Download MBA-Marketing (Master of Business Administration) 3rd Semester Marketing Research Notes
Unit - I
Chapter 1
MARKETING RESEARCH
Objective:
The objective of this chapter is to understand:
The meaning of Marketing Research
The difference between basic and applied research
The various classifications of Marketing Research
The scope of Marketing Research
The various methods of Marketing Research
Definition of Research
Research always starts with a question to which we seek an answer using
scientific methods. We define the question as a Problem.
Research is often described as an active, diligent, and systematic process of
inquiry aimed at discovering, interpreting and revising facts.
The word research is derived from the French language; its literal meaning is 'to
investigate thoroughly'.
Undertaking research is basically applying scientific methods to find solution to
a problem. It is a systematic and explorative study carried out to analyse and
apply various solutions to a defined problem.
Research can be classified into two broad categories:
1. Basic Research and
2. Applied Research
Basic research
Basic research is also called fundamental or pure research. As the name itself
refers, Basic Research is of basic nature which is not carried out in response to a
problem. It is more educative, towards understanding the fundamentals and aim
at expanding the knowledge base of an individual or organisation. It does not
have any commercial potential.
Applied research
Applied Research on the otherhand is carried out to seek alternate solutions for a
problem at hand. Applied research is done to solve specific, practical questions;
its primary aim is not to gain knowledge. It specifies possible outcomes of each
of the alternatives and its commercial implications.
Applied research can be carried out by academic or industrial institutions. Often,
an academic instituion such as a university will have a specific applied research
program funded by an industrial partner interested in that program. Electronics,
informatics, computer science, process engineering and drug design are some of
the common areas of applied research.
Applied research can further be divided into:
1. Problem-solving research: It involves research oriented towards a
crucial problem facing the organisation which may be issue specific.
Ex: How do we improve the communication skills of our employees?
2. Problem-oriented research: The research is oriented towards a crucial
problem facing the organisation. It is undertaken inside the organisation
or by an external consultant on its behalf. This research is conceptual in
nature and newer innovative techniques of problem-solving are applied.
Ex: How to improve the production yield from machine X using modern
techniques?
Activity 1:
Classify the following on the basis of basic research or aplied research:
1. Research carried out to understand the disease Typhoid.
2. Research carried out to understand the methods to improve the
productivity of people working on Machine X.
3. Research carried out to study the impact of absenteism on productivity.
4. Understanding a new software programme which has been launched.
Defining Marketing Research:
Marketing research (also called consumer research) is a form of business
research. The field of marketing research as a statistical science was pioneered
by Arthur Nielsen with the founding of the ACNielsen Company in 1923.
Marketing research is a systematic and objective study of problems pertaining to
the marketing of goods and services. It is applicable to any area of marketing.
Research is the only tool an organization has to keep in contact with its external
operating environment. In order to be proactive and change with the
environment simple questions need to be asked:
What are the customer needs and how are they changing? How to
meet these changing needs? What do the customers think about
existing products or services? What more are they looking at?
What are the competitors doing to retain customers in this
environment? Are their strategies exceeding or influencing
yours? What should you do to be more competitive?
How are macro and micro environmental factors influencing your
organisation? How will you react t this environment?
Authors have defined Marketing Research in many ways:
Kotler (1999) defines marketing research as systematic problem
analysis, model-building and fact-finding for the purpose of improved
decision-making and control in the marketing of goods and services`.
The American Marketing Association (AMA, 1961) defines it as the
systematic gathering, recording and analyzing of data relating to the
marketing of goods and services`.
Green and Tull have defined marketing research as the systematic and
objective search for and analysis of information relevant to the
identification and solution of any problem in the field of marketing.
The aim of marketing management is to satisfy the needs of the consumer.
Marketing research helps in achieving this. Marketing research is a systematic
and logical way of assessing ways of satisfying customer needs.
According to all the above definitions, Marketing Research starts by stating the
problem or the issue to be investigated; indicate what kind of information is
required to resolve the problem; identify where and how to get it; specify the
methodology for analyzing the research findings; sum up the research findings
and then suggest the best solution for marketing decision making.
Scope of marketing research:
Marketing research can be used in:
Product Management: One of the major scope of marketing research is
to manage the current products and new products. In product
management Marketing Research is helpful in
o Competitive Intelligence ? To understand the competitive
product stretegy.
o Prelaunch strategy for new products
o Test Marketing ? To monitor the performance of the brand by
launching in a select area and then taking it across the country. In
other words it is a small-scale product launch used to determine
the likely acceptance of the product when it is introduced into a
wider market.
o Concept testing - to test the acceptance of a concept by target
consumers.
Sales analyis: Marketing research is used to study the sales trend and
make suitable strategies when required. It is used to
o Assess market potential
o Estimation of demand for a product
o Market share estimation
o Study seasonal variation for a product
o Market segmentation studies
o Estimate size of the market
o Need analysis to find out where the product fits in
Corporate Research: Marketing Research is used to analyse the
corporate effectiveness. Some examples are:
o Assessing the image of the company
o Knowledge of the company activities
Advertising Research: Advertising is an arena in which Marketing
Research is extensively used. Some scope are:
o Readership feedbacks ? Mainly carried out for newspapers and
magazines
o Advertising Recall ? To assess the recall of telivision or other
advertising and thereby assess its effectiveness.
Syndicated Research: This is compiled by agencies on a regular basis
and sold to organisations on subscription basis.
All of these forms of marketing research can be classified as either problem-
identification research or as problem-solving research.
A similar distinction exists between exploratory research and conclusive
research.
Exploratory research provides insights into and comprehension of an
issue or situation. It should draw definitive conclusions only with
extreme caution.
Conclusive research draws conclusions: the results of the study can be
generalized to the whole population.
Research can also be:
Primary Marketing Research: It is research conducted by an
organisation for its own purpose which addresses its requirements. It is
generally expensive but is specific and objective to the organisation`s
requirement.
Secondary Marketing Research: This is used if the organisation is
considering extending its business into new markets or adding new
services or product lines. This type of research is based on information
obtained from studies previously performed by government agencies,
chambers of commerce, trade associations and other organizations. This
also includes Census Bureau information.
In other terms this is research published previously and usually by
someone else. Secondary research costs less than primary research, but
seldom comes in a form that exactly meets the needs of the researcher. It
can cater to anyone who wishes to use the data.
This data can be found in local libraries or on the Web, but books and
business publications, as well as magazines and newspapers, are also
great sources.
Hence, Primary research delivers more specific results than secondary research,
which is an especially important while launching a new product or service. In
addition, primary research is usually based on statistical methodologies that
involve sampling as small as 1 percent of a target market. This tiny
representative sample can give an accurate representation of a particular market.
With the advance in technology a lot of software have been developed which
help in primary market research online and offline thereby making analysis and
interpretation easier.
The ideal way to conduct Marketing Research is to do secondary research first
and then do the primary research for the data not available form secondary
sources.
Hence, secondary research lays the groundwork and primary research helps fill
in the gaps. By using both types of market research, organisations get a better
picture of their market and have the information they need to make important
business decisions.
Activity 2:
List down areas in marketing where Marketing Research would be helpful:
............................................................................................................
............................................................................................................
............................................................................................................
Marketing Research Methods
Methodologically, marketing research uses four types of research designs,
namely:
Qualitative marketing research ? This is generally used for
exploratory purposes. The data collected is qualitative and focuses on
people`s opinions and attitudes towards a product or service.
The respondents are generally few in number and the findings cannot be
generalised tot eh whole population. No statistical methods are generally
applied.
Ex: Focus groups, In-depth interviews, and Projective techniques
Quantitative marketing research ? This is generally used to draw
conclusions for a specific problem. It tests a specific hypothesis and uses
random sampling techniques so as to infer from the sample to the
population. It involves a large number of respondents and analysis is
carried out using statistical techniques.
Ex: Surveys and Questionnaires
Observational techniques - The researcher observes social phenomena
in their natural setting and draws conclusion from the same. The
observations can occur cross-sectionally (observations made at one
time) or longitudinally (observations occur over several time-periods)
Ex: Product-use analysis and computer cookie tracing
Experimental techniques ? Here, the researcher creates a quasi-
artificial environment to try to control spurious factors, then manipulates
at least one of the variables to get an answer to a research
Ex: Test marketing and Purchase laboratories
More than one research designs could be used at a time. They may start with
secondary research to get background information, then conduct a focus group
(qualitative research design) to explore the issues. Finally they might do a full
nation-wide survey (quantitative research design) in order to devise specific
recommendations for the client organisation.
Difference between Market Research and Marketing Research
Generally Market Research and Marketing Research are confused to be the
same. But there is a clear distinction between the both.
Market Research: Market Research involves researching a specific industry or
market.
Ex: Researching the automobile industry to discover the number of competitors
and their market share.
Marketing Research: Marketing Research analyses a given marketing
opportunity or problem, defines the research and data collection methods
required to deal with the problem or take advantage of the opportunity, through
to the implementation of the project. It is a more systematic method which aims
to discover the root cause for a specific problem within an organisation and put
forward solutions to that problem.
Ex: Research carried out to analyze and find solution for increasing turnover in
an organisation.
Summary:
The meaning of research is to investigate thoroughly. It can be divided into basic
and applied research. Basic research is the pure research which is more
educative. Applied research is carried out to seek alternate solution for a
problem.
Applied research can further be classified as problem-solving research and
problem-oriented research depending upon the research problem at hand.
Marketing research is a systematic and objective study of problems pertaining to
the marketing of goods and services. It is applicable to any area of marketing.
For ex. Product management, sales, advertising research, etc.
Marketing Research can be Primary Market Research or Secondary Market
research depending on the data source used. It can be qualitative or quantitative
research depending upon the nature of the research.
Marketing Research is different from Market Research wherein the former is
oriented towards solving marketing problems and the latter is market related.
Questions:
Answer the following questions:
1. What is Marketing Research?
2. Define Marketing Research as stated by various authors.
3. Why should you conduct Marketing Research?
4. Differentiate between basic research and applied research.
5. List down the scope of Marketing Research.
6. What is the difference between problem-solving research and problem-
oriented research?
7. Give an example of situations where you will use exploratory research.
8. Enumerate the differences between primary and secondary marketing
research.
9. Give examples of Qualitative and Quantitative Marketing Research.
10. Is there any difference between Market Research and Marketing
Research? Explain
Chapter 2
THE MARKETING RESEARCH PROCESS
Objectives:
To understand the Marketing Research Process.
To learn in detail about the various steps in the Marketing Research
Process
The Marketing Research Process
As we saw earlier Marketing Research is very much essential to make strategic
decisions which are important for the growth of the organisation. It helps in
making the right decisions systematically using statistical methods.
Marketing Research reduces the uncertainty in the decision-making process and
increase the probability and magnitude of success if conducted in a systematic,
analytical, and objective manner. Marketing research by itself does not arrive at
marketing decisions, nor does it guarantee that the organization will be
successful in marketing its products. It is only a tool which helps in the decision
making process.
The Marketing Research Process involves a number of inter-related activities
which have bearing on each other. Once the need for Marketing Research has
been established, broadly it involves the steps as depicted in Figure 1 below:
Define the research problem
Determine research design
Identify data types and sources
Design data col ection forms
Determine sampling design and size
Collect the data
Analyze and interpret the data
Prepare the research report
Figure1: The steps in Marketing Research Process
Let us now know in detail about the various steps involved in the Marketing
Research Process.
1. Define the research problem
The first step in Marketing is to define the research problem. A problem well
defined is half-solved. If a problem is poorly defined, a good research design
cannot be developed.
The decision problem faced by the organisation must be translated into a market
research problem in the form of questions. These questions must define the
information that is required to make the decision and how this information can
be obtained. This way, the decision problem gets translated into a research
problem.
For example, a decision problem may be whether to launch a new product. The
corresponding research problem might be to assess whether the market would
accept the new product.
In order to define the problem more precisely, an exploratory research can be
carried out. Survey of secondary data, pilot studies or experience surveys are
some of the popular methods.
2. Determine research design
The research design specifies the method and procedure for conducting a
particular study.
As studied already, marketing research and hence the research designs can be
classified into one of three categories
Exploratory research
Descriptive research
Causal research
This classification is based on the objective of the research. In some cases the
research will fall into one of these categories, but in other cases different phases
of the same research project will fall into different categories.
Problems are formulated clearly in exploratory research. It aims at clarifying
concepts, gathering explanations, gaining insight, eliminating impractical ideas,
and forming hypotheses. Exploratory research can be performed using a
literature search, surveying certain people about their experiences, focus groups,
and case studies. During the survey, exploratory research studies would not try
to acquire a representative sample, but rather, seek to interview those who are
knowledgeable and who might be able to provide insight concerning the
relationship among variables. Case studies can include contrasting situations or
benchmarking against an organization known for its excellence. Exploratory
research may develop hypotheses, but it does not seek to test them. Exploratory
research is characterized by its flexibility.
A descriptive study is undertaken when the researcher wants to know the
characteristics of certain groups such as age, sex, educational level, income,
occupation, etc. Descriptive research is more rigid than exploratory research and
seeks to describe users of a product, determine the proportion of the population
that uses a product, or predict future demand for a product. Descriptive research
should define questions, people surveyed, and the method of analysis prior to
beginning data collection. In other words, the who, what, where, when, why, and
how aspects of the research should be defined. Such preparation allows one the
opportunity to make any required changes before the costly process of data
collection has begun.
There are two basic types of descriptive research: longitudinal studies and cross-
sectional studies. Longitudinal studies are time series analyses that make
repeated measurements of the same individuals, thus allowing one to monitor
behavior such as brand-switching. However, longitudinal studies are not
necessarily representative since many people may refuse to participate because
of the commitment required. Cross-sectional studies sample the population to
make measurements at a specific point in time. A special type of cross-sectional
analysis is a cohort analysis, which tracks an aggregate of individuals who
experience the same event within the same time interval over time. Cohort
analyses are useful for long-term forecasting of product demand.
Causal research seeks to find cause and effect relationships between variables. It
accomplishes this goal through laboratory and field experiments.
3. Identify data types and sources
The next step is to determine the sources of data to be used. The researcher has
to decide whether to go for primary data or secondary data. Sometimes a
combination of both is used.
Before going through the time and expense of collecting primary data, one
should check for secondary data that previously may have been collected for
other purposes but that can be used in the immediate study. Secondary data may
be internal to the firm, such as sales invoices and warranty cards, or may be
external to the firm such as published data or commercially available data. The
government census is a valuable source of secondary data.
Secondary data has the advantage of saving time and reducing data gathering
costs. The disadvantages are that the data may not fit the problem perfectly and
that the accuracy may be more difficult to verify for secondary data than for
primary data.
Many a time the secondary data might have to be supplemented by primary data
originated specifically for the study at hand. Some common types of primary
data are:
Demographic and socioeconomic characteristics
Psychological and lifestyle characteristics
Attitudes and opinions
Awareness and knowledge - for example, brand awareness
Intentions - for example, purchase intentions. While useful, intentions
are not a reliable indication of actual future behavior.
Motivation - a person's motives are more stable than his/her behavior, so
motive is a better predictor of future behavior than is past behavior.
Behavior
Primary data can be obtained by communication or by observation.
Communication involves questioning respondents either verbally or in writing.
This method is versatile, since one needs to only ask for the information;
however, the response may not be accurate. Communication usually is quicker
and cheaper than observation. Observation involves the recording of actions and
is performed by either a person or some mechanical or electronic device.
Observation is less versatile than communication since some attributes of a
person may not be readily observable, such as attitudes, awareness, knowledge,
intentions, and motivation. Observation also might take longer since observers
may have to wait for appropriate events to occur, though observation using
scanner data might be quicker and more cost effective. Observation typically is
more accurate than communication.
Personal interviews have an interviewer bias that mail-in questionnaires do not
have. For example, in a personal interview the respondent's perception of the
interviewer may affect the responses.
4. Design data collection forms
Once it has been decided to obtain primary data, the mode of collection needs to
be decided. Two methods are available for data collection:
1. Observational methods
2. Survey methods
Observational methods: As the name itself suggests, the data are collected
through observation. An observer observes and records the data faithfully and
accurately. This may be suitable in case of some studies but is not useful to
observe attitudes, opinions, motivations and other intangible states of mind.
Also in this method, the data collected is non-reactive, as it does not involve the
respondent.
Surveys: It is one of the most common methods of collecting data for primary
marketing research. Surveys can be:
Personal: The information is sought through personal interviews. A
questionnaire is prepared and administered to the respondent during the
interview. This is a detailed method of collecting information.
Telephonic: This is suitable if limited information is sought in a fixed
time frame.
Mail: Here, the questionnaire is sent out in mail and the response is
sought. Timely response cannot be sought in this method as there is no
control over the survey. All the people to whom the mail was sent may
not respond.
Sometimes a combination of two or more methods may be used. Whatever be
the method, a structured questionnaire is required to be used. The questionnaire
is an important tool for gathering primary data. Poorly constructed questions can
result in large errors and invalidate the research data, so significant effort should
be put into the questionnaire design. The questionnaire should be tested
thoroughly prior to conducting the survey.
5. Determine sampling design and size
A sampling plan is a very important part of the research process. The marketing
researcher has to decide whether it will be a sample survey or a census.
Definitely a sample survey has its distinct merits.
The population from which the sample has to be drawn has to be well defined. A
broad choice is to be made between probability sampling and non-probability
sampling. The sample design is then chosen depending on the suitability and the
availability of the sample frame.
The size of the sample chosen is based on statistical methods. This is well
defined and also reproduces the characteristics of the population. In practice,
however, this objective is never completely attained on account of the
occurrence of two types of errors ? errors due to bias in the selection and sapling
errors.
6. Collect the data
The next step is to collect the data for which the research process has been
spelled out. The interviewing and the supervision of field work should be looked
into. One of the most difficult tasks is interviewing for marketing research.
Many a time the respondents may not part with crucial information unless
approached with tact and intelligence. Supervision of field work is important to
ensure timely and proper completion of the field survey.
If this is not carried out properly, then there results an interview error which
may be detrimental to marketing research.
7. Analyze and interpret the data
The next step is to analyze the data that has been collected from the field survey.
The raw data is transformed into the right format. First, it is edited so that errors
can be corrected or omitted. The data is then coded; this procedure converts the
edited raw data into numbers or symbols. A codebook is created to document
how the data is coded. Finally, the data is tabulated to count the number of
samples falling into various categories.
Simple tabulations count the occurrences of each variable independently of the
other variables. Cross tabulations, also known as contingency tables or cross
tabs, treats two or more variables simultaneously.
Cross tabulation is the most commonly utilized data analysis method in
marketing research. Many studies take the analysis no further than cross
tabulation.
Once the tabulation is done, the following analysis can be carried out.
Conjoint Analysis: The conjoint analysis is a powerful technique for
determining consumer preferences for product attributes.
Hypothesis Testing: The null hypothesis in an experiment is the
hypothesis that the independent variable has no effect on the dependent
variable. The null hypothesis is expressed as H0. This hypothesis is
assumed to be true unless proven otherwise. The alternative to the null
hypothesis is the hypothesis that the independent variable does have an
effect on the dependent variable. This hypothesis is known as the
alternative, research, or experimental hypothesis and is expressed as H1.
Once analysis is completed, make the marketing research conclusion. In order to
analyze whether research results are statistically significant or simply by chance,
a test of statistical significance can be run.
8. Prepare the research report
All the research findings have to be compiled in a report to be then presented to
the organization. The format of the marketing research report varies with the
needs of the organization. The report often contains the following sections:
Authorization letter for the research
Table of Contents
List of illustrations
Executive summary
Research objectives
Methodology
Results
Limitations
Conclusions and recommendations
Appendices containing copies of the questionnaires, etc.
The report has to be written with objectivity, coherence, clarity in the
presentation of the ideas and use of charts and diagrams. Sometimes, the study
might also throw up one or more areas where further investigation is required.
Summary:
Marketing Research reduces the uncertainty in the decision-making process and
increase the probability and magnitude of success if conducted in a systematic,
analytical, and objective manner.
The Marketing Research Process involves a number of inter-related activities
which have bearing on each other. Each and every step plays an important role
in the research process.
Questions:
1. List out the various steps involved in the Marketing Research Process.
2. It is very important to define the research problem, explain.
3. Classify research designs and explain the relevance of each.
4. What are the types of data sources?
5. Enumerate the methods available for data collection
6. Is it important to determine the sample size? Explain.
7. How will you analyze the data collected from the research?
8. How will you prepare a research report?
9. The various steps in the Marketing Research Process are inter-related.
Explain.
10. Does the Market Research Process serve as a framework for finding
solution to the research problem at hand? Evaluate critically.
Chapter 3
RESEARCH DESIGN
Objective:
To understand the meaning of Research Design.
To study about the various types of Research designs.
To understand the type of research design to use for specific problems.
The Research Design
Research design provides the glue that holds the research project together. A
design is used to structure the research, to show how all of the major parts of the
research project -- the samples or groups, measures, treatments or programs, and
methods of assignment -- work together to try to address the central research
questions.
According to Green and Tull: A Research Design is the specification of methods
and procedures for acquiring the information needed. It is the over-all
operational pattern or framework of the project that stipulates what information
is to be collected from which sources by what procedures.
Hence it is clear that Research design is the blueprint for research. It lays down
the methodology involved in the collection of information and arriving at
meaningful conclusions from the same.
There are many methods for studying and tackling a problem, but there are no
perfect methods. Many times more than one method could be used in the
research process.
There are many classifications accepted for a Research Design. One of the most
accepted classification is grouping it under three types:
1. Exploratory
2. Descriptive and
3. Causal
This can be depicted as in figure 2 given below:
Researc
Resear h Desi
h D
gn
Explorato
Ex
r
plorato y R
y ese
R
arch Desi
arch D
gn
Conc
Con lus
lu iv
s e Rese
e R
arch D
ese
esi
arch D
gn
Descriptiv
ripti e Rese
e R
arch
ese
Caus
Cau al research
Cross
Cr
-Sec
-
tional Desi
tional
gn
Longitudinal Des
Longitudin
i
al Des gn
Figure 2: Classification of Research Designs
Exploratory Research Design
As the term suggests, exploratory research is often conducted because a problem
has not been clearly defined as yet, or its real scope is as yet unclear. It is a
process of discovery wherein you uncover as many ideas as possible. It allows
the researcher to familiarize him/herself with the problem or concept to be
studied, and perhaps generate hypothesis to be tested. It expands knowledge. It
is the initial research, before more conclusive research is undertaken.
Exploratory research helps determine the best research design, data collection
method and selection of subjects.
Another common reason for conducting exploratory research is to test concepts
before they are launched in the marketplace, always a very costly endeavor. In
concept testing, consumers are provided either with a written concept or a
prototype for a new, revised or repositioned product, service or strategy.
Exploratory research relies more on secondary data. It does not have a rigid
design as the researcher themselves are not very well versed with the subject and
are trying to gain knowledge of the same. Hence it can be quite informal, relying
on secondary research such as reviewing available literature and/or data, or
qualitative approaches such as informal discussions with consumers, employees,
management or competitors, and more formal approaches through in-depth
interviews, focus groups, projective methods, case studies or pilot studies.
The results of exploratory research are not usually useful for decision-making by
themselves, but they can provide significant insight into a given situation. The
findings of this can be used to develop the research further. Points can be arrived
at which requires to apply the other methodologies.
Conclusive Research Design
Descriptive research is also used to generate hypotheses but generally has more
information available than in exploratory research. Descriptive research is
usually conducted to characterize one or more variables within a population,
particularly in relation to person, place, and time.
As the name indicates, conclusive research is meant to provide information that
is useful in reaching conclusions or decision-making. It is mostly quantitative in
nature, in the form of numbers that can be quantified and summarized. It relies
on both secondary data, particularly existing databases that are reanalyzed to
shed light on a different problem than the original one for which they were
constituted, and primary research, or data specifically gathered for the current
study.
The purpose of conclusive research is to provide a reliable or representative
picture of the population through the use of a valid research instrument. In the
case of formal research, it will also test hypothesis.
Conclusive research can be sub-divided into two categories:
1. Descriptive or statistical research, and
2. Causal research
Descriptive Research
Descriptive research or statistical research provides data about the population or
universe being studied. It describes the "who, what, when, where and how" of a
situation and not what caused it. Therefore, descriptive research is used when
the objective is to provide a systematic description that is as factual and accurate
as possible. It provides the number of times something occurs, or frequency,
lends itself to statistical calculations such as determining the average number of
occurrences or central tendencies.
One of its major limitations is that it cannot help determine what causes a
specific behaviour, motivation or occurrence. In other words, it cannot establish
a causal research relationship between variables.
The two most common types of descriptive research designs are
1. Observation: Observation is a primary method of collecting data by
human, mechanical, electrical or electronic means. The researcher may
or may not have direct contact or communication with the people whose
behaviour is being recorded. Observation techniques can be part of
qualitative research as well as quantitative research techniques. The
commonly used observation methods are:
Participant and non participant observation: This depends on whether the
researcher chooses to be part of the situation s/he is studying.
(e.g. studying team dynamics by being a team member would be
participant observation)
Obtrusive and unobtrusive observation: Depends on whether the subjects
being studied can detect the observation
(e.g. hidden microphones or cameras observing behaviour)
Observation in natural or contrived settings: Observing the behaviour in
its natural setting and in a condition where the natural settings are
created.
Disguised and non-disguised observation: Depends on whether the
subjects being observed are aware that they are being studied or not. In
disguised observation, the researcher may not disclose his true identity
and pretend to be someone else to keep away the bias in the findings.
Structured and unstructured observation: This refers to guidelines or a
checklist being used for the aspects of the behaviour that are to be
recorded; for instance, noting who starts the introductory conversation
between the group members and what specific words are used by way of
introduction.
Direct and indirect observation: This depends on whether the behaviour
is being observed during the time it occurs or after the occurrence, as in
the case of TV viewing, for instance, where choice of program and
channel flicking can all be recorded for later analysis.
One distinct advantage of the observation technique is that it records actual
behaviour, not what people say they said/did or believe they will say/do. On the
other hand, the observation technique does not provide us with any insights into
what the person may be thinking or what might motivate a given
behaviour/comment. This type of information can only be obtained by asking
people directly or indirectly.
2. Surveys: The survey technique mainly involves the collection of primary
data about subjects, usually by selecting a representative sample of the
population or universe under study, through the use of a questionnaire. It
is a very popular since many different types of information can be
collected, including attitudinal, motivational, behavioral and perceptive
aspects. It allows for standardization and uniformity in the questions
asked and in the method of approaching subjects, making it easier to
compare and contrast answers by respondent group. It also ensures
higher reliability than some other techniques.
If properly designed and implemented, surveys can be an efficient and accurate
means of determining information about a given population. Results can be
provided relatively quickly, and depending on the sample size and methodology
chosen, they are relatively inexpensive. However, surveys also have a number of
disadvantages, which must be considered by the researcher in determining the
appropriate data collection technique.
Since in any survey, the respondent knows that s/he is being studied, the
information provided may not be valid insofar as the respondent may wish to
impress (e.g. by attributing him/herself a higher income or education level) or
please (e.g. researcher by providing the kind of response s/he believes the
researcher is looking for) the researcher. This is known as response error or bias.
The willingness or ability to reply can also pose a problem. If the information
sought is considered sensitive or intrusive the respondent may hesitate to reply,
leading to a high rate of refusal. This can be overcome by framing such
questions carefully.
There can be an interviewer error or bias as the interviewer can (inadvertently)
influence the response elicited through comments made or by stressing certain
words in the question itself. This is seen through facial expressions, body
language or even the clothing that is worn.
Another consideration is response rate. Depending on the method chosen, the
length of the questionnaire, the type and/or motivation of the respondent, the
type of questions and/or subject matter, the time of day or place, and whether
respondents were informed to expect the survey or offered an incentive can all
influence the response rate obtained. Proper questionnaire design and question
wording can help increase response rate.
Descriptive studies are also classified into:
1. Cross-sectional studies: It deals with a sample of elements from a given
population. Number of characteristics from the sample elements are
collected and analyzed. It is of two types: field studies and surveys.
2. Longitudinal studies. This is based on panel data and panel methods. A
panel constitutes a group of respondents who are interviewed and
reinterviewed from time to time. Hence the same variable is repeatedly
measured. This helps in studying a particular behaviour over a period of
time.
Causal Research
Causal research is undertaken to see if there is a cause and effect relationship
between variables. In order to determine causality, it is important to hold the
variable that is assumed to cause the change in the other variable(s) constant and
then measure the changes in the other variable(s). This type of research is very
complex and the researcher can never be completely certain that there are not
other factors influencing the causal relationship, especially when dealing with
people`s attitudes and motivations. There are often much deeper psychological
considerations that even the respondent may not be aware of.
There are two research methods for exploring the cause and effect relationship
between variables:
1. Experimentation or natural experimentation: This highly controlled method
allows the researcher to manipulate a specific independent variable in order to
determine what effect this manipulation would have on other dependent
variables. Experimentation also calls for a control group as well as an
experimentation group, and subjects would be assigned randomly to either
group. The researcher can further decide whether the experiment should take
place in a laboratory or in the field, i.e. the "natural" setting as opposed to an
"artificial" one. Laboratory research allows the researcher to control and/or
eliminate as many intervening variables as possible.
2. Simulation: Another way of establishing causality between variables is
through the use of simulation.
A sophisticated set of mathematical formula are used to simulate or imitate a
real life situation. By changing one variable in the equation, it is possible to
determine the effect on the other variables in the equation.
For the natural experiments there are three classes of designs:
1. Time-series and trend designs
2. Cross-sectional designs and
3. A combination of the above two.
Time series and trend designs: In a time series design, data is collected from
the sample or population at successive intervals. The trend data relate to
matched samples drawn from the same population at successive intervals. It can
be of many types.
A simple design can be represented as below:
X
O
Where X indicates the exposure of a group to an experimental treatment and O
indicates the observation or measurement taken on the subject or group after an
experimental treatment. Another method also involves a control group. This can
be represented as below:
O1
O2
O3
X
O4
O5
O6
O`1
O`2
O`3
O`4
O`5
O`6
Where O`s represent measurement of the control group. This is termed as
multiple time-series design.
Cross-sectional designs: It studies the effect of different levels of treatments on
several groups at the same time. It can be represented as below:
X1
O1
X2
O2
X3
O3
X4
O4
An example would be different kind of incentives given for the same product in
various territories. This would help in understanding the effect of varying the
incentive on the sales performance across territories.
Combinational Design: This design combines both the time-series and cross-
sectional designs.
This design is generally seen while measuring advertising effectiveness in a
panel. An advertisement is run and the respondents are asked if they have seen it
earlier. Those who have seen it earlier constitute the test group and those who
have not constitute the control group. The purchase made before and after the
advertisement by the test and the control group marks the advertising
effectiveness.
So many research designs have been listed. The one that is ultimately selected
should help in solving the problem. It should help in arriving at the desired
conclusions.
Summary:
A Research Design is the specification of methods and procedures for acquiring
the information needed. It is the blueprint for a research process.
There are many classifications accepted for a Research Design. One of the most
accepted classification is grouping it under three types: Exploratory, Descriptive
and Causal.
An exploratory research is often conducted because a problem has not been
clearly defined as yet, or its real scope is as yet unclear. Conclusive research on
the other hand is meant to provide information that is useful in reaching
conclusions or decision-making. It is mostly quantitative in nature. A Causal
research is undertaken to see if there is a cause and effect relationship between
variables.
Causal research again can be: Time-series and trend designs, Cross-sectional
designs and a combination of the above two.
Questions:
1. What is a Research Design? Explain.
2. Classify the Research Designs and define each of them.
3. Exploring helps in knowledge growth`. Explain with relevance to
Exploratory Research design.
4. Conclusive research helps in drawing conclusions`. Explain.
5. What is descriptive Research? Classify and explain the same.
6. What are cross-sectional and longitudinal studies? Explain.
7. What is causal research? Explain the causal research methods.
8. What is Time series design? When is it used?
9. A good Research design is essential for solving a research problem.
Explain.
10. Are more than one research designs used to find a solution to a problem?
Critically evaluate the same.
Chapter 4
DATA SOURCES
Objective:
To understand the meaning and importance of data sources.
To read in detail about the sources of Primary data and secondary data
sources.
To understand the relevance of these data sources while solving a
research problem.
Data sources
One of the most important components of Marketing Research is collection of
data required to solve a defined research problem. The general tendency of the
researcher is to organize a survey and collect the data from the field.
The most important point to be considered before this is to research the
secondary sources and gather data already available. This gives a logical
perspective to problem solving. Only then the actual data required to be
collected from the primary survey can be well defined.
Hence it is imperative to know the advantages and drawbacks of Secondary and
Primary data.
Secondary Data
Secondary data is defined as the data that has been collected by individuals or
agencies for purposes other than those of the particular research study. For
example, if a government department has conducted a survey of, say, school
going children, then a uniform manufacturer might use this data for his research
purpose.
As mentioned earlier, it is ideal to undertake a marketing research study after a
prior search of secondary sources (also termed desk research). The reasons for
this are summed up below.
conclusions and answer to solve the problem. Primary data collection
may not be required.
Secondary data is economical than collecting primary data. A thorough
examination of secondary sources can yield a great deal more
information than through a primary data collection exercise which needs
to be critically evaluated.
Searching secondary sources is much less time consuming than primary
data collection.
Secondary sources of information may at times consider a large sample
and hence can yield more accurate data than that primary research. This
is especially true for census information or syndicated reports by
government departments which are generally large scale. This is likely to
yield far more accurate results than custom designed surveys that are
based on relatively small sample sizes.
Secondary data can play a substantial role in the exploratory phase of the
research when the main objective is to define the research problem and
to generate hypotheses. The assembly and analysis of secondary data
helps in better understanding of the marketing problem. This also gives
an idea about the course of action and missing links which can be got
from the primary research.
Secondary sources are very useful to structure the sample and define the
population.
Disadvantages of secondary data:
Even though the secondary data offers a lot of advantages; it also has its own
shortcomings. This corresponds to both the source and the quality of the data.
The main disadvantages may be listed as follows:
The researcher has to be careful while using the units defined in the data.
It is better to study the definitions used prior to accepting the same for
research purpose.
For ex, the meaning of family might differ from urban and rural as it may
consider the nuclear or the joint family system especially in India. Hence
while considering secondary data on the size of family, these definitions
need to be kept in mind.
It should be noted that definitions may change over time and if this is not
evaluated the conclusion derived may be wrong.
The errors of measurements are not generally published in secondary
sources and hence this should be considered while looking at data from
secondary sources. The solution is to try to speak to the individuals
involved in the collection of the data to obtain some guidance on the
level of accuracy of the data. This is especially crucial if the stake is high
in terms of commercial implications.
The data has to be validated for source biases as it may have been
prepared to appear exaggerated or otherwise. Hence it is better to go
through details of the purpose for which the data had been collected.
The reliability of published data may vary over time. Hence the data
needs to be checked for time validity. For ex: New states have been
formulated in India. Data pertaining to population studies conducted
before the formation of the new state needs to be evaluated for its current
application.
Many a time the data collected may be outdated and hence it needs to be
refreshed again. This may otherwise hinder the analysis.
All these drawbacks do exist, but still secondary data has it own merits. It is
ideal to use multiple sources of secondary data. In this way, these different
sources can be cross-checked and validated for the source of information. It is
better to disregard the data whenever any controversy exists.
The below flowchart (figure 3) depicts the evaluation procedure for using
secondary data. As can be seen, the flowchart divides into two phases. The early
stages relate to the relevance of the data to the research objectives and the later
stages of the flowchart are concerned with questions about the accuracy of
secondary data.
Is the data h
at
el
e pful
pfu to a
t
nswer
nsw the
h
questi
st ons of the
t
resea
r
r
esea ch obj
ch o ecti
bj
ve?
NO
ve?
ST
S OP
YES
Does
D
i
oes t app
a l
pp y to the
t
popu
p
l
opu ati
at on o
o
f
inter
nt est a
er
nd t
n he titme p
m
eri
er od se
o
l
d se ected?
NO
ected?
ST
S OP
YES
Can
C
the u
th
ni
n ts cl
ts c assi
ass fifed be
ed
app
be
l
app ied or
ed
rewor
ew ked as pe
ked
r
as pe requ
e i
qu rement
em
?
NO
?
ST
S OP
YES
Is the cost of data
d
wor
w th acqu
th
i
acqu rirng
n i
g t?
NO
ST
S OP
YES
Is the data a
at
ccu
cc r
u ate an
d u
an
nb
n i
b ased?
as
NO
ST
S OP
YES
US
U E
S T
E HE
H DA
E
T
DA A
Figure 3: Evaluation of secondary data
Secondary sources of information:
Secondary sources of information can be collected from two sources: internal
sources and external sources.
Internal sources of secondary information:
Lot of data is available within an organization regarding day to day operations.
These data can be utilized wherever required. These include:
1. Sales data: Sales orders are received, invoiced and delivered. Cost of the
goods supplied is also recorded. Sales across different territories are
recorded via the reports received from the field. Most of these reports
can be used for making marketing decisions. These resources are
generally overlooked while deciding on critical issues. Lot of
information pertaining to sales by territory, sales by customer type,
prices and discounts, average size of order by customer, customer type,
geographical area, average sales by sales person and sales by pack size
and pack type, etc.
This data can be used to identify the most profitable product and
customers, tracking sales trends, analysis on discounts given, scattering
pattern of sales orders, effect of seasonality on sales, etc.
2. Financial data: This relates to data on various costs involved in
procurement of raw materials, production of goods, distribution of
goods, conversion costs, labor costs, transportation cost, storage cost,
etc. With such data the efficiency of operation can be determined. It
helps in assessing the cost of production of a new product, analysis the
cost of free capacity, etc.
3. Transportation data: A good record of the data relating to transport
operations determine which route to use, which transporter to use, cost of
effective routing patterns, etc. This helps in determining whether it
would be sensible to have your own vehicle or hire a vehicle. This
enables decision towards a trade off analysis towards a better
profitability.
4. Human Resource data: Enormous information could be collected from
an organizational perspective from the human resource department. Data
on employee turnover, absenteeism, strength of employees could be
obtained. This would help in man power planning for the present and
future, succession planning, training and development for better
productivity, etc.
5. Storage data: This may help in calculating the direct product
profitability by calculating the rate of stock turn; stock handling costs,
assessing the efficiency of certain marketing operations and the
efficiency of the marketing system as a whole.
External sources of secondary information
Lot of secondary data is now available further to the discovery of World Wide
Web and lot of institutions looking at such analysis.
Large number of organizations provides marketing information including
national and local government agencies, quasi-government agencies, trade
associations, universities, research institutes, financial institutions, specialist
suppliers of secondary marketing data and professional marketing research
enterprises. Dillon et al advise that searches of printed sources of secondary data
begin with referral texts such as directories, indexes, handbooks and guides.
These sorts of publications rarely provide the data in which the researcher is
interested but serve in helping him/her locate potentially useful data sources.
The main sources of external secondary sources are (1) government (federal,
state and local) (2) trade associations (3) commercial services (4) national and
international institutions.
1. Government statistics: These may include all or some of the following:
Population census, Social surveys, family expenditure surveys,
Import/export statistics, Production statistics, and Agricultural statistics.
Some of the Indian government bodies are:
Population Statistics of Govt. of India ? Provides statistics related
to general population of India
Central Bureau of Health Intelligence ? Provides health related
statistics
Indian Council of medical Research ? Provides information on
research being conducted on major diseases
Policy Reform Options Database ? Provides data on policy
reforms
Ministry of Health and Family Welfare ? Provides information
on family welfare
Ministry of Statistics and Programme implementation ? Gives
information on various statistical indicators of Indian economy
India Brand Equity Foundation ? Provides information on Indian
economy and Industry
Insurance Regulatory and Development Authority ? Data on
Health Insurance in India.
2. Trade associations - They might produce a wide range of data. Normally
it may produce a trade directory and, perhaps, a yearbook.
3. Syndicated reports ? These are published market research reports from
various organizations which charge for their information. These data
relate to consumer information and media information. These are
generally prepared to cater to all interested and not to any specific
organizational requirement. Hence the relevant data is extracted from
this.
4. National and international institutions: Economic reviews, Research
reports, journals and articles are all useful sources to contact. A lot of
secondary data can be obtained from World Bank, WHO, International
Monetary Fund, International Fund for Agricultural Development,
United Nations Development Programme, Food and Agriculture
Organization and International Labor Organization.
Advances in telecommunications technology have combined to allow people
around the world to exchange information quickly and easily. Internet has made
access to information relatively easier and quicker.
Primary Data:
Primary data collection begins when a researcher is not able to find the data
required for his research purpose from the secondary sources. Market
researchers are interested in a variety of primary data about
demographic/socioeconomic
characteristics,
attitudes/opinions/interests,
awareness/knowledge, intentions, motivation, and behavior.
There are two basic means of obtaining primary data depending on the nature of
the problem and by the availability of time and money:
1. Observation
As the name implies, the researcher observes the situation of interest and records
the relevant facts, actions, or behaviors. Observation provides accurate data
about what consumers do in certain situations but do not provide details of why
it happened.
There are several methods of observation:
Structured ? unstructured observation: In structured observation, the
observer is given a set of behaviors to observe. In unstructured
observation, the observer is allowed to observe anything that may be
relevant to the research objective. In the first instance there may be a bias
and in the second the result may vary from observation to observation.
Disguised ? undisguised observation: In a disguised observation, the
subjects do not know that they are being observed. This is a better way to
observe as the subjects are not conscious that they are being observed
and behave freely. In an undisguised observation the subjects are aware
that they are being observed and tend to be cautious.
Observation under normal setting ? Laboratory setting: Normal setting
would be a field survey; laboratory setting would be under a fixed roof
or venue. The latter allows a prompt and economical way of collecting
data and permit the use of more objective measurements.
Direct ? Indirect Observation ? In the case of direct observation, the
event or the behaviour of a person is observed as it occurs. In an indirect
observation, a record of a past behaviour is observed.
Human ? Mechanical Observation ? The observations are recorded
manually in Human Observations. In Mechanical Observation, the
research is carried out through hidden cameras and audiometers; hence
there is no subjective bias.
2. Questionnaire
Questionnaires are data collecting instruments used to ask respondents questions
to secure the desired information. Questionnaires may be administered by mail,
over the telephone, by computer, or in person.
The design of a questionnaire depends on whether the researcher wishes to
collect exploratory information (i.e. qualitative information for the purposes of
better understanding or the generation of hypotheses on a subject) or
quantitative information (to test specific hypotheses that have previously been
generated).
The questionnaires can be classified into four types:
1. Structured ? non-disguised
2. Structured disguised
3. Non-structured - non- disguised
4. Non-structured ? disguised
Non-disguised are the direct questions and the object of enquiry is revealed to
the respondent. Disguised are the indirect questions where the object of enquiry
is not revealed to the respondent. In a structured questionnaire the questions are
asked in a pre-determined order.
Formal standardised questionnaires: If the data is required to be analysed
statistically, a formal standardised questionnaire is designed. The points to be
remembered while designing such questionnaires are:
The questionnaire has to be framed in such a manner that each
respondent receives the same stimuli
The questionnaire has to be well-defined so that the interviewer is able to
answer the respondent`s clarifications if necessary.
The response format must be easy to complete during the interviewing
process.
A well-designed questionnaire should primarily meet the research
objectives.
A questionnaire should obtain the most complete and accurate
information possible.
The questionnaire should be brief and to the point and be so arranged
that the respondent remains interested throughout the interview.
Development of a questionnaire:
The following steps are involved in the development of a questionnaire:
1. Choose and decide on the information required
The objective behind the survey should be kept in mind while designing a
questionnaire. Though the researcher has an idea about the kind of information
to be collected, additional help can be obtained from secondary data. In respect
of secondary data, the researcher should be aware of what work has been done
on the same or similar problems in the past, what factors have not yet been
examined, and how the present survey questionnaire can build on what has
already been discovered.
2. Define the target respondents
The researcher must define the population that he wishes to collect the data
from. Firstly, in marketing research, researchers often have to decide whether
they should cover only existing users of the generic product type or whether to
also include non-users. Secondly, researchers have to draw up a sampling frame.
Thirdly, in designing the questionnaire we must take into account factors such as
the age, education, etc. of the target respondents.
3. Selection of methodology to reach the target segment
This will influence not only the questions the researcher is able to ask but the
phrasing of those questions. The main methods available in survey research are:
personal interviews, group or focus interviews, mailed questionnaires and
telephone interviews.
Among these, the first two are used more extensively than the second pair. The
general rule is that the more sensitive or personal the information, the more
personal the form of data collection should be.
4. Decide on question content
There will be lot of temptation to use questions without critically evaluating
their contribution towards the achievement of the research objectives. Hence
researchers must proactively look at whether the question is really required and
if it can be used in testing one or more of the hypotheses established during the
research design.
5. Decide on type of questions
The questions can be classified into two forms, i.e. closed and open-ended. So
In a closed type of question, the respondent chooses between an alternative
already stated. He does not get a chance to answer in a descriptive manner.
For ex.: Do you use Brand X? Yes ________ No________.
The closed type of questioning has a number of important advantages:
It is easy for the respondent to answer. He does not have to think and
answer.
It 'prompts' the respondent so that the respondent has to rely less on
memory while answering a question.
Analysis is easier as responses can be easily classified
It permits categorization of the response to specify the answer categories.
It also has some disadvantages:
It does not allow the respondent the opportunity to give a different
response to those suggested.
They 'suggest' answers that respondents may not have considered before.
In an open-ended question the respondent is asked to give a reply to a question
in his/her own words. No answers are suggested. These responses are
explanatory in nature and give some insights from the respondents end.
Example: "What do you like most about this Product?"
Open-ended questions have a lot of advantages when used in a questionnaire:
They allow the respondent to answer in his own words, with no influence
by any specific alternatives suggested by the interviewer.
They often reveal the issues which are most important to the respondent,
and this may reveal findings which were not originally anticipated when
the survey was initiated.
Respondents can 'qualify' their answers or emphasize their opinions.
The inherent problem of an open-ended question is that they must be treated
with caution as:
Respondents may find it difficult to 'articulate' their responses i.e. to
properly and fully explain their attitudes or motivations.
Respondents may not give a full answer simply because they may forget
to mention important points. Some respondents need prompting of the
types of answer they could give.
Data collected is in the form of verbatim comments - it has to be coded
and reduced to manageable categories. This can be time consuming for
analysis and there are numerous opportunities for error in recording and
interpreting the answers given on the part of interviewers.
Respondents will tend to answer open questions in different 'dimensions'.
Such responses need to be probed further for clarity on response.
There are a lot of advantages of this type of questions as
The researcher can avoid the potential problems of poor memory or poor
articulation by then subsequently being able to prompt the respondent into
considering particular response options.
Recording of the responses during the interview is relatively easy.
The one disadvantage of this form of question is that it requires the researcher to
have a good prior knowledge of the subject in order to generate realistic/likely
response options before printing the questionnaire.
In many situations a questionnaire will need to incorporate all these forms of
question, because some forms are more appropriate for seeking particular forms
of response.
6. Putting questions into a meaningful order and format
Opening questions: Opening questions generally should be easy to answer and
not in any way threatening to the respondents. This is crucial because it is the
respondent's first exposure to the interview and sets the tone for the nature of the
task to be performed. If they find the first question difficult to understand, or
beyond their knowledge and experience, or embarrassing in some way, or
uninteresting they are likely to break off immediately. If, on the other hand, they
find the opening question easy and pleasant to answer, they are encouraged to
continue.
Question flow: Questions should flow in some kind of psychological order, so
that one leads easily and naturally to the next. There could be a continuity
maintained on the flow of the questions where the response from on leads to
another. This helps in creating a sequence and the respondent`s interest is
maintained. Questions on one subject, or one particular aspect of a subject,
should be grouped together. Respondents may feel it disconcerting to keep
shifting from one topic to another, or to be asked to return to some subject they
thought they gave their opinions about earlier.
Question variety: Respondents become bored quickly and restless when asked
similar questions for half an hour or so. Hence the questions need to be carefully
keyed in to maintain the interest throughout the interview.
7. Closing questions
By the time the respondent comes to the end of the questionnaire it is quite
natural for a respondent to become increasingly indifferent to the questionnaire.
This is mainly due to impatience or fatigue. He might give careless answers to
the later questions. Hence such questions should be included in the earlier part
of the questionnaire. Potentially sensitive questions should be left to the end, to
avoid respondents cutting off the interview before important information is
collected.
8. Physical appearance of the questionnaire
The physical appearance of a questionnaire has a significant effect upon both the
quantity and quality of marketing data obtained. Data quality can also be
affected by the physical appearance of the questionnaire with unnecessarily
confusing layouts making it more difficult for interviewers, or respondents in the
case of self-completion questionnaires, to complete this task accurately.
In general it is best for a questionnaire to be as short as possible. A long
questionnaire leads to a long interview and this may lead to decreasing interest
on the part of the respondent.
Piloting the questionnaires
Piloting is very mush essential to test whether the desired responses are being
obtained for the questions. Many a time, the perception of the respondents varies
from those of the researcher. Hence these issues can be corrected in the initial
stage itself so that the research process is facilitated. The purpose of pretesting
the questionnaire is to determine:
If the wordings used help in achieving the desired results
Are the questions in the right order?
Are the questions easy to understand?
If any questions needed to be added or deleted.
Are the instructions to interviewers are adequate?
The few respondents selected for the pilot survey should be broadly
representative of the type of respondent to be interviewed in the main survey.
If the questionnaire has been subjected to a thorough pilot test, the research
objective is easy to achieve. This solves the difficulties that may arise later in
terms of invalid and inadequate responses which might lead to wastage of time
and resources.
Summary:
One of the most important components of Marketing Research is collection of
data required to solve a defined research problem. There are two types of data
sources: Secondary and Primary.
Secondary data is defined as the data that has been collected by individuals or
agencies for purposes other than those of the particular research study.
Secondary data can be obtained from internal sources and external sources. It
has to be validated before use to ensure that it suits the purpose of research.
Primary data is the data collected from the field for finding the answer to the
research problem. It can be collected by two methods: Observation and
Questionnaire. Designing a questionnaire is very important as it determines the
quality of response sought and analyzed for finding solutions to a research
problem at hand.
Questions:
1. Tabulate the difference between Primary and Secondary data.
2. Sum up the advantages of secondary data.
3. Does secondary data have some disadvantages? Explain.
4. Give a brief description about the internal sources of secondary
information.
5. List out some of the sources of secondary data known to you.
6. What is primary data? How is it collected?
7. Explain the methods of observation used to collect primary data.
8. Describe the various steps involved in the development of a
questionnaire.
9. Enumerate and tabulate the differences between open and closed
questions.
10. Give the advantages and disadvantages of open and closed questions.
Chapter 5
MARKETING INFORMATION SYSTEM
Objective:
To understand the concept of Marketing Information System or MIS.
Learn about the components of an effective MIS.
List out the types of decisions facilitated by MIS.
Marketing Information System
There are five important functions of managers: Planning, Organizing,
Coordinating, Deciding and Controlling.
A marketing information system (MIS) is intended to bring together a lot of data
into an understandable body of information. An MIS provides processed data
which is suitable for decision making.
According to Kotler, an MIS is more than a system of data collection or a set of
information technologies:
"A Marketing Information System is a continuing and interacting structure of
people, equipment and procedures to gather, sort, analyse, evaluate, and
distribute pertinent, timely and accurate information for use by marketing
decision makers to improve their marketing planning, implementation, and
control".
According to Smith, Brien and Stafford MIS is:
A structured, interacting complex of persons, machines and procedures
designed to generate an orderly flow of pertinent information, collected from
both intra-and extra ? firm sources, for use as the basis for decision-making in
specified responsibility areas of marketing management.
Hence, such a system facilitates decision-making in different areas of marketing
management.
Marketing Research and Marketing Information System:
There are differences between Marketing Research and Marketing Information
System.
Marketing Research is about seeking information from external sources. Its
main purpose is to solve the research problem. It tends to focus on past
information and is not computer based. It is one source of information for
Market Information System.
Marketing Information System handles both data from internal sources like
orders, sales, inventory levels, payables, etc and also data from external sources
like developments in the macro environment.
Components of an effective Marketing Information System:
According to Kotler, an effective Marketing Information System has the
following components:
1. Internal Accounting System: Maintains data pertaining to sales,
receivables, costs, etc that are internal to the organisation.
2. Market Intelligence System: As the name itself implies, it speaks about
the external changes occurring in the macro environment and prepares
mangers to make effective strategies.
3. Marketing Research System: This undertakes studies on specific
marketing problems and provides solution to the management.
4. Marketing Management Science System: It mainly helps in building
models for better understanding of the marketing process.
The below diagram (figure 4) illustrates the major components of an MIS, the
environmental factors monitored by the system and the types of marketing
decision which the MIS seeks to help in.
Data
Information
Marketing
Internal
Marketing
Strategic
Environment
Report
Research
Decisions
System
System
Markets
Channels
Control
Competitors
Marketing
Decisions
Marketing
Political
Intel igence
Models
Legal
System
Operational
Economy
Decisions
Technology
Marketing Decisions and Communications
Figure 4: Components of an MIS
With the help of this model all the four important constituents can be explained.
A fully fledged MIS should have the following components, the methods (and
technologies) of collection, storing, retrieving and processing data.
Three levels of decision making can be observed here: strategic, control (or
tactical) and operational. MIS has to support all these three decision making.
Strategic decisions are very important as they have implications on changing the
structure of an organisation. Therefore the MIS must provide precise and
accurate information.
Control decisions deal with broad policy issues.
Operational decisions concern the management of the organisation's marketing
mix.
MIS should help the manager in his decision making process for problem
identification, generation and evaluation of alternative courses of action, to
acquire necessary feedback on implementing his decision and help him to take
corrective action.
Summary:
There are five important functions of managers: Planning, Organizing,
Coordinating, Deciding and Controlling.
A marketing information system (MIS) is intended to bring together a lot of data
into an understandable body of information. An MIS provides processed data
which is suitable for decision making.
Three levels of decision making can be observed here: strategic, control (or
tactical) and operational. MIS has to support all these three decision making.
Questions:
1. How is a MIS beneficial to a Manager?
2. Define MIS according to various authors.
3. What is the difference between Marketing Research and Marketing
Information System?
4. What are the components of an effective MIS? Explain.
5. What are the types of decision making facilitated by MIS?
References for this Unit:
1. Kotler, P., (1988) Marketing Management: Analysis Planning and
Control, Prentice-Hall
2. Agnilar, F., (1967) Scanning The Business Environment, Macmillan,
New York
3. Green, P.E. Tull, D.S. and Albaum G (1993) Research methods for
marketing decisions, 5th edition, Prentice Hall
4. Joselyn, R. W. (1977) Designing the marketing research,
Petrocellis/Charter, New York
5. Dillon, W.R., Madden, T. and Firtle, N. H., (1994) Marketing Research
in a Research Environment, 3rd edition, Irwin.
6. Sudman, S. and Bradburn, N. M. (1973), Asking Questions
7. G. C. Beri, Marketing Research, second edition, Tata McGraw-Hill
Publishing Company Limited.
8. Wikepedia ? The free encyclopedia
Unit 2 Sampling
________________________________________________________________
_______
Objectives
In this section, we will introduce you to the concept of sampling, sampling
methods, sample size and sampling error. After you go through this unit, you
should be able to understand :
the concept of sampling.
the differences between census and sampling.
various sampling terminologies
how an appropriate sampling design can be determined.
sampling plan and the steps in developing it.
various probability and non-probability sampling methods.
the concept of Random-Digit Dialing.
the concept of sample size and the methods of determining it.
the concept of sampling error and the types.
In this section, we have discussed the following :
2.1. Introduction to sampling
2.1.1. When a Census Is Appropriate
2.1.2. When a Sample Is Appropriate
2.2. Sampling terminology
2.3. Determining the Appropriate Sampling Design
2.
4. Sampling Plan
2.4.1. Steps in Developing a Sampling Plan
2.5. Sampling methods
2.5.1. Types of probability sampling designs
2.5.2. Types of non-probability sampling designs
2.6. Random-Digit Dialing (RDD)
2.7. Sample size
2.8. Sampling error
2.9. Non-response Problems
2.1. Introduction
Sampling is an important concept that we practice in our every day life.
Sampling involves selecting a relatively small number of elements from a larger
defined group of elements and expecting that the information gathered from the
small group will allow judgments to be made about the larger group. If all the
respondents in a population are asked to provide information, such survey is
called a census. Information obtained from a subset of the population is known
as the statistic (from sample). Researchers then attempt to make an inference
about the population parameter with the knowledge of the relevant sample
statistic. Sampling is often used when conducting a census is impossible or
unreasonable. When using a census, the researcher is interested in collecting
primary data about or from every member of the defined target population.
2.1.1. When a Census Is Appropriate
A census is appropriate if the population size itself is quite small. A
census also is conducted if information is needed from every individual or object
in the population. For example, if the researcher is interested in determining the
number of foreign students enrolled in a university, it is necessary to get
information from all the departments in the university because of possible
variations within each department. If the cost of making an incorrect decision is
high or if sampling errors are high, then a census may be more appropriate than
a sample.
2.1.2. When a Sample Is Appropriate
Sampling may be useful if the population size is large and if both the
cost and time associated with obtaining information from the population is high.
The opportunity to make a quick decision may be lost if a large population must
be surveyed. With sampling, in a given time period, more time can be spent on
each personal interview, thereby increasing the response quality. It is easy to
manage surveys of smaller samples and still exercise quality control in the
interview process. Sampling may be sufficient in many instances. If the
population being dealt with is homogeneous, then sampling is fine. If taking a
census is not possible, then sampling is the only alternative.
2.2. Sampling terminology
2.2.1. Population
A population is an identifiable total group or aggregation of elements
that are of interest to the researcher and pertinent to the specified problem. A
defined target population consists of the complete group of elements that are
specifically identified for investigation according to the objectives of the
research project.
2.2.2. Element
An element is a person or object from which data and information are
sought. In research, the element is a particular product or group of individuals.
Elements must be unique, countable and when added together, make up the
whole target population. Target population elements must include a particular
consumer product, specific group of people or specific organisations.
2.2.3. Sampling units
Sampling units are the target population elements available for selection
during the sampling process. In a simple, single-stage sample, thee sampling
units and the population elements may be the same. But many studies involve
complex problems that require the use of a multi-stage sampling process.
2.2.4. Sampling frame
It is the list of all eligible sampling units. Some common sources of
sampling frames are list of registered voters and customer lists from magazine
publishers, credit card companies and the like. There are specialized commercial
companies that are in the business of developing databases that contain names,
addresses and telephone numbers of potential population elements. It is usually
very difficult and expensive for a researcher to
gain access to truly accurate or representative , current sampling frames. In such
situations, a researcher would have to employ an alternate method such as
random-digit dialing or a location survey in order to generate a sample of
prospective respondents.
The sampling frame contains the operational population from which the sample
will be drawn. In an ideal situation, the operational population, the defined
target population frame and the sampling frame are identical. In those situations
where a sampling frame contains all of the eligible sampling units of the defined
population plus additional ones,
Then it is said to have over-registration.
But if the eligible sampling units are accidentally left out of the sampling
frame, hen the frame has an under-registration condition.
2.2.5. Sampling gap
Both over-registration and under-registration factors create sampling
gaps. A sampling gap is the representation difference between the population
elements and sampling units in the sampling frame. A sampling gap can also be
viewed as sampling frame error and occurs when certain sample units are not
excluded or complete segments of the defined target population are not
accurately represented in the sampling frame. The larger the sampling frame
error, the greater the chance of misleading and inaccurate data results.
2.2.6. Sampling distribution
One important assumption that underlies sampling theory is that the
population elements are randomly distributed. If a researcher were able to do a
census of the entire target population elements, then the probability distribution
of the population, or the relative frequencies of a population`s parameters would
depict a normal bell-shaped distribution pattern.
Sampling distribution is the frequency distribution of a specific sample
statistic ( sample mean or sample proportion ) from repeated random samples of
the same size.
2.2.7. Central Limit Theorem
The Central Limit Theorem becomes the backbone for doing survey
research and data collection through experimental designs. The theorem states
that for almost all defined target populations, the sampling distribution of the
mean or the percentage value derived from a simple random sample will be
approximately normally distributed, provided that the sample size is sufficiently
large. When n` is greater than or equal to 30, the sample is a large sample.
Activity 2.1.
Develop a population list or sampling frame for an attitude study, when the target
population is :
i. the students of your college.
ii. High-income families in your residential area.
iii. Shops that sell tennis rackets in your city.
2.3. Determining the Appropriate Sampling Design
Selection of the most appropriate sampling design should incorporate the
seven factors :
Research objectives
A full understanding of the overall information problem situation and the
research objectives provides the initial guidelines for determining the
appropriate sampling design. If the research objectives include the desire to
generalise the sample data results to the defined target population, then the
researcher must seriously consider using some type of probability sampling
method rather than a non-probability sampling method. The stage of the research
project and type of research (e.g., exploratory, descriptive, casual will influence
the researcher`s selection of sampling method.
Degree of accuracy
The degree of accuracy required or the researcher`s tolerance for error
may vary from project to project. If the researcher wants to make predictions or
inferences about the true position of all members of the defined target
population, then he or she must choose some type of probability sampling
method. In contrast, if the researcher is solely trying to identify and obtain
preliminary insights into the defined target population, non-probability methods
might prove to be more appropriate.
Availability of resources
If the researcher`s financial and human resources are restricted, these
limitations will certainly eliminate some of the more time-consuming, complex
probability sampling methods. Researchers who are influenced by the cost
concerns versus the value of the information will often opt for a non-probability
sampling method rather than conduct no research at all.
Critical factors in selecting the appropriate sampling design
__________________________________________________________________
F
ac t
or
Questions
__________________________________________________________________
Research objectives
Do the research objectives call for the use of qualitative or
quantitative research designs ?
Degree of accuracy
Does the research call for making predictions about the
defined target population or only preliminary insights ?
Availability of resources
Are there budget constraints with respect to the resources that
can be allocated to the research project ?
Time frame
How quickly does the research project have to be completed ?
Advanced knowledge of
Are there complete lists of the defined target population
Target population
elements ?
Scope of the research
Is the research going to be international, national, regional or
Time frame
Researchers who need to meet a short deadline will be more likely to
select a simple, less time-consuming sampling method rather than a more
complex and accurate method. Researchers tend to opt for using some form of
convenience sampling to gather data necessary to test the reliability of a newly
developed construct or scale measurement.
Advanced knowledge of the target population
In many cases, a complete list of the population elements will not be
available to the researcher. A lack of adequate lists may automatically rule out
systematic random sampling, stratified random sampling, or any other type of
probability sampling method. A preliminary study may be conducted to
generate information to build a sampling frame for the study. The researcher
must gain a strong understanding of the key descriptor factors that make up the
true members of any target population.
Scope of the research
Whether the scope of the research project is to be international, national,
regional, or local will influence the choice of the sampling method. The
projected geographic proximity of the defined target population elements will
influence the researcher`s ability to compile needed lists of sampling units.
When the target population elements are known or viewed to be unequally
distributed geographically, a cluster sampling method may become much more
attractive than other available methods. The broader the geographical scope of
the research project, the more extensive and complex the sampling method
becomes to ensure proper representation of the population.
Perceived statistical analysis needs
The need for statistical projections (i.e., estimates) based on the sample
results is often a criterion. Only probability sampling techniques allow the
researcher to appropriately use statistical analysis for estimates. While statistical
analysis methods can be performed on data structures obtained from non-
probability samples of people and objects, the researcher`s ability to accurately
generalize the results and findings to the larger defined target population is very
suspect and technically inappropriate.
2.4. Sampling Plan
A sampling plan is the blueprint or frame work needed to ensure that the
raw data collected are representative of the defined target population. A good
sampling plan will include, the following steps: (1) define the target population,
(2) select the data collection method, (3) identify the sampling frames needed,
(4) select the appropriate sampling method, (5) determine necessary sample
sizes and overall contact rates, (6) create an operating plan for selecting
sampling units, and (7) execute the operational plan.
2.4.1. Steps in Developing a Sampling Plan
Step 1: Define the target population
In any sampling pan, the first task of the researcher is to
determine and identify the complete group of people or objects that should be
investigated in the project. The target population should be given its identity by
the use of descriptors that represent the characteristics of elements that make the
target population`s frame. These elements become the prosperity sampling units
from which a sample will be drawn. Clear understanding of the target population
will help the researcher successfully draw a representative sample. Devoting
effort to identifying the target population usually will pay off. The following
guidelines should be considered :
Look to the research objectives
If the research objectives are well thought out, the target population
definition will be clear as well. The Research objectives include the research
question, the research hypothesis and a statement of the research boundaries.
Each of these elements contributes to refining the definition of the target
population.
Consider alternatives
It is rare to find a study for which there are no alternative, reasonable,
target population definitions. The task is to identify and evaluate several of the
alternatives. The key point is to recognise that alternative definitions exist.
Know your market
If the research objective is to learn about the market response to some
element of the marketing program, it is necessary to know something about the
market. Without it, the population definition will have to be unnecessarily broad
and, therefore, will lead to an unnecessary increase in research expenses.
Consider the appropriate sampling unit
The target population consists of sampling units. A sampling unit may
contain people, households, or products. One task is to specify which sampling
unit is appropriate.
Specify clearly what is excluded
The specification of target population should make clear what is
excluded. For example, a study of voting intentions on certain candidates and
issues might restrict the sampling population to those of voting age and even to
those who intend to vote or to those who voted in the last election.
Don't over-define
The population, should be compatible with the study purpose and the
research questions; but, the research should not over-define the population.
Consider convenience
When there is a choice, preference should be given to populations that
are convenient to sample.
Step 2: Select the data collection method
Using the information problem definition, the data requirements, and the
established research objectives, the researcher must choose a method for
collecting the required raw data from the target population elements. Choices
include interviewing approach or a self-administered survey. The method of data
collection guides the researcher in identifying and securing the necessary
sampling frame(s) for conducting the research.
Steps in developing a Sampling plan
Step 1
Define the target population
Step 2
Select the data collection method
Step 3
Identify the sampling frame(s) needed
Step 4
Select the appropriate sampling method
Step 5
Determine the necessary sample sizes and overall
contact rates
Step 6
Create an operating plan for selecting sampling units
Step 7
Execute the operational plan
Step 3: Identifying the sampling frame(s) needed
After gaining an understanding of whom or what should be investigated,
the researcher must assemble a list of eligible sampling units. This list needs to
contain enough information about each prospective sampling unit so that the
researcher can successfully contact them. An incomplete sampling frame
decreases the likelihood of drawing a representative sample. Sampling frame
lists can be created from a number of different sources. In creating the necessary
sampling frames, the researcher must be aware of possible conditions of over-
registration and under-registration of the prospective sampling units. These
conditions will create sampling gaps or sampling frame errors that decrease the
likelihood of being able to draw a representative sample.
Creating lists
The biggest problem in simple random sampling is obtaining appropriate
lists. Lists do not exist for specialised populations. A solution for this problem is
just to use a convenient list. When lists that do not match the population are
used, biases are introduced. Sometimes several lists are combined in the hope of
obtaining a more complete representation of the population. This approach
introduces the problem of duplication. Those appearing on several lists will have
an increased chance of being selected. Removing duplication can be expensive
and must be balanced against the bias that is introduced. Another problem with
lists is simply that of keeping them current. These lists can become outdated
quickly as people move and change jobs within an organisation.
Creating lists for telephone interviewing
Telephone directories are used extensively as a basis for generating a
sample. The concern with the use of directories is that population members may
be omitted because they have changed residences, requested an unlisted number,
or simply do not have a telephone. The incidence of unlisted numbers is
extensive and varies dramatically from area to area. Another approach is to buy
lists from magazines, credit-card firms, mail-order firms, or other such sources.
One problem is that each such list has its own type of biases.
Dealing with population sampling frame differences
When a sampling frame does not coincide with a population definition,
three types of problems arise : the subset problem, the superset problem and the
intersection problem. A subset problem occurs when the sampling frame is
smaller than the population. In other words, some of the elements in the
population will not be present in the sampling frame. A superset problem occurs
when the sampling frame is larger than the population but contains all the
elements of the population. An intersection problem occurs when some elements
of the population are omitted from the sampling frame, and when the sampling
frame contains more elements than the population.
Step 4: Select the appropriate sampling method
The researcher must choose between two types of sampling
orientations : probability and non-probability. Using a probability sampling
method will always yield better and more accurate information about the target
population`s parameters than will any of the available non-probability sampling
methods. Probability sampling has several advantages over non-probability
sampling. First, it permits the researcher to demonstrate the sample's
representativeness. Second, it allows an explicit statement as to how much
variation is introduced, because a sample is used instead of a census of the
populations. Finally, it makes possible the more explicit identification of
possible biases.
Step 5: Determine necessary sample sizes and overall contact rates
In this step of a sampling plan, the researcher must consider how precise
the sample estimates must be and how much time and money are available to
collect the required raw data. To determine the appropriate sample size,
decisions have to be made concerning (1) the variability of the population
characteristic under investigation, (2) the level of confidence desired in the
estimates, and (3) the degree of precision desired in estimating the population
characteristic. The researcher must decide how many completed surveys will
need to enter the data analysis activities of the overall research project.
Step 6 : Create an operating plan for selecting sampling units
In this step, the researcher wants to clearly lay out, in detail, the actual
procedures to use in containing each of the prospective respondents who were
drawn into the sample. All instructions should be clearly written so that
interviewers know exactly what to do and how to handle any problems in the
process of contacting prospective respondents.
Step 7 : Execute the operational plan
In some research projects, this step is similar to actually conducting the
data collection activities. (e.g., actual calling of a prospective respondent to do a
telephone interview). The important thing in this stage is to maintain
consistency and control.
Activity 2.2.
A telephone survey is planned to determine the day-after recall of several test
commercials to be run in Pondicherry. Design a sampling plan.
2.5. Sampling methods
There are two basic sampling designs : Probability and non-probability
sampling methods. In probability sampling, each unit in the defined target
population has a known, non-zero probability of being selected for the sample,
The actual probability of selection for each sampling unit may or may not be
equal depending on the type of probability sampling design used. It allows the
researcher to judge the reliability and validity of raw data collected by
calculating the probability to which the findings based on the sample would
differ from the defined target population. The results obtained by the
probability method can be generalized to the target population within a specified
margin of error through the use of statistical methods. In non-probability
sampling, the probability of selection of each sample unit is not known.
Therefore, potential sampling error cannot be accurately known either. The
selection of sampling units is based on some type of intuitive judgments, desire
or knowledge of the researcher. The degree to which the sample nay or may not
be representative of the defined target population depends on the sampling
approach and how well the researcher executes and controls the selection
activities. There is always a temptation to generalize non-probability sample
data results to the defined target population.
Comparative differences of probability and non-probability sampling methods
________________________________________________________________
Factor
Probability sampling
Non-probability
sampling
_____________________________________________________________________________
1.List of the population elements Complete list necessary
Not necessary
2. Information about the-
Each unit identified
need detail on habits,
sampling units
activities, traits etc.
3. Sampling skill
skill required
little skill required
4. Time requirement
More time-consuming
Less time consuming
5. Cost per unit sampled
Moderate to high
Low
6. Estimate of population -
Unbiased
Biased
parameters
7. Sample representativeness
Assured
Undeterminable
8. Accuracy and
Computed with
Unknown
Relaiability
confidence intervals
2.5.
1. Types of probability sampling designs
9. Measurement of sampling error Statistical measures
No true measure
Sim
ple ran dom s ampling
available
________________________________________________________________
Simple Random Sampling is a probability sampling procedure which
ensures that every sampling unit making up the defined target population has a
known, equal, non-zero chance of being selected. For example, let`s say an
instructor decided to draw a sample of 10 students (n=10), from among all the
students in a Marketing Research class that consisted of 30 students ( N=30).
The instructor could write each student`s name on a separate, identical piece of
paper and place all of the names in a jar. Each student would have an equal,
known probability of selection for a sample of a given size that could be
expressed by the formula :
Size of sample
Probability of selection = -----------------------------------
Size of population
Here, each student would have a 10 /30 (or 0.33) chance of being
randomly selected in the drawn sample. When the defined target population
consists of a larger
Types of sampling methods
Samplin g Methods
Probability sampling
Non-probability
methods
sampling methods
Simple Random sampling
Convenience sampling
sampling
Systematic Random sampling
Judgment sampling
Stratified Random sampling
Quota sampling
Cluster sampling
Snowball sampling
number of sampling units, a more sophisticated method would be used to
randomly draw the necessary sample. One of the procedures commonly used in
marketing research is to incorporate a printed or computer generated table of
random numbers to select the sampling units. A table of random numbers is a
table that lists randomly generated numbers. Many computer programs have the
ability to generate a table random numbers.
With the marketing research students defined above as the target
population a random sample could be generated by assigning each students a
unique two-digit code ranging from 01 to 30. Then we could go to the table of
random numbers and select a starting point, which can be anywhere on the table.
Using the partial table of random numbers given below, say we select the upper-
left-hand corner of the table (31) as our starting point. We would then begin to
read down the first column (or across the first row) and select those two-digit
numbers that matched the numbers within the acceptable range until 10 students
had been selected. Reading down the first column, we would start with 31, then
go to 14, 49,99, 54 and so on.
A partial table of random numbers
_________________________________________
31 25
81 44
54 34
67 03
14 96
99 80
14 54
30 74
49 05
49 56
35 51
68 36
99 67
57 65
14 46
92 88
54 14
95 34
93 18
78 27
57 50
34 89
99 14
57 37
98 67
78 25
06 90
39 90
40 99
00 87
90 42
88 18
20 82
09 18
84 91
64 80
78 84
39 91
16 08
14 89
_________________________________________
Source : M.G.Kendall and B. Babington Smith, Table of Random Sampling Numbers,
Tracts for Computers, 24 (Cambridge, England : Cambridge University Press, 1946), p.33.
Use only those random numbers that matched the numbers within the
acceptable range of 01 to 30. Numbers that fall outside the acceptable range
would be disregarded. Thus, we would select students with numbers 14, 20, 25,
05, 09, 18, 06, 16, 08, and 30. If the overall research objectives call for
telephone interviews, drawing the necessary sample can be achieved using the
random-digit dialing (RDD) technique.
Advantages and disadvantages
The simple random sampling technique has several advantages. The
technique is easily understood and the survey`s data results can be generalised to
the defined target population with a pre-specified margin of error e`.
Another advantage is that simple random samples allow the researcher to
gain unbiased estimates of the population`s characteristics. This method
basically guarantees that every sampling unit of the population has a known and
equal chance of being selected, no matter the actual size of the sample, resulting
in a valid representation of the
Defined target population. The disadvantage of this method is the difficulty of
obtaining a complete, current, and accurate listing of the population elements.
Simple random sampling requires that all sampling units be identified. For this
reason, simple random sampling often works best for small populations or those
where computer-derived lists are available.
Activity 2.3.
By using a standard random number table, generate random numbers for
selecting 50 samples from a population of 800 students who are doing a
course on Marketing Research under ABC University.
Systematic Random Sampling
Systematic random sampling (SYMRS) is similar to simple random
sampling but requires that the defined target population be ordered in some way,
usually in the form of a customer list, taxpayer roll, or membership roster. In
research practices, SYMRS has become a very popular alternative probability
method of drawing samples. Compared to simple random sampling, systematic
random sampling is potentially less costly because it can be done relatively
quickly. When executed properly, SYMRS can create a sample of objects or
prospective respondents that is very similar in quality to a sample drawn using
simple random sampling.
To employ SYMRS, the researcher must be able to secure a complete
listing of the potential sampling units that make up the defined target population.
Individual sampling units are selected according to their position using a skip
interval. The skip interval is determined by dividing the number of potential
sampling units in the defined target population by the number of units desired in
the sample. The required skip interval is calculated using the formula :
Skip interval = Defined target population list size
Desired sample size
For instance, if the researcher wants a sample of 100 to be drawn from a
defined target population of 1,000, the skip interval would be 10 (ie. 1,000/100).
Once the skip interval is determined, the researcher would then randomly select
a starting point and take every 10th unit until he proceeded through the entire
target population list.
There are two important considerations when using systematic random
sampling. First, it is important that the natural order of the defined target
population list be unrelated to the characteristic being studied. Second, the skip
interval must not correspond to a systematic change in the target population.
Activity 2.4.
Co
m
pa r e and contrast simple random sampling and systematic random sampling.
Take 10 samples from a population of all the students in your class.
Write a short report on how differently the samples came out of the two methods
and which method generated a truly representative sample of your class.
Steps in drawing a Systematic Random Sample
Step 1
Obtain a list of potential sampling units that contains an acceptable
frame of the target population elements.
Step 2
Determine the total number of sampling units making up the list of the
defined target population`s elements and the desired sample size.
Step 3
Compute the needed skip interval by dividing the number of potential
sampling units on the list by the desired sample size.
Step 4
Using a random number generation system, randomly determine a
starting point to sample the list of names.
Step 5
Apply the skip interval to determine the remaining items to be included
in the sample.
Advantages and disadvantages
Systematic sampling is frequently used because, if done correctly, it is a
relatively easy way to draw a sample while ensuring randomness. The
availability of lists and the shorter time required to draw a sample makes
systematic sampling an attractive, economical methods for researchers. The
greatest weakness of systematic random sampling is the potential for there to be
hidden patterns in the data that are not found by the researcher. This could result
in a sample that is not truly representative of the defined target population.
Another difficulty is that the researcher must know exactly how many sampling
units make up the defined target population. In research situations in which the
size of the target population is extremely large or unknown, identifying the true
number of units is difficult, and even estimates may not be accurate.
Stratified Random Sampling
Stratified random sampling (STRS) requires the separation of the defined
target population into different groups, called strata, and the selecting of samples
from each stratum. The goal in stratifying is to minimize the variability (or
skewness) within each stratum and maximize the differences between strata. In
some ways, STRS can be compared to segmentation of the defined target
population into smaller, more homogeneous sets of elements.
To ensure that the sample maintains the required precision of the total
population, representative samples must be drawn from each of the smaller
population groups. Drawing a stratified random sample involves three basic
steps:
i. Dividing the target population into homogeneous sub-groups or
strata.
ii. Drawing random samples from each stratum.
iii. Combining the samples from each stratum into a single sample of the
target population.
There are two common methods for deriving samples from the strata :
proportionate and disproportionate. In proportionate stratified sampling, the
sample size from each stratum is depended on the stratum`s size relative to the
defined target population. Therefore, the larger strata are sampled more heavily
using proportionate stratified sampling because they make up a larger
percentage of the target population. In disproportionate stratified sampling, the
sample size selected from each stratum is independent of that stratum`s
proportion of the total defined target population. This approach is used when
stratification of the target population produces sample sizes for sub-groups that
contradict their relative importance to the study.
An alternative type of disproportionate stratified method is optimal
allocation. In this method, consideration is given to the relative size of the
stratum as well as variability within the stratum to determine the necessary
sample size of each stratum. The basic logic underlying optimal allocation is
that the greater the homogeneity of the prospective sampling units within a
particular stratum, the fewer the units that would have to be selected to the
estimate the true population parameter (0 or P) accurately for that sub- group.
The different types of stratified sampling are :
Proportional Stratified Sampling
In this type of sampling procedure the number of objects or sampling
units chosen from each group is proportional to the number in the population.
Proportional stratified sampling can further be classified as directly proportional
and inversely proportional stratified sampling.
Directly Proportional Stratified Sampling
Assume that a researcher is evaluating customer satisfaction for a
beverage that is consumed by a total of 600 people. Among the 600 people, 400
are brand-loyal and 200 are variety-seeking. Past research indicates that the level
of customer satisfaction is related to consumer characteristics, such as being
either brand-loyal or variety-seeking. Therefore, it should be beneficial to divide
the total population of 600 consumers into two groups 400 and 200 each and
randomly sample from within each of the two groups. If a sample size of 60 is
desired, then a 10 percent directly proportional stratified sampling is employed.
10 Percent Directly Proportional Stratified Sample
Consumer Type
Group Size
Size
Brand-loyal
400
40
Variety-seeking
200
20
Total
600
60
Steps in drawing a Stratified Random Sample
Step 1
Obtain a list of potential sampling units that contains an acceptable
frame of the target population elements.
Step 2
Using some type of secondary information or past experience with the
defined target population, select a stratification factor for which the
population`s distribution is skewed and can be used to determine that
the total defined target population consists of separate subpopulations of
elements
Step 3
Using the selected stratification factor, segment the defined target
population into strata consistent with each of the identified separate
subpopulations.
Step 4
Determine whether there is a need to apply a disproportionate or
optimal allocation method to the stratification process. Otherwise, use
the proportionate method and then estimate the desired sample size.
Step 5
Select a probability sample from each stratum.
Inversely Proportional Stratified Sampling
Assume that among the 600 consumers in the population, say 200 are
very heavy drinkers and 400 are light drinkers. If a researcher values the opinion
of the heavy drinkers more than that of the light drinkers, more people will have
to be sampled from the heavy drinkers group. In such instances, one can use an
inversely proportional stratified sampling. If a sample size of 60 is desired, a 10
percent inversely proportional stratified sampling is employed.
10 Percent Inversely Proportional Stratified Sample
Consumer Type
Group Size
Size
Heavy Drinkers
400
40
Light Drinkers
200
20
Total
600
60
In inversely proportional stratified sampling, the selection probabilities
are computed as follows:
Denominator = 600/200 + 600/400 = 3 + 1.5 = 4.5
Heavy Drinkers proportional sample size = 3/ 4.5 = 0.667; 0.667 x 60 =
40
Light Drinkers proportional sample size = 1.5/4.5 = 0.333; 0.333 x 60 =
20
Disproportional Stratified sampling
In stratified sampling, when the sample size in each group is not
proportional to the respective group sizes, it is known as disproportional
stratified sampling. When multiple groups are compared and their respective
group sizes are small, a proportional stratified sampling will not yield a sample
size large enough for meaningful comparisons, and disproportional stratified
sampling is used. One way of selecting sample sizes within each group is to
have equal group sizes in the sample. In the example of heavy and light drinkers,
a researcher could select 30 people from each of the two groups.
In general, stratified sampling is employed in many research projects,
because it is easy to understand and execute.
Advantages and Disadvantages
Dividing the defined target population into homogeneous strata provides
several advantages, including:
i. the assurance of representativeness in the sample;
ii. the opportunity to study each stratum and make relative comparison
between strata; and
iii. the ability to take estimates for the target population with the
expectation of greater precision or less error in the overall sample.
The primary difficulty encountered with stratified sampling is determining the
basis for stratifying. It is imperative that the basis for stratifying be directly
associated with the target population`s characteristics of interest. Normally, the
larger the number of relevant strata, the more precise the results. The inclusion
of excess or irrelevant strata will only waste time and money without providing
meaningful results.
Cluster Sampling
While cluster sampling is similar to stratified random sampling, it is
different in that the sampling units are divided into mutually exclusive and
collectively exhaustive sub-populations, called clusters. Each cluster is assumed
to be representative of the heterogeneity of the target population. Examples of
possible divisions for cluster sampling include the customers who patronise a
store on a given day, the audience for a movie shown at a particular time (e.g.,
the matinee), or the invoices processed during a specific week. Once the cluster
has been identified, the prospective sampling units are drawn into the sample by
either using a simple random sampling method or canvassing all the elements
within the defined cluster.
Area sampling
A popular form of cluster sampling is area sampling. In area sampling,
the clusters are formed by geographic designations. Examples include cities, sub
divisions and blocks. Any geographical unit with identifiable boundaries can be
used. When using area sampling, the researcher has to additional options : the
one-step approach or the two -step approach. When deciding on using one-step
approach, the researcher must have enough prior information about the various
geographic clusters. By assuming that all the clusters are identical, the
researcher can focus his attention on surveying the sampling units within one
designated cluster group and then generalize the data results to the full target
population. The probability aspect of this particular sampling method is
executed by randomly selecting one geographic cluster and performing a census
on all the sampling units located within that selected cluster.
Alternatively, the researcher may execute a two-step cluster sampling
approach. First, the researcher would randomly sample a set of cluster and then
would decide on the most appropriate probability method to sample individuals
within each of the selected clusters. The two-step approach is preferable over the
one-step approach, because there is a strong possibility that a single cluster will
not be as representative of all other clusters as the researcher thinks.
Advantages and Disadvantages
The cluster sampling method is widely used in marketing research due to
its overall cost-effectiveness and feasibility of implementation, especially in area
sampling situations. In many cases, the only reliable sampling unit frame
available to researcher is one that describes and lists clusters. These lists of
geographic regions, telephone exchanges, or blocks of residential dwellings can
normally be easily compiled. Clustering method tends to be a cost-efficient way
of sampling and collecting raw data from a defined target population.
One primary disadvantage related to cluster sampling is the tendency for
clusters to be homogeneous. The more homogeneous the cluster, the less
precise the derived sample estimate in representing the defined target
population`s parameters. The actual object or people within a cluster should be
as heterogeneous as those in the target population itself. Another concern with
cluster sampling methods is the appropriateness of the designated cluster factor
used to identify the sampling units within clusters.
A comparison between the stratified sampling process and the cluster
sampling process is given in the following table :
Stratified Sampling
Cluster Sampling
Homogeneity within group
Homogeneity between groups
Homogeneity between groups
Homogeneity within group
All groups are included
Random selection of groups
Sampling efficiency improved by increasing Sampling efficiency improved by decreasing
accuracy at a faster rate than cost
cost at a faster rate than accuracy
Activity. 2.5.
Compare and contrast various probability sampling methods. Prepare a chart
showing their relative merits and demerits.
2.5.2. Types of non-probability sampling designs
Non-probability sampling typically is used in situations such as :
i. the exploratory stages of a research project,
ii. pre-testing a questionnaire,
iii. dealing with a homogeneous population,
iv. when a researcher lacks statistical knowledge, and
i. when operational ease is required.
Convenience Sampling
Convenience sampling (or accidental sampling) is a method in which
samples are drawn at the convenience of the researcher or interviewer. The
assumptions are that the target population is homogeneous and the individuals
interviewed are similar to the overall defined target population with regard to the
characteristics being studied.
Advantages and Disadvantages
Convenience sampling allows a large number of respondents to be
interviewed in a relatively short time. For this reason, it is commonly used in
the early stages of research. The use of convenience samples in the
development phases of constructs and scale measurements can have a seriously
negative impact on the overall reliability and validity of those measures and
instruments used to collect raw data. Another major disadvantage of
convenience samples is that the raw data and results are not generalized to the
defined target population with any measure of precision. It is not possible to
measure the representativeness of the sample, because sampling error estimates
cannot be accurately determined.
Judgment Sampling
In judgment sampling, (also referred to as purposive sampling),
participants are selected according to an experienced individual`s belief that they
will meet the requirements of the study. Judgmental sampling is associated with
a variety of biases. For example, shopping center intercept interviewing can
over-sample those who shop frequently, who appear friendly, and who have
uncertainty, because the sampling frame is unknown and the sampling procedure
is not well specified.
There are situations where judgmental sampling is useful and even advisable.
First, there are times when probability sampling is either not feasible or
expensive. For example, a list of sidewalk vendors might be impossible to
obtain, and a judgmental sample might be appropriate in that case.
Second, if the sample size is to be very small - say, under 10 - a
judgmental sample usually will be more reliable and representative than a
probability sample. Suppose one or two cities of medium size are to be used to
represent 200 such cities. Then it would be appropriate to pick judgmentally two
cities that appeared to be most representative with respect to such external
criteria as demographics, media habits, and shopping characteristics.
Third, sometimes it is useful to obtain a deliberately biased sample. If,
for example, a product or service modification is to be evaluated, it might be
possible to identify a group that, by its very nature, should be disposed toward
the modification.
Advantages and Disadvantages
If the judgment of the researcher or expert is correct, then the sample
generated from judgment sampling will be much better than one generated by
convenience sampling. But, it is not possible to measure the representativeness
of the sample. The raw data and information collected from sampling units
generated though the judgment sampling method should be interpreted as
nothing more preliminary insights.
Quota Sampling
The quota sampling method involves the selection of prospective
participants according to pre-specified quota regarding either demographic
characteristics (e.g., age, race, gender, income), specific attitudes (e.g.,
satisfied/dissatisfied, liking/disliking, great/marginal/no quality), or specific
behaviours (e.g., regular/occasional/ rare customer, product user/non user). The
underlying purpose of quota sampling is to provide an assurance that pre-
specified sub-groups of the defined target population are represented on
pertinent sampling factors that are determined by the researcher or client.
Surveys frequently use quotas that have been determined by the specific nature
of the research objectives.
In order to meet the quotas, researcher using quota sampling sometimes
overlook the problems associated with adhering to the quotas. Assume that an
oil company is interested in finding out if women assume responsibility for
vehicle maintenance. The company is interested in interviewing women aged
below 35 and with age equal to and above 35, as well as working women and
nonworking women. Suppose the distribution of the population of women in a
city (N=1,000) is as follows:
Population Characteristics
<35 years
35 years and Above
Total
Percentage
Working women
300
200
500
50
Non-working women
200
300
500
50
Total
500
500
1,000
100
Percentage
50
50
100
Assume that the researcher is interested in interviewing 100 women from
this city and develops a quota system such that 50 percent of the sample should
be working women and 50 percent of the sample should also be under 35 years
old. A quota matrix can be developed for a sample size of 100.
Sample Characteristics
<35 years
35 years and Above
Total
Percentage
Working women
50
0
50
50
Non-working women
0
50
50
50
Total
50
50
100
100
Percentage
50
50
Advantages and Disadvantages
The greatest advantage of quota sampling is that the sample generated
contains specific sub-groups in the proportions desired by researchers. In those
research projects that require interviews, the use of quotas ensures that the
appropriate sub-groups are identified and included in the survey. The quota
sampling method may eliminate or reduce selection bias on the part of the field
workers. An inherent limitation of quota sampling is that the success of the
study will be dependent on subjective decisions made by the researchers. Also,
it is incapable of measuring the true representativeness of the sample or
accuracy of the estimate obtained. Hence, attempts to generalise data results
beyond those respondents who were sampled and interviewed become very
questionable and may misrepresent the defined target population.
Snowball Sampling
Snowball Sampling involves the practice of identifying and qualifying a
set of initial prospective respondents who can, in turn, help the researcher
identify additional people to be included in the study. This method of sampling
is also called referral sampling, because one respondent refers other potential
respondents. Snowball sampling is typically used in research situations where :
i. the defined target population is very small and unique, and
ii. compiling a complete list of sampling units is a nearly impossible
task.
The snowball method would yield better results at a much lower cost. Here the
researcher would identify and interview one qualified respondent, then solicit
his or her help in identifying other people with similar characteristics. The main
underlying logic of this method is that rare groups of people tend to form their
own unique social circles.
Advantages and Disadvantages
Snowball sampling is a reasonable method of identifying and selecting
prospective respondents who are members of small, hard-to-reach, uniquely
defined target population. It is most useful in qualitative research practices, like
focus group interviews. Reduced sample sizes and costs are primary advantages
to this sampling method. Snowball sampling definitely allows bias to enter the
overall research study.
Activity. 2.6.
Co
m
pa r e
a
nd contrast various non-probability sampling methods. Prepare a chart
showing their relative merits and demerits.
2.6. Random-Digit Dialing (RDD)
If truly accurate or representative, current sampling frames are not
available, a researcher would have to employ an alternate method such as
random-digit dialing, if conducting telephone interviews. The only solution to
the problem of unlisted telephone numbers is to generate phone numbers by
some random process. This practice, referred to as Random-Digit Dialing
involves generation of random numbers at random. But practical considerations
complicate the problem greatly. The foremost is the relatively small proportion
of the working numbers among all possible 10-digit telephone numbers. Only
about 1 in 170 of all possible telephone numbers is actually in use. The
proportion of the working residential numbers in RDD samples can be increased
dramatically by selecting from only the 103 working area codes ( first three
digits). The approach yields approximately 1 working residential number for
every 17 randomly generated. From a cost stand point, this rate is still low,
entailing too many unproductive dialings while including a proportionate
number of unlisted phone homes in the sample. There are three alternative
approaches built around the use of a telephone book.
Four-Digit approach
Taking the four-digit approach, the researcher must, in addition to
restricting the sample to the 103 working area codes, select numbers only from
working central offices or exchanges. The lat four digits of the number are
generated via some process that approaches randomness. Problem with this
approach is that all exchanges have an equal probability of being selected, while
some have a high proportion of all possible numbers in service and others have
only a small proportion in service.
Three-Digit approach
The next logical progression in RDD is the three-digit dialing approach.
This method increases the proportion of working numbers to better than one in
three. Consulting the section of a criss-cross directory where phone numbers are
listed numerically will show that within a particular exchange, certain sequences
of 1000 numbers are totally unused while other groups of 1000 are in use.
Generate the last three digits of each exchange by means of some random
process. This method is more efficient in eliminating non working numbers, but
increases bias due to missing new exchanges that have been activated.
The four digit method is safer from the standpoint of avoiding bias, but
more expensive due to the greater number of calls that must be made. It is
suggested that three-digit method is most appropriate when the directories for
the area of interest are relatively current or when there has been little growth in
the area since the publication of the re cent directory. In other cases, the four-
digit method should be used.
Using telephone books
RDD samples can also be generated from the telephone book. This is
accomplished by selecting numbers at random from the book and adding a
random number as the sixth or seventh digit. Somewhere between one in two
and one in three of the numbers generated will be working residential numbers.
This is a viable approach because, all exchanges are proportionately represented
in the book. The phone book is recommended as an RDD sample source only in
those cases where the appropriate computer hardware and software are not
available. There are two major reasons for making this recommendation. First,
the construction of sample by this approach is time-consuming and expensive.
Second, if the interviewers are given directions and left to generate numbers
themselves, the researcher loses all control over the validity of the sample.
Computer programmes can incorporate three- or four- digit approaches and
generate RDD samples at a very low cost. The print-out can be set up to capture
additional data and to help the researcher control field costs and proper
execution of the sampling plan.
Activity 2.7.
Identify a situation where you would be in favour of using a non-probability
sampling method over probability sampling method.
2.7. Sample size
Determining the appropriate sample size is not an easy task. The
researcher must consider how precise the estimates must be and how much time
and money are available to collect the required data, since the data collection is
one of the most expensive components of a study. Three factors play an
important role in determining appropriate sample sizes. They are :
i. The variability of the population characteristic under consideration : The
greater the variability of the characteristic, the larger the size of the sample
necessary.
ii. The level of confidence desired in the estimate : The higher the level of
confidence desired, the larger the sample size needed.
iii. The degree of precision desired in estimating the population characteristic :
The more precise the required sample results, the larger the necessary sample
size.
2.7.1. Estimating the sample size by traditional methods
There are four traditional approaches to determine the sample size. They
are :
i. Judgementally / arbitrarily : The researcher can simply select a sample size
arbitrarily or on the basis of some judgementally based criterion.. There may
be instances where the sample size represents all that where available at a
particular point of time.
ii. Analysis considerations : Analysis considerations may decide the sample size.
Sample size may be determined from the minimum cell size needed.
iii. The budget : In certain cases, the budget may determine the sample size.
iv. Applying standard error : Sample size determination is based on specifying
the desired precision in advance and then applying the appropriate standard
error formula.
Two major classes of procedures are available for estimating the sample size
1. Confidence ? interval approach : This is based on the idea of
constructing confidence intervals around sample means or proportions.
2. Hypothesis-testing approach : This makes use of both type I error
(rejecting a true null hypothesis) and Type II error (accepting a false null
hypothesis).
Confidence ? interval approach
In this method, a confidence interval is constructed around sample based
mean or proportion. The standard error formulae are used for this purpose. This
can be explained with an example. Consider a researcher my have taken a
sample of 100 consumers and noted that their average per-capita consumption of
orange juice was 2.6 litres per week. Pat studies indicate that the population
standard deviation T can be assumed to be 0.3 litre. With this information, we
can find a range around the sample mean level of 2.6 litres for which some pre-
specified probability statement can be made about the process underlying the
construction of such confidence intervals. Suppose that we want to set up a 95%
confidence interval around the sample mean of 2.6 litres. The standard error of
the mean can be computed as :
0.3
x= ----- = ---------- = 0.03
n 100
From the table, we can find that the central 95% of the normal distribution lies
within C 1.96 Z variates.
95% confidence interval ranges from 2.54 to 2.66 litres. Thus the pre- assigned
chance of finding the true population mean to be within 2.54 and 2.66 litres is
95%.
The case of sample mean
Following are the steps involved :
1. Specify the amount of error (E) that can be allowed. This is the maximum
allowable difference between the sample mean and the population mean. 8 C
E defines the interval within which 0 will lie with some pre-specified level
of confidence.
2. Specify the desired level of confidence. It can be 95%.
3. Determine the number of standard errors (Z) associated with the confidence
level.
4. Estimate the standard deviation of the population. The standard deviation can
be estimated by judgment, by reference to other studies or by the use of a
pilot sample.
5. Calculate the sample size using the formula for the standard error of the
mean.
E
x = -------
Z
6. Neglacting the finite multiplier, we solve for n in the formula
E
x = ------- = ------
Z
n
7. In general we can find n from the formula
2 Z2
n = --------------
E2
The case of sample proportion
The procedure for determining sample size for interval estimates of proportion
are :
1. Specify the amount of error that can be allowed. Suppose that the desired
reliability is such that an allowable interval of p ? A = ? 0.05 is set, the
allowable error E is 0.05.
2. Specify the desired level of confidence. Suppose that the level of confidence
is 95%.
3. Determine the number of standard errors (Z) associated with the confidence
level.
4. Estimate the population proportion. (A). The population proportion can be
estimated by judgment, by reference to other studies or by the use of a pilot
sample.
5. Calculate the sample size using the formula for the standard error of the
proportion.
E
p= -----
Z
6. Neglacting the finite multiplier, we solve for n in the formula
E ______________
p = ----- = A ( 1- A ) / n
Z
7. We can find n from the formula
A ( 1- A ) Z2
n = -------------------------------
E2
Hypothesis testing approach
Sample size can also be determined by the hypothesis testing approach. For this,
an assumed probability of making Type I error ( called alpha risk) and the
probability of making Type II error ( called beta risk) are needed. These risks
are based on the hypotheses :
H0 : the null hypothesis
H1 : the alternate hypothesis
In hypothesis testing, the sample results sometimes lead us to reject H0 when its
is true. This is a type I error. On other occasions, the sample findings may lead
us to accept H0 when it is false. This is a Type II error.
The case involving means
The steps are :
1. Specify the values for the null (H0) and the alternate (H1) hypotheses to
be tested in terms of population means, 0 and 1 respectively.
2. Specify the allowable probabilities ( and respectively) of Type I and
Type II errors. The Type I error is the error of rejecting a true null
hypothesis. The Type II error is made if the alternate hypothesis is
rejected when it is true. and are the allowable probabilities of making
those two types of errors respectively.
3. Determine the number of standard errors associated with each of the
error probabilities and .
4. Estimate the population standard deviation .
5. Calculate the sample size that will meet the and error requirements.
Since two sampling distributions are involved, a simultaneous solution of
two equations is required to determine the sample size and critical value
that will satisfy both equations. These equations are :
Critical value = 0 + Z [ ----- ]
n
Critical value = 1 - Z [ ----- ]
n
6. Setting the right hand side of these two equations equal and solving for
n gives
(Z + Z)2 2
n = -----------------
(1 - 0)2
The case involving proportions
The steps are :
1. Specify the values for the null (H0) and the alternate (H1) hypotheses to
be tested in terms of population proportions, 0 and 1 respectively.
2. Specify the allowable probabilities ( and respectively) of Type I and
Type II errors.
3. Determine the number of standard errors associated with each of the
error probabilities Z and Z.
4. Calculate the desired sample size n from the formula :
2
Z [0 (1- 0)] + Z [1 (1- 1)]
n = -------------------------------------------------
1 - 0
2.7.2. Bayesian approach to sample size determination
Bayesian procedures are based on the central principle that one should
select the sample size that results in the largest positive difference between the
expected payoff of sample information and the estimated cost of sampling. The
difference between the expected payoff of the sample information and the
estimated cost of sampling is frequently referred to as the expected net gain
from sampling. An equivalent way of stating the principle is that one should
select the sample size that leads to the largest expected net gain from sampling.
In a decisional situation in which one of the primary objectives is to
maximize payoff, this rule is appropriate. The general approach to applying it
requires the decision maker to :
- Determine the expected value of the sample information for a given
sample size.
- Estimate the sampling cost for that specific option.
- Find the expected net gain from sampling under that option.
- Search through other sample sizes to find the one that leads to the
highest expected net gain from sampling.
While logically sound concept, the Bayesian approach is difficult to implement.
The primary problem comes in operationalising the first of the steps stated
above. In order to determine the expected value of the sample information for a
given sample size, one must relate the sample size being considered to the
conditional probabilities of making errors, including the effects of non-sampling
errors. In real life situations, this may become very difficult to do.
2.8. Sampling error
Several potential sources of error can affect the quality of a research
process. The errors can influence the various stages of the research process and
result in inaccurate or useless research findings. Researchers have numerous
opportunities to make mistakes or errors in judgment, that result in creating
some type of bias in any research study. All types of errors can be logically
classified as either sampling or non-sampling errors. Random sampling errors
can be detected by observing the difference between the sample results and the
results of a census conducted using identical procedures. Two difficulties are
associated with detection of the sampling error :
i. the fact that very seldom is a census conducted in survey research
and
ii. sampling error can be determined only after the sample is drawn
and data collection has been completed.
The Total Error in a research study is the difference between the true mean value
(within the population) of the variable being studied and the observed mean
value obtained through the research study.
To tal Error
Sampling Error
Non-sampling
Error
Design Errors
Administering
Response Errors
Non response
Errors
Error
Sampling error is any type of bias that is attributable to mistakes made in
either the selection process of prospective sampling units or determining the
sample size. Sampling error is the difference between a measure obtained from a
sample representing the population and the true measure that can be obtained
only from the entire population. This error occurs because no sample is a perfect
representation of a given population, unless the sample size equals the
population. Random sampling error tends to occur because of chance variations
in the scientific selection of the needed sample units. Even if the sampling units
were properly selected according to the guidelines of sampling theory, those
units still might not be a perfect representation of the defined target population.
When there is a discrepancy between the statistic calculated or estimated from
the sample and the actual value from the population, a sampling error has
occurred. Based on the principles of the central limit theorem, the size of the
sampling error and its impact can be reduced by increasing the size of the
sample.
Ways to minimize sampling error
1. Increase the sample size.
2. Use a statistically efficient sampling plan. That is making the sample as representative of
the population as possible.
Non-sampling errors are those types of biases that occur in a research
study regardless of whether a sample or census is used. There may be several
reasons for these errors. Non-sampling errors can be broadly classified into four
groups :
ii.
Design errors
iii.
Administering errors
iv.
Response errors
v.
Non response errors.
i. Design errors : Design errors also called researcher- induced errors are
mainly
due to flaws in the research design. These errors can take various
forms.
Selection error : This occurs when a sample obtained through a non
probability sampling method is not representative of the population.
Population specification error : This occurs when an inappropriate
population
is chosen from which to obtain data for the research study.
Sampling frame error : It occurs when the sample is drawn from an
inaccurate sampling frame.
Surrogate information error : It is the difference between the information
required for a research study and the information being sought by the
researcher.
Measurement error : It is the difference between the information sought
by the researcher for a study and the information generated by a
particular measurement method employed by the researcher.
Experimental error : Any error caused by the improper design of the
experiment induce an experimental error into the study.
Data analysis error : This can occur when the data from the questionnaires
are coded, edited, analysed or interpreted.
ii. Administering errors : All errors that occur during the administration of a
survey instrument to the respondents are classified as administering
errors. They are caused by the person administering the questionnaire.
They may be caused by three major factors :
Questioning error : This error arises while addressing questions to the
respondents.
Recording error : This arises from improperly recording the respondent`s
answers.
Inference error : This error occurs when an interviewer interferes with or
fails to follow the exact procedure while collecting data.
iii. Response errors : Also called data errors occur when the respondent
intentionally or unintentionally provides inaccurate answers to the survey
questions.
iv. Non-response errors : These occur if some members of a sample were not
contacted or some of the members contacted provide no response to the
survey.
2.9. Non-response Problems
The object of sampling is to obtain a body of data that are representative
of the population. Unfortunately, some sample members become non-
respondents because they :
i. refuse to respond,
ii. lack the ability to respond,
iii. are not at home, or
iv. are inaccessible.
Non-response can be a serious problem. If a sample size of 1,000 is needed and
only a 50 percent response rate is expected, then 2,000 people will need to be
identified as possible sample members. The seriousness of non-response bias
depends on the extent of the non-response. If the percentage involved is small,
the bias is small. The non-response problem depends on how the non-
respondents differ from the respondents, particularly on the key questions of
interest. To deal with non-response problem tendency will be to replace each
non-respondent with a "matched" member of the sample. The difficulty is that
the replacement cannot be matched easily on the characteristic that prompted the
non-response, such as being employed or being a frequent traveller. Three
approaches are :
i. to improve the research design to reduce the number of non-responses,
ii. to repeat the contact one or more times (call-backs) to try to reduce
non-responses, and
iii. to attempt to estimate the non-response bias.
Improving the research Design
The challenge in personal and telephone interviewing is to gain initial
interest and to generate rapport through interviewer skill and the design and
placement of questions. In mail surveys, the task is to motivate the respondent to
respond, through incentives and other devices. The number of not-at-homes can
be reduced by scheduling calls with some knowledge of the respondents` likely
activity patterns.
Call-Backs
Call-backs refer to new attempts to obtain responses. The use of call-
backs is predicated on the assumption that they will generate a useful number of
additional responses and that the additional responses will reduce meaningfully
a non-response bias. If the non-response is due to refusals or the inability to
respond, call-backs may not reduce significantly the number of non-respondents.
They are most effective for the not-at-home non-respondent. The efficiency of
the call-backs will be improved by scheduling them at different times of the day
and week. In a mail survey, the call-back is particularly important, because the
non-response level can be high. It is common practice to remind non-
respondents at regular intervals.
Estimating the Effects of Non-response
One approach is to make an extra effort to interview a sub-sample of the non-
respondents. In the case of a mail survey, the sub-sample might be interviewed
by telephone. In a telephone or personal survey, an attractive incentive, such as a
worthwhile gift, might be employed to entice a sample of the non-respondents to
co-operate.
Summary
Sampling is an important concept that we practice in our every day life.
Sampling involves selecting a relatively small number of elements from a larger
defined group of elements and expecting that the information gathered from the small
group will allow judgments to be made about the larger group. If all the respondents
in a population are asked to provide information, such survey is called a census.
Information obtained from a subset of the population is known as the statistic (from
sample).
Selection of the most appropriate sampling design should incorporate the
seven factors. They are : Research objectives, Degree of accuracy, Availability of
resources, Time frame, Advanced knowledge of the target population, Scope of the
research and Perceived statistical analysis needs. A sampling plan is the blueprint or
frame work needed to ensure that the raw data collected are representative of the
defined target population.
There are two basic sampling designs : Probability and non-probability
sampling methods. In probability sampling, each unit in the defined target population
has a known, non-zero probability of being selected for the sample, The actual
probability of selection for each sampling unit may or may not be equal depending on
the type of probability sampling design used. It allows the researcher to judge the
reliability and validity of raw data collected by calculating the probability to which
the findings based on the sample would differ from the defined target population.
The results obtained by the probability method can be generalized to the target
population within a specified margin of error through the use of statistical methods.
In non-probability sampling, the probability of selection of each sample unit is not
known. Therefore, potential sampling error cannot be accurately known either. The
selection of sampling units is based on some type of intuitive judgments, desire or
knowledge of the researcher. The degree to which the sample may or may not be
representative of the defined target population depends on the sampling approach and
how well the researcher executes and controls the selection activities. There is always
a temptation to generalize non-probability sample data results to the defined target
population.
Various probability sampling methods are : a. Simple random sampling, b.
Systematic Random Sampling, c. Stratified Random Sampling, d. Cluster Sampling ,
e. Area sampling. Non-probability sampling methods can be classified as : a.
Convenience Sampling, b. Judgment Sampling, c. Quota Sampling, d. Snowball
Sampling.
Random-Digit Dialing involves generation of random numbers at random. If
truly accurate or representative, current sampling frames are not available, a
researcher would have to employ an alternate method such as random-digit dialing, if
conducting telephone interviews.
The traditional approaches in determining the sample size are :
i. Judgementally / arbitrarily, ii. Analysis considerations, iii. The budget, iv.
Applying standard error. Bayesian procedures are based on the central principle that
one should select the sample size that results in the largest positive difference
between the expected payoff of sample information and the estimated cost of
sampling.
Sampling error is the difference between a measure obtained from a sample
representing the population and the true measure that can be obtained only from the
e n t i r e
po pulation. The Total Error in a research study is the difference between the
true mean value (within the population) of the variable being studied and the
observed mean value obtained through the research study. Total error can be
classified into sampling error and non-sampling error.
Some sample members become non-respondents because they : i. refuse to
respond, ii. lack the ability to respond, iii. are not at home, or iv. are inaccessible.
Self Assessment Questions
1. Define sampling.
2. Give the differences between census and sampling.
3. Define the following :
Population
Element
Sampling units
Sampling frame
Sampling gap
Sampling distribution
4. What is Central Limit Theorem ?
5. What are the factors influencing the selection of a sampling design ?
6. Define sampling plan. What are the steps in developing a sampling plan ?
7. Differentiate between probability and non-probability sampling methods.
8. What are the types of probability sampling methods ?
9. Define the following :
Simple random sampling
Systematic Random Sampling
Stratified Random Sampling
Cluster Sampling
Area sampling
10. What are the types of non-probability sampling methods ?
11. Define the following :
Convenience Sampling
Judgment Sampling
Quota Sampling
Snowball Sampling
12. What is Random-Digit Dialing (RDD) ?
13. What are the factors in determining appropriate sample sizes ?
14. What are the traditional approaches in determining the sample size ?
15. What is Bayesian approach to sample size determination ?
16. What is sampling error ? What are the types of total error ?
17. What are the types of non-sampling errors ?
18. What are the reasons for non-response of respondents ?
Answer Key
1. Sampling involves selecting a relatively small number of elements from a larger
defined group of elements and expecting that the information gathered from the small
group will allow judgments to be made about the larger group.
2. If all the respondents in a population are asked to provide information, such survey is
called a census. A subset of the population is known as the sample.
3. a. Population : A population is an identifiable total group or aggregation of elements
that are of interest to the researcher and pertinent to the specified problem.
b. Element : An element is a person or object from which data and information are
sought.
c. Sampling units : Sampling units are the target population elements available for
selection during the sampling process.
d. Sampling frame : Sampling frame is the list of all eligible sampling units.
e. Sampling gap : A sampling gap is the representation difference between the
population elements and sampling units in the sampling frame.
f. Sampling distribution : Sampling distribution is the frequency distribution of a
specific sample statistic ( sample mean or sample proportion ) from repeated random
samples of the same size.
4. The theorem states that for almost all defined target populations, the sampling
distribution of the mean or the percentage value derived from a simple random sample
will be approximately normally distributed, provided that the sample size is sufficiently
large.
5. Research objectives, Degree of accuracy, Availability of resources, Time frame,
Advanced knowledge of the target population, Scope of the research and Perceived
statistical analysis needs.
6. A sampling plan is the blueprint or frame work needed to ensure that the raw data
collected are representative of the defined target population. A good sampling plan will
include, the following steps: (1) define the target population, (2) select the data
collection method, (3) identify the sampling frames needed, (4) select the appropriate
sampling method, (5) determine necessary sample sizes and overall contact rates, (6)
create an operating plan for selecting sampling units, and (7) execute the operational
plan.
7. ________________________________________________________________
Factor
Probability sampling
Non-probability
sampling
_____________________________________________________________________________
a. List of the population elements Complete list necessary
Not necessary
b. Information about the-
Each unit identified
need detail on habits,
sampling units
activities, traits etc.
c. Sampling skill
skill required
little skill required
d. Time requirement
More time-consuming
Less time consuming
e. Cost per unit sampled
Moderate to high
Low
f. Estimate of population -
Unbiased
Biased
parameters
g. Sample representativeness
Assured
Undeterminable
h.
A c
c u
r a
c y a
nd
Computed with
Unknown
Reliability
confidence intervals
i. Measurement of sampling error Statistical measures
No true measure
available
8.
a. Simple random sampling
b. Systematic Random Sampling
c. Stratified Random Sampling
d. Cluster Sampling
e. Area sampling.
9.
a. Simple random sampling : Simple Random Sampling is a probability sampling
procedure which ensures that every sampling unit making up the defined target
population has a known, equal, non-zero chance of being selected.
b. Systematic Random Sampling : Systematic random sampling (SYMRS) is similar to
simple random sampling but requires that the defined target population be ordered in
some way, usually in the form of a customer list, taxpayer roll, or membership roster.
c. Stratified Random Sampling : Stratified random sampling (STRS) requires the
separation of the defined target population into different groups, called strata, and the
selecting of samples from each stratum.
d. Cluster Sampling : The sampling units are divided into mutually exclusive and
collectively exhaustive sub-populations, called clusters. Each cluster is assumed to be
representative of the heterogeneity of the target population.
e. Area sampling :The clusters are formed by geographic designations. Examples
include cities, sub divisions and blocks.
10.
a. Convenience Sampling
b. Judgment Sampling
c. Quota Sampling
d. Snowball Sampling
11.
a. Convenience Sampling : Convenience sampling (or accidental sampling) is a method
in which samples are drawn at the convenience of the researcher or interviewer.
b. Judgment Sampling : In judgment sampling, (also referred to as purposive sampling),
participants are selected according to an experienced individual's belief that they will
meet the requirements of the study.
c. Quota Sampling : The quota sampling method involves the selection of prospective
participants according to pre-specified quota regarding either demographic
characteristics, specific attitudes, or specific behaviours.
d. Snowball Sampling : Snowball Sampling involves the practice of identifying and
qualifying a set of initial prospective respondents who can, in turn, help the researcher
identify additional people to be included in the study.
12. Random-Digit Dialing involves generation of random numbers at random. If truly
accurate or representative, current sampling frames are not available, a researcher
would have to employ an alternate method such as random-digit dialing, if conducting
telephone interviews.
13.
i. The variability of the population characteristic under consideration.
ii. The level of confidence desired in the estimate.
iii. The degree of precision desired in estimating the population characteristic.
14.
i. Judgementally / arbitrarily.
ii. Analysis considerations.
iii. The budget.
iv. Applying standard error.
15. Bayesian procedures are based on the central principle that one should select the
sample size that results in the largest positive difference between the expected payoff of
sample information and the estimated cost of sampling.
16. Sampling error is the difference between a measure obtained from a sample
representing the population and the true measure that can be obtained only from the
entire population. The Total Error in a research study is the difference between the true
mean value (within the population) of the variable being studied and the observed mean
value obtained through the research study. Total error can be classified into sampling
error and non-sampling error.
17.Non-sampling errors can be broadly classified into four groups :
i. Design errors
ii. Administering errors
iii. Response errors
iv. Non response errors
18. Some sample members become non-respondents because they :
i. refuse to respond,
ii. lack the ability to respond,
iii. are not at home, or
iv. are inaccessible.
Glossary
Area sampling :The clusters are formed by geographic designations. Examples include cities,
sub divisions and blocks.
Cluster Sampling : The sampling units are divided into mutually exclusive and collectively
exhaustive sub-populations, called clusters. Each cluster is assumed to be representative of the
heterogeneity of the target population.
Convenience Sampling : Convenience sampling (or accidental sampling) is a method in which
samples are drawn at the convenience of the researcher or interviewer.
Element : An element is a person or object from which data and information are sought.
Judgment Sampling : In judgment sampling, (also referred to as purposive sampling),
participants are selected according to an experienced individual's belief that they will meet the
requirements of the study.
Population : A population is an identifiable total group or aggregation of elements that are of
interest to the researcher and pertinent to the specified problem.
Quota Sampling : The quota sampling method involves the selection of prospective participants
according to pre-specified quota regarding either demographic characteristics, specific attitudes,
or specific behaviours.
Random-Digit Dialing : Random-Digit Dialing involves generation of random numbers at
random. If truly accurate or representative, current sampling frames are not available, a
researcher would have to employ an alternate method such as random-digit dialing, if conducting
telephone interviews.
Sampling distribution : Sampling distribution is the frequency distribution of a specific sample
statistic from repeated random samples of the same size.
Sampling frame : Sampling frame is the list of all eligible sampling units.
Sampling gap : A sampling gap is the representation difference between the population elements
and sampling units in the sampling frame.
Sampling units : Sampling units are the target population elements available for selection during
the sampling process.
Simple random sampling : Simple Random Sampling is a probability sampling procedure which
ensures that every sampling unit making up the defined target population has a known, equal,
non-zero chance of being selected.
Snowball Sampling : Snowball Sampling involves the practice of identifying and qualifying a
set of initial prospective respondents who can, in turn, help the researcher identify additional
people to be included in the study.
Stratified Random Sampling : Stratified random sampling (STRS) requires the separation of the
defined target population into different groups, called strata, and the selecting of samples from
each stratum.
Systematic Random Sampling : Systematic random sampling (SYMRS) is similar to simple
random sampling but requires that the defined target population be ordered in some way, usually
in the form of a customer list, taxpayer roll, or membership roster.
________________________________________________________________
_______
Reference
1. Paul. E. Green, Donald. S. Tull, Gerald Albaum, Research for Marketing Decisions,
Prentice Hall of India Pvt. Ltd. New Delhi.
2. Joseph. F. Hair, Robert. P. Bush, David. J. Ortinan, Marketing Research, Tata
McGraw-Hill Publishing Company Ltd, New Delhi.
3. David. A. Aaker, V. Kumar, George.S.Day, Marketing Research, John Wiley & Sons
Inc, Singapore.
4. William. M.K. Trochim, Research Methods, Biztantra, New Delhi.
Unit III
THE MEASUREMENT
3.1 Introduction
3.2 Definition of measurement
3.3 Functions of measurement
3.4 Advantages and disadvantages of measurement
3.5 Uses of measurement in research
3.6 Fundamentals of measurement
3.7 Types of measurement
Learning objectives
To understand the concept of measurement
To learn about the measurement process and how to develop a good
measurement scale
To understand the four levels of scales and their typical usage
To explore the concepts of reliability and validity
The business research revolves around measurement, which refers to obtaining
symbols to represent properties of objects, events or states. Measurement helps
in identifying attitudes of individuals .When we measure an object, essentially
all we are doing is counting the number of standard pieces it takes to be the
same size as the object.
Measurement is defined as:
The determination of size in relation to some observed standard, e.g. metre,
kilogram, second, ampere, degree Kelvin, candela, mole, or some unit derived
from these seven basic units.
According to the famous British scientist, Lord Kelvin, "When you measure
what you are speaking about and express it in numbers, you know something
about it, but when you cannot (or do not) measure it, when you cannot (or do
not) express it in numbers, then your knowledge is of a meager and
unsatisfactory kind."
Measurement is the process observing and recording the observations.
Measurement is the estimation or determination of extent, dimension or
capacity, usually in relation to some standard or unit of measurement. The
measurement is expressed as a number of units of the standard, such as distance
being indicated by a number of kilometers.
The process of measuring involves estimating the ratio of the magnitude of a
quantity to the magnitude of a unit of the same type length, time etc. A
measurement is the result of such a process, expressed as the product of a real
number and a unit, where the real number is the estimated ratio. Example is 100
meters, which is an expression of an object`s length relative to a unit of length,
the meter etc.,
The fundamental ideas involved in measuring are nominal, ordinal, interval and
ratio. We consider four broad categories of measurements. Survey research
includes the design and implementation of interviews and questionnaires.
Scaling involves consideration of the major methods of developing and
implementing a scale. Qualitative research provides an overview of the broad
range of non-numerical measurement approaches.
Measurement is the assignment of numbers to objects in such a way that
physical relationships and operations among the objects correspond to arithmetic
relationships and operations among the numbers.
According to Dr D.D. Sharma, Measurement is concerned with correspondence
between empirical entities and a formal model of abstract elements or numbers
,the relationship among these elements and the operations which can be
performed on them. Such a rule of correspondence determines a scale.
According to S.S.Stevens, Measurement is the assignment of numerals to
objects or events according to rules.
According to Campbell,`` Measurement is the assignment of numbers to
represent properties.
CONCEPT OF MEASUREMENT
Measurement in quantitative research
Views regarding the role of measurement in quantitative research are somewhat
divergent. Measurement is often regarded as being only a means by which
observations are expressed numerically in order to investigate causal relations or
associations. However, it has been argued that measurement often plays a more
important role in quantitative research. For example, Thomas Kuhn (1961)
argued that results that appear anomalous in the context of accepted theory
potentially lead to the genesis of a search for a new, natural phenomenon. He
believed that such anomalies are most striking when encountered during the
process of obtaining measurements, as reflected in the following observations
regarding the function of measurement in science:
When measurement departs from theory, it is likely to yield mere numbers, and
their very neutrality makes them particularly sterile as a source of remedial
suggestions. However, numbers register the departure from theory with an
authority and finesse that no qualitative technique can duplicate, and that
departure is often enough to start a search (Kuhn, 1961, p. 180).
In classical physics, the theory and definitions, which underpin measurement,
are generally deterministic in nature. In contrast, probabilistic measurement
models known as the Rasch model and Item response theory models are
generally employed in the social sciences. Psychometrics is the field of study
concerned with the theory and technique for measuring social and psychological
attributes and phenomena. This field is central to much quantitative research that
is undertaken within the social sciences.
Measurement may be defined as the assignment of numerals to
characteristics of objects , persons, states, or events, according to rules.
What is measured is not the objects, person, state, or event itself but
some characteristic of it. When objects are counted, for example, we do
not measure the object itself but only its characteristic of being present.
We never measure people, only their age, height, weight, or some other
characteristic.
LEVELS OF MEASUREMENT
There are four levels of measurement: nominal, ordinal, interval and ratio.
We describe below the characteristics of each of these levels.
(1) Nominal measurement: Nominal measurement is the most
elementary method of measurement which classifies persons,
objects or events into a number of mutually exclusive
categories on the basis of the simple presence or absence,
applicability or inapplicability, possession or non-possession
of certain property. This is have` us. have nots`
measurement which assigns mutually exclusive labels to
identify objects. Thus, the population of a town may be
classified according to gender into males` and females`
or according to religion into Hindus, Jains, Parsis, Muslims,
Sikhs and Christians and each category of persons given
certain labels either in the form of numerals (0,1,2,3) or in
the form of letters (A,B,C,D).
These labels only tell us that the categories are
qualitatively different from each other. They have no
quantitative significance, i.e., they cannot be added,
subtracted, multiplied or divided. Once can, if one likes,
interchange the labels of various categories for they do not
signify any ranking or ordering of categories . The numeral
1 given to a certain category does not imply its superior
position to other category which is given numeral 0.
The only arithmetic operation possible in case of a nominal
measurement is counting. Thus, mode is the only legitimate
measure of central tendency. One can also calculate the
percentage of objects failing within each category. But, it
will not make sense to calculate the arithmetic mean of
gender in a sample consisting of 45 men and 55 women.
All we can say is that there are more females than males
in the sample or that 45% of the sample is male.
(2) Ordinal measurement. In ordinal measurement numerals,
letters or other symbols are used to rank objects. This is
essentially an advances form of categorization. Objects in this
measurement are classified not only as to whether they
share some characteristic with another object but also
whether they have more or less of this characteristics than
some other object on some characteristic. A significant
amount of marketing research, particularly consumer-
oriented research, relies on this type of data. Their most
common use is in obtaining preferences measurements. For
example, a consumer or a sample of experts may be asked
to rank preference for several brands, flavors, or package
designs. Attitude measures are also often ordinal in nature.
However, ordinal measurements do not provide information
on how much more or less of the characteristic various
objects possess. For example, if in respect of a certain
characteristic two objects have the ranks 5 and 8 and two
other objects the ranks 3 and 6, we cannot say that the
differences between the two pairs are equal. There is also
no way to know that any object has none of the
characteristic being measured.
The king of descriptive statistics that can be calculated
from these data are mode, median and percentages. For
example, in the following table which shows the quality
ratings given by 600 housewives to one brand of coffee
one can usefully calculate the median and modal quality
ratings. Both are 2` in this case. One can also calculate the
percentages of the total appearing in each rank. But it is
meaningless to calculate a mean because the differences
between ordinal scales values are not necessarily, the same.
Quality Rating
No. of respondents giving
rating
1
200
2
200
3
100
4
50
5
50
(2) Interval measurement. Interval measurements represent
numerals to rank objects such that numerically equal
distances on the scale represent equal distances in the
property being measured. The distances between numerals
are thus, very meaningful because, by comparing these
distances we can know how far apart the objects are with
respect to the property in question. For example, if we are
measuring the achievements of 4 students W, X, Y, and Z on
an interval scale and obtain the values 1, 4, 5 and 8
respectively ( as Shown below) then by comparing the
intervals, we can legitimately say that the difference
between W and Y in their achievements is the same as the
difference between X and Z and That the difference
between X and Z is four times the difference between W
and Y.
W X Y Z
0 1 2 3 4 5 6 7 8 9 10
One very important drawback of this measurement, however, is that We
cannot compare objects on the basis of ratios of their absolute scores.
Thus, in our example, we cannot say that the achievements of X is twice
as great as that of Z. The reason is that in our measurement the zero
point is arbitrary so that any change in this point will change the
absolute scores and the ratios between scores. For example, if the
arbitrary zero point of the above scale is shifted four scale-points
downward, the position of X and Z on the scale will change to 8 and 12
and the previous ration of 1: 2 will now become 2:3
The most common examples of interval scales are the centigrade and
Fahrenheit temperature scales which start with different points of origin.
The point of origin (for the same natural phenomenon, the freezing point
of water) is zero on the Centigrade scale and 32 on Fahrenheit. Because
of this difference the ratio between any two readings on the Centigrade
scale (e.g., between 10 to 30) is not the same as that on the Fahrenheit
scale as shown below. One can only talk of a 20. C rise in temperature
but not of 30. C as being three times as hot as 10.C.
Centigrade 0 10 30 100
Fahrenheit 32 53 86 212
The most frequent of interval measurement in social sciences is index
numbers which are calculated on the basis of an arbitrary zero point.
Another common form of interval measurement is a likert scale which is
used in the measurement of attitudes and personality.
Virtually the entire range of statistical analysis can be applied to data
measured on interval scales. One can use such descriptive measures as
the mean, median, mode, range and standard deviation. Bivariate
correlation analysis, t-test, analysis of variance test, and most multivariate
techniques applied for purposes of drawing inferences can also be used.
(4) Ratio measurement. This measurement, besides possessing the property
of the interval measurement, possesses one additional property , viz., it
has a true, natural or absolute zero point, one for which there is
universal agreement as to its location A true zero means that the object
measuring zero possesses none of the property in question. Height and
weight are obvious example. In general, wherever the objects, are being
counted` for possessing or not possessing a certain characteristic ratio
measurement is being done. In all such cases the number 0 has a true
meaning ? the object possesses none of the characteristic being measured.
Other common examples of this type of measurement are sales, costs,
number of purchasers, length, time, etc.
With a ratio measurement, the comparison between ratios of the absolute
magnitude of the numbers becomes possible. Thus, a person weighing
100kg . is said to be twice as heavy as one weighing 50kg. and a person
weighing 150kg is three times as heavy. Further, with a ratio scale we can
compare intervals, rank objects according to magnitude, or use the
numbers to identify the objects. All descriptive measures and inferential
techniques are applicable to ratio-measured data.
The following table summaries the above discussion about four types of
measurement;
Types Of Measurement
Type
Basic empirical
Typical
Typical Statistics
operation
usage
Descriptive Inferential
1. Nominal
Determination
Classification
Percentage chi-square,
Of
equality
(0, Male-female
Binomial test
1,2,..........9)
purchaser
non-
purchaser, Team
A Team-B
2. Ordinal
Determination
Rankings:
Median Rank-order
Of
greater
or
less preference data, correletion
(o<1<2...........<9)
market position,
attitude measures,
many
psychological
measures
3. Interval Determination of equality Index
numbers, Mean, Range, Product-moment
of intervals (2-1=7-6)
attitude measures Standard
Determination of equality Sales,
units Deviation
4. Ratio
o f r
a t ios (2/4 = 4/8)
produced, number Geometric Coefficient of
of
customers. variation
Costs, age
mean
COMPONENTS OF MEASUREMENT
Ideally speaking there should be only one component of a measurement
and this component should be a direct reflection of the characteristic being
measured. Unfortunately, such a situation seldom exists. Very often, a
measurement consists of not one but two components- one representing the
influence of the characteristic being measured and the other representing
the influence of those characteristics which the researcher is not interested
in measuring but which still creep in against his wishes. The
characteristics are as follows:
1.
Additional stable characteristics of the object or event, for
example, the respondent`s tendency to give only favorable
responses independent of his true feelings.
2.
Short-term characteristics of the object, for example, fatigue,
health, hunger, and emotional state.
3.
Situational characteristics, for example, the presence or absences
or some other person or of location under which the
measurement is taken
4.
Characteristics of the measurement process, for example, Gender,
age, and ethnic background. In addition, style of dress of the
interviewer or the method of interviewing-telephone, mail,
personal interview, etc.
5.
Characteristics of the measuring instruments, for example,
unclear instructions, ambiguous questions, confusing terms,
omitted questions, etc.
6.
Characteristics of the response process, for example, mistakes
caused by checking a wrong responses
7.
Characteristics of the analysis, for example, mistakes caused by
wrong coding, tabulating, etc.
Six out of these seven characteristics (2 to 7) give rise to variable errors
in a measurement, i.e., errors which occur randomly each time something
is measured. First characteristics gives rise to a systematic error, i.e., error
that occurs in a consistent manner each time something is measured. It is
also caleed bias. Thus, the general situation is:
M=C+VE+SE where M stands for the measurement , C stands for the
characteristic being measured, VE stands, for variable errors and SE
stands for systematic errors.
ACCURACY MEASUREMENT
Accuracy of measurement depends upon the extent to which it is free
from systematic and variable errors. Freedom from variable errors is
known as the reliability of a measurement and freedom from systematic
errors is known as the validity of a measurement. Reliability and Validity
every researcher to know how they are measured.
The Measurement Process
The following are the steps involved in the measurement process
1. Identifying the concept of interest
2. Development of a construct
3. A constitutive definition
4. Operational definition
5. Measurement scales
6. Evaluating reliability and validity of the scale
7. utilization of scale
8. Research findings
We shall discuss the four numerical scales such as Nominal Scale, Ordinal scale,
interval scale and ratio scale, functions of measurement and measurement error,
benefits of measurement as described below.
Nominal Level of Measurement
a. Nominal Scales
i. Partitions data into categories that is mutually exclusive
and collectively exhaustive. Examples
ii. Gender:
(1) Male (2) Female
iii. Geographic Area:
(1) Urban (2) Rural (3) Suburban(4) Semi-urban
Ordinal Scales
b. Used strictly to indicate rank order.
Ranking
Please rank the following brands of Colour Television from 1-5 with 1 being the
most and 5 the least preferred
Samsung________ LG_______
Thomson ________
BPL _______
Videocon ________
? Interval Scales
? Contains all the features of ordinal scales
? Added dimension that the intervals between the data
points on the scale are equal.
? Ratio Scales
? All powers of those mentioned as well as a meaningful
absolute zero or origin.
? Accurate Data Imply Accurate Measurement
M = A + E
Where M = measurement
A = accuracy
E = errors-random or systematic
Nominal Scale
Nominal Scale is defined as measurement of a variable which results in the
classification of phenomena into a set of consistent and non-overlapping
attributes (yes, no, male, female, etc.). Nonparametric statistics: Tests of
significance that require few assumptions about the population. Use of these
statistics should occur when samples are small (fewer than 30 subjects), when
subjects were not randomly sampled, or data are not interval level. ...
A nominal scale allows for the classification and labeling of elements or
objects into mutually exclusive and exhaustive categories based on defined
features.
A scale of measurement with two or more categories that have no numerical
(less than, greater than) properties. Proportion of map distance to ground
distance at a point or line
Nominal Data
Classification data, e.g. male/ female
No ordering, e.g. it makes no sense to state that
Female >Male
Arbitrary labels, e.g., male/female, 0/1, etc
Nominal Scale. We can examine if a nominal scale data is equal to some
particular value or to count the number of occurrences of each value. For
example, gender is a nominal scale variable. We could examine if the gender
of a person is Female or to count the number of males in a sample.
At the first level of measurement, numbers are used to classify data. In fact,
words or letters would be equally appropriate. For example when we wanted
to classify a cricket team into left footed and right footed players, we
could put all the left footed players into a group classified as 1 and all the
right footed players into a group as 2.
The only mathematical or statistical operation that can be performed on
nominal scales is a frequency run or count. We cannot determine an average,
except for the mode ? that number which holds the most responses -, nor can
we add and subtract numbers.
The purpose of this set of notes is to briefly summarize several aspects of
scales of measurement including:
(a) the measurement principle involved for each scale,
(b) examples of the measurement scales,
(c) permissible arithmetic operations for each scale, and
(d) examples of statistics that are appropriate for each scale.
People or objects with the same scale value are the same on some attribute.
The values of the scale have no 'numeric' meaning in the way that you
usually think about numbers. People or objects with a higher scale value
have more of some attribute.
The intervals between adjacent scale values are indeterminate. Scale
assignment is by the property of "greater than," "equal to," or "less than."
Intervals between adjacent scale values are equal with respect the the
attribute being measured.
E.g., the difference between 8 and 9 is the same as the difference between 76
and 77.
There is a rationale zero point for the scale.
Ratios are equivalent, e.g., the ratio of 2 to 1 is the same as the ratio of 8 to
4.
Each "higher" level of measurement includes the measurement principle of
the "lower" level of measurement. For example, the numbers 8 and 9 in an
interval scale indicate that the object assigned a 9 has more of the attribute
being measured than does the object assigned an 8 (ordinal property) and
that all persons assigned a 9 have equivalent amounts of the attribute being
measured (nominal property). This also implies that you can do lower level
statistics on higher-level measurement scales.
Guttmann Scales
For thinking about numeric types, measurement, truth, social structures
.Fiske argues that all social life is composed of patterns of interaction that
are based on four types of scale: categorical, ordinal, interval, and ratio. The
scales provide ways of perceiving, and thereby of organizing social
interaction. The four scales are not mere manifestations of a single culture,
but are different primary mathematical structures. They are different
axiomatically. They are Trans cultural. In 1944, Louis Guttmann pointed out
that all forms of measurement belong to one of four types of scale:
categorical, ordinal, interval, and ratio.
Why is Level of Measurement Important?
First, knowing the level of measurement helps us to decide how to interpret the
data from that variable. When we know that a measure is nominal then we know
that the numerical values are just short codes for the longer names. Second,
knowing the level of measurement helps us to decide what statistical analysis is
appropriate on the values that were assigned.
In nominal measurement, the numerical values just "name" the attribute
uniquely. No ordering of the cases is implied. For example, the numerical
numbers in football team l are measures at the nominal level. A player with
number 7 is not more of anything than a player with number 5.
In ordinal measurement, the attributes can be rank-ordered. Here, distances
between attributes do not have any meaning. For example, on a survey you
might code Educational Qualification as 0=less than Higher .Secondary
1=Higher .Secondary. 2=Under Graduate degree; 3=Post Graduate degree;
4=professional degree; 5=doctoral degree . In this measure, higher numbers
mean more education. But is distance from 0 to 1 same as 3 to 4? Of course not.
The interval between values is not interpretable in an ordinal measure.
In interval measurement the distance between attributes does have meaning. For
example, when we measure temperature (in Fahrenheit), the gap between 30 and
40 is same as gap between 70 and 80. The interval between values is
interpretable. Because of this, it makes sense to compute an average of an
interval variable, where it doesn't make sense to do so for ordinal scales. But we
should note that in interval measurement ratios don't make any sense - 80
degrees is not twice as hot as 40 degrees (although the attribute value is twice as
large).
Finally, in ratio measurement there is always an absolute zero that is
meaningful. This means that we could construct a meaningful fraction (or ratio)
with a ratio variable. Weight is a ratio variable. In applied business research,
most "count" variables are ratio, for example, the number of clients in past six
months. Because we could have zero clients and because it is meaningful to
say that "...we had twice as many clients in the past six months as we did in the
previous six months."
The categories of the variable:
have
an
inherent are
numbers are numbers that
are
order
with
have
a
Level
names
from more to less equal intervals theoretical
or higher to lower
between them
zero point
Nominal X
level
Ordinal X
X
level
Interval X
X
X
level
Ratio
X
X
X
X
level
Source: Prof Candance Clark, social statistics
It's important to recognize that there is a hierarchy implied in the level of
measurement idea. At lower levels of measurement, assumptions tend to be less
restrictive and data analyses tend to be less sensitive. At each level up the
hierarchy, the current level includes all of the qualities of the one below it and
adds something new. In general, it is desirable to have a higher level of
measurement (e.g., interval or ratio) rather than a lower one (nominal or
ordinal).
Functions of Measurement
? Standardization
? Sharpen arguments and distinctions
? Enhance the precision of statements
? Allow application of mathematics
Causality
Ultimately, we seek to establish causal relationships among the phenomena we
study. Necessary condition for causality: one that must be present for an event to
occur, although its presence does not guarantee occurrence; its absence
guarantees nonoccurrence.
Sufficient condition for causality: one that guarantees that the event will occur
whenever it is present; but the event may still occur in its absence .
Ideal outcome of a scientific study - a set of conditions that are simultaneously
necessary and sufficient for an event that would truly explain the event .The
classic scientific technique for establishing causality is the controlled
experiment. Social experiments rarely achieve the same kind of controlled
conditions that are obtainable in a laboratory.
The cause must precede the effect in time.
The two variables must be empirically correlated. Otherwise stated as:
association is a necessary condition for causality. The correlation between the
two variables cannot be explained by the existence of a third variable. We could
find out the cause and effect relationships. Correlation does not equal causality.
Causal attributions always involve an interpretation that is relative to the
categories and understandings of mechanisms available when the study is
written. Causal inferences are based on assumptions. Human actions and
interactions are rarely so simple that a single source of causality can be
identified. Social science theories do help to provide possible explanations of
tendencies or actions of groups with common characteristics. Yet there is always
a possibility of alternative inferences. Even assuming that some common set of
explanations could be identified among selected subjects, are they representative
of the whole population? The standard of acceptability for any experimental or
observational evidence is: can someone else replicate the process such that the
same results are produced.
Survey research is one of the most important areas of measurement in applied
business research. The broad area of survey research encompasses any
measurement procedures that involve asking questions of respondents. A
"survey" can be anything from a short paper-and-pencil feedback form to an
intensive one-on-one in-depth interview.
Summary
In the early 1940`s, the Harvard psychologist S.S. Stevens coined the terms
nominal, ordinal, interval, and ratio to describe a hierarchy of measurement
scales used in psychophysics, and classified statistical procedures according to
the scales for which they were permissible. Measurement may be defined as
the assignment of numerals to characteristics of objects , persons, states, or
events, according to rules. What is measured is not the objects, person,
state, or event itself but some characteristic of it. When objects are
counted, for example, we do not measure the object itself but only its
characteristic of being present. We never measure people, only their age,
height, weight, or some other characteristic. There are four levels of
measurement: nominal, ordinal, interval and ratio.
KEY TERMS
Measurement
Nominal scale
Ordinal scale
Interval scale
Ratio scale
QUESTIONS FOR REVIEW
1. Define: Measurement
2. What are the steps involved in the Measurement Process.
Define the following: Nominal scale, Ordinal scale, Interval scale and Ratio
scale
3. Indicate whether each of the following measures uses a nominal scale,
ordinal scale, Interval scale and ratio scale.
a. Prices of commodities
b. Married and unmarried
c. Employed and unemployed
d. Professorial rank Professor, Associate Professor, Assistant professor.
4. What are the characteristics of Measurement?
5. What do you mean by the term Measurement error?
Further Reference
1. William G.Zigmand, 2003, Business Research Methods, Thompson Asia pvt
ltd.,
2.D.D.Sharma , Marketing research,1999,SultanChand&Sons
3. Cooper, DonaldR, Schindler, Pamela.s 2006 Businees Research Methods
4.P.C.Tripathi,2005.A
Text
Book
of
Research
Methodology,
SultanChand&Sons.
5.O.R.Krishnaswami,M.Ranganatham,2005.Methodology of Research in Social
Sciences,Himalaya Publishing House,Mumbai.
Unit-IV
DATA INSTRUMENTS
LEARNING OBJECTIVES
After reading this chapter , you should be able to understand
Features of a research instrument.
Types of research instruments.
Selection of research instrument for the study.
INTRODUCTION
In today's unpredictable business environment, customer behavior and attitudes
are changing, the economy is shifting, stakeholders are demanding, and the
competition is getting fierce. Business firms can capitalize these changes by
conducting marketing research. And good marketing research leads to great
ideas that build business results. The foundation of good research is good data.
Traditionally, the term Marketing Research evokes images of huge volumes of
data collected by armies of interviewers. And the same images always involve
data relating exclusively to responses to marketing initiatives, advertising,
pricing, new product launches etc., Perhaps there might be some awareness of
data on Customer Satisfaction and Retention.
There are very few hard and fast rules to define the task of data collection. Each
research project uses a data collection technique appropriate to the particular
research methodology. The two primary goals for both quantitative and
qualitative studies are to boost response and maximize accuracy. To adequately
address the research questions a variety of data collection methods and
instruments will be used. These methods and instruments are not specific to any
one question but provide data that when used in combinations will address the
research questions. In general researcher can opt for any instrument or
combinations of various instruments for the study chosen.
i) TYPE OF INSTRUMENTS
Data instruments are the instruments employed by the researcher to collect the
required data. Data collection instruments are designed to collect standard
information from a large number of study participants. After deciding which
type of information is needed, the methods of accessing data must be
determined. There are different methods of collecting data. The actual design of
the instrument, the data collection form that is used to ask and record the
information is critical to the success of the research. Based on the objectives of
the study researcher can design the instruments or forms to gather the data. The
questionnaire is an example of research instrument, in an observation based
study, the instrument can also be a mechanical or electronic device of
observation like Tape recorder and camera. Survey methods may include mailed
questionnaires, telephone interviews, or face-to-face interviews. Information
gathered may be qualitative as well as quantitative. In experimental types of
research some times an observation device is used, some times a questionnaire is
used.
DATA COLLECTION INSTRUMENTS
a. Questionnaires
b. Interview schedule
c. Observational / Scanning tools
d. Archival Data Collection Tools
e. Interviews / Focus Group Protocols
f. Experimentation
g. Multi-Modal Data Collection
a. Questionnaires :
Questionnaire is a common instrument for collecting data beyond the physical
reach of the researcher, that is, from a large size sample of people. It is an
impersonal instrument for collecting information and must, therefore, contain
clear questions, worded as simply as possible to avoid any confusion or
ambiguity since the researcher probably will not be present to explain what was
meant by any one particular question. Questionnaires are an inexpensive way to
gather data from a potentially large number of respondents. Often they are the
only feasible way to reach a number of reviewers large enough to allow
statistically analysis of the results. A well-designed questionnaire that is used
effectively can gather information on both the overall performance of the test
system as well as information on specific components of the system. If the
questionnaire includes demographic questions on the participants, they can be
used to correlate performance and satisfaction with the test system among
different groups of users.
It is important to remember that a questionnaire should be viewed as a multi-
stage process beginning with definition of the aspects to be examined and
ending with interpretation of the results. Every step needs to be designed
carefully because the final results are only as good as the weakest link in the
questionnaire process. Although questionnaires may be cheap to administer
compared to other data collection methods, they are every bit as expensive in
terms of design time and interpretation.
b. Interview schedule:
Interview schedule is a set of questions prepared by the researcher to conduct
the survey through personal contact. The research worker or enumerator fills the
schedule. With respect to primary research, the foremost tool is the personal
interview. The face-to-face contact between researcher and respondent is not
equal in terms of the potential quality of data that can be obtained. In the face-
to-face interview it is possible to record more than the verbal responses of the
interviewee, which are often superficial. When human beings communicate
directly with each other much more information is communicated between them.
When two people face one another, the dialogue is conducted on several levels.
It goes beyond verbal expression. The nature of words used, facial expressions
and body language all communicate what the other party means.
c. Observational / Scanning tools:
Tools used to guide visual inspection and documentation of characteristics of
people or settings of interest. The scanning process can include recording
observations in writing or using checklists, videotaping, or other methods.
Scanning processes may be supplemented by interviews, surveys, and other
techniques.
d. Archival Data Collection Tools :
Forms or other procedures designed to collect or abstract data from existing
records for research or evaluation purposes. Sometimes called indicator` data,
examples of archival data sources include program files and records from
student life programs, academic administration, social services agencies, police,
courts, or other community-based agencies.
e. Interviews / Focus Group Protocols:
A set of questions and procedures designed to guide one-on-one or group
interviews. Procedures fall on a continuum from unstructured, in which the
questions are used as a general guide, to very structured, in which the questions
are asked word-for-word in a set order.
f. Experimentation:
The popularity of experimentation in marketing research has much to do with
the possibilities of establishing cause and effect. Experiments can be configured
in such a way as to allow the variable causing a particular effect to be isolated.
Other methods commonly used in marketing research, like surveys, provide
much more ambiguous findings. In fact, experimentation is the most scientific
method employed in marketing research
g. Multi-Modal Data Collection
Many clients require more than one method of data collection to optimally
execute their research. An example of this is a very complex survey that would
be cost prohibitive on the telephone, but would be easy to complete on paper.
Reaching the respondent by different roots certainly increases the response rate.
Construction of a good data instrument
In general researcher designs the instrument catering to the nature of the study
and other environment. The following points may be considered for constructing
a good instrument
Reliability of the instrument .
Relevancy
Brevity
Nonambiguity
Specificity
Test the validity of the data.
unbiased reply.
Cost effective.
Possibility of pre-test .
ii) STEPS INVOLVED IN DESIGNING DATA INSTRUMENT
Data collection tools plays vital role in research. They serve two very important
purposes. Firstly it helps the researcher to identify varies elements required to
address for achieving the objectives of the study. Secondly it roots for how data
should be collected. While preparing the instrument due importance must be
given to the environment. Companies desire and background of the respondent
are to be considered for improving the effectives of the selected method. Many
researchers elect to use instruments that were used previously by other
researchers. An advantage of the strategy is that information about reliability
and validity of the instruments may be established already. Instruments often
have been revised many times and are designed to facilitate easy data collection.
Disadvantages to using existing instruments are that the instruments may not
have been used with the populations to be studied or may not address a
particular issue .The following steps can be used to design a data instrument.
1. Decide what information is needed
2. Know the environment
3. Decide what method to use
4. Format the instrument
5. Pilot test
6. Get feedback on the instrument
7. Revise
IV.B. DATA COLLECTION METHODS
In marketing research collection of data plays a dominant role. Data is required
to make decision in any business situation. the researcher is faced with one of
the most difficult problems of obtaining suitable, accurate and adequate data. As
soon as the research design has been decide, the next step is that of selecting the
sources of data. Based on the nature and purpose of the study researcher can go
for various modes of data collection .Data sources can be classified into two
categories namely primary and secondary sources .Discussion about sources are
as follows.
TYPES OF DATA
PRIMARY
SECONDARY
INTERNAL
EXTERNAL
INTERNAL
EXTERANL
SOURCES
SOURCES
SOURCES
SOURCES
INTERVIEW
CENSES /
SURVEY/
COMPANY
PERIODICAL AND
KEY PEOPLE
INTERVIEW
RECORDS
REPORTS
OBSERVATIO
N
IV.B. a. COLLECTION OF SECONDARY DATA
OR
DESK RESEARCH
Secondary data is data which has been collected by individuals or agencies for
purposes other than those of their particular research study. These secondary
data source may also be termed as paper source. These secondary sources could
include previous research reports, newspaper, magazine and journal content,
government and NGO statistics. Sometimes secondary research is required in
the preliminary stages of research to determine what is known already and what
new data are required, or to inform research design. At other times, it may make
be the only research technique used. For example, if a government department
has conducted a survey of, say, family medical expenditures, and then medicine
manufacturer might use this data in the organization`s evaluations of the total
potential market for a new product. Similarly, statistics prepared by a ministry of
health will prove useful to a whole host of people and organizations, including
those marketing medicine supplies.
i).THE NATURE OF SECONDARY SOURCES OF INFORMATION
No marketing research study should be undertaken without a prior search of
secondary sources (also termed desk research). There are several grounds for
making such a bold statement.
1. Secondary data may be available which is entirely appropriate and
wholly adequate to draw conclusions and answer the question or
solve the problem.
2. It is far cheaper to collect secondary data than to obtain primary
data. For the same level of research budget a thorough examination
of secondary sources can yield a great deal more information than
can be had through a primary data collection exercise.
3. The time involved in searching secondary sources is much less than
that needed to complete primary data collection.
4. Secondary sources of information can yield more accurate data than
that obtained through primary research. This is not always true but
where a government or international agency has undertaken a large
scale survey, or even a census, this is likely to yield far more
accurate results than custom designed and executed surveys when
these are based on relatively small sample sizes.
5. It should not be forgotten that secondary data can play a substantial
role in the exploratory phase of the research when the task at hand is
to define the research problem and to generate hypotheses.
6. Secondary sources help define the population. Secondary data can be
extremely useful both in defining the population and in structuring
the sample to be taken. For instance, government statistics on a
country's agriculture will help decide how to stratify a sample and,
once sample estimates have been calculated, these can be used to
project those estimates to the population.
ii). CLASSIFICATION OF SECONDARY DATA
The sources of secondary data can be classified into four categories. They are:
1. Published sources: data available from various national and international
institutions
2. Unpublished sources: in few cases data maintained by the organizations are
not available to the public/researchers in published forms. Researchers can get
this information through personal contacts.
3. Internal source: internal data are prepared by summaries of companies
internal operations. Sales report, invoice, shipment records and consumer
database maintained by the companies are the major sources.
4. External data: which means data collected from outside of the company.
Collection of external data is more complex because the data have much
greater variety and the sources are more.
Sources of secondary data
Trade associations
National and local press Industry magazines
National/ international governments
Informal contacts
Trade directories
Published company accounts
Business libraries
Professional institutes and organizations
Omnibus surveys
Previously gathered marketing research
Census data
Public records
Web sites
Government publications: Official statistics are statistics collected by
governments and their various agencies, bureaus, and departments. These
statistics can be useful to researchers because they are an easily obtainable and
comprehensive source of information that usually covers long periods of time.
However, because official statistics are often characterized by unreliability,
data gaps, over-aggregation, inaccuracies, mutual inconsistencies, and lack of
timely reporting, it is important to critically analyze official statistics for
accuracy and validity. There are several reasons why these problems exist:The
scale of official surveys generally requires large numbers of enumerators
(interviewers) and, in order to reach those numbers enumerators contracted are
often under-skilled. The size of the survey area and research team usually
prohibits adequate supervision of enumerators and the research process; and
resource limitations (human and technical) often prevent timely and accurate
reporting of results.
Technical Reports: Technical reports are accounts of work done on research
projects. They are written to provide research results to colleagues, research
institutions, governments, and other interested researchers. A report may come
from completed research or on-going research projects.
Scholarly Journals: Scholarly journals generally contain reports of original
research or experimentation written by experts in specific fields. Articles in
scholarly journals usually undergo a peer review where other experts in the
Same field review the content of the article for accuracy, originality, and
relevance.
Literature Review Articles: Literature review articles assemble and review
original research dealing with a specific topic. Reviews are usually written by
experts in the field and may be the first written overview of a topic area.
Review articles discuss and list all the relevant publications from which the
information is derived.
Trade Journals: Trade journals contain articles that discuss practical
information concerning various fields. These journals provide people in these
fields with information pertaining to that field or trade.
Syndicate: The growing demand for marketing data has produced a number of
companies that collect and sell standardized data designed to serve information
needs shared by a number of organization; the most common are information
needs associated with performance-monitoring research. Syndicated data
sources can be classified as
1. Consumer data-collected from consumers regarding purchases.
2. Retail data-focus on products or services sold through retailers .
3. Wholesale data- report on movement of different brands .
4. Industrial data-data on companies to construct sales prospect
5. Advertising evaluation data- to measure the effectiveness of the
advertisement
6. Media and audience data- to match market and media
Reference Books: Reference books provide secondary source material. In many
cases, specific facts or a summary of a topic is all that is included. Handbooks,
manuals, encyclopedias, and dictionaries are considered reference books .
International publications: The United Nations Organizations(UNO),
International
Labor
Organization(ILO),International
Monetary
Fund(IMF),World Bank, etc., are publishing numerous items of data relating to
the socio-economic conditions of different countries.
iii). QUALITY OF INFORMATION SOURCES
One of the advantages of secondary data review and analysis is that individuals
with limited research training or technical expertise can be trained to conduct
this type of analysis. Key to the process, however, is the ability to judge the
quality of the data or information that has been gathered. The following tips
will help you to assess the quality of the data.
Determine the Original Purpose of the Data Collection: Consider the
purpose of the data or publication. Is it a government document or statistic, data
collected for corporate and/or marketing purposes, or the output of a source
whose business is to publish secondary data (e.g., research institutions).
Ascertain the Credentials of the Source(s) or Author(s) of the Information :
What are the author`s or source`s credentials -- educational background, past
works/writings, or experience -- in this area?
Are the methods sound?: Does the article have a section that discusses the
methods used to conduct the study? If it does not, you can assume that it is a
popular audience publication and should look for additional supporting
information or data.
Date of Publication : When was the source published? Is the source current or
out-of-date? Topic areas of continuing or rapid development, such as the
sciences, demand more current information.
Who is the Intended Audience? Is the publication aimed at a specialized or a
general audience? Is the source too elementary -- aimed at the general public?
What is the Coverage of the Report or Document? Does the work update
other sources, substantiate other materials/reports that you have read, or add new
information to the topic area?
After judging the competency of the data based on the above tests, researcher
can chose the right one.
SECONDARY DATA
ADVANTAGES
DISADVANTAGES
1. Low cost
1. Lack of consistency of
2. speed
perspective
3. Often the only resource, for
2. Biases and inaccuracies can
example historical
not be checked
documents
3. Published statistics often
4. Only way to examine
raise more questions than
large-scale trends
they answer (for example,
5. Saves time
what does church
6. Wide range of availability
attendance tells us about
7. Serves as a source of
religious beliefs?)
comparative data
4. The concern over whether
any data can be totally
separated from the context
of its collection
Marketing Research Agencies in India
Operations Research Group (ORG)
National Council of Applied Economic Research (NCAER)
Indian Market Research Bureau (IMRB)
Marketing and Research Group (MARG)
Marketing Research and Advisory Services (MRAS)
Marketing and Business Associate (MBA)
IV.B.b. COLLECTION OF PRIMARY DATA
Primary data is data collected for the first time. It is original and collected for a
specific purpose, or to solve a specific problem. It is expensive, and time
consuming, but is more focused than secondary research. There are many ways
to conduct primary ( data collection) research. The following are the prime
methods:
1. Observation
2. Questionnaire
3. Interview
4. Focus groups
5. Projective techniques
6. Product tests
7. Diaries
8. Omnibus Studies
9. Clinics
10. Surveys
11. Case study
I-OBSERVATION METHODS
Observation involves the process of physically or mechanically recording some
specific aspect of a consumer`s activity or behavior .Observational techniques
are methods by which an individual or individuals gather firsthand data on
programs, processes, or behaviors being studied. They provide evaluators with
an opportunity to collect data on a wide range of behaviors, to capture a great
variety of interactions, and to openly explore the evaluation topic. By directly
observing operations and activities, the evaluator can develop a holistic
perspective, i.e., an understanding of the context within which the project
operates. This may be especially important where it is not the event that is of
interest, but rather how that event may fit into, or be impacted by, a sequence of
events. Observational approaches also allow the evaluator to learn about things
the participants or staff may be unaware of or that they are unwilling or unable
to discuss in an interview or focus group.
When to use observations. Observations can be useful during both the
formative and summative phases of evaluation. For example, during the
formative phase, observations can be useful in determining whether or not the
project is being delivered and operated as planned. In the hypothetical project,
observations could be used to describe the faculty development sessions,
examining the extent to which participants understand the concepts, ask the right
questions, and are engaged in appropriate interactions. Some research designs
call for this type of data. Some contented that observation is objective than
communication .How ever , observation whether in a field setting or a
laboratory setting ,is not very versatile.
Recording Observational Data
Observations are carried out using a carefully developed set of steps and
instruments. The observer is more than just an onlooker, but rather comes to the
scene with a set of target concepts, definitions, and criteria for describing events.
While in some studies observers may simply record and describe, in the majority
of evaluations, their descriptions are, or eventually will be, judged against a
continuum of expectations.
Observations usually are guided by a structured protocol. The protocol can take
a variety of forms, ranging from the request for narrative describing events seen
to a checklist or a rating scale of specific behaviors/activities that address the
evaluation question of interest. The use of a protocol helps assure that all
observers are gathering the pertinent information and, with appropriate training,
applying the same criteria in the evaluation .The protocol goes beyond a
recording of events, i.e., use of identified materials, and provides an overall
context for the data. The protocol should prompt the observer to,
Describe the setting of program delivery, i.e., where the observation
took place and what the physical setting was like;
Identify the people who participated in those activities, i.e.,
characteristics of those who were present;
Describe the content of the intervention, i.e., actual activities and
messages that were delivered;
Document the interactions between implementation staff and project
participants;
Describe and assess the quality of the delivery of the intervention; and
Be alert to unanticipated events that might require refocusing one or
more evaluation questions.
METHODS OF OBSERVATION
There are several methods of observation of which any one or a combination of
some of them can be used by the observer.
1. Structured-unstructured observation:
Structured observation is used when the research problem has been formulated
precisely and the observers have been told specifically what is to be observed.
They may be given a simple form to record their observations. Unstructured
observation implies that observers are free to observe what ever they think is
relevant and important .While structured observations are free from subjective
bias, unstructured observations are subject to this limitation .The extent of the
bias may vary to the extent an observation is unstructured.
2. Disguised-undisguised observation:
In the case of disguised observation, the subjects do not know that they are
being observed. In some cases, disguised observation may be made by the
observer by posing as one of the shoppers who are being observed. This type of
observation is often preferred because it is feared that people may behave
differently when they know they are being observed. It may be difficult to
completely disguise an observation, though this apart, it poses an ethical
question of its desirability when those who are being observed are kept in the
dark.
3. Observation under natural setting-laboratory setting:
Observations can also be classified on the bases of their setting, i.e. natural or
laboratory. Observations in field studies are in their natural setting and are,
therefore undertaken in extremely realistic conditions. Sometimes, an
experimental manipulation may be introduced in a field study. Observation in a
laboratory setting, on the other hand, enables the observer to control extraneous
variables which influence the behavior of people. Observational studies in
laboratory settings have certain advantages over field studies. They enable the
collection of data promptly and economically and in addition, permit the use of
more objective measurements.
4. Direct-indirect observation:
In the case of direct observation, the event or the behavior of a person is to be
observed as it occurs .In contrast; indirect observation implies that some record
of past behavior is observed. In other words ,the behavior itself is not observed
,rather its effects are observed. An observer engaged in indirect observation
generally looks for physical traces of behavior or occurrence of an event.
Suppose he is interested in knowing about the liquor consumption of a
household, he would like for empty liquor bottles in the garbage. Similarly, the
observer may seek the permission of the housewife to see the pantry. He may
carry a pantry audit to ascertain the consumption of certain types of products. It
may be noted that the success of an indirect observation largely depends on how
best the observer is able to identify physical traces of the problem under study.
Direct observation is far more common than indirect observation.
5. Human-mechanical observation:
Another way of classifying observations is whether they are made manually or
by machines. Most of the studies in marketing research based on human
observation wherein trained observers are required to observe and faithfully
record their observations. In some cases, mechanical devices such as eye
cameras and audiometers are use for observation. One of the major advantages
of electrical/mechanical devices is that their recordings are free from subjective
bias. As against this advantage, such observations may be less valid than human
observations. This is because the observer`s power of integration can lead to a
more valid evaluation of the observation.
OBSERVATIONS
ADVANTAGES
DISADVANTAGES
1. Expensive and time consuming
1. Provide direct information about
2. Need well-qualified, highly trained
behavior of individuals and groups
observers; may need to be content
experts
2. Permit evaluator to enter into and
3. May affect behavior of participants
understand situation/context
4. Selective perception of observer
may distort data
3. Provide good opportunities for
5. Investigator has little control over
identifying unanticipated outcomes
situation
6. Behavior or set of behaviors
4. Exist in natural, unstructured, and
observed may be unusual
flexible setting
II-QUESTIONNAIRE METHOD OF DATA
COLLECTION
This method of data collection is quite popular, particularly in case of big
enquiries. It is being adopted by private individuals, research workers, private
and public organizations and even by governments. In this method a
questionnaire is sent (usually by post) to the persons concerned with a request to
answer the questions and return the questionnaire. A questionnaire consists of a
number of questions written or printed in a definite order on a form or set of
forms. The questionnaire is send to respondents who are expected to read and
understand the questions and write down the reply in the space meant for the
purpose in the questionnaire itself. The respondents have to answer the questions
on their own. The method of collecting data by mailing the questionnaire to
respondents is most extensively employed in various economic and business
surveys.
The merits claimed on behalf of this method are as follows:
1. There is low cost even when the universe is large and is widely spread
geographically.
2. It is free from the bias of the interviewer; answers are in respondent`s
own words.
3. Respondents have adequate time to give well thought out answers.
4. Respondents, who are not easily approachable, can also be reached
conveniently.
5. Large samples can be made use of and thus results can be made more
dependable and reliable.
The major demerits of this system are listed here:
Low rate of return of the duly filled in questionnaire; bias due to no-response is
often indeterminate.
1. It can be used only when respondents are educated and cooperating.
2. The control over questionnaire may be lost once it is sent.
3. There is inbuilt inflexibility because of the difficulty of amending the
approach once questionnaire have been dispatched.
4. There is also the possibility of ambiguous replies or omission of replies
altogether to certain questions; interpretation of omissions is difficult.
5. It is difficult to know whether willing respondents are truly
representative.
6. This method is likely to be the slowest of all.
Before using this method, it is always advisable to conduct pilot study` (Pilot
Survey) for testing the questionnaires.
Main aspects of questionnaire:
Quite often questionnaire is considered as the heart of a survey operation. Hence
it should be very carefully constructed. If it is not properly set up, then the
survey is bound to fail. This fact requires us to study the main aspects of a
questionnaire viz., the general form, question sequence and question formulation
and wording. Researcher should note the following with regard to these three
main aspects of a questionnaire:
1. General form:
structured questionnaire.
unstructured questionnaire.
So far as the general form of a questionnaire is co concerned it can either be
structured or unstructured questionnaire. Structured questionnaires are those
questionnaires in which there are definite, concrete and pre-determined
questions. The questions are presented with exactly the same wording and in the
same order to all respondents. Structured questions may also have fixed
alternative questions in which responses of the informants are limited to the
stated alternatives. Thus a highly structured questionnaire is one in which all
questions and answers are specified and comments in the respondent`s own
words are held to the minimum. When these characteristics are not present in a
questionnaire, it can be termed as unstructured or non-structured questionnaire.
2. Question sequence:
In order to make questionnaire effective and to ensure quality to the replies
received, a researcher should pay attention to the question-sequence in preparing
the questionnaire. A proper sequence of questions reduces considerably the
chances of individual questions being misunderstood. The question-sequence
must be clear and smoothly-moving, meaning thereby the relation of one
question to another should be readily apparent to the respondent, with questions
that are easiest to answer being put in the beginning. The first few questions are
particularly important because they are likely to influence the attitude of the
respondent and in seeking his human interest. The following type of questions
should be generally avoided as opening questions in the questionnaire:
o Questions that put too great a strain on the memory or intellect of the
respondent.
o Questions of a personal character.
o Questions related to personal wealth etc,
Following the opening questions, we should have questions that are really vital
to research problem and a connecting thread should run through successive
questions. Ideally, the question-sequence should conform to the respondent`s
way of thinking.
2. Question formulation and wording:
With regard to this aspect of questionnaire, the researcher should note that each
question must be very clear for any sort of misunderstanding can do irreparable
harm to survey. Question should also be impartial in order not to give a biased
picture of the true state of affairs. Questions should be constructed with a view
to their forming a logical part of a well thought out tabulation plan. In general all
questions should meet the following standards :
(a) Should be easily understood.
(b) Should be simple i.e., should convey only one thought at a time.
(c) Should be concrete and should conform as much as possible to the
respondent`s ways of thinking.
4. Kind of questions:
Open end question (format)
Closed-end question
In general, there are two types of questions one will ask, open format or closed
format. Open format questions are those that ask for unprompted opinions. In
other words, there are no predetermined set of responses, and the participant is
free to answer however he chooses. Open format questions are good for
soliciting. An obvious advantage is that the variety of responses should be wider
and more truly reflect the opinions of the respondents. This increases the
likelihood of you receiving unexpected and insightful suggestions, for it is
impossible to predict the full range of opinion. It is common for a questionnaire
to end with and open format question asking the respondent for her unabashed
ideas for changes or improvements.
Open format questions have several disadvantages. First, their very nature
requires them to be read individually. There is no way to automatically tabulate
or perform statistical analysis on them. This is obviously more costly in both
time and money, and may not be practical for lower budget or time sensitive
evaluations. They are also open to the influence of the reader, for no two people
will interpret an answer in precisely the same way. This conflict can be
eliminated by using a single reader, but a large number of responses can make
this impossible
Finally, open format questions require more thought and time on the part of the
respondent. Whenever more is asked of the respondent, the chance of tiring or
boring the respondent increases.
Closed format questions usually take the form of a multiple-choice question.
They are easy for the respondent, There is no clear consensus on the number of
options that should be given in an closed format question. Obviously, there
needs to be sufficient choices to fully cover the range of answers but not so
many that the distinction between them becomes vague.
Qualities of a Good Question:
There are good and bad questions. The qualities of a good question are as
follows:
1. Questions should evoke the truth.
2. Asks for an answer on only one dimension.
3. Can accommodate all possible answers.
4. Don`t imply a desired answer.
5. Don`t use emotionally loaded or vaguely defined words.
6. Don`t ask the respondent to order or rank a series of more than few
items.
III-INTERVIEW METHODS OF DATA COLLECTION
This is the technique most associated with marketing research. An interview is a
specialized type of communication, usually verbal, between two or more people
and is carried out for a specific purpose. It is different from an ordinary
conversation in that its form and purpose is structured. Interviews can be
telephone, face-to-face, or over the Internet. The use of interviews as a data
collection method begins with the assumption that the participants` perspectives
are meaningful, knowable, and able to be made explicit, and that their
perspectives affect the success of the project. An interview, rather than a paper
and pencil survey, is selected when interpersonal contact is important and when
opportunities for follow-up of interesting comments are desired.
Generally two types of interviews are used in marketing research: structured
interviews, in which a carefully worded questionnaire is administered; and in-
depth interviews, in which the interviewer does not follow a rigid form. In the
former, the emphasis is on obtaining answers to carefully phrased questions.
Interviewers are trained to deviate only minimally from the question wording to
ensure uniformity of interview administration. In the latter, however, the
interviewers seek to encourage free and open responses, and there may be a
tradeoff between comprehensive coverage of topics and in-depth exploration of
a more limited set of questions.
PREPARING FOR INTERVIEWS
If you are going to carry out an interview, you need to think about the steps
involved. Good preparation must be done before the interview to make sure that
you get what you need from it, and some thought given afterwards to the
information gathered. The following steps can be followed for conducting the
interview
1. Decide on the purpose of the interview:
Clear understanding of the purpose of the interview will definitely enhances
the efficiency of the method. Before throwing questions the researchers
themselves must sure of their motto. For example, Are you trying to find out
someone's opinion about something? Are you trying to help someone with a
problem? , are you interviewing someone for a job?
2. Decide what kind of information you need to get from the interview to
achieve the purpose:
Every interview must be preplanned to ensure the desired outcome and to use
time efficiently. Begin with an opening statement of purpose, goals, timing,
confidentiality, and format. The researcher may have a list of specific goals.
When the interview is conducted to study the level of customer satisfaction, the
parameters must be specifically selected.
3. Decide what questions you are going to ask:
Researcher should draw up a list of questions so that the answers help to achieve
the goals. Before selecting interviewing as a data collection method, the
researcher must determine whether the research question can be answered
appropriately by interviewing people who have experienced the phenomenon of
interest. A hypothetical study will be used to illustrate one process the
researcher could use to facilitate interviewing.
4. Carry out the interview:
Starting of an interview it self an art, the approaches used by the researcher
heavily influences the respondent in the way of answering. Spend a little time
setting the scene and getting acquainted with the interviewee (ice breaker) will
certainly open the mind of the respondent. Giving assurance to the respondent
about confidentiality and criticism positively increases the reliability.
5. Study the answers to the questions: The closing of the interview is very
important. Researcher must ensure that the answers heard are the answers that
were given. Restate the answers and get acceptance from the interviewee. Allow
clarification of any misstatements or misunderstandings. Give reassurance of the
purpose of the interview being data collection not oversight. Share any
suggestions for improvement if considered advisable Before completing the
interview the researcher make sure that , whether the goals have been achieved
or not. If not, uncovered areas are to be identified and enough time must be
asked from the respondent to continue the interview.
6. Make a decision about the purpose of the interview:
As soon as the interview is over if possible , the out come of the interview may
be reported to the respondent. In marketing research for various reasons this
may not be advisable.
The role of the interviewer:
The interviewer is really the "jack-of-all-trades" in marketing. The interviewer's
role is complex and multifaceted. The researcher must posse`s capability of
extracting the answer from the respondent .In most cases they are the backbone
of the functioning of the interview. To perform the task the following things are
essential:
1. Locate and enlist cooperation of respondents
2. Motivate respondents to do good job
3. Clarify any confusion/concerns
4. Observe quality of responses
5. Conduct a good interview
TYPES OF INTERVIEW
a. Telephone interviews
Telephone ownership is very common in developed countries. It is ideal for
collecting data from a geographically dispersed sample. The interviews tend to
be very structured and tend to lack depth. Telephone interviews are cheaper to
conduct than face-to-face interviews (on a per person basis).
Advantages of telephone interviews
1. Phone interviews can have a shorter data collection period than face-
to-face interviews.
2. Phone interviews may have a better response rate .
3. Phone interview reduces travel costs while permitting interaction
with remote participants.
4. Phone interviews can supplement site visits, face-to-face interviews,
and other methods. For example, after a site visit or a survey you
could gather additional data by conducting a phone interview
5. Cheaper than face-to-face interviews
Disadvantages of telephone interviews
1. Phone interviews can be quite tiring, so they are often shorter than
face-to-face interviews.
2. Phone interviews can be difficult if the interviewer or interviewee
has a strong accent.
3. Phone interviews are not as good as face-to-face interviews when
you are dealing with complex issues.
4. If you have multiple interviewers, you have to worry about consistent
approaches to the interviews.
5. Phone interviews are often conducted at times that are convenient to
the participant, but not for the interviewer (evenings, early mornings,
weekends)
6. Phone interviews at a person`s office or home can involve many
potential distractions like colleagues stopping by, calls on other lines,
background noise, and the lure of using the computer to work during
the phone interview.
7. Respondents can simply hang up
8. Interviews tend to be a lot shorter
9. Visual aids cannot be used
10. Researchers cannot see the behavior or body language of the
rspondet.
b. Face-to-face interviews
When marketers want to discuss questions in great depth, they generally use
personal interview, which are face-to-face interviews conducted with one
person or a group of people at one time. personal interviews can be large and
more detailed than other types of study. The research will probe and develop
points of interest.
Advantages of face-to-face interviews
1. They allow more 'depth'
2. Physical prompts such as products and pictures can be used
3. Body language can emphasize responses
4. Respondents can be 'observed' at the same time
Disadvantages of face-to-face interviews
1. Interviews can be expensive
2. It can take a long period of time to arrange and conduct.
3. Some respondents will give biased responses when face-to-face with
a researcher.
c. In-depth interviews
An in-depth interview is a dialogue between a skilled interviewer and an
interviewee. Its goal is to detailed material that can be used in analysis. Such
interviews are best conducted face to face, although in some situations telephone
interviewing can be successful. Extensive probing and open-ended questions
characterize In-depth interviews. Typically, the project evaluator prepares an
interview guide that includes a list of questions or issues that are to be explored
and suggested probes for following up on key topics.
ADVANTAGES AND DISADVANTAGES OF IN-DEPTH INTERVIEW
Advantages
Disadvantages
Usually yield richest data,
Expensive
and
time-
details, new insights
consuming
Permit face-to-face contact
Need well-qualified, highly
with respondents
trained interviewers
Provide opportunity to
Interviewee may distort
explore topics in depth
information through recall
Afford
ability
to
error, selective perceptions,
experience the affective as
desire to please interviewer
well as cognitive aspects of
Flexibility can result in
responses
inconsistencies
across
Allow
interviewer
to
interviews
explain or help clarify
Volume of information too
questions, increasing the
large; may be difficult to
likelihood
of
useful
transcribe and reduce data
responses
Allow interviewer to be
flexible in administering
interview to
particular
individuals
or
circumstances
d. Computer Direct Interviews
These are interviews in which the Interviewees enter their own answers directly
into a computer. They can be used at malls, trade shows, offices, and so on. The
Survey System's optional Interviewing Module and Interview Stations can easily
create computer-direct interviews. Some researchers set up a Web page survey
for this purpose.
Advantages
1. The virtual elimination of data entry and editing costs.
2. You will get more accurate answers to sensitive questions.
3. The elimination of interviewer bias. Different interviewers can ask
questions in different ways, leading to different results. The
computer asks the questions the same way every time.
4. Ensuring skip patterns are accurately followed. The Survey System
can ensure people are not asked questions they should skip based on
their earlier answers. These automatic skips are more accurate than
relying on an Interviewer reading a paper questionnaire.
5. Response rates are usually higher. Computer-aided interviewing is
still novel enough that some people will answer a computer interview
when they would not have completed another kind of interview.
Disadvantages
1. The Interviewees must have access to a computer or one must be
provided for them.
2. As with mail surveys, computer directs interviews may have serious
response rate problems in populations of lower educational and
literacy levels. This method may grow in importance as computer use
increases.
IV-FOCUS GROUPS
Focus groups are made up from a number of selected respondents based together
in the same room. Highly experienced researchers work with the focus group to
gather in depth qualitative feedback. Groups tend to be made up from 10 to 18
participants. Discussion, opinion, and beliefs are encouraged, and the research
will probe into specific areas that are of interest to the company commissioning
the research.
Typically focus groups are used as a means of testing concepts, new products,
and messages. A focus group is qualitative research, which means that you do
not obtain results with percentages, statistical testing, or tables. Instead, this
methodology is less structured than surveys or other quantitative research and
tends to be more exploratory as well. Rather than providing quantifiable
responses to a specific question obtained from a large sampling of the
population, focus group participants provide a flow of input and interaction
related to the topic or group of topics that the group is centered around. While
they appear to be less formal than a survey, focus groups do provide an
important source of information for making business decisions. It is important,
however, to ensure that persons using the results of such a qualitative study
understand how to correctly interpret the resulting information.
Quantitative research provides results that can be generalized to a specific
population, because it is based on a statistical sampling of the target population.
The results of qualitative research, such as focus groups, however, are not
quantifiable. They reflect only a very small segment of the target market in
question. Given the number of focus group participants, results are not
necessarily representative of the general population from which participants are
recruited and should not be considered as such.
When to Use and When to Avoid Focus Groups
Focus groups are best used when the concept or idea you wish to evaluate is new
and when the best evaluation comes from letting the target customer (who can
be either a consumer or a customer in the business-to-business sense) view the
concept directly. A good example of this is a new advertising campaign.
Typically, an advertising agency will want to test a new advertisement with the
consumers it hopes to reach with the campaign. The agency needs to know
whether the message is clear, whether consumers view the advertisement
positively or negatively, whether the advertisement would prompt them to
purchase the product, and so forth. The agency can explore these questions by
showing the advertisement to consumers and discussing their reactions and their
likes and dislikes. Their comments will allow the agency to fine-tune the
advertisement or, if it is not well liked, to go back to the drawing board.
Another common use of focus group research is product concept testing.
Participants in a group can discuss detailed diagrams and product descriptions,
or they may test a product prototype in a hands-on fashion. This allows
researchers to identify customer needs and, attitudes regarding a concept before
investing in further product development.
A focus group study also is often used to design the questionnaire for a
quantitative survey. The focus group covers general issues on a topic, and
respondents' comments often help researchers identify pertinent issues that
might otherwise be left out of a survey. Hypotheses generated by focus groups
frequently lead to further testing using quantitative methods. Alternatively,
focus groups can be used to further interpret quantitative research. For example,
if a telephone survey produces a significant percentage of unexpected comments
on a certain topic, a focus group study could investigate the issue with greater
depth.
Another particularly useful application of focus group research is as a
brainstorming mechanism. If you have a problem to solve, this type of
methodology often provides fresh insights regarding the issue at hand. It can
also provide an excellent forum for generating creative ideas or new-product
ideas for new markets, as well as generating new ideas for your established
markets.
Focus groups can be used for the following purposes
To test new concepts
To position a product/ service
To assess product usability
To evaluate advertising/ copy
To evaluate promotions
To develop questionnaires
To generate ideas or support brainstorming
V. PROJECTIVE TECHNIQUES
The technique of inferring a subjects attitudes or values based on his or her
description of vague objects requiring interpretation. Projective techniques are
borrowed from the field of psychology. They will generate highly subjective
qualitative data. There are many examples of such approaches including: Inkblot
tests - look for images in a series of inkblots Cartoons - complete the 'bubbles'
on a cartoon series Sentence or story completion Word association - depends on
very quick (subconscious) responses to words .Psychodrama - Imagine that you
are a product and describe what it is like to be operated, warn, or used.
Projective techniques consist of the following:
a. Word association tests: In word association test the subject is presented
with a list of words, one at a time, and asked to respond with first that
comes to the respondents mind. Both verbal and nonverbal responses are
recorded. Word association frequently used to test potential brand names
and company images among public .interpreting word association tests is
difficult, and the marketing researcher should make sure to avoid subjective
interpretations .when there is considerable agreement in the free-association
process, the researcher assumes that the test has revealed the consumers
inner feelings about the subject
b. Sentence completion test: A projective technique in which respondents are
required to complete a number of partial sentences with the first word or
phrase that comes to mind. The sentence completion method is also based
on the principles of free association. For example:
People who smoke are --------
A man who smoke filter cigarette are------
Imported cigarette is most liked by-------
The women in politics are------
When comparing to word association tests the answers given to sentence
completion tests tends to more extensive .the intent of sentence completion
questions are more apparent.
VI- PRODUCT TESTS
Product tests are often completed as part of the 'test' marketing process. Products
are displayed in a mall of shopping center. Potential customers are asked to visit
the store and their purchase behavior is observed. Observers will contemplate
how the product is handled, how the packing is read, how much time the
consumer spends with the product, and so on.
VII-DIARIES
Diaries are used by a number of specially recruited consumers. They are asked
to complete a diary that lists and records their purchasing behavior of a period of
time (weeks, months, or years). It demands a substantial commitment on the part
of the respondent. However, by collecting a series of diaries with a number of
entries, the researcher has a reasonable picture of purchasing behavior.
VIII-OMNIBUS STUDIES
An omnibus study is where an organization purchases a single or a few
questions on a 'hybrid' interview (either face-to-face or by telephone). The
organization will be one of many that simply want to a straightforward answer
to a simple question. An omnibus survey could include questions from
companies in sectors as diverse as heath care and tobacco. The researches are far
cheaper, and commit less time and effort than conducting own research.
IX-CLINICS
The Clinic methodology constitutes a hybrid approach of both qualitative and
quantitative, primarily for testing product or service concepts. A significant
number of respondents are screened and recruited to a central facility (this may
be a focus group facility or it may be a hotel conference room). Concepts are
presented to the large audience of respondents; videos may be shown; product
prototypes may be passed around; client technical people may make parts of the
presentation. Each respondent is provided with a wireless remote answer device
and, during the session, they are asked specific questions and reply in real-time
through their wireless devices. Results can, as appropriate, be fed back
instantaneously and are used at the end of the presentation phase to develop a
list of those respondents who have answered in a certain way, and these are
asked to stay for a focus group. Thus the focus group might have people who
liked a new product concept but were only prepared to pay a little for it, or
people who did not like some or all aspects of the product or offer. The result is
a set of quantitative data from the broader audience, and qualitative insight from
the highly focused focus group. The process can be undertaken quickly across
multiple countries and the results made available within a few weeks.
X-SURVEYS
A survey is a research technique in which data are systematically collected
directly from the people being studied by the questionnaire. Surveys are a form
of questioning that is more rigid than interviews and that involve larger groups
of people. Surveys will provide a limited amount of information from a large
group of people and are useful when you want to learn what a larger population
thinks.
The Steps in a Survey
1. Establish the goals of the survey - What you want to learn
2. Determine your sample - Whom you will interview
3. Choose interviewing methodology - How you will interview
4. Create your questionnaire - What you will ask
5. Pre-test the questionnaire, if practical - Test the questions
6. Conduct interviews and enter data - Ask the questions
7. Analyze the data - Produce the reports
The first step in any survey is deciding what you want to learn. The goals of the
project determine whom you will survey and what you will ask them. If your
goals are unclear, the results will probably be unclear. Some typical goals
include learning more about:
The potential market for a new product or service
Ratings of current products or services
Employee attitudes
Customer/patient satisfaction levels
Reader/viewer/listener opinions
Association member opinions
Opinions about political candidates or issues
Corporate images
Selecting the Sample
There are two main components in determining whom you will interview. The
first is deciding what kind of people to interview. Researchers often call this
group the target population. If you conduct an employee attitude survey or an
association membership survey, the population is obvious. If you are trying to
determine the likely success of a product, the target population may be less
obvious. Correctly determining the target population is critical. If you do not
interview the right kinds of people, you will not successfully meet your goals.
The next thing to decide is how many people you need to interview. Statisticians
know that a small, representative sample will reflect the group from which it is
drawn. The larger the sample, the more precisely it reflects the target group.
However, the rate of improvement in the precision decreases as your sample size
increases. For example, to increase a sample from 250 to 1,000 only doubles the
precision. You must make a decision about your sample size based on factors
such as: time available, budget and necessary degree of precision.
Avoiding a Biased Sample
A biased sample will produce biased results. Totally excluding all bias is almost
impossible; however, if you recognize bias exists you can intuitively discount
some of the answers. The consequences of a source of bias depend on the nature
of the survey. For example, a survey for a product aimed at retirees will not be
as biased by daytime interviews as will a general public opinion survey. A
survey about Internet products can safely ignore people who do not use the
Internet.
Quotas
A Quota is a sample size for a sub-group. It is sometimes useful to establish
quotas to ensure that your sample accurately reflects relevant sub-groups in your
target population. For example, men and women have somewhat different
opinions in many areas. If you want your survey to accurately reflect the general
population's opinions, you will want to ensure that the percentage of men and
women in your sample reflect their percentages of the general population.
If you are interviewing users of a particular type of product, you probably want
to ensure that users of the different current brands are represented in proportions
that approximate the current market share. Alternatively, you may want to
ensure that you have enough users of each brand to be able to analyze the users
of each brand as a separate group.
TELEPHONE SURVEYS
Surveying by telephone is the most popular interviewing method in advanced
countries. Where the time is short and distance is too far the research questions
will be asked through telephone .
Advantages
1. People can usually be contacted faster over the telephone than with
other methods. If the Interviewers are using CATI (computer-assisted
telephone interviewing), the results can be available minutes after
completing the last interview.
2. You can dial random telephone numbers when you do not have the
actual telephone numbers of potential respondents.
Disadvantages
1. Many people are reluctant to answer phone interviews and use their
answering machines to screen calls.
2. You cannot show your sample products by phone.
MAIL SURVEYS
Mail and telephone surveys are a method of collecting information by sending
surveys via email or postal mail. Participants return completed forms to the
researcher. Surveys may ask respondents to rate items on a scale .Some surveys
also allow respondents to write their feelings or attitudes about a particular event
or to elaborate in more detail on an item, or to express suggestions, etc.
Advantages
1. Mail surveys are among the least expensive.
2. This is the only kind of survey you can do if you have the names and
addresses of the target population, but not their telephone numbers.
3. The questionnaire can include pictures - something that is not
possible over the phone.
4. Mail surveys allow the respondent to answer at their leisure, rather
than at the often inconvenient moment they are contacted for a phone
or personal interview. For this reason, they are not considered as
intrusive as other kinds of interviews.
Disadvantages
1. Time! Mail surveys take longer than other kinds. You will need to
wait several weeks after mailing out questionnaires before you can be
sure that you have gotten most of the responses.
2. In populations of lower educational and literacy levels, response rates
to mail surveys are often too small to be useful.
EMAIL SURVEYS
Email surveys are both very economical and very fast. More people have email
than have full Internet access. This makes email a better choice than a Web page
survey for some populations. On the other hand, email surveys are limited to
simple questionnaires, whereas Web page surveys can include complex logic.
Advantages
1. Speed. An email questionnaire can gather several thousand
responses within a day or two.
2. There is practically no cost involved once the set up has been
completed.
3. You can attach pictures and sound files.
4. The novelty element of an email survey often stimulates higher
response levels than ordinary snail mail surveys.
Disadvantages
1. You must possess (or purchase) a list of email addresses.
2. Some people will respond several times or pass questionnaires along
to friends to answer. Many programs have no check to eliminate
people responding multiple times to bias the results. The Survey
System`s Email Module will only accept one reply from each address
sent the questionnaire. It eliminates duplicate and pass along
questionnaires and checks to ensure that respondents have not
ignored instructions (e.g., giving 2 answers to a question requesting
only one).
3. Many people dislike unsolicited email even more than unsolicited
regular mail. You may want to send email questionnaires only to
people who expect to get email from you.
4. You cannot use email surveys to generalize findings to the whole
populations. People who have email are different from those who do
not, even when matched on demographic characteristics, such as age
and gender.
5. Email surveys cannot automatically skip questions or randomize
question or answer choice order or use other automatic techniques
that can enhance surveys the way Web page surveys can.
INTERNET/INTRANET (WEB PAGE) SURVEYS
Web surveys are rapidly gaining popularity. They have major speed, cost, and
flexibility advantages, but also significant sampling limitations. These
limitations make software selection especially important and restrict the groups
you can study using this technique.
Advantages
1. Web page surveys are extremely fast. A questionnaire posted on a
popular Web site can gather several thousand responses within a few
hours. Many people who will respond to an email invitation to take a
Web survey will do so the first day, and most will do so within a few
days.
2. There is practically no cost involved once the set up has been
completed. Large samples do not cost more than smaller ones
(except for any cost to acquire the sample).
3. You can show pictures. Some Web survey software can also show
video and play sound.
4. Web page questionnaires can use complex question skipping logic,
randomizations and other features not possible with paper
questionnaires or most email surveys. These features can assure
better data.
5. Web page questionnaires can use colors, fonts and other formatting
options not possible in most email surveys.
6. A significant number of people will give more honest answers to
questions about sensitive topics, such as drug use or sex, when giving
their answers to a computer, instead of to a person or on paper.
7. On average, people give longer answers to open-ended questions on
Web page questionnaires than they do on other kinds of self-
administered surveys.
Disadvantages
1. Current use of the Internet is far from universal. Internet surveys do
not reflect the population as a whole. This is true even if a sample of
Internet users is selected to match the general population in terms of
age, gender and other demographics.
2. People can easily quit in the middle of a questionnaire. They are not
as likely to complete a long questionnaire on the Web, as they would
be if talking with a good interviewer.
3. If your survey pops up on a web page, you often have no control over
who replies - anyone from Antarctica to Zanzibar, cruising that web
page may answer.
4. Depending on your software, there is often no control over people
responding multiple times to bias the results.
Things to consider when conducting surveys:
Who are you planning on surveying? Decide what group you are going to
focus on surveying based on who you have access to and what your research is
focused on.
How many people are you going to survey? You want to choose a target
number of surveys to conduct. You don't want too few surveys because you
won't have enough answers to support any generalizations or findings you may
make. At the same time, you do not want too many surveys because you will be
overwhelmed with analyzing your data.
How are you going to survey people? You can choose to conduct your survey
in person (i.e. walk up to people and ask them questions); on paper (i.e. hand out
surveys and ask people to return them); or even via the Internet. The survey
method should be chosen based on the length of your survey and types of
questions.
How long is your survey going to be? The answer to this question depends on
what information you are attempting to discover and how much you want to find
out. Longer surveys sometimes involve the same question asked in multiple
ways to see if people are consistent in their answering strategies. For your first
survey, however, it is better to keep things simple. Short questions are usually
more effective than longer ones.
What type of questions are you going to ask? Do you want open-ended
questions or closed questions? Open-ended questions are questions that allow
the participant any type of response. An example of an open-ended question is:
How are you feeling today? A closed question is one with a set of possible
responses or yes/no responses. An example is: Did you feel that the new campus
regulation about parking is fair? While closed questions are much easier to
analyze they do not provide the rich responses you may get with open-ended
questions. Ultimately, what type of question you ask depends on what you want
to discover.
SURVEY CHARACTERISTICS
Survey
Mail survey
Telephone
Personal
characteristics
survey
interview survey
Lowest
Moderate
High
Cost
Wide
Wide
Moderate
Geographic
distribution
Low
Moderate
High
Flexibility and
questioning
No
Moderate
High
Interviewer bias
Slowest
Fastest
Moderate
Speed and data
Lowest
Moderate
Highest
collection
Control and data
Poor
Moderate
Highest
collection
Response rates
XI-CASE STUDY
Case study is a vaery popular method of qualitative research.It involves a
careful and complete observation rather than using large samples and following
a rigid protocol to examine a limited number of variables, case study methods
involve an in-depth, longitudinal examination of a single instance or event: a
case. They provide a systematic way of looking at events, collecting data,
analyzing information, and reporting the results. As a result the researcher may
gain a sharpened understanding of why the instance happened as it did, and what
might become important to look at more extensively in future research. Case
studies lend themselves especially to generating (rather than testing) hypothesis
The design involves the intense investigation of situations which are relevant to
the problem situation.The concept is to select several target cases where an
intensive analysis will help to
Know the environment
Identify relevant variables
Indicate nature of relationship among the variable
Identify the naure of roblems and opportunies in the orginal case.
The case history method is especially useful in situations in which a complicated
series of variables interact to produce the problem or opportunity .
PRIMARY vs. SECONDARY SOURCE OF DATA
Primary sources are the raw material of the research process, they represent the
records of research or events as first described. Secondary sources are based on
primary sources. These sources analyze, describe, and synthesize the primary or
original source. Usage of primary source and secondary source should be
complementary. Savvy entrepreneurs will do secondary research first and then
conduct primary research. For example, the owner of a video-rental shop would
want to know all about a neighborhood before opening a new store there. Using
information gleaned from secondary sources, the owner can find all kinds of
demographic data, including detailed income data and spending patterns. They
can then send out a questionnaire to a sampling of households to find out what
kinds of movies people like to rent. That primary-research technique will help
when it comes time to stock the store with the latest Hollywood releases.
Secondary research lays the groundwork and primary research helps fill in the
gaps. By using both types of market research, business owners get a well-
rounded view of their market and have the information they need to make
important business decisions. The selection of methods depends upon various
factors like
1. Scope of the study.
2. Objectives of the study.
3. Time availability for the study.
4. Degree of accuracy needed.
5. Study environment, respondents.
6. Budget estimated.
IV.C. FIELD OPERATIONS
Learning objectives
After reading this part, you may be able familiar with the following topics
Objectives of field operations in marketing research.
Functions involved in field operation.
The actual data collection process is rarely done by the person who designs the
research. An irony of marketing research is that highly educated and trained
individuals will design the research, but the people who gather the data typically
have little research training and experience. Field operation is that phase of the
project during which researchers make contact with the respondents,
administrator the data collection instruments, record the data, and return the data
to a central location for processing. The wisdom behind the research design and
skill involved in developing the data collection instrument will be wasted in the
field operation is poorly administrated. The planning of the field operation is
highly influenced by data collection method employed.
Marketing managers preparing for the fieldwork operation should concentrate
on the following areas.
1. Briefing sessions for field workers.
2. Supervision of the fieldwork
3. Verification
Briefing for field workers:
A proper research design will eliminate numerous sources of error, but careful
execution of the fieldwork is necessary to produce results without substantial
error. For the best execution the fieldworkers must be properly trained. Whether
the fieldworkers/interviewers have just completed their training in basics or are
already experienced, there is a need to inform them about the current research.
Both experienced and inexperienced fieldworkers must be briefed on the
background of the project, study method sampling technique. Instruction s for
handling certain key situations are always important. To train the fieldworkers
about the questionnaire, a field supervisor can elaborate the questions and its
purpose. The general cautions to be taken by the fieldworkers can also be
outlined
The briefing session will also cover the sampling procedure. A number of
research projects allow the interviewer to be at least partially responsible for
selecting sample. When this is the case, the potential for selection bias exists.
For example in quota sample the field worker may select a wrong sample due to
poor awareness about the importance of the quota representation. In addition to
technical procedure few general guidelines can also be given to these
fieldworkers, like the ethics in conducting field work.
Supervision of Fieldwork:
Irrespective of the training given to the field workers, they may commit few
errors in execution. Direct supervision of personal interviewers, telephone
interviewers and other fieldworkers is necessary to ensure that the technique
communicated in the training sessions are implemented in the field. The
supervision of fieldworkers, like other forms of supervision, refers to controlling
the efforts of workers. Field supervision of fieldworkers requires checking to see
that field procedures are being properly followed. The supervisor checks field
operations to ensure that the interviewing schedule is being met. In addition to
quality control, continual training may be provided. For example, a telephone
supervisor may notice that interviewers are allowing the phone to ring
continuously before considering the call a no answer. The supervisor can
instruct interviewers that if a telephone is allowed to ring too long and then
answered, the respondent may be annoyed
Verification: It is the quality control procedures in fieldwork to ensure that
interviewers are following the field plan. At the field the most important job of
the supervisor is to verify that the interviews are being conducted according to
the schedule and the plan. An interviewer might be tempted to skip the process
for various reasons. Careful recording of the number of completed interviews
will help ensure that the sampling procedure is being properly conducted. It is
the supervisors function to motivate interviewers to carefully follow the plan.
Where ever the deficiency is found the supervisor should take remedial actions
to solve the complication.
IV.D. ERRORS AND DIFFICULTIES
Learning objectives
After reading this part, you may be able familiar with the following topics
Sources error in marketing research
Impact of errors on research
Types of errors
i) ERRORS :
Since the usability of any market research depends upon the accuracy of the
results, error control plays a critical role in the research process. Every step in
the marketing research progression can produce serious errors. The control of
these errors is of critical concern in marketing research. Avoiding many of the
simpler marketing research errors takes only common sense, but avoiding many
of the more complex mistakes requires a much deeper level of awareness. The
common marketing research errors are highlighted below.
Types of errors
Sampling errors
Non sampling errors
Sampling errors
Most marketing research studies utilize samples of people, product or stores.
Based upon these sample results, the researcher and the manager make
conclusion of the whole population from which the sample was selected. For
example the attitudes of all Maruti car owners could be inferred from a sample
of a 1000 owners because the
ERRORS
NON
SAMPLING
ERRORS
SAMPLIN
G
ERRORS
SAMPLING
NON RESSPONSE
DATA ERRORS
FRAME ERRORS
ERRORS
NONCOVE
OVERCOVE
MEASEUREMENT
PROCES
ERRORS
INTERVIE
RAGE
RAGE
SING
WER
ERRORS
ERRORS
ERRORS
ERRORS
sample is used to estimate the population. The difference between the sample
value and the corresponding population value is called as sampling error .
Non sampling errors
Non sampling errors are all the errors that may occur in the marketing research
process except the sampling error. This concept simply includes all the aspects
of the research process where mistakes and deliberate deceptions can occur..
Unfortunately these mistakes and deceptions occur with great frequency in the
marketing research process. An important point to note is that sampling error is
measurable while it is not easy to measure as non-sampling error.
Hence, the researcher should take care of the following points to deal with non-
sampling error.
What kind of non- sampling errors may occur.
What effects these errors may have on our results, and
What steps we can take to reduce these errors.
The effect of non sampling errors:
Sampling error has two properties that make it useful to the researcher.
measurable
relevance with sample size
Unfortunately non sampling errors are not easily measurable and they do not
decrease with sample size. Infact in all likelihood non sampling errors increase
as sample size increases. What non sampling errors do is put a bias in results of
unknown direction and magnitude
TYPES OF NON SAMPLING ERRORS:
Defective problem definition:
Problem of the study should be clearly stated so that they can be linked directly
to the research results. Research objectives should always be clear so that
research results can be presented in relation to specific objectives. A product
manager requests a study to test a media mix. If the true problem is pricing
strategy then any research that is conducted no matter how technically correct
will not be helpful to the manager.
Defective population definition:
The study population must be defined to fit the study objectives. A universe
which is relevant to the problem being studied, and a sample which adequately
represents that universe, are vital requirements of high quality research.
Consider the case of the manager of one of the restaurants in a major
metropolitan airport would like to know what sort of image the restaurant has
among those who have some likelihood of eating in the airport. The population
is defined as people over 18 years old, getting off planes in the third week of
October. If the sample is selected from this population one might get misleading
results. It does not include significant numbers of potential customers that is
people who are just taking off. Also the sample included people who has no
chance of eating in the restaurants. That is people who change planes without
going into the main terminal where the restaurants are located. Conclusions from
this study are questionable.
Frame non representative of the population:
There are many acceptable methods of sampling, each with advantages and
disadvantages in specific situations. The following guidelines are not meant to
specify how sampling must be designed or managed, but rather are aimed at
insuring that sample design and management are disclosed in sufficient detail to
allow clear judgments of a sample's adequacy for the stated research purpose.
The sampling frame must match the defined population. Consider the case of
the investment company that uses the telephone book (the frame)to select a
sample of potential stock buyers this frame would not cover the defined
population well as a
significant number of high income people have unlisted phone numbers. These
high income people are the prime potential stock buyers. Again conclusions are
suspects.
Non responsive errors:
Errors occur because people in selected sample either refuse to be a part of the
sample or they are not at home during the sample periods. a sample is a
representative sample as selected. Some of the selected elements do not form
part of the realized sample, it is not a truly representative sample. The resulting
error is called non?response error. As an example of this problem consider the
case of a resort developer who attempts to interview people during the day. The
study yields some refusals and a lot of not at homes. One must wonder
whether the refusals as a group hold different attitudes of the development from
those who respond.
Questionnaire structure error:
The error made when the structure and layout of the survey instrument leads to
inaccurate responses. For example asking probing questions regarding
viewpoints on potentially negative experiences before asking an overall
satisfaction question, where overall satisfaction would be incorrectly affected by
the recent recall of potentially bad experiences.
Measurement error :
This is caused when information gathered is different from the information
sought. For example, respondents are asked to indicate whether they own a car
or not. Some of them may respond in the positive just to boost their image
before an interviewer, even though they may not be owning a car. Such
responses will result in measurement error.
Data analysis error:
The error that occurs when analysis is incorrectly executed. Simple
mathematical errors are common, which is why data analysis should be checked
over by more than one qualified person for quality. A more significant data
analysis error is when simple frequency reporting (straight number percentage
reporting) is executed when far greater information can be mined from the
results (often inexpensively) through additional analysis such as cross-tabulation
analysis, multiple regression (driver analysis), cluster analysis, factor analysis,
perceptual mapping (multidimensional scaling), structural equation modeling
tests, etc.
Reporting error:
The best approach and program design combined with the best analysis is only
as good as the researcher`s capability to synthesize and report on the results. The
most common reporting error by far is the improper representation of the
significant findings in a format conducive to creating management
understanding and buy-in of survey results. It could be something as simple as
poor language syntax to as complex as choosing the wrong results to report or
not choosing the best way to graphically represent the results. More common in
the current environment is not selecting the best delivery vehicle. For example, a
quality online reporting system is much preferred when distributing results
across a company that is geographically spread out.
Common sources of error in field work:
Five common sources of error in field work are identified in the following
discussion:
1. errors in selecting respondents
2. non-response errors (i.e. failure to get data from selected respondents)
3. errors created by the method of seeking data
4. errors resulting from interviewer`s misinterpreting or misrecording
answers; and
5. interviewer cheating
Respondent selection errors:
Telephone errors:
In telephone surveys the interviewer is typically given a list of numbers to call,
or numbers are dialed on the basis of random-digit dialing. If the interviewer in
the former case is also given the name of the individual at each number to whom
she is to speak, there is no problem, as the interviewer simply asks for that
individual by name. Unfortunately, names are seldom available, and the
interviewer must select the individual to be interviewed at the number called-
usually a household.
Mail intercept interviews:
In the case of shopping center interviews, respondents are often selected by
convenience. It cannot be said that errors are made in respondent selection when
this is the case, but the procedure is often biased because interviewers are likely
to select those individuals who look friendly and appear easy to interview. To
introduce more objectivity into the process, specific times and locations within
the shopping center can be selected randomly, and it can be specified that the
first person passing a given point after each interview or attempted interview
will be sought as the next respondent.
Door-door surveys:
When quota samples are used in door-door surveys, interviewers select the
individuals to be interviewed subject only to quotas for various population
groups such as sex, age and income. This interviewer control of the selection of
the respondents is unlikely to result in the equivalent of a random sample.
Interviewers tend to follow the paths of least resistance and greatest
convenience.
Not -at -homes:
The percentage of not-at-homes varies by city size, day of the week, time of day,
season of the year, age and the sex of the respondent, as well as with the
provisions made to control not-at-homes in the individual studies. It is, however,
almost always surprisingly large. Failure to obtain data from not-at-homes may
bias survey results because population groups vary in the probability of being at
home.
Refusals:
Refusal rates vary from project to project and may range up to 20%. Since
refusals are often the result of personality and mood, it can be argued that they
will occur randomly and will not bias results. Moreover, refusals are a matter of
degree and of circumstances such as convenience at time of call. Repeated
efforts to obtain compliance can reduce the refusal rate, but only to a degree. In
addition to general refusals, refusals may occur on specific questions,
particularly those relating to income and other personal questions.
ii) MINIMIZING OF ERRORS:
In research it is not possible to eliminate all the errors .However, to the extent
that can be minimized. The following precautions can be followed to reduce the
errors :
Selection of suitable study method
Selection of appropriate instrument
Adequate sample size
Using of trained and experiences researchers
Planned data processing
iii) DIFFICULTIES IN DATA COLLECTION
Marketing research basically a problem solving tool. Starting from the definition
of the problem till the end of the presentation of the report researchers are
facing numerous problems. The major problems faced by the researchers are
mentioned below:
1. Volatile changes in the business environment make marketing research
more complicated. Continuous changes in market make the results of the
study unsuitable.
2. Lack of scientific training and application in marketing research
methodology is a major problem in our country .
3. The Research and Development Department has become a common
feature in many organizations. But decisions makers do not appear to be
very eager on implement the findings of the study.
4. Many of the organizations are not reach conscious and feel that
investment in research is wastage of resources and does not encourage
research.
5. Many people largely depend on customs, traditions and routine practices
in their decision making, as they feel research does not have any useful
purpose to serve in the management of their business.
6. The secrecy of business information is sacrosanct to business
organizations. Most of the business organizations in our country do not
part with information to researchers.
7. There is insufficient support between the business organizations and
research institutions, which essential for the development of good and
meaningful research.
8. In India companies are not in a position to allocate huge funds for the
research.
9. Lack of availability secondary data makes marketing research as
baseless start and time consuming one.
10. Poor awareness among the consumers about research makes the study
more burning.
11. Researchers in India are not familiar with the new research instruments
available for conducting market research,
12. poor library facilities at any places, because of which researchers have to
spend much of time and energy in tracing out relevant material and
information.
13. there is a difficulty of timely availability of upto date data from
published sources
14. Lack of code of conduct among the researchers brings bad image on
research. There is a need for developing code of conduct for researchers
to educate them about ethical aspects of research.
IV.E. DATA PROCESSING
Learning objectives
In this chapter you will learn about
Objective of data processing
Functions involved in data processing
Process of editing and coding.
Once the collection of data is over, the next step is to organize data so that
meaningful conclusions may be drawn. The information content of the
observations has to be reduced to a relatively few concepts and aggregates. The
data collected from the field has to be processed as laid down in the research
plan. This is possible only through systematic processing of data. Data
processing involves editing, coding, classification and tabulation of the data
collected so that they are amenable to analysis. This is an intermediary stage
between the collection of data and their analysis and interpretation.
i) EDITING OF DATA
Editing is the first stage in data processing. Editing may be broadly defined to be
a procedure, which uses available information and assumptions to substitute
inconsistent values in a data set. In other words, editing is the process of
examining the data collected through various methods to detect errors and
omissions and correct them for further analysis. While editing, care has to be
taken to see that the data are as accurate and complete as possible, units of
observations and number of decimal places are the same for the same variable.
The following practical guidelines may be handy while editing the data:
1. The editor should have a copy of the instructions given to the
interviewers.
2. The editor should not destroy or erase the original entry. Original entry
should be crossed out in such a manner that they are still legible.
3. All answers, which are modified or filled in afresh by the editor, have to
be indicated.
4. All completed schedules should have the signature of the editor and the
date.
For checking the quality of data collected, it is advisable to take a small sample
of the questionnaire and examine them thoroughly. This helps in understanding
the following types of problems:
1. Whether all the questions are answered,
2. Whether the answers are properly recorded,
3. Whether there is any bias,
4. Whether there is any interviewer dishonesty,
5. Whether there are inconsistencies.
At times, it may be worthwhile to group the same set of questionnaires
according to the investigators (whether any particular investigator has specific
problems) or according to geographical regions (whether any particular region
has specific problems) or according to the sex or background of the
investigators, and corrective actions may be taken if any problem is observed.
Mechanics of editing:
For editing purpose the researcher must draw a proper mechanism which will
simplify the process as well as reduces the duplication work. Data frequently are
written in with a colored pen or pencil .When space on the questionnaire
permits, the original data usually are left in to permit a subsequent edit and to
identify the originals concepts. Before tabulation of data it may be good to
prepare an operation manual to decide the process for identifying inconsistencies
and errors and also the methods to edit and correct them.
Types of editing
1.Field editing: Preliminary editing by a field supervisor on the interview to
catch technical omissions, check legibility of hand writing and clarify responses
that are logically inconsistent
2.In-house editing: A rigorous editing job performed by a centralized office
staff. The researcher normally has centralized office staff to perform editing and
coding. The researcher must setup a centralized office with all facilities for
editing and coding purposes by which coordination can be accomplished
ii ) CODING OF DATA
Coding refers to the process by which data are categorized into groups and
numerals or other symbols or both are assigned to each item depending on the
class it falls in. Hence, coding involves:
Deciding the categories to be used, and
Assigning individual codes to them.
For example, for the open-ended question Do you enjoy milkshakes, if so,
how much would you say you enjoy milkshakes. The researcher in the coding
process will then have to observe different answers and give them a numeric
value. For example, If we use a five-point scale with 1 being low (don`t
enjoy`) and 5 being high (favorite treat`), and the response is I like
milkshakes ? the researcher would code the response as a 3, if the response
was I absolutely love milkshakes! the researcher would code the response as a
5.
In general, coding reduces the huge amount of information collected into a form
that is amenable to analysis. A careful study of the answers is the starting point
of coding. Next, a coding frame is to be developed by listing the answers and by
assigning the codes to them. A coding manual is to be prepared with the details
of variable names, codes and instructions. Normally, the coding manual should
be prepared before collection of data, but for open-ended and partially coded
questions. These two categories are to be taken care of after the data collection.
The following are the broad general rules for coding:
1. Each respondent should be given a code number (an identification
number).
2. Each qualitative question should have codes. Quantitative variables
mayor may not be coded depending on the purpose. Monthly income
should not be coded if one of the objectives is to compute average
monthly income- But if it is used as a classificatory variable it may
be coded to indicate poor, middle or upper Income group.
3. All responses including "don't know", "no opinion" "no response"
etc., are to be coded.
4. Sometimes it is not possible to anticipate all the responses and some
questions are not coded before collection of data. Responses of all
the questions are to be studied carefully and codes are to be decided
by examining the essence of the answers. In partially coded
questions, usually there is an option " Any other (specify)".
Depending on the purpose, responses to this question may be
examined and additional codes may be assigned.
Production coding:
The actual process of transferring the data from the questionnaire or data
collection form after the data have been collected is called production coding.
Based on the nature of the data collection tool, codes may be written directly on
the instrument or special coding sheet. Coding should be done in a central
location so that a supervisor can help to solve interpretation problems.
Computerized coding:
Studies having large sample size uses a computer system for coding as well as
data processing. The process of transferring data from a research project, such as
answers to survey questionnaire, to computer is referred to as data entry.
Several alternative means of entering data in to the computer are also available.
A
researcher using computer?assisted telephone interviewing or with on-line
direct data entry equipment automatically stores and tabulates responses as they
are collected. Optical Scanning Systems may also be used to directly read
material from marked sheets.
CHAPTER SUMMARY
In marketing research the process of data collections plays a vital role .In this
chapter you have studied about data instruments, features of different data
collection methods and data processing. Based on the study selected the
researcher either can go for secondary source or primary data collection and for
both .The process of editing and coding shapes the data to get the maximum
results.
KEY WORDS
Data: Quantitative or/ and qualitative information, collected for study and
analysis.
Case study: the exploratory research technique that intensively investigates one
or a few situations similar to the problem situation.
Coding: The process of identifying and assigning score or other character
symbol to previously edited data.
Editing: the process of checking completeness, consistency, and legibility of
data and making the data ready for coding and transfer to storage.
Interview: A method of collecting primary data by meeting the informants and
asking the questions.
Observation: The process of observing individuals in controlled situations.
Questionnaire: is a device for collection of primary data containing a list of
questions pertaining to enquiry, sent to the informants, and the informant
himself writes the answers.
Primary Data: Data that is collected originally for the first time.
Secondary Data: Data which were collected and processed by someone else but
are being used in the present study.
Published Sources: Sources which consist of published statistical information.
Schedule: is a device for collection of primary data containing a list of
questions to be filled in by the enumerators who are specially appointed for that
purpose.
Survey : a method of primary data collection in which information is gathered
by communicating with a representative sample of people.
Questions for practice
1. What are the qualities of a good research instrument?
2. Enumerate the different methods of data collection.
3. Explain the merits and limitations of using secondary data.
4. What precautions would you take while using the data from secondary
sources?
5. Explain the various sources of primary data.
6. What is observation? Explain the role of observer in the process of
observation.
7. Examine the merits and limitations of the observation method.
8. Explain different methods of conducting interviews.
9. How does the case study method differ from the survey method? analyse
the merits and limitations of case study method in marketing research
10. What are the guidelines in the construction of questionnaire? Explain.
11. What is field operations? Explain the cautions to be taken by the
researcher in field operations.
12. What is data processing? Explain the various functions of data
processing.
13. Explain the importance of data editing in marketing research.
14. Explain the main sources of error in field operations.
15. its is never safe to take published statistics at their face value without
knowing their meaning and limitations. Elucidate this statement by
enumerating and explaining the various points which you would consider
before using any published data. Illustrate your answer by examples
wherever possible.
G.Thamizh chelvan
S.G.Lecturer in commerce and International Business
Dr.G.R.D.Col ege of science
Avinashi road
Coimbatore
UNIT V - MARKETING RESEARCH
CHAPTER I
Data Analysis ? Univariate Analysis
Objective:
The objective of this chapter is to understand:
Methods of analysing Market Research data
Methods of Univariate Analysis
Introduction:
After collection of data from marketing research, the data has to be analyzed.
This can be carried out by various statistical methods.
Data analysis begins with univariate analysis. Univariate analysis also is the
foundation for the bivariate and multivariate analysis.
Univariate analysis is the assessment of the distributional properties of a
variable. It serves two broad purposes: description and preparation for
multivariate analysis. These functions correspond to the two primary forms of
univariate analysis, the assessment of central tendency and of dispersion, or
convergence and divergence. This mainly deals with the meaning of a typical
value and to what extent do values differ from this typical value. Descriptive
research focuses on identifying what is most characteristic of a set of
observations. Any variation from this typical value usually is the most important
concern with regard to subsequent multivariate analysis.
There is a possibility that many times univariate analysis itself is the research
goal. For example, we might calculate the percentage of women going for work.
Descriptive research emphasizes what are most typical using estimates of central
tendency.
When univariate analysis is preliminary to multivariate analysis, dispersion
takes center stage. This analysis often uncovers at least some technical problems
that need to be resolved before other forms of analysis can proceed.
Measures of Central Tendency
Measures of central tendency summarize the entire distribution of values as one
single quantity or quality that can be thought of as the average value. Measures
of central tendency are measures of the location of the middle or the center of a
distribution. The mean is the most commonly used measure of central tendency.
The three most commonly-used measures of central tendency are:
1. Mean
a. Arithmetic Mean
The arithmetic mean is commonly called the average. The sum of the
values divided by the number of values--often called the "average."
The mean is the sum of all the scores divided by the number of
scores.
This is denoted as
= X/N
where is the population mean and N is the number of scores. If the
scores are from a sample, then the symbol M refers to the mean and
N refers to the sample size. The formula for M is the same as the
formula for .
Example: The mean of 5, 10, 22, 25, 17 is (5 + 10 + 22 + 25 + 17) / 5
= 15.8.
The mean is a good measure of central tendency for roughly
symmetric distributions. This can be misleading in skewed
distributions since it can be greatly influenced by extreme scores.
Therefore, other statistics such as the median may be more
informative for distributions such as reaction time or family income
that are frequently much skewed.
b. Geometric Mean
The geometric mean is the nth root of the product of the scores.
Thus, the geometric mean of the scores: 1, 2, 3, and 10 is the fourth
root of 1 x 2 x 3 x 10 which is the fourth root of 60 which equals
2.78. The formula can be written as:
Geometric mean =
where X means to take the product of all the values of X.
The geometric mean can also be computed by:
1. taking the logarithm of each number
2. computing the arithmetic mean of the logarithms
3. raising the base used to take the logarithms to the arithmetic
mean.
c. Harmonic Mean
The harmonic mean is used to take the mean of sample sizes. If there
are k samples each of size n, then the harmonic mean is defined as:
For the numbers 1, 2, 3, and 10, the harmonic mean is:
= 2.069. This is less than the geometric mean of 2.78 and the
arithmetic mean of 4.
The arithmetic mean in the case of a frequency distribution is calculated using
the following steps:
The middle point of each class interval is found and is multiplied by the
number of observations (frequencies) in that class.
The resultant values are summed up.
The total is divided by the number of observations.
h
fix
i i
x
X = i=1
=
n
where,
X = the sample mean
fi = the frequency of the i th class
xi= the mid-point of the i th class
h = the number of classes
n= the total number of observations in the sample
2. Median
The median is the middle score of a distribution: half the scores are above
the median and half are below the median. The median is less sensitive to
extreme scores than the mean and this makes it a better measure than the
mean for highly skewed distributions. The median income is usually more
informative than the mean income, for example.
Example: The median of the same five numbers (5, 10, 22, 25, and 17) is 17.
The sum of the absolute deviations of each number from the median is lower
than is the sum of absolute deviations from any other number. The mean,
median, and mode are equal in symmetric distributions. The mean is higher
than the median in positively skewed distributions and lower than the
median in negatively skewed distributions
For a grouped series, the median is calculated as:
l -l
-
M =
M l
= + 2 1
+
(m
( -
m c
- )
c
1
f
where,
M = Median
l1 = the lower limit of the class in which the median lies
l2 = the upper limit of the class in which the median lies
f = the frequency of the class in which the median lies
m = the middle item or n/2
c = the cumulative frequency of the class preceding the one in which the median
lies.
3. Mode
The third common measure of central tendency is the mode. The advantage
of the mode as a measure of central tendency is that its meaning is obvious.
Further, it is the only measure of central tendency that can be used with
nominal data.
Some of the important characteristics of the mode are:
It can be applied to both qualitative and quantitative distribution
It is not affected by the extreme values in the distribution
It can be ascertained in an open-ended distribution
Mode denotes the most frequently-occurring value (or values). The mode is
the value (or values) with the highest frequency.
Example: For men having the following ages -- 19, 18, 21, 23, 23, 23, 24
and 21, the mode is 23.
In case of a grouped data, the mode is calculated using the following
formula:
f -f
-
1 0
Mode = l
M
+
X i
X
1
(f -f
- ) ?
) (f -f
- )
1 0
1 2
where,
l1 = the lower value of the class in which the mode lies
f1 = the frequency of the class in which the mode lies
f0 = the frequency of the class preceding the modal class
f2 = the frequency of the class succeeding the modal class
i = the class interval of the modal class
The mode is greatly subject to sample fluctuations and is therefore not
recommended to be used as the only measure of central tendency. A further
disadvantage of the mode is that many distributions have more than one mode.
These distributions are called "multimodal." In contrast to the mode, the median
and the mean both pertain exclusively to ordered data.
The choice of a measure of central tendency depends upon the level of
measurement (nominal, ordinal, interval, or ratio) of the variable and the shape
of its distribution. The mode is the only indicator of central tendency for a
nominal variable. It may be computed for other types of variables as well, but is
not especially useful unless there is a distinct peak, that is, when one value
clearly predominates. At least an ordinal level of measurement is required for
the median; the mean additionally requires an interval level of measurement. In
general, the mean is preferred over the median and the median over the mode
because the mean utilizes the most information about the distribution whereas
the mode preserves only one piece of information. This also depends on the
manner in which the variable is distributed.
It is necessary to examine the entire set of values to determine which value best
typifies the set as a whole. It is inadvisable to calculate a measure of central
tendency for a variable without first examining its distribution.
Dispersion
As seen earlier, the measures of central tendency are used to estimate "normal"
values of a dataset. Measures of dispersion are important for describing the
spread of the data, or its variation around a central value. Two distinct samples
may have the same mean or median, but completely different levels of
variability, or vice versa. A proper description of a set of data should include
both of these characteristics. There are various methods that can be used to
measure the dispersion of a dataset, each with its own set of advantages and
disadvantages.
1. Range
It`s calculation is one of the simplest. It is defined as the difference between the
largest and smallest sample values. The range depends only on extreme values
and provides no information about how the remaining data is distributed.
2. Variance and Standard Deviation
The standard deviation is the square root of the sample variance. These
measures of dispersion are very important.
What the formula means:
(1) xr - means take each value in turn and subtract the mean from each value.
(2) (xr - )? means square each of the results obtained from step (1). This is to
get rid of any minus signs.
(3) (xr - )? means add up all of the results obtained from step (2).
(4) Divide step (3) by n, which is the sum of the numbers
(5) For the standard deviation, square root the answer to step (4).
Grouped Data
The formula for the standard deviation when the data is grouped is:
Example:
Find out the variance for the following. The table shows scores (out of 10)
obtained by 20 people in a test
Scores (x) Frequency (f)
1
0
2
1
3
1
4
3
5
2
6
5
7
5
8
2
9
0
10
1
Solution:
Frequency
Scores (x) (f)
fx
fx2
1
0
0
0
2
1
2
4
3
1
3
9
4
3
12
48
5
2
10
50
6
5
30
180
7
5
35
245
8
2
16
128
9
0
0
0
10
1
10
100
20
118
764
Variance = 764 - (118)?
20 ( 20 )
= 38.2 - 34.81 = 3.39
Summary:
The chapter dealt with the measures of central tendency which are the arithmetic
mean, the median and the mode. This was followed by the measures of
dispersion; the standard deviation and the coefficient of variation.
The measures of central tendency are used to estimate "normal" values of a
dataset. Measures of dispersion are important for describing the spread of the
data, or its variation around a central value. Two distinct samples may have the
same mean or median, but completely different levels of variability, or vice
versa.
Questions:
1. What are the measures of central tendency?
2. What is a mean? Explain the various forms of mean.
3. What are the characteristics of the median?
4. What are the characteristics of the mode?
5. What are the measures of the central tendency?
6. Explain standard deviation.
7. How you measure variance?
8. How are the measures of central tendency different from the measures of
dispersion?
CHAPTER II
HYPOTHESIS TESTING
Objective:
The objective of this chapter is to understand:
Meaning of hypothesis
Hypothesis testing
Types of hypothesis
Two types of error in hypothesis testing
One-tailed and two-tailed tests
Introduction:
A statistical hypothesis is an assumption about a population parameter. This
assumption may or may not be true.
The best way to determine whether a statistical hypothesis is true is to examine
the entire population. Since this is often impractical, researchers typically
examine a random sample from the population. If the sample data are consistent
with the statistical hypothesis, the hypothesis is accepted; if not, the hypothesis
is rejected.
There are two types of statistical hypothesis.
Null hypothesis. The null hypothesis is usually the hypothesis that
sample observations result purely from chance effects.
Alternative hypothesis. The alternative hypothesis is the hypothesis that
sample observations are influenced by some non-random cause.
For example, suppose we wanted to determine whether a coin was fair and
balanced. A null hypothesis might be that half the flips would result in Heads
and half, in Tails. The alternative hypothesis might be that the number of Heads
and Tails would be very different. Suppose we flipped the coin 50 times,
resulting in 40 Heads and 10 Tails. Given this result, we would be inclined to
reject the null hypothesis and accept the alternative hypothesis.
The term null means nothing or invalid. It may be written as:
H0: = 0
Where, H0 is the null hypothesis and 0 is the mean of the population.
The alternative hypothesis is
HA: 0
The rejection of the null hypothesis will show that the mean of the population is
not 0. If this happens, then it implies that the alternative hypothesis is accepted.
There can be more than two or more alternative hypothesis though only one can
be tested at a time against the null hypothesis.
Hypothesis Testing
Statisticians follow a formal process to determine whether to accept or reject a
null hypothesis, based on sample data. This process, called hypothesis testing,
consists of four steps.
Formulating the hypothesis: The first step in hypothesis testing is to
formulate the hypothesis to be tested. This means stating the null
hypothesis and the alternative hypothesis.
Identifying the test statistic: The test statistic is a statistic that will be
used by the researcher to determine whether the null hypothesis should
be accepted or rejected. Typically, the test statistic is the sample estimate
of the population parameter in the null hypothesis. Therefore, since we
are testing a hypothesis about a population mean, the test statistic will be
the sample mean. When the hypothesis pertains to a large sample (30 or
more), the z-test implying normal distribution is used. When a sample is
small (less than 30), the use of the z-test will be inappropriate. Instead
the t-test will be more suitable. The test criteria frequently used in
hypothesis testing are Z, t. F and 2.
Formulating a decision rule: The decision rule consists of two parts:
(1) a test statistic and (2) a range of values, called the region of
acceptance. The decision rule determines whether a null hypothesis is
accepted or rejected. If the test statistic falls within the region of
acceptance, the null hypothesis is accepted; otherwise, it is rejected.
Accepting or rejecting the null hypothesis: Once the region of
acceptance is defined, the null hypothesis can be tested against sample
data. The test statistic is computed. For example consider that the test
statistic is the sample mean. If the sample mean falls within the region of
acceptance, the null hypothesis is accepted; if not, it is rejected.
Decision Rules
There are four possibilities that can arise when a hypothesis is tested:
1. The hypothesis is true but our test leads to its rejection.
2. The hypothesis is false but our test leads to its acceptance.
3. The hypothesis is true and our test leads to its acceptance.
4. The hypothesis is false and our test leads to its rejection.
Out of these four, the first two lead to an error in decision. The first possibility
leads to a Type I error and the second possibility leads to a Type II error. This
can be shown as follows:
STAT
ST
E
AT OF
O
F NATU
N
R
ATU E
R
H Tru
r e
u
H Fa
F lse
a
o
o
Cor
Co r
r e
r c
e t tRe
t
Re en
e t
n io
t n
Typ
y e
p
e III IEr
r
Er or
Ret
Re a
t in
a
in H
o
Ho
DEC
D
I
EC SI
I ON
O
Typ
y e
p
e I IEr
r
Er or
o
Cor
Co r
r e
r c
e t tRe
j
Re ec
e t
c io
t n
Reje
Re c
je t
c tH
io
Ho
The decision rule is a procedure that a researcher uses to decide whether to
accept or reject the null hypothesis. There are two types of errors that can result
from a decision rule.
Type I error. A Type I error occurs when the researcher rejects a null
hypothesis when it is true. The probability of committing a Type I error
is called the significance level. This probability is also called alpha, and
is often denoted by .
Type II error. A Type II error occurs when the researcher accepts a null
hypothesis that is false. The probability of committing a Type II error is
called Beta, and is often denoted by . The probability of not committing
a Type II error is called the Power of the test.
In practice, the decision rule has two parts: (1) a test statistic and (2) a range of
values. The range of values is called the region of acceptance. The region of
acceptance is defined so that the chance of making a Type I error is equal to the
significance level. If the test statistic falls within the region of acceptance, the
null hypothesis is accepted.
Note: The set of values outside the region of acceptance is called the region of
rejection. If the test statistic falls within the region of rejection, the null
hypothesis is rejected. In such cases, we say that the hypothesis has been
rejected at the level of significance.
One-Tailed and Two-Tailed Tests
To understand the difference between these look at the table below. It shows
three sets of hypothesis. Each makes a statement about how the population mean
is related to a specified value M.
Set Null
Alternative Number of
hypothesis hypothesis tails
1
= M
M
2
2
> M
< M
1
3
< M
> M
1
The first set of hypotheses (Set 1) is an example of a two-tailed test, since an
extreme value on either side of the sampling distribution would cause a
researcher to reject the null hypothesis. The other two sets of hypotheses (Sets 2
and 3) are one-tailed tests, since an extreme value on only one side of the
sampling distribution would cause a researcher to reject the null hypothesis.
A test of a statistical hypothesis, where the region of rejection is on only one
side of the sampling distribution, is called a one-tailed test. For example,
suppose the null hypothesis states that the mean is less than or equal to 10. The
alternative hypothesis would be that the mean is greater than 10. The region of
rejection would consist of a range of numbers located on the right side of
sampling distribution; that is, a set of numbers greater than 10.
A test of a statistical hypothesis, where the region of rejection is on both sides of
the sampling distribution, is called a two-tailed test. For example, suppose the
null hypothesis states that the mean is equal to 10. The alternative hypothesis
would be that the mean is less than 10 or greater than 10. The region of rejection
would consist of a range of numbers located on both sides of sampling
distribution; that is, the region of rejection would consist partly of numbers that
were less than 10 and partly of numbers that were greater than 10.
Other Considerations
There may be other considerations to be made while testing the null hypothesis.
These are:
Assumptions need to be made about the sampling distribution of the
mean score. If the sample is relatively large (i.e., greater than or equal to
30), you can assume, based on the central limit theorem, that the
sampling distribution will be roughly normal.
On the other hand, if the sample size is small (less than 30) and if the
population random variable is approximately normally distributed (i.e., has a
bell-shaped curve), you can transform the mean score into a t score. The t
score will have a t distribution.
Assume that the mean of the sampling distribution is equal to the test
value M specified in the null hypothesis.
In some situations, the standard deviation needs to be computed from the
sampling distribution sx. If the standard deviation of the population is
known, then
sx = * sqrt[( 1/n ) - ( 1/N )]
where n is the sample size and N is the population size. On the other
hand, if the standard deviation of the population is unknown, then
sx = s * sqrt[( 1/n ) - ( 1/N )]
where s is the sample standard deviation.
Hypothesis testing in respect of interval data:
1. Test of sample mean (Single population):
On the basis of the sample drawn from the population one needs to infer about
the population parameter which could be the population mean. We need to carry
out an appropriate statistical test of significance for testing the hypothesis
concerning the population mean. We can consider two cases, sample size being
large (n>30) and sample size small (n30).
If n is large, the sample distribution of mean follows normal distribution as per
central limit theorem and we can go for Z test. However, when sample size is
small, we may have to choose between Z and t test depending upon whether
standard deviation () is known or not. In case is known, we should go for Z
test or t test.
In case of a normal distribution of sample means (x bar) with mean and
standard deviation is represented by
=
= / n
x
The standard normal distribution is calculated as:
x -
Z =
x
This is used for testing the hypothesis.
2. Test of Proportion:
In order to test the hypothesis that population proportion (P) takes a specified
value Po against a two sided alternative, we test the hypothesis as follows:
Ho (Null Hypothesis): = o
H1 (Alternative Hypothesis): o
Here it can be shown that:
E (P) = and
(1-)
=
P
n
Using the Z-test, we find that
P -
Z =
P
Summary:
A statistical hypothesis is an assumption about a population parameter. This
assumption may or may not be true.
There are two types of statistical hypothesis.
Null hypothesis. The null hypothesis is usually the hypothesis that
sample observations result purely from chance effects.
Alternative hypothesis. The alternative hypothesis is the hypothesis that
sample observations are influenced by some non-random cause.
This process, called hypothesis testing, consists of four steps: Formulating the
hypothesis, identifying the test statistic, formulating a decision rule and
accepting or rejecting the null hypothesis
There are two types of errors that can result from a decision rule: Type I error
which occurs when the researcher rejects a null hypothesis when it is true and
Type II error which occurs when the researcher accepts a null hypothesis that is
false.
A test of a statistical hypothesis, where the region of rejection is on only one
side of the sampling distribution, is called a one-tailed test. A test of a statistical
hypothesis, where the region of rejection is on both sides of the sampling
distribution, is called a two-tailed test.
Questions:
1. What is a hypothesis? Explain.
2. Differentiate between Null hypothesis and alternative hypothesis.
3. What are the steps involved in testing a hypothesis?
4. Is hypothesis testing useful in marketing decision making?
5. What are Type I error and Type II errors?
6. Explain the relationship between Type I and Type II errors.
7. What are one-tailed and two-tailed tests?
8. What are the other considerations to be made while testing the null
hypothesis?
9. How will you test the sample mean from a single population?
10. How will you test the hypothesis of a population proportion?
CHAPTER III
Bivariate Analysis
Objective:
The objective of this chapter is to understand:
Association between Dependent Variable (DV) and Independent
Variable (IV).
Importance and methodology of correlation analysis.
Importance and methodology of regression analysis.
Introduction:
In bivariate analysis, the hypothesis of "association" and causality are tested. In
its simplest form, association simply refers to the extent to which it becomes
easier to know/predict a value for the Dependent Variable (DV) if we know a
case's value on the Independent Variable (IV).
This association could be understood by a measure of association. A measure of
association often ranges between ?1 and 1. Where the sign of the integer
represents the "direction" of correlation (negative or positive relationships) and
the distance away from 0 represents the degree or extent of correlation ? the
farther the number away from 0, the higher or "more perfect" the relationship is
between the IV and DV.
Measures of association and statistical significance that are used vary by the
level of measurement of the variables analyzed.
For Nominal Variables:
Measure of association is Lambda
The formula for Lambda is
(Reduction in Error from guessing to predicting based on IV)
Number of Original Error
This gives a ratio of how much improvement your prediction has by knowing
values on the IV. Lambda ranges from 0 to 1. The higher the number, the
stronger the relationship between the two variables.
The measure of statistical significance for nominal variables (and limited scale
ordinal variables) is Chi Square. In fact, Chi Square can measure the statistical
significance of any crosstab. It tells us how different the values in the cells of a
crosstab are from expected values (or values predicted if no real relationship
between the two variables existed ? uses marginals to calculate these expected
values).
The Chi Square is based on these factors:
1. The distribution of cases among the cells (can show the extent to which
differences are observed;
2. The number of cells (degrees of freedom), and
3. The size of the sample (n).
The Chi Square requires the following steps:
1. State the null hypothesis and calculate the number in each category
assuming that the null hypothesis is correct.
2. Determine the level of significance.
3. Calculate the Chi Square (x2) as follows
k (O
(
? E )2
)
i
i
X2
X =
i=1
Ei
Where, Oi is the observed frequency in the ith category
Ei is the expected frequency in the ith category
k is the number of categories
4. Determine the number of degrees of freedom. For the specified level of
significance and the degrees of freedom, find the critical or theoretical
value of x2.
5. The calculated value of the x2 is then calculated with the theoretical
value of x2 and the region of rejection is determined.
For ordinal variables:
The appropriate measures of association all attempt to measure how values of
ordered variables relate for the sample of cases. For instance, how many times
high values are associated with high values, how many times are they associated
with low values. They each use discordant and concordant pairs to create a
value between ?1 and 1. O indicates no relationship between how the values for
the cases pair up. The closer to ?1 means the stronger the negative (inverse)
relationship, and the closer to 1 the more "perfect" the positive relationship.
There are several measures of association which measure ordinal variables'
relationships. Somer's d, tau b, tau c, and gamma are the most usual ones. All
are slight variations of the formula related in layman's terms above. While
Somer's d, tau b, and tau c will all have very close to the same value, gamma
usually will appear to have a slightly higher (stronger) relationship.
For interval variables
Regression and correlation analysis are used for this. The two terms correlation
and regression are distinct from each other and many times are used
interchangeably. Correlation is a statistical technique used for measuring the
relationship or interdependence of two or more variables. Correlation does not
necessarily indicate a causal relationship between two or more variables.
Regression analysis refers to the technique for deriving an equation that relates
the dependent variable to one or more variables. It is used to predict one variable
on the basis of another and hence helps in bringing out the causal relationship
between two or more variables.
Correlation Analysis:
Correlation is a statistical technique used for measuring the relationship or
interdependence of two or more variables. The measurement scales used should
be at least interval scales, but other correlation coefficients are available to
handle other types of data. Correlation coefficients can range from -1.00 to
+1.00. The value of -1.00 represents a perfect negative correlation while a value
of +1.00 represents a perfect positive correlation. A value of 0.00 represents a
lack of correlation.
In statistics, the Correlation coefficient (r) is a measure of how well a linear
equation describes the relation between two variables X and Y measured on the
same object or organism. It is defined as the sum of the products of the standard
scores of the two measures divided by the degrees of freedom:
The result obtained is equivalent to dividing the covariance between the two
variables by the product of their standard deviations. In general the correlation
coefficient is one of the two square roots (either positive or negative) of the
coefficient of determination (r2), which is the ratio of explained variation to total
variation:
where:
Y = a score on a random variable Y
Y' = corresponding predicted value of Y, given the correlation of X and Y and
the value of X
= sample mean of Y (i.e., the mean of a finite number of independent
observed realizations of Y, not to be confused with the expected value of Y)
The correlation coefficient adds a sign to show the direction of the relationship.
This applies when the relationship is linear.
The coefficient ranges from -1 to 1. A value of 1 shows that a linear equation
describes the relationship perfectly and positively, with all data points lying on
the same line and with Y increasing with X. A score of -1 shows that all data
points lie on a single line but that Y increases as X decreases. A value of 0
shows that a linear model is inappropriate ? that there is no linear relationship
between the variables.
If we have a series of n measurements of X and Y written as xi and yi where i
= 1, 2, ..., n, then the coefficient can be used to estimate the correlation of X and
Y . This is also known as the "sample correlation coefficient". It is especially
important if X and Y are both normally distributed. The Pearson correlation
coefficient is then the best estimate of the correlation of X and Y . The Pearson
correlation coefficient is written:
where and are the sample means of xi and yi , sx and sy are the sample
standard deviations of xi and yi and the sum is from i = 1 to n.
This can be written as:
The absolute value of the sample correlation must be less than or equal to 1
Regression Analysis:
In statistics, regression analysis is used to model relationships between random
variables, determine the magnitude of the relationships between variables, and
can be used to make predictions based on the models.
Regression analysis models the relationship between one or more response
variables (also called dependent variables) usually named Y, and the predictors
(also called independent variables) usually named X1,...,Xp.
A line in a two dimensional or two-variable space is defined by the equation
Y=a+b*X; in full text: the Y variable can be expressed in terms of a constant (a)
and a slope (b) times the X variable. The constant is also referred to as the
intercept, and the slope as the regression coefficient or B coefficient.
Regression analysis can predict the outcome of a given key business indicator
(dependent variable) based on the interactions of other related business drivers
(independent variable). For example, it allows predicting sales volume, using the
amount spent on advertising and number of sales people that a company
employs.
The coefficient of determination, r 2, is useful because it gives the proportion of
the variance (fluctuation) of one variable that is predictable from the other
variable. It is a measure that allows us to determine how certain one can be in
making predictions from a certain model/graph. The coefficient of determination
is the ratio of the explained variation to the total variation. The coefficient of
determination is such that 0 < r 2 < 1, and denotes the strength of the linear
association between x and y. The coefficient of determination represents the
percent of the data that is the closest to the line of best fit.
Every sample has some variation in it. The total variation is made up of two
parts, the part that can be explained by the regression equation and the part that
can't be explained by the regression equation.
In regression analysis, the coefficient of determination is a measure of goodness-
of-fit (i.e. how well or tightly the data fit the estimated model). The coefficient
is defined as the ratio of two sums of squares:
r2 = SSR / SST
where SSR is the sum of squares due to regression, SST is the total sum of
squares. By "sum of squares" we mean the sum of squared deviations between
actual values and the mean (SST), or between predicted values and the mean
(SSR). The coefficient of determination takes on values between 0 and 1, with
values closer to 1 implying a better fit.
The objectives for which market researchers use regression analysis are:
1. Pattern connecting the dependent variable and the independent variable
by establishing a functional relationship between the two.
2. For solving problems involving prediction and forecasting.
3. To study the quantum of variation in the dependent variable using the set
of independent variables.
The Scatter Diagrams:
The Scatter Diagram is a tool for determining the potential correlation between
two different sets of variables, i.e., how one variable changes with the other
variable. This diagram simply plots pairs of corresponding data from two
variables. The scatter diagram does not determine the exact relationship
between the two variables, but it does indicate whether they are correlated or
not. It also does not predict cause and effect relationships between these
variables.
The scatter diagram is used to:
1) Quickly confirm a hypothesis that two variables are correlated;
2) Provide a graphical representation of the strength of the relationship between
two variables; and
3) Serve as a follow-up step to a cause-effect analysis to establish whether a
change in an identified cause can indeed produce a change in its identified
effect.
To make a scatter diagram for two variables requiring confirmation of
correlation, the following simple steps are usually followed:
1) 50-100 pairs of data are collected for this and tabulated;
2) The x- and y-axes of the diagram, along with the scales that
increase to the right for the x-axis and upward for the y-axis are
drawn;
3) The data for one variable is assigned to the x-axis (the
independent variable) and the data for the other variable is
assigned to the y-axis (the independent variable);
4) Plot the data pairs on the scatter diagram, encircling (as many
times as necessary) all data points that are repeated.
Interpretation of the resulting scatter diagram is very simple and can be analyzed
looking at the pattern formed by the points. If the data points plotted on the
scatter diagram are all over the place with no discernible pattern whatsoever,
then there is no correlation at all between the two variables of the scatter
diagram. An example of a scatter diagram that shows no correlation is shown
below
The below scatter diagrams show positive correlation (Strongly positive and
weak positive correlation):
The below scatter diagrams show negative correlation (Strongly negative and
weak negative correlation):
If the variables are correlated, the points will fall along a line or curve. The
better the correlation, the tighter the points will hug the line.
Summary:
In bivariate analysis, the hypothesis of "association" and causality are tested. In
its simplest form, association simply refers to the extent to which it becomes
easier to know/predict a value for the Dependent Variable (DV) if we know a
case's value on the Independent Variable (IV).
The two terms correlation and regression are distinct from each other and many
times are used interchangeably. Correlation is a statistical technique used for
measuring the relationship or interdependence of two or more variables.
Regression analysis refers to the technique for deriving an equation that relates
the dependent variable to one or more variables. It is used to predict one variable
on the basis of another and hence helps in bringing out the causal relationship
between two or more variables.
The Scatter Diagram is a tool for determining the potential correlation between
two different sets of variables, i.e., how one variable changes with the other
variable. This diagram simply plots pairs of corresponding data from two
variables.
Questions:
1) What is a chi-square test? What are the uses of the same?
2) What is correlation analysis?
3) What is regression analysis?
4) What is the difference between correlation and regression analysis?
5) What are scatter diagrams? What do they depict?
CHAPTER IV
Multivariate Analysis
Objective:
The objective of this chapter is to understand:
The process of Multivariate analysis
Study about analysis of variation, Multiple Regression Analysis,
Discriminant Analysis, Conjoint Analysis, Factor Analysis, Cluster
Analysis and Multi-dimensional Scaling
Introduction:
Multivariate analysis is the analysis of the simultaneous relationships among
three or more phenomena. In Univariate analysis the focus is on the level
(average) and distribution (variance) of the phenomenon, in a bivariate analysis
the focus shifts to the degree of relationships (correlations or co variances)
between the phenomena. In a multivariate analysis, the focus shifts from paired
relationships to more complex simultaneous relationships among phenomena.
The various methods of Multivariate analysis are:
1. Analysis of Variance
2. Dependence Methods
Multiple Regression Analysis
Discriminant Analysis
Conjoint Analysis
3. Inter Dependence Methods
Factor Analysis
Cluster Analysis
Multi-dimensional Scaling
Analysis of Variance
This can be studied under ANOVA, ANCOVA, MANOVA and MANCOVA.
ANOVA: An ANOVA (Analysis of Variance), sometimes called an F test, is
closely related to the t test. The t test measures the difference between the means
of two groups whereas an ANOVA tests the difference between the means of
two or more groups.
ANOVA can be a one-way ANOVA or a factorial ANOVA.
One-way or single factor ANOVA, tests differences between groups that are
only classified on one independent variable. A factorial ANOVA can show
whether there are significant main effects of the independent variables and
whether there are significant interaction effects between independent variables
in a set of data. Interaction effects occur when the impact of one independent
variable depends on the level of the second independent variable.
The advantage of using ANOVA rather than multiple t-tests is that it reduces the
probability of a type-I error. Making multiple comparisons increases the
likelihood of finding something by chance--making a type-I error.
An F indicates that there is a significant difference between groups, not which
groups are significantly different from each other. This is one potential
drawback to an ANOVA, which is loss of specificity.
A factorial ANOVA can examine data that are classified on multiple
independent variables. More than two independent variables can be compared in
an ANOVA (e.g., three-way, four-way).
ANCOVA: In ANCOVA, we can analyze both qualitative (class) and
quantitative (continuous) independent variables. The mixed procedure allows
the user to model both class and continuous variables. In ANOVA-type models,
hypotheses about class or interactions among class variables are tests of means
or differences among means. In regression-type models, in which all factors
are continuous variables (rather than categories), hypothesis tests are tests about
regression coefficients.
As one might expect, the assumptions of ANCOVA combines both the
assumptions of regression and ANOVA. In addition, the tests of adjusted means
are based on the assumption that the class variable by covariate interaction is
negligible, that is the regression lines are parallel.
MANOVA: Multivariate analysis of variance (MANOVA) is simply an
ANOVA with several dependent variables. For example, we may conduct a
study where we try two different textbooks, and we are interested in the students'
improvements in Physics and Chemistry. In that case, improvements in physics
and chemistry are the two dependent variables, and our hypothesis is that both
together are affected by the difference in textbooks. A multivariate analysis of
variance (MANOVA) could be used to test this hypothesis. Instead of a
univariate F value, we would obtain a multivariate F value (Wilks' lambda)
based on a comparison of the error variance/covariance matrix and the effect
variance/covariance matrix. The "covariance" here is included because the two
measures are probably correlated and we must take this correlation into account
when performing the significance test.
MANOVA is useful in experimental situations where at least some of the
independent variables are manipulated. It has several advantages over ANOVA.
1. By measuring several dependent variables in a single experiment, there
is a better chance of discovering which factor is truly important.
2. It can protect against Type I errors that might occur if multiple
ANOVA`s were conducted independently. Additionally, it can reveal
differences not discovered by ANOVA tests.
However, there are several cautions as well. It is a substantially more
complicated design than ANOVA, and therefore there can be some ambiguity as
to which independent variable affects each dependent variable. Moreover, one
degree of freedom is lost for each dependent variable that is added. Finally, the
dependent variables should be largely uncorrelated. If the dependent variables
are highly correlated, there is little advantage in including more than one in the
test given the resultant loss in degrees of freedom.
Some of the assumptions made here are:
Normal Distribution: The dependent variable should be normally
distributed within groups. Overall, the F test is robust to non-normality
if it is caused by skewness rather than outliers. Tests for outliers should
be run before performing a MANOVA, and outliers should be
transformed or removed.
Homogeneity of Variances: Homogeneity of variances assumes that the
dependent variables exhibit equal levels of variance across the range of
predictor variables.
Homogeneity of Variances and Covariance`s: In multivariate designs,
with multiple dependent measures, the homogeneity of variances
assumption described earlier also applies. However, since there are
multiple dependent variables, it is also required that their covariances are
homogeneous across the cells of the design. There are various specific
tests of this assumption.
Two Special Cases arise in MANOVA:
Unequal sample sizes: As in ANOVA, when cells in a factorial
MANOVA have different sample sizes, the sum of squares for effect
plus error does not equal the total sum of squares. This causes tests of
main effects and interactions to be correlated.
Within-subjects design: Problems arise if the researcher measures several
different dependent variables on different occasions.
MANCOVA: MANCOVA is an extension of ANCOVA. It is simply a
MANOVA where the artificial direct variables are initially adjusted for
differences in one or more covariates. This can reduce error "noise" when error
associated with the covariate is removed.
MULTIPLE REGRESSION ANALYSIS
Multiple regression is used to account for (predict) the variance in an interval
dependent, based on linear combinations of interval, dichotomous, or dummy
independent variables. Multiple regression can establish that a set of
independent variables explains a proportion of the variance in a dependent
variable at a significant level (through a significance test of R2), and can
establish the relative predictive importance of the independent variables.
The multiple regression equation takes the form
y = a + b1x1 + b2x2 + ... + bnxn
where y is the dependent variable which is to be predicted, x1, x2 and xn are the n
known variables on which the predictions are to be based and a, b1, b2, ....bn are
parameters, the values of which are to be determined by the methods of least
squares.
Associated with multiple regression is r2 (multiple correlation), which is the
percent of variance in the dependent variable explained collectively by all of the
independent variables.
(Y ? Y)2
Y) ? (Y ? Y )2
)
r2
r =
i
i
c
=
(Y ? Y)2
Y)
i
where r2 is the co-efficient of determination, Yi is the value of ith item in Y
series, Y (bar) is the mean of the Y series and Yc is the computed value of the ith
item in Y series on the basis of the regression.
Multiple regression shares all the assumptions of correlation: linearity of
relationships, the same level of relationship throughout the range of the
independent variable, interval or near-interval data, absence of outliers, and data
whose range is not truncated. In addition, it is important that the model being
tested is correctly specified. The exclusion of important causal variables or the
inclusion of extraneous variables can change the interpretation of the importance
of the independent variables.
Multiple regressions with dummy variables yield the same inferences as
multiple analysis of variance (MANOVA), to which it is statistically equivalent.
When the dependent variable is a dichotomy the assumptions of multiple
regressions cannot be met, discriminant analysis or logistic regression is used
instead. Partial least squares regression is sometimes used to predict one set of
response variables from a set of independent variables
DISCRIMINANT ANALYSIS
Discriminant analysis is used to classify the sample into two or more categories.
Example: Consumers may be classified as heavy and light users; Sales people
can be classified as successful and unsuccessful and so on.
Discriminant function analysis is used to determine which variables discriminate
between two or more naturally occurring groups.
For example, a researcher may want to investigate which variables discriminate
between engineers who decide
(1) To seek employment in private companies,
(2) To take up government services, or
(3) To seek opportunities abroad.
For that purpose the researcher could collect data on numerous variables after
the graduation of the engineers. Discriminant Analysis could then be used to
determine which variable(s) are the best predictors of the engineers` choice of
employment.
The objectives of two group discriminant analysis are to find a linear composite
of the predictor variable to help the analyst to separate the groups, establishing
procedures for assigning new individuals, testing for significant differences
between the mean predictor variables and determining the variable which
accounts for the most intergroup differences.
This is commonly carried out with the help of a computer program.
CONJOINT ANALYSIS
Conjoint analysis deals with the measurement of the combined effect of two or
more attributes that are important from the view of the consumer. The use of the
conjoint analysis will be more appropriate in a situation where a company would
like to know the most desirable attribute for a new product or service.
For example, a hotel would like to know whether choice of menu or prompt
service would attract a customer to visit them frequently.
For this, it will seek data from the consumer in the firm of response to identify
product attributes. The various options available for this are direct interview
with the customer or focus group interviews. All the attributes are weighed and
compared.
The main steps involved in the application of conjoint analysis are
1. Determination of salient attributes: the attributes have to be selected
based on the marketers experience or through interviews. Only valuable
attributes need to be considered.
2. Assigning levels to the selected attributes: this can vary from most
preferred to the least preferred.
3. Fractional factorial design of experiments: during the comparison of
the profile of different products, it is essential to have a minimum
number of designs which provide us all the information required. This
will ensure easy management of the design.
4. Physical design of the stimuli: a prototype or a picture of the concept
may be given to the consumer or customer to get a realistic picture.
5. Data collection: The customers are asked to rank all the alternatives
using a rating scale. This will ensure ease of data collection and analysis.
6. Determination of part-worth utilities: Regression methods,
mathematical programming methods, econometric methods may be used
for the part-worth utility values.
The applications of conjoint analysis are
It can be used for optimum product design based on the attributes
considered. Simulations can be created to represent competitors`
action or a fresh scenario.
Consumers can be segmented based on their sensitivity to
product attributes.
It may help a manager to conduct SWOT analysis of the brand as
the part-worth utility speaks about the relative brand strength.
There are certain limitations of conjoint analysis. It may not be perfect and
convincing and may fail to capture utility functions and decision roles.
FACTOR ANALYSIS
Factor analysis is a name given to a class of techniques whose purpose is data
reduction and summarization. The data from market research are vast and factor
analysis helps in reducing the number of variables. Factor analysis is an
explorative technique.
Factor analysis was invented nearly 100 years ago by psychologist Charles
Spearman, who hypothesized that the enormous variety of tests of mental
ability--measures of mathematical skill, vocabulary, other verbal skills, artistic
skills, logical reasoning ability, etc.--could all be explained by one underlying
"factor" of general intelligence that he called g. He hypothesized that if g could
be measured and you could select a subpopulation of people with the same score
on g, in that subpopulation you would find no correlations among any tests of
mental ability. In other words, he hypothesized that g was the only factor
common to all those measures.
The objectives of factor analysis are simplifying the data by reducing a large
number of variables to a set of a small number of variables and analyzing the
interdependence of relationship among a total set of variables.
Factor analysis can be used in several ways as given below:
1. It brings out the hidden dimensions relevant to a researcher among
product preferences.
2. Helps to find out relationships among observed values.
3. Used when the data is large and has to be simplified and condensed.
The limitations of factor analysis are:
It is a complicated tool and should be used if the researcher has a
good understanding of the technique.
The reliability of the results is some times questionable.
It suitability depends on the judgment of the researcher.
Factor analysis is used in the case of exploratory research and has to be used
were the concepts are well formulated and tested.
CLUSTER ANALYSIS
Cluster analysis is a technique used to segment a market. It is used to classify a
person or object into a small number or mutually exclusive and exhaustive
groups. Its object is to sort cases (people, things, events, etc) into groups, or
clusters, so that the degree of association is strong between members of the same
cluster and weak between members of different clusters. Each cluster thus
describes, in terms of the data collected, the class to which its members belong;
and this description may be abstracted through use from the particular to the
general class or type.
CA lacks an underlying body of statistical theory and is heuristic in nature.
Cluster analysis requires decisions to be made by the user relating to the
calculation of clusters, decisions which have a strong influence on the results of
the classification. CA is useful to classify groups or objects and is more
objective than subjective.
Cluster analysis, like factor analysis and multi dimensional scaling, is an
interdependence technique: it makes no distinction between dependent and
independent variables. The entire set of interdependent relationships is
examined. It is similar to multi dimensional scaling in that both examine inter-
object similarity by examining the complete set of interdependent relationships.
The difference is that multi dimensional scaling identifies underlying
dimensions, while cluster analysis identifies clusters. Cluster analysis is the
obverse of factor analysis. Whereas factor analysis reduces the number of
variables by grouping them into a smaller set of factors, cluster analysis reduces
the number of observations or cases by grouping them into a smaller set of
clusters.
In marketing, cluster analysis is used for:
Segmenting the market and determining target markets
Product positioning and New Product Development
Selecting test markets
Example: A supermarket might gather data on all of their existing customers
and survey them regarding their buying criteria relative to the product line. They
could then use cluster analysis to group customers with similar buying patterns
together. This type of cluster analysis, also known as market segmentation, is
performed at increasing rates, due to the advent of high-speed computers and the
ready availability of demographic data. Based on the broader descriptions of
individuals within each cluster, the retail managers could make decisions that
would be appropriate for the individuals within.
Clusters for this example might include:
Price-sensitive shoppers
Indifferent shoppers
Quality-focused shoppers
High-end status shoppers
Monthly shopper
Clustering methods may be top down and employ logical division, or bottom up
and undertake aggregation. Aggregation procedures which are based upon
combining cases through assessment of similarities are the most common and
popular will be the focus of this section.
Care should be taken that groups (classes) are meaningful in some fashion and
are not arbitrary or artificial. To do so the clustering techniques attempt to have
minimal internal variation as compared to maximal variation between groups.
Homogeneous and distinct groups are delineated based upon assessment of
distances or an F-test.
Steps in Cluster Analysis:
The two key steps within cluster analysis are the measurement of distances
between objects and to group the objects based upon the resultant distances
(linkages).
The distances provide for a measure of similarity between objects and may be
measured in a variety of ways, such as Euclidean and Manhattan metric
distance. The criteria used to then link (group) the variables may also be
undertaken in a variety of manners, as a result significant variation in results
may be seen.
Linkages are based upon how the association between groups is measured. For
example, simple linkage or nearest neighbor distance, measures the distance to
the nearest object in a group while furthest neighbor linkage or complete
linkage, measures the distance between furthest objects. These linkages are both
based upon single data values within groups, whereas average between group
linkages is based upon the distance from all objects in a group. Centroid linkage
has a new value, representing the group centroid, which is compared to the
ungrouped point to weigh inclusion.
Ward's method is variance based with the groups variance assessed to enable
clustering. The group which sees the smallest increase in variance with the
iterative inclusion of a case will receive the case. Ward's is a popular default
linkage which produces compact groups of well distributed size. Standardization
of variables is undertaken to enable the comparison of variables to minimize the
bias in weighting which may result from differing measurement scales and
ranges. Z score format accounts for differences between mean values and
reduces the standard deviation when variables have multivariate normality
Choosing number of groups:
The ideal number of groups to establish may be assessed graphically or
numerically. Graphically the number of groups may be assessed with an icicle
plot or dendrogram. The dendrogram bisected at a point which will divide the
cases into a cluster based upon groupings up to the point where the bisection
occurred. Numerically the number of cases may be assessed on the
agglomeration schedule, by counting up from the bottom to where a significant
break in slope (numbers) occurs.
The optimal number of groups may be assessed based upon knowledge of the
data set. Discriminant analysis may also be employed to assess optimality and
efficiency of computed groups, by imputing the cluster analysis derived classes
for analysis with the original data.
Like the other techniques, cluster analysis presents the problem of how many
factors, or dimensions or clusters to keep. One rule followed here is to choose a
place where the cluster structure remains stable for a long distance. Some other
possibilities are to look for cluster groupings that agree with existing or expected
structures, or to replicate the analysis on subsets of the data to see if the
structures emerge consistently.
MULTI-DIMENSIONAL SCALING
Multi-dimensional scaling (MDS) or perceptual map or positioning map is used
for measuring human perception and preferences. It is spatial representation of
relationships. It helps in the identification of attributes and the positioning of
different products or brands on the basis of these attributes.
MDS is of two types, metric MDS and non metric MDS.
Given below is an example of a perceptual map. AC represents the x-axis and
BD represents the y-axis. The value of the variable may be low to high from one
end of the axis to the other.
C
D
B
A
Two approaches can be used for analyzing multi dimensional data. It can be
done by measuring the attributes or distance between objects.
For MDS, a set of number called proximities and a computer based algorithms
must be available.
The applications of MDS in marketing are in market segmentation, vendor
evaluation, attitude scaling, advertisement evaluation, product repositioning,
new product development and test marketing. Thus MDS measures the
psychological distance or the dis-similarities to evaluate the external
environment.
The limitations of MDS are
Concepts of similarity and preferences may differ
The selection of attributes are subjective
It is at time difficult to interpret the results
Different computer programmes may produce different results
Summary
This chapter deals with the different methods used in multivariate analysis and
briefly describes their application. The methods are chosen based on the data
and requirements of analysis.
Questions:
1. What is multivariate analysis? Explain.
2. Briefly explain the methods of multivariate analysis.
3. Explain variance of analysis. Write a note on the different types.
4. What is multiple regression analysis?
5. What is discriminant analysis? How is it useful in Marketing?
6. What are the steps involved in the application of conjoint analysis?
7. What are the limitations and uses of factor analysis?
8. What is cluster analysis? What are the steps involved in cluster
analysis?
9. What is Multi-dimensional scaling? Where is it used?
10. How do you choose a method for analysis?
CHAPTER V
REPORT WRITNG
Objective:
The objective of this chapter is to understand:
Importance of report writing
Types of reports
Considerations for oral and written reports
Steps in the preparation of the report
Presentation of data and feedback
Report Writing
It is not sufficient if the market researcher has collected information on the
research problem, he also has to interpret the data and draw specific conclusions
from it. The results of marketing research must be effectively communicated to
management. This report has to be clear and concise.
Types of Report:
Reports can be broadly classified into two types: oral and written.
Presenting the results of a marketing research study to management generally
involves a formal written report as well as an oral presentation. The report and
presentation are extremely important. This is because:
1. The results of marketing research are often intangible (there is very little
physical evidence of the resources, such as time and effort, that went into
the project); the written report is usually the only documentation of the
project.
2. The written report and the oral presentation are the only aspect of the
study that marketing executives are exposed to, and consequently the
overall evaluation of the research project rests on how well this
information is communicated. They might not have been part of the
marketing research study and hence will not know what information was
collected.
3. Since the written research report and oral presentation are typically the
responsibility of the marketing research supplier, the communication
effectiveness and usefulness of the information provided plays a crucial
role in determining the choice of the particular marketing research
supplier for the future.
Differences between oral and written reports:
An oral report is any presentation that is verbally done to the management.
Written reports are documentation of the research findings.
There are three major differences between oral and written reports.
1. Oral reports are difficult to interpret as they lack visual advantage.
Charts, diagrams or pictures cannot be used to stress on important points.
They have to rely only on pauses and volume emphasis.
2. The pace of the presentation cannot be controlled and regulated by the
audience who is being presented with the oral report. In a written report,
the reader can clarify a certain point by reading it two or more times, if
needed slowly and carefully.
3. A researcher will write the report very precisely and with more accuracy
since he is aware that a written report is bound to receive considerable
attention and scrutiny from the readers. In contrast, an oral report will
not be so precise nor will the researcher give as much time in its
presentation since it cannot be subjected to the same degree of scrutiny
as a written report.
Considerations for Oral Reporting
A researcher has to consider the following when he has been asked to make an
oral presentation:
Consider the audience to whom the report has to be presented and
prepare the same so as to incorporate the technical requirements.
It should be properly planned. The researcher has to rehearse what he is
going to say or recommend. He has to collect and organize the data in a
logical manner.
Suitability of the language is another point worth considering. The
reporting has to be simple, concise and clear.
Visual aids like charts, graphs and handouts can be used judiciously if
required to create a better impact. However this also needs to be used
sparingly as it might disturb the proceedings.
Consideration for Written Reporting
Every person has a different style of writing. There is not really one right style
for a report, but there are some basic principles for writing a research report
clearly.
Types of Report:
There are many classifications available for reports. Some of them are listed
below:
Sl. No.
Classification
Types
Daily, Weekly, Fortnightly, Monthly or
1 Time interval
Annual
2 Functional Basis
Informational, Examinational and Analytical
Economics, Finance, Industry, and other
3 Subject Matter
subjects of interest
4 Physical-form
Short-form and Long-form
Relationship between the reader
Administrative, Professional and
5 and the writer
Independent
6 Employment status of authors
Private and Public reports
7 Formality
Formal and Informal
None of these are mutually exclusive. A report may be a combination of some of
the above.
Preparation of the Report
Preparing a research report is just not writing alone, but involves other activities
besides writing. In fact, writing is actually the last step in the preparation
process. Before writing can take place, the results of the research project must
be fully understood and thought must be given to what the report will say. The
objectives of the research should be matched with the results and the report has
to be written.
Thus, preparing a research report involves three steps:
1. Understanding: Understand the objective of the research and then
provide the solution.
2. Organizing: Organize your thoughts and findings and build a logical
flow.
3. Writing: Now, draft the outline of your report and then write the same.
The general guidelines that should be followed for any report or research paper
are as follows:
Consider the audience: The information resulting from the study is
going to be used by the marketing managers, who will use the results to
make decisions. Thus, the report has to be understood by them; the report
should not be too technical and not too much jargon should be used.
This is a particular difficulty when reporting the results of statistical
analysis where there is a high probability that few of the target audience
have a grasp of statistical concepts. Hence, this needs to be translated
into simple language.
Be concise, but precise: Many a time, the researcher in order to convey
his effort, tends to overcrowd the report with data. This leads to loss of
focus. On the one hand, a written report should be complete in the sense
that it stands by itself and that no additional clarification is needed. On
the other hand, the report must be concise and must focus on the critical
elements of the project and must exclude unimportant issues. Hence, a
research report has to be concise and precise.
Understand the results and draw conclusions: The managers who read
the report are expecting to see interpretive conclusions in the report. The
researcher must therefore understand the results and be able to interpret
these. He should analyze the results and present them to the managers.
Research Report Format
The following outline is the suggested format for writing the research report:
1. Title page
2. Letter of Authorization
3. Summary of findings
4. Table of contents
o List of tables
o List of figures
5. Introduction
o Background to the research problem
o Objectives
o Hypotheses
6. Methodology
o Data collection
o Sample and sampling method
o Statistical or qualitative methods used for data analysis
o Sample description
7. Findings
8. Limitations
9. Results, interpretation and conclusions.
10. Recommendations
11. Appendices
12. Bibliography
The title page indicates the topic on which the report has been prepared, the date
of submission, Prepared by and Prepared for details.
The letter of authorization is provided to facilitate the research process.
The summary of findings is perhaps the most important component of the
written report, since many of the management team who are to receive a copy of
the report will only read this section. The summary of findings is usually put
right after the title page, or is bound separately and presented together with the
report.
The table of contents guides the reader as to what it contains and helps him to
navigate the document. A separate table has to be provided for the charts and
graphs if they are part of the research report.
The introduction should describe the background of the study and the details of
the research problem. Following that, automatically the broad aim of the
research can be specified, which is then translated into a number of specific
objectives. Furthermore, the hypotheses that are to be tested in the research are
stated in this section.
In the methodology chapter the sampling methods and procedures are described,
as well as the different statistical methods that are used for data analysis.
Finally, the sample is described, giving the overall statistics, usually consisting
of frequency counts for the various sample characteristics.
Once the sample has been described, the main findings are to be presented in
such a way that all objectives of the study are achieved and the hypotheses are
tested. As mentioned before, it is essential that the main findings are well
interpreted and conclusions are drawn wherever possible.
The limitations or the caveats of the study should also be mentioned so that the
manager understands the pitfalls in the research process and he can make
suitable assumptions for his decision making.
Recommendations are required if the manager has specifically asked the agency
to suggest certain measures at the end of the research.
Appendix contains supporting materials for the report which cannot be given in
the body of the text. The bibliography details provide the details of the sources
which were referenced for the preparation of the report. The index shows the
various topics and the relevant page numbers in the report in an alphabetical
order.
Data presentation
Easy-to-understand tables and graphics will greatly enhance the readability of
the written research report. As a general rule, all tables and figures should
contain:
1. Identification number corresponding to the list of tables and the list of
figures
2. A title that conveys the content of the table or figure, also corresponding
to the list of tables and the list of figures, and
3. Appropriate column labels and row labels for tables, and figure legends
defining specific elements in the figure.
There are a number of ways to produce tables and figures. When typing a report
on a typewriter or word-processor, it is sometimes easiest to type a table out by
hand. However, when complicated tables have to be produced, it is advisable to
use spreadsheet software like Lotus 123 or Excel.
Feedback on the Report
A feedback should be sought from the manger to whom the report has been
submitted. This will help the researcher to understand the drawbacks of his
report and improve on the same. Feedback is generally not given and it has to be
sought. This will help the researcher to improve his services during the future
projects.
Summary:
It is not sufficient if the market researcher has collected information on the
research problem, he also has to interpret the data and draw specific conclusions
from it.
Presenting the results of a marketing research study to management generally
involves a formal written report as well as an oral presentation. The report and
presentation are extremely important.
Preparing a research report involves three steps:
1. Understanding: Understand the objective of the research and then
provide the solution.
2. Organizing: Organize your thoughts and findings and build a logical
flow.
3. Writing: Now, draft the outline of your report and then write the same.
The following outline is the suggested format for writing the research report:
1. Title page
2. Letter of Authorization
3. Summary of findings
4. Table of contents
o List of tables
o List of figures
5. Introduction
o Background to the research problem
o Objectives
o Hypotheses
6. Methodology
o Data collection
o Sample and sampling method
o Statistical or qualitative methods used for data analysis
o Sample description
7. Findings
8. Limitations
9. Results, interpretation and conclusions.
10. Recommendations
11. Appendices
12. Bibliography
Questions:
1. List out the differences between an oral and a written report.
2. When oral reports preferred and what are the considerations for it?
3. What are the points to be remembered which making a written report?
4. How are reports classified? Explain.
5. What is an outline? How is it useful for the preparation of a report?
6. Describe the various steps involved in writing a report.
7. Is feedback important for a report? Why?
8. How important do you think are writing skills and language important in
presenting the research findings?
9. Imagine that you are presenting a report to a head of an education
institution. Write a report on the current educational system and how
your institute can make a difference.
10. Writing a report is different from conducting a research. Explain.
CHAPTER VI
PARETO ANALYSIS
Objective:
The objective of this chapter is to understand:
The meaning of Pareto analysis
Defining the Pareto effect
Steps involved in creating a Pareto chart
Importance of Pareto analysis
Introduction:
Pareto analysis is a very simple technique that helps a manager to choose the
most effective changes to make. It is represented as a bar graph used to arrange
information in such a way that priorities for process improvement can be
established. Pareto charts are constructed to provide a before-and-after
comparison of the effect of control or quality improvement measures.
The 80-20 theory was first developed in 1906, by an Italian economist, Vilfredo
Pareto, who observed an unequal distribution of wealth and power in a relatively
small proportion of the total population. Joseph M. Juran adapted Pareto's
economic observations to business applications.
Pareto Analysis is also used in inventory management through an approach
called "ABC Classification". The ABC classification system works by grouping
items by annual sales volume. This helps identify the small number of items that
will account for most of the sales volume and that are the most important ones to
control for effective inventory management. Thus it helps in effective inventory
management.
The Pareto effect
In practically every industrial country a small proportion of all the factories
employ a disproportionate number of factory operatives. In some countries 15
percent of the firms employ 70 percent of the people. This same state of affairs
is repeated time after time. In retailing for example, one usually finds that up to
80 percent of the turnover is accounted for by 20 percent of the lines.
This effect, known as the 80 : 20 rule, can be observed in action so often that it
seems to be almost a universal truth. As several economists have pointed out, at
the turn of the century the bulk of the country`s wealth was in the hands of a
small number of people.
This fact gave rise to the Pareto effect or Pareto`s law: a small proportion of
causes produce a large proportion of results. Thus frequently a vital few causes
may need special attention wile the trivial many may warrant very little. It is this
phrase that is most commonly used in talking about the Pareto effect ? the vital
few and the trivial many`. A vital few customers may account for a very large
percentage of total sales. A vital few taxes produce the bulk of total revenue. A
vital few improvements can produce the bulk of the results.
The Pareto effect is named after Vilfredo Pareto, an economist and sociologist
who lived from 1848 to 1923. Originally trained as an engineer he was a one
time managing director of a group of coalmines. Later he took the chair of
economics at Lausanne University, ultimately becoming a recluse. Mussolini
made him a senator in 1922 but by his death in 1923 he was already at odds with
the regime. Pareto was an elitist believing that the concept of the vital few and
the trivial many extended to human beings.
This method stems in the first place from Pareto`s suggestion of a curve of the
distribution of wealth in a book of 1896. Whatever the source, the phrase of the
vital few and the trivial many` deserves a place in every manager`s thinking. It
is itself one of the most vital concepts in modern management. The results of
thinking along Pareto lines are immense.
For example, there may be a lot of customer complaints, a lot of shop floor
accidents, a high percentage of rejects, and a sudden increase in costs etc. The
first stage is to carry out a Pareto analysis. This is nothing more than a list of
causes in descending order of their frequency or occurrence. This list
automatically reveals the vital few at the top of the list, gradually tailing off into
the trivial many at the bottom of the list. Management`s task is now clear and
unavoidable: effort must be expended on those vital few at the head of the list
first. This is because nothing of importance can take place unless it affects the
vital few. Thus management`s attention is unavoidably focused where it will do
most good.
Another example is stock control. You frequently find an elaborate procedure
for stock control with considerable paperwork flow. This is usually because the
systems and procedures are geared to the most costly or fast-moving items. As a
result trivial parts may cost a firm more in paperwork than they cost to purchase
or to produce. An answer is to split the stock into three types, usually called A,
B and C. Grade A items are the top 10 percent or so in money terms while grade
C are the bottom 50-75 percent. Grade B are the items in between. It is often
well worthwhile treating these three types of stock in a different way leading to
considerable savings in money tied up in stock.
Production control can use the same principle by identifying these vital few
processes, which control the manufacture, and then building the planning around
these key processes. In quality control concentrating in particular on the most
troublesome causes follows the principle. In management control, the principle
is used by top management looking continually at certain key figures.
Thus it is clear that the Pareto concept ? the vital few and the trivial many` ? is
of utmost importance to management.
Pareto Charts:
Pareto charts provide a tool for visualizing the Pareto principle, which states that
a small set of problems (the "vital few") affecting a common outcome tend to
occur much more frequently than the remainder (the "useful many"). A Pareto
chart can be used to decide which subset of problems should be solved first, or
which problems deserve the most attention.
The Pareto Chart is used to illustrate occurrences of problems or defects in a
descending order. It is used for making decisions at critical points in different
processes. This means that it can be used both during the development process
as well as when products are in use, e.g. customer complaints.
How to create a Pareto Chart:
First step is to list all elements of interest
Use the same unit of measurement and measure each of the elements.
Order the elements according to their measure
Calculate the percentage for each element out of the total measurement
Accumulate the percentage from top to bottom to equal 100 %
Create a bar and line graph, line representing cumulative percentage
Work on the most important element first
Advantages of Pareto Charts:
Pareto charts are a key improvement tool because they help us identify patterns
and potential causes of a problem. Several Pareto charts can be created out of
the same set of data. This will help a manager to quickly scan a number of
factors that might contribute to a problem and focus on those with the greatest
potential payback for his efforts
It is difficult to choose which issue to work on first when faced with a range of
issues. To resolve this dilemma, the most useful thing to do is to apply Pareto's
rule. This rule says - "eighty percent of your troubles will come from 20 per cent
of your problems". In other words, problems will rarely have equal impact, so it
is best to first concentrate on the most important.
This does not provide a scientifically accurate estimation of the weightage with
respect to the range of alternatives. This reminds a manager to always look for
'the vital few' issues, and to separate them from 'the trivial many', before
attempting to solve problems. The next step is to identify which particular
problems are the most important. This is done by collecting appropriate data and
displaying it in the form of a histogram with each measured characteristic shown
in descending order of magnitude. Such a histogram is known as a Pareto chart.
Following is an example of a Pareto chart.
40
30
20
10
0
The high value items to the left hand side of the chart are the ones that are
needed to be concentrated on first.
Some problems and difficulties associated with Pareto Analysis:
Misrepresentation of the data.
Inappropriate measurements depicted.
Lack of understanding of how it should be applied to particular
problems.
Knowing when and how to use Pareto Analysis.
Inaccurate plotting of cumulative percent data.
Overcoming the difficulties.
In conclusion
Even in circumstances which do not strictly conform to the 80: 20 rule the
method is an extremely useful way to identify the most critical aspects on which
to concentrate. When used correctly Pareto Analysis is a powerful and effective
tool in continuous improvement and problem solving to separate the vital few`
from the many other` causes in terms of cost and/or frequency of occurrence.
It is the discipline of organizing the data that is central to the success of using
Pareto Analysis. Once calculated and displayed graphically, it becomes a selling
tool to the improvement team and management, raising the question why the
team is focusing its energies on certain aspects of the problem.
Summary:
Pareto analysis is a very simple technique that helps a manager to choose the
most effective changes to make. . This effect, known as the 80: 20 rules, can be
observed in action so often that it seems to be almost a universal truth.
It is represented as a bar graph used to arrange information in such a way that
priorities for process improvement can be established. Pareto charts are
constructed to provide a before-and-after comparison of the effect of control or
quality improvement measures
Pareto charts provide a tool for visualizing the Pareto principle, which states that
a small set of problems (the "vital few") affecting a common outcome tend to
occur much more frequently than the remainder (the "useful many"). A Pareto
chart can be used to decide which subset of problems should be solved first, or
which problems deserve the most attention.
Once calculated and displayed graphically, it becomes a selling tool to the
improvement team and management, raising the question why the team is
focusing its energies on certain aspects of the problem.
Questions:
1. What is Pareto principle?
2. What is the importance of this in Marketing?
3. What is Pareto effect? Explain.
4. How is Pareto charts created?
5. What are the advantages of creating Pareto charts?
6. What are the difficulties associated with the creation of Pareto charts?
CHAPTER VI
ISHIKAWA DIAGRAMS
Objective:
The objective of this chapter is to understand:
Origin of Ishikawa Diagrams
Concept of Cause and Effect
Steps involved in creating an Ishikawa Diagram
Introduction:
Ishikawa Diagrams are a graphical method for finding the most likely causes for
an undesired effect. The method was first used by Kaoru Ishikawa in the 1960s.
The Cause and Effect diagram also known as the "fishbone" or "Ishikawa"
diagram after its creator Kaoru Ishikawa is used to systematically list all the
different causes that can be attributed to a specific problem (or effect). A cause-
and-effect diagram can help identify the reasons why a process goes out of
control.
Because of its shape, it is also known as the fishbone diagram. Another name for
this technique is: the cause-and-effect diagram. The fishbone diagram is a
method/tool used in a root cause analysis.
The Ishikawa diagram is one of the seven basic tools of quality control, which
include the histogram, Pareto chart, check sheet, control chart, cause-and-effect
diagram, flowchart, and scatter diagram.
The purpose of this diagram is to arrive at a few key sources that contribute
most significantly to the problem being examined. These sources are then
targeted for improvement. The diagram also illustrates the relationships among
the wide variety of possible contributors to the effect.
The figure below shows a simple Ishikawa diagram
Envi
E
ron
r men
m t
Wor
o ker
k s
Too hot
h
Trai
a ning
n
Unh
U
app
a y
Wor
o ker
k s
old
Man
M agemen
em t
Machine
M
s
The basic concept in the Cause-and-Effect diagram is that the name of a basic
problem of interest is entered at the right of the diagram at the end of the main
bone". The main possible causes of the problem (the effect) are drawn as
bones off of the main backbone. The "Four-M" categories are typically used as
a starting point: "Materials", "Machines", "Manpower", and "Methods".
Different names can be chosen to suit the problem at hand, or these general
categories can be revised. The key is to have three to six main categories that
encompass all possible influences. Brainstorming is typically done to add
possible causes to the main "bones" and more specific causes to the "bones" on
the main "bones". This subdivision into ever increasing specificity continues as
long as the problem areas can be further subdivided. The practical maximum
depth of this tree is usually about four or five levels. When the fishbone is
complete, one has a rather complete picture of all the possibilities about what
could be the root cause for the designated problem.
The Cause-and-Effect diagram can be used by individuals or teams; probably
most effectively by a group. A typical utilization is the drawing of a diagram on
a blackboard by a team leader who first presents the main problem and asks for
assistance from the group to determine the main causes which are subsequently
drawn on the board as the main bones of the diagram. The team assists by
making suggestions and, eventually, the entire cause and effect diagram is filled
out. Once the entire fishbone is complete, team discussion takes place to decide
on the most likely root causes of the problem. These causes are circled to
indicate items that should be acted upon, and the use of the tool is complete.
How to Construct an Ishikawa Diagram:
Place the main problem under investigation in a box on the right.
Generate and clarify all the potential sources of variation.
Use an affinity diagram to sort the process variables into naturally
related groups. The labels of these groups are the names for the major
bones on the Ishikawa diagram.
The process variables are then placed on the appropriate bones of the
Ishikawa diagram.
Combine each bone in turn, insuring that the process variables are
specific, measurable, and controllable. If they are not, branch or
"explode" the process variables until the ends of the branches are
specific, measurable, and controllable.
The Ishikawa diagram, like most quality tools, is a visualization and knowledge
organization tool. Simply collecting the ideas of a group in a systematic way
facilitates the understanding and ultimate diagnosis of the problem. Several
computer tools have been created for assisting in creating Ishikawa diagrams.
Summary:
Ishikawa Diagrams are a graphical method for finding the most likely causes for
an undesired effect.
The Cause and Effect diagram also known as the "fishbone" or "Ishikawa"
diagram after its creator Kaoru Ishikawa is used to systematically list all the
different causes that can be attributed to a specific problem (or effect). A cause-
and-effect diagram can help identify the reasons why a process goes out of
control.
The Ishikawa diagram, like most quality tools, is a visualization and knowledge
organization tool. Simply collecting the ideas of a group in a systematic way
facilitates the understanding and ultimate diagnosis of the problem.
Questions:
1. What are Ishikawa diagrams?
2. Ishikawa diagrams are useful tools in Marketing. Explain.
3. What are the considerations for creating an Ishikawa diagram?
4. List out the steps involved in the creation of Ishikawa diagram.
References for this Unit:
1. Naresh K. Malhotra: MARKETING RESEARCH: AN APPLIED
ORIENTATION, Pearson Education, Asia.
2. Aaker, Kumar & Day: MARKETING RESEARCH, John Wiley & Sons,
1998.
3. Paul E. Green & Donald S. Tull: RESEARCH FOR MARKETING
DECISIONS.
4. Carol S. Aneshensel; University of California, Los Angeles;
UNIVARIATE ANALYSIS: CENTRAL TENDENCY, SPREAD,
AND ASSOCIATIONS
5. Cooley, W.W. and P. R. Lohnes. 1971. Multivariate Data Analysis. John
Wiley & Sons, Inc.
6. George H. Dunteman (1984). Introduction to multivariate analysis.
Thousand Oaks, CA: Sage Publications. Chapter 5 covers classification
procedures and discriminant analysis.
7. Morrison, D.F. 1967. Multivariate Statistical Methods. McGraw-Hill:
New York.
8. Overall, J.E. and C.J. Klett. 1972. Applied Multivariate Analysis.
McGraw-Hill: New York.
9. Tabachnick, B.G. and L.S. Fidell. 1996. Using Multivariate Statistics.
Harper Collins College Publishers: New York.
This post was last modified on 14 March 2022