative research methods are used primarily as a prelude to quantitative research. They are used to define a problem,
generate hypotheses, identify determinants, and develop quantitative research designs. They are inexpensive and fast.
Because of the low number of respondents involved, these exploratory research methods cannot be used to generalize to
the whole population. They are however, very valuable for exploring an issue and are used by almost all researchers.
They can be better than quantitative research at probing below the surface for affective drives and subconscious
motivations.

2 Approaches

Most qualitative methods use a direct approach : they clearly disclose the purpose of the study and the organization that
commissioned it. Questions are direct and to the point. Many other qualitative techniques use an indirect approach. The true
intent of the research is disguised, either by claiming a false purpose or by omitting any reference to the study’s purpose.
Some researchers have ethical misgivings about the deceit involved in this approach. Those researchers that use this
approach feel that it provides the more honest and accurate responses. If disguised methods are used, all respondents
should, on completion, attend a debriefing session in which the true purpose of the research is given and the reason for the
deception explained.

3 The main types of qualitative research are:

 Depth Interviews
o interview is conducted one-on-one, and lasts between 30 and 60 minutes
o best method for in-depth probing of personal opinions, beliefs, and values
o very rich depth of information
o very flexible
o probing is very useful at uncovering hidden issues
o they are unstructured (or loosely structured)- this differentiates them from survey interviews in which the
same questions are asked to all respondents
 o can be time consuming and responses can be difficult to interpret
o requires skilled interviewers - expensive - interviewer bias can easily be introduced
o there is no social pressure on respondents to conform and no group dynamics
o start with general questions and rapport establishing questions, then proceed to more
purposive questions
o laddering is a technique used by depth interviewers in which you start with questions about external
objects and external social phenomena, then proceed to internal attitudes and feelings
o hidden issue questioning is a technique used by depth interviewers in which they concentrate on
deeply felt personal concerns and pet peeves
o symbolic analysis is a technique used by depth interviewers in which deeper symbolic meanings are
probed by asking questions about their opposites
 Focus Groups
o an interactive group discussion lead by a moderator
o unstructured (or loosely structured) discussion where the moderator encourages the free flow of ideas
o usually 8 to 12 members in the group
o usually last for 1 to 2 hours
o usually recorded on video
o the room usually has a large window with one-way glass - participants cannot see out, but the
researchers can see in
o inexpensive and fast
o can use computer and internet technology for on-line focus groups
o respondents feel a group pressure to conform
o group dynamics is useful in developing new streams of thought and covering an issue thoroughly
o see focus group for a more detailed description
 Projective Techniques
o these are unstructured prompts or stimulus that encourage the respondent to project their underlying
motivations, beliefs, attitudes, or feelings onto an ambiguous situation
o they are all indirect techniques that attempt to disguise the purpose of the research
o examples of projective techniques include:
 word association - say the first word that comes to mind after hearing a word - only some of the
words in the list are test words that the researcher is interested in, the rest are fillers - is useful
in testing brand names - variants include chain word association and controlled word
association
 sentence completion - respondents are given incomplete sentences and asked to complete
them
 story completion - respondents are given part of a story and are asked to complete it
 cartoon tests - pictures of cartoon characters are shown in a specific situation and with dialogue
balloons - one of the dialogue balloons is empty and the respondent is asked to fill it in
 thematic apperception tests - respondents are shown a picture (or series of pictures) and asked
to make up a story about the picture(s)
 role playing - respondents are asked to play the role of someone else - researchers assume
that subjects will project their own feelings or behaviours into the role
 third-person technique - a verbal or visual representation of an individual and his/her situation is
presented to the respondent - the respondent is asked to relate the attitudes or feelings of that
person - researchers assume that talking in the third person will minimize the social pressure to
give standard or politically correct responses

Features of Qualitative & Quantitative Research

Qualitative Quantitative

"All research ultimately has  "There's no such thing as

 qualitative data.
a qualitative grounding"
Everything is either 1 or 0"
- Donald Campbell
- Fred Kerlinger

The aim is to classify

features, count them, and
The aim is a complete,
construct statistical models in
detailed description.
an attempt to explain what is
observed.

Researcher may only know Researcher knows clearly in

roughly in advance what advance what he/she is
he/she is looking for. looking for.

Recommended during earlier Recommended during latter

phases of research projects. phases of research projects.

All aspects of the study are

The design emerges as the
carefully designed before
study unfolds.
data is collected.

Researcher uses tools, such

Researcher is the data as questionnaires or
gathering instrument. equipment to collect
numerical data.

Data is in the form of words, Data is in the form of

pictures or objects. numbers and statistics.

Subjective - individuals�
Objective � seeks precise
interpretation of events is
measurement & analysis of
important ,e.g., uses
target concepts, e.g., uses
participant observation, in-
surveys, questionnaires etc.
depth interviews etc.

Quantitative data is more

Qualitative data is more
efficient, able to test
'rich', time consuming, and
hypotheses, but may miss
less able to be generalized. 
contextual detail.

Researcher tends to become Researcher tends to remain

subjectively immersed in the objectively separated from
 subject matter. the subject matter.

 (the two quotes are from Miles & Huberman (1994, p. 40). Qualitative
Data Analysis)

Main Points
 Qualitative research involves analysis of data such as
words (e.g., from interviews), pictures (e.g., video), or
objects (e.g., an artifact).
 Quantitative research involves analysis of numerical
data.
 The strengths and weaknesses of qualitative and
quantitative research are a perennial, hot debate,
especially in the social sciences.  The issues invoke classic
'paradigm war'.
 The personality / thinking style of the researcher
and/or the culture of the organization is under-recognized
as a key factor in preferred choice of methods.
 Overly focusing on the debate of
"qualitative versus quantitative" frames the methods in
opposition.  It is important to focus also on how the
techniques can be integrated, such as in mixed methods
research.  More good can come of social science
researchers developing skills in both realms than debating
which method is superior.

Introduction
Nonparametric tests are sometimes called distribution free statistics because they do
not require that the data fit a normal distribution. More generally, nonparametric tests
require less restrictive assumptions about the data. Another important reason for using
these tests is that they allow for the analysis of categorical as well as rank data.
 1. Why Not Used All the Time? 
Since nonparametric tests require fewer assumptions and can be used with a
broader range of data types, the question becomes, "Why not use them all of
the time?" Parametric tests are often preferred because:
1. They are robust.
2. They have greater power efficiency, in other words, they have
greater power relative to the sample size.
3. They provide unique information (e.g., the interaction in a factorial
design).
4. Parametric and nonparametric tests often address two different types
of questions.

Parametric vs. non-parametric tests

 
 
There are two types of test data and consequently different types of analysis. As the table
below shows, parametric data has an underlying normal distribution which allows for more
conclusions to be drawn as the shape can be mathematically described. Anything else is
non-parametric.
 

  Parametric Non-parametric

Assumed distribution Normal Any

Assumed variance Homogeneous Any

Typical data Ratio orInterval Ordinal orNominal

Data set relationships Independent Any

Usual central measure Mean Median

Can draw more Simplicity; Less

Benefits
conclusions affected by outliers

Tests    
 Choosing parametric Choosing a non-
Choosing
test parametric test

Correlation test Pearson Spearman

Independent Independent-
Mann-Whitney test
measures, 2 groups measures t-test

One-way,
Independent
independent- Kruskal-Wallis test
measures, >2 groups
measuresANOVA

Repeated measures,
Matched-pair t-test
2 conditions Wilcoxon test

Repeated measures, One-way, repeated

Friedman's test
>2 conditions measuresANOVA

As the table shows, there are different tests for parametric and non-parametric data.

McNemar's test 
In statistics, McNemar's test is a non-parametric method used on nominal data. It is applied to
2 × 2 contingency tables with a dichotomous trait, with matched pairs of subjects, to determine whether
the row and column marginal frequencies are equal ("marginal homogeneity"). It is named after Quinn
McNemar, who introduced it in 1947.
The McNemar test assesses the significant difference between the two types of dependent samples in cases where
the variables of the interest is dichotomous. The McNemar test is basically used by the researcher during the pre and
the post studies in order to test for an experimental type of effect. The Marginal homogeneity test introduced for
testing the significance of two dependent samples is nothing but an extended form of the McNemar test.

Example
In the following example, a researcher attempts to determine if a drug has an effect on a particular
disease. Counts of individuals are given in the table, with the diagnosis (disease:present or absent)
before treatment given in the rows, and the diagnosis after treatment in the columns. The test requires the
same subjects to be included in the before-and-after measurements (matched pairs).

After: present After: absent Row total

Before: presen
101 121 222
t

Before: absent 59 33 92

Column total 160 154 314

In this example, the null hypothesis of "marginal homogeneity" would mean there was no effect of the
treatment. From the above data, the McNemar test statistic

has the value 21.01, which is extremely unlikely from the distribution implied by the null hypothesis.
Thus the test provides strong evidence to reject the null hypothesis of no treatment effect.

sign test 
In statistics, the sign test can be used to test the hypothesis that there is "no difference in medians"
between the continuous distributions of two random variables X and Y, in the situation when we can
draw paired samples from X and Y. It is a non-parametric test which makes very few assumptions
 about the nature of the distributions under test - this means that it has very general applicability but
may lack the statistical power of other tests such as the paired-samplesT-test.

The sign test is probably the simplest of all the nonparametric methods. It is used to
compare a single sample with some hypothesized value, and it is therefore of use in
those situations in which the one-sample or paired t-test might traditionally be applied. 

In statistics, the Mann–Whitney U test (also called the Mann–Whitney–Wilcoxon (MWW) orWilcoxon

rank-sum test) is a non-parametric statistical hypothesis test for assessing whether two
independent samples of observations have equally large values. It is one of the most well-known non-
parametric significance tests. It was proposed initially by Frank Wilcoxon in 1945,for equal sample sizes,
and extended to arbitrary sample sizes and in other ways by Henry Mannand his student Donald Ransom
Whitney in 1947

THE METHOD
The Mann-Whitney U-test is used to test whether two independent samples of observations are
drawn from the same or identical distributions. An advantage with this test is that the two
samples under consideration may not necessarily have the same number of observations.

This test is based on the idea that the particular pattern exhibited when ‘m’ number of X
random variables and ‘n’ number of Y random variables are arranged together in increasing
order of magnitude provides information about the relationship between their parent
populations.

The Mann-Whitney test criterion is based on the magnitude of the Y’s in relation to the X’s, i.e.
the position of Y’s in the combined ordered sequence. A sample pattern of arrangement where
most of the Y’s are greater than most of the X’s or vice versa would be evidence against random
mixing. This would tend to discredit the null hypothesis of identical distribution.

ASSUMPTIONS
The test has two important assumptions. First the two samples under consideration are
random, and are independent of each other, as are the observations within each sample.
Second the observations are numeric or ordinal (arranged in ranks).
HOW TO CALCULATE THE MANN-WHITNEY U
In order to calculate the U statistics, the combined set of data is first arranged in ascending
order with tied scores receiving a rank equal to the average position of those scores in the
ordered sequence.

Let T denote the sum of ranks for the first sample. The Mann-Whitney test statistic is then
calculated using U = n1 n2 + {n1 (n1 + 1)/2} – T , where n1 and n2 are the sizes of the first and
second samples respectively.

AN EXAMPLE
An example can clarify better. Consider the following samples.

Sample A

Observation 25 25 19 21 22 19 15

Rank 15.5 15.5 9.5 13 14 9.5 3.5

Sample B

Observation 18 14 13 15 17 19 18 20 19

Rank 6.5 2 1 3.5 5 9.5 6.5 12 9.5

Here, T = 80.5, n1 = 7, n2 = 9.Hence, U = (7 * 9) + [{7 * (7+1)}/2] – 80.5 = 10.5.

We next compare the value of calculated U with the value given in the Tables of Critical Values
for the Mann-Whitney U-test, where the critical values are provided for given n 1 and n2 , and
accordingly accept or reject the null hypothesis. Even though the distribution of U is known,
the normal distribution provides a good approximation in case of large samples.