BUSINESS RESEARCH

THEORY NOTES
Before You Start
These notes are prepared from Research Methodology
by Chawla & Sondhi, and various online sources to
which I am truly thankful.
The notes are made from examination point of view and
are restricted to the syllabus. However, for a greater
learning experience, it is better to study this subject
in-depth. Nevertheless, these notes might serve as a
starting point. I hereby do not take any responsibility
for any consequences that may result from this set of
information. (25 Pages of Material)
For any reviews or recommendations, contact me at
8826811598.
Hope these serve your academic needs in any manner
possible. If helpful do circulate to your juniors in
coming years.
Resources
BMS-BFIA: ​bit.ly/NOTES-COUNCIL
BMS SEM-1 Resources:​ ​bit.ly/BMS-SEM1
BMS SEM-2 Resources:​ ​bit.ly/BMS-SEM2
MARKETING Notes: ​bit.ly/Marketing-BMS
HRM Notes: ​bit.ly/HRM-BMS
BR Notes: ​bit.ly/BR-BMS4
POM Notes: ​bit.ly/POM-BMS
 Unit 1

Role of Business Research in Decision Making

Management Research: It is an unbiased, structured and sequential method of
enquiry directed towards a clear implicit or explicit business objective. This
enquiry might lead to validating the existing postulates or arriving at new
theories and models.
Ex. Consider a food chain is not able to break even on its business. This serves
as a management problem which must be solved. On the other hand, a similar
food store is able to earn well, thereby you undertake a consumer survey and
also plan a feedback session with a group of customers where you conduct a
focus group activity to analyse what might be the causes of preferring the other
store. You as a researcher identify certain issues in the customer service area
that is causing this problem. You call up a meeting with the employees to
analyse the gaps in their dealing with the customer, thus you are able to identify
the training needs. You compare the training models that exist and implement
after comparing them on various aspects. Finally, after a month you compare the
sales figure and find that you earn profits over a above break even. You then
plan on undertaking further intensive training for growth of the chain.
As we can see that the whole research process is undertaken in a structured
format in order to find a solution to the business problem thus meeting the
objective. Through the above case we can conclude that research can act as a
tool for decision making used by the business.

Types of Research
1. Fundamental/Basic Research: It is undertaken for a purely academic
reason, to investigate certain best practices or to find relationship between
various business variables. The purpose of the study is vast and the
benefits generated would be applicable to entire business community. The
time period is flexible.
E.g. Work-family conflict and impact on turnover rates
 2. Applied Research: Research undertaken to solve an immediate business
problem. The scope is limited to the organisation undertaking the research, and
time period is limited as per the objectives and immediateness of the study. E.g.
Finding the reason behind a product’s failure despite positive results in pilot
testing.

3. Exploratory Research: A form of research aimed at resolving ambiguity, and

helping the researcher gain a better understanding and direction for the study, to
evaluate whether it is worthwhile to undertake a full-scale investigation.
It is important for the researcher to observe and record all possible information which
will in turn help in accuracy of the results.
E.g. Employee preference for Various Incentives.

4. Conclusive Research: It tests and authenticates the propositions revealed by

exploratory research and is usually quantitative in nature. It is more structured
and definite and the validity of research assumes critical significance as the
results might need to be implemented.
E.g. Impact of Gain Share incentives on Employee Motivation.

5. Descriptive Research: It aims at describing the primary characteristics and

data related to the object/situation/concept under study.
It is thereby used to test specific hypothesis and trends, and is more structured in its
means of sampling, data collection, analysis etc.
(It indicates the existing symptoms of a particular situation without establishing the
causality of relationship.)
E.g. Annual Census carried by Government.

6. Causal Research: It is concerned with exploring the effects of one variable on

another. It requires a a rigid sequential approach for sampling, data collection
and analysis.
E.g. Impact of trade discounts on sales.

Difference B/w Exploratory & Conclusive - Loosely Structured vs. Well Structured,
Flexible & Investigative Methodology vs. Formal Methodology, no testing of
hypothesis vs. Testing of Hypothesis, Findings might be topic specific not
having relevance outside the researcher’s domain, Findings are significant with
theoretical or applied implication.

Hierarchy:
Business Research
 1. Basic Research
2. Applied Research
Both further classify into
i.Exploratory Research
ii.Conclusive Research
a. Descriptive Research
b. Causal Research

Applications of Research for Business Decisions

Business managers in each field are confronted with problem situations that
require effective decision making. These decisions required additional
information which can be addressed by research.
1. Marketing Function:
i.Product Research: Product Testing and Development.
ii.Pricing Research: Price Determination Strategy
iii.Promotional Research: Designing Communication Mix
iv.Place Research: Location Analysis

2. HR Research
It involves fundamental research as with many macro level researches implemented by
organisations. It also involves applied research to find solutions.
E.g. Incentive & Benefit Studies such as Recognition & Reward Studies, Training &
Development Studies such as Need Gap analysis.

3. Financial & Accounting Research: Stock Market’s Response to Corporate Action

& Financial Reporting, Analysis of Corporate Financial Reporting Behaviour.

4. Production & Operation Management: Operation Planning includes Product &

Service Design, Resource Allocation, while POM also includes Demand
Forecasting & Decision Analysis.

5. Cross Functional Research: It Requires an open orientation where experts from

across the departments/disciplines contribute to the study, working towards a
common goal. E.g. New Product Development requires commitment of
Marketing, Production and Finance.

Steps in Research Process

 1. Management Dilemma: Research is triggered by the need and desire to
know more, with an academic purpose to uncover some new phenomenon
(basic research) or to find the solution to an immediate business problem
(applied research).
2. Defining the research problem: A research problem is the first step
towards the end result that one hope to achieve and requires considerable
thought and analysis.
3. Formulating Research Hypothesis: It is the statement whose validity is to
be tested by the means of the research process.
4. Developing Research Proposal: It is a framework of whole research
which shows the proposed methodology of the study. It provides a time
and objective bound commitment that the researched must make for
himself or the manager. Once its accepted, the researcher initiates.
5. Research Design Formulation: Based on the objectives of the study we
find out which form of Research Design is appropriate. (Described in
Unit 2) It’s main task is to guide the technique in which the research
problem will be investigated.
6. Sampling Design: It involves selecting representativeness (of population)
of the sample selected from the population under study and avoiding the
probability of error by selecting a bias free sample whose error is
measurable.
7. Data Collection: The data may be collected via:
i.Primary Data: It is original and collected first hand.
ii.Secondary Data: It is the information collected and compiled earlier. (Both
described in Unit 3 & 2 Respectively)

8. Data Refining: The data should be edited and refined to suit the study. It
should then be coded and tabulated for analysis.

9. Data Analysis: Selection of analytical tools is done, followed by running

them to find conclusive results. Univariate, Bivariate and multivariate analysis
are done to examine a one, two or more than two variables.

10. Research Report: Here the results of the study are linked with the research
objectives, to find whether the study answers the questions sought.
Research Proposal
The question would ask to make a proposal, it’s recommended to see multiple
proposals to get an idea.
Format:
 1. Executive Summary: Summary about the Management Problem and the
probable research questions which need to be answered to derive
conclusive results.
2. Background of Problem: Systematic build up to the research questions
and a compelling reason for pursuing the study.
3. Problem Statement & Research Objectives: Clear Definition of the
problem broken down into specific objectives.
4. Research Design: It includes the systematic approach to be followed in
order to achieve the objectives, including the population to be studies,
sample size and selection technique, sources of data & instrument design.
There must be clear and logical justification of using the design over
other methods available.
5. Hypothesis: It is the statement whose validity is to be tested by the means
of the research process.
6. Results & Outcomes of Study: How the researcher intends to link the
findings of the proposed study design to the stated research objectives.
Optional: Scheduling of Research, Costing & Budgeting of Research, Literature
Review (in Background of Problem), Glossary & References.

Formats: (For Reference)

(​https://www.bcps.org/offices/lis/researchcourse/develop_write_sample.html
Management Decision problem to Business Research Problem
 1. The entire process begins with the acknowledgement and identification of
the difficulty encountered by the business manager/researcher. The
problem can either be solved by the manager on his own or is outsourced
to a researcher.
Management Problem Research Problem
What should be done to increase What is the awareness and purchase
consumer base of Pepsi? intention of soft drink consumers for
Pepsi?
How to reduce turnover (leaving What is the impact of inflexible timings on
jobs) in govt. colleges? Govt. College Employees?
How to improve Delivery of How does Dominos manage its supply
Dominos? chain?

2. The researcher gains perspective on the basis of discussion with experts

(industry or subject experts).
● Getting perspective of other researchers by a comprehensive
examination of their earlier studies.
● In case the research is intended to be carried out then the detailed
history along with current practices of organisation is needed.
● Brief and loosely structured survey with respondents from population
to be studied can be done.
3. We can then arrive at a clearly stated research problem.
● It should include the unit of analysis i.e. the individual or group to be
studied.
● It should have the definition of the concepts or constructs (idea
containing elements) to be studied. This involves determining the
independent (causal), dependent and external variables.
● It is advisable to construct a model or theoretical framework based on
above variables.
4. Finally, we develop the statement that is to be authenticated via the research
process, which is the Research Hypothesis.
It could be Descriptive (assumption about the probability of occurrence) or Relational
(probability of relation between two or more variables) in nature.
(The hypothesis may be reduced to its statistical form, for testing purpose.)
 Unit 2

Research Design
Research Design: It is the framework that has been created to seek answers to
research questions. One the other hand, research method is the technique to
collect the information required. It provides direction to the investigation being
conducted in the most efficient manner.
Relevance: Research Design gives the necessary direction to the research study.
It includes the systematic approach to be followed in order to achieve the
objectives, including the population to be studies, sample size and selection
technique, sources of data & instrument design. There must be clear and logical
justification of using the design over other methods available.
Without a research design, the researcher would be left with a lot of ambiguity
in terms of how the research would be conducted, thereby he might not be able
to initiate the research, far less likely to meet his desired research objectives.
Types
1. Exploratory Research Design: It is flexible in its approach, loosely
structured and involves a qualitative investigation in most cases. The
objective is to gain clarity about the problem situation.
The essential purpose is to:
● define the research problem
● explore and evaluate multiple research opportunities
● formulate the research hypothesis
● define the variables
● Identify nature of relationship b/w variables and impact of external
factors.
E.g. The researcher might be interested in doing an exploratory analysis of the
Political Opinions of Delhi for the coming 2019 elections, to begin the research
a structured methodology may not be essential as one can take the interview of
people from various location, identify the factors that favour one party over
another and rate the factor by the level of importance. Once, he finds the study
meaningful enough to be pursued at a larger and structured scale. He may
pursue Descriptive Research.
 2. Descriptive Research Design: They are a structured and formal form of
design, where the objective is to provide a comprehensive and detailed
explanation of the phenomenon under study. It however lacks the precision and
accuracy of experimental designs
The objective here is to:
● Give a detailed profile of the population being studied.
(Who-what-when-where-why and how of a certain population)
● Description might be in stagnant time period or stretched across
different stages in a stipulated time period.
● Studies are carried out to measure the simultaneous occurrence of
certain phenomenon or variables.
E.g. To establish the relation between Cost of Capital & Market Value, a correlation
between two variables can be undertaken.

3. Causal Research Design: It is concerned with exploring the effects of one

independent variable on another dependent variable while keeping in control the
extraneous variable which impact the effects.
It requires a a rigid sequential approach for sampling, data collection and analysis. E.g.
Impact of trade discounts on sales.
Secondary Data
Secondary Sources of Data contain the details of previously collected findings
which can be found by easier and inexpensive means.
The data can point to the researcher that his proposed research is redundant, or it
can help him find a gap in research that has not been addressed which he can
conduct a research on.
Sources
i.Databases: Data related to an entire industry or a business are available for free
on public domains or through a structured acquisition process and cost. These
are both government and non-government publications. E.g. Capitaline.
ii.Comprehensive Case Method: It is intricately designed data, which reveals a
complete presentation of facts as they occur in a single entity and focuses on a
single unit of analysis (employee, customer, organisation) E.g. A system of HR
policies implemented by a company would serve as a Comprehensive case for
another company planning to implement the same.
iii.Expert Opinion Survey: It is conducted when no previous information or data is
available on a topic of research. The information is collected from significant
erudite (knowledgeable) people. (not a secondary source)
 E.g. Contacting a 5-star hotel Chef, for developing a new recipe, as per consumer
preference.
iv. Focus Group Discussions: This involves carrying out discussions with
significant individuals associated with a problem under study. It is used
especially for consumer and motivational research studies. (Described in
Unit 3) (not secondary source)
i.Internal Sources of Data (organisation specific): Company Records, Employee
Records, Sales Data, Sales Invoices.
Organisation of large volumes of information into clusters of data as per user
requirement is called data mining.
vi. External Sources of Data (available outside an institution): Published
Data: Governmental Data (Census Data, Planning Commission etc.) &
Other Data Sources (Books-Periodicals- Directories-Databases)

Advantages

i.Resource Advantage: Use of secondary information save cost and time

ii.Accessibility of Data: The secondary data once compiled by a publication, it can
then be easily accessed by individuals for research purposes rather than
collecting the same for a single study.
iii.Accuracy and stability of data: Data collected by recognized bodies on larger
scale have advantage of accuracy and reliability as well.
iv.Assessment of Data: Information can be used to compare and support the
primary research findings of investigators. If the study was conducted on a
sample, we can use the same and test their applicability on a larger population.
Disadvantages
i.Applicability of Data: If it is unique to a particular study, it will be difficult to
fit it to the current research. And therefore, the data might not be applicable for
the current objective. Differences occur in terms of the variables; sample and
the units being used.
ii.Accuracy of Data: Level of accuracy depends on who undertook the
investigation, and why the research was undertaken. E.g. Research undertaken
by a political party may be skewed to show favourable outcomes of the work
done by them.
Criteria for Evaluating Secondary Sources of Data (Indian Context)
A. Methodology Check
The process or design used to collect the data.
i.Sample: Respondent Selection, Response Rate, Quality of Data Recording.
 ii.Quality of instrument design (used for data collection)
iii.Analytical Tools used (to avoid personal bias of investigator)

B. Accuracy Check
It determines the significance of the source of information from where the data was
collected.

C. Topical Check
Aims at investigating the information that is being used or cited in the research study
for periodical upgradations. For instance, for a study on trends and fashion, a
time period of 5 years would lead to a lot of changes in trends, thus the
hypothesis made basing such data would be errored.

D. Cost-Benefit Analysis: Ensure that the data procured is not costly to make
the benefits derived from it meaningless.
Syndicated Research (In India)
A research study which is conducted and funded by a market research firm (not
for specific client) is called a syndicated research. The result of the data is
provided in the form of reports, presentations which is made available in open
market for anyone to purchase.
The research problem and the scope of research is formulated by the market
research company based on their experience. They often need refinement when
used for customised market research.
It may be conducted to:
i.To update the knowledge base on key industries
ii.To utilise human resource during slack period
iii.To improve visibility of the company and use it for publicity
iv.To show off the skills of the company
The above reasons are why a market research company may invest its funds and
time in conducting a syndicate research without any assurance of covering its
costs. In terms of cost, syndicate research is almost 10 times cheaper than
custom market research studies
It is available with University libraries, Industry associations, Market Research
Portals.
In India Companies like IMRB International, RNB Research undertake
Marketing Research.
 (It can be used during the exploratory research period)

Unit 3

Survey vs. Observations

Survey Observation
It is a method of Primary Data The action or process of closely
collection that uses questions to monitoring something or someone, to
collect information about how derive information.
people think or act.
Types: Types:
i.Personal Administration i.Participant Observation: Researcher may
i.Telephone interact with participants and becomes a
i.Mail (Offline) part of their community. It is a form of
v.E-Mail undisguised observation where
v.Web-Based participant knows he is being observed.
i.Unobtrusive Observation: You do not
Comparison given below.
interact with participants and simply
observe their behaviour. It can be of
disguised format where people are
unaware they are being observed.
Bias can be avoided by using Bias can be avoided using a
neutral questions, using rightly “double-entry notebook” where
designed answer options, having observations and Personal thoughts of
the right survey structure. researcher are kept separate.
Survey is relatively complex and Observation is relatively simple, and thus
the development and costs relatively cheaper, although is more
implementation cost can be higher time consuming.
if Personal Administration
technique is used.
Survey involves a greater Observation involves lesser participation
participation of the respondent. of the respondent especially in
Unobtrusive Observation.
Both are primary data collection techniques, both can be structured
(predetermine format) or unstructured (no predetermined format), both are
prone to researcher’s bias.
 Comparison of Self-Administered Techniques

Depth Interview
Personal Interview is a one-to-one interaction between the
investigator/interviewer and interviewee. The purpose of dialogue is research
specific and ranges from completely unstructured to highly structured.
● The interview method can be used for exploratory research.
● Interview can be used a medium for primary data collection, when the
area to be investigated is high on affective (moods, feelings) and
structured method is not apt.
● Unstructured interview has no defined guidelines. The direction of
interview is decided spontaneously, but the probability of subjectivity
is high.
Structured – This format has high validity. The questions are sequenced &
pre-decided.
● Interviewing Skills – The depth of information depends upon the
listening and probing skills of interviewer. He must be
well-acquainted with the study objectives and aware about the
objectives of the study. His attitude must be as objective as possible,
to avoid biasing the results.
● Analysis & Interpretation: Information does not undergo statistical
analysis. The data is in a narrative form, and the data can be subjected
 to content analysis to achieve a better structure for the results
obtained.
● Types – At-home Interviews, Mall-intercept interviews,
Computer-assisted Personal Interviewing (CAPI).

Focus Groups ​(short note)

● It is highly versatile and dynamic method of collecting information
from a representative group of respondents. The process involves a
moderator who manoeuvres the group of respondents who are
specifically invited and gathered at a neutral setting. The moderator
initiates the discussion and then the group carries it forwards by
holding a focused and interactive discussion. Some consider the group
dynamics to be important while others consider it detrimental.
● Size: It shouldn’t be too large or tool small. 8-12 members is
recommended.
● Nature – Individuals from similar backgrounds must be included to
involve certain agreement.
● Acquaintance – It can be disruptive to the free flow of discussion as
people might shy from revealing themselves.
● Setting: Space or Setting should be neutral, informal and comfortable.
● Time-Period: Held in a single session. (not exceeding 1.5 hour)
● Recording – Cameras are to be used, to review the discussion for
analysis and interpretation.
● Moderator – The nature, content and validity of data rely largely on
the moderator’s skills. He may or may not be a participant and has the
role of initiating the discussion and steering it forward.
● Focus groups may be dual-moderator, fencing moderator (two
moderators taking opposite side), Friendship Groups (to ensure
comfort level if study requires), Online Focus Group.
● Focus Groups help in: Idea generation, Group Dynamics (ensure free
flowing discussion but group conformity can be a disadvantage). But
it’s analysis can be cumbersome and the moderator bias may also
exist.
 Projective Techniques
Projective Techniques uncover the different levels of consciousness of an
individual’s mind and reveal the data which is inhibited by socially-desirable
and correct responses.
● It involves asking ambiguous stimuli and indirect questions related to
imaginary situations or people. A simulation can be presented to the
person to project their underlying beliefs, attitude on to the projection.
● It aims at uncovering different levels of consciousness.
● Association Technique – Presenting a stimulus to respondent and he
responds with the first thing that comes to his mind. The responses
and time taken is noted and compared with the norms and scores
available.
● Completion Technique – Presenting an incomplete object to the
respondent which can be completed by the respondent in any way.
Story Completion – The subject is supposed to complete an incomplete story and
provide a conclusion.
● Construction Technique – Focus here is on a completed object which a
story may be, a picture, a dialogue or a description, and the respondent
is asked questions related to it.
● Choice or Ordering Technique – Presenting an assortment of stimuli
and subject is to sort them into categories, based on study instruction
given for instance good, bad, ugly.
● Expressive Technique – Here the method or means used are
significant, where the subject needs to express not his feelings but that
of the protagonist in a given situation. The main advantage here is that
people are less inhibitive when they give their opinion in third person.
The technique can be applied for Business research in understand the consumer
perception, thoughts and feelings and behaviour for a particular product,
company or brand. (primarily in Market Research)
The analysis of this data is difficult and the test is expensive and time consuming.
 Primary Scales of Measurement
Measurement means assigning numbers or other symbols to the characteristics
of certain objects. We don’t measure objects but their characteristics. Numbers
permit statistical analysis and help in communication of results.
Scaling is an extension of measurement and involves creating a continuum on
which measurements of objects are located.
1. Nominal Scale – Lowest level of measurement, here the numbers are
assigned for the purpose of identification of the objects. An object which
is assigned a higher number is in no way superior to the one with the
lower number. (Are you married 69 = Yes, 96 = No.)
(The nos. on nominal scale cannot be +/-/x/divided)

2. Ordinal Scale – It tells whether an object has more or less of a characteristic

than some other object. It cannot tell the difference between the magnitude of
the characteristics.
(Customer Service characteristics like Politeness, Quickness, Assistance can be ranked
over what is more important)
(Similar to nominal scale in arithmetic, but you can calculate median,
percentiles and quartiles of distribution)

3. Interval Scale – Here the difference of the score on the scale has meaningful
interpretation i.e. it can measure magnitude of differences of characteristics.
It has a non-zero origin.
Ex. Relationship b/w Celsius & Fahrenheit temperature is an interval scale data.
(Difference b/w 70 & 80 Celsius is same as b/w 80 & 90 degrees)
C = 5/9 (F-32) [ y = a+bx form] (but we can’t say that if Delhi is 40 degrees, Shimla is
20 degrees, then Delhi is twice as hot as Shimla)

We can run multiple mathematical and statistical techniques on interval scale, and it
can be converted into a nominal and ordinal scale.
 4. Ratio Scale
Here the difference as well as the ratio of the measurements on scale hold meaning. It
has a natural (zero) origin. Ex. Weighing Scale Height Scale. (Y = bX i.e. 40 kg
is twice (ratio = 2) as heavy as 80 kg)
All mathematical operations can be carried out. (It can be converted to any of the
previously described scales)
Scaling Techniques
Comparative Scale – It is assumed that the respondent makes use of a standard
frame of reference before answering the questions.
i.Paired Comparison: A respondent is presented with two objects and is asked to
select one according to whichever criterion he/she wishes to use. Resulting data
is ordinal in nature.
(if there are 5 brands, it would result in 5x4/2 = 10 pairs whose comparison would
be undertaken following which the probabilities of selection one item over
the other in a pair maybe determined)

ii. Rank Order Scaling: Respondents are presented with several objects
simultaneously and are asked to order or rank them as per certain
criterion.
E.g. Ranking several soft drinks on basis of fizziness.

iii. Constant Sum: The respondents are asked to allocate a total of 100 points
between various objects and brands. The higher the allocation of points,
greater is characteristic associated with the object.
Colleg Points (Charac. – Building Design)
e
KMV 30
SSCBS 40
DDU 30
Non-Comparative Scale – The respondents don’t make use of any frame of
reference before answering the questions.
I.Graphic Rating Scale: It’s a continuous scale, where the respondent is asked to
tick his preference on a graph.

II. Itemized Rating Scale: Here the respondents are provided a scale, that has
a number of brief descriptions associated with each of the response
categories.
a. Likert Scale: It is a multiple item agree-disagree 5-point scale, where the
respondents are given a certain number of items on which they are asked to
express their degree of agreement and disagreement.
It is also called a summated scale because the scores on individual items can
added together to produce a total score for the respondent.
 b. Semantic Differential Scale: The respondent is required to rate
each characteristic on a five or seven-point rating scale. Here
bipolar adjectives are used. Once the scale is administered, the
mean score for each characteristic is calculated, by summing up its
total points.

Questionnaire Form & Design

A Questionnaire is a means of meeting the research objectives converted into
clear questions to extract information from the respondent. It should be
designed to engage the respondent and to encourage a meaningful response.
Lastly, it should be self-explanatory and not confusing as it leads to less
accuracy of answers.
Formalized – Structured, with pre-determined questions to be answered in a
defined sequence.
Un-Formalized/Unconstructed Questions – Allows a respondent to express his
attitude in a free manner
Concealed – Tries to reveal the latent cause of a behaviour or opinion which
cannot be revealed by direct questions.
Formalized & Unconcealed – Used by many management researchers,
predefined format, and easy to analyse and administer to large no. of people.

Method of Forming a Questionnaire​: (preparing one can be asked)

1. Convert the research objectives into information needed
 2. Decide the method of administration (5-6 pages back)
3. Content of Questionnaire: Time of respondent must not be wasted,
analyse how essential it is to ask the question., Motivating the respondent
is important – questions must engage.
4. Types of Questions -
i.Open-ended Questions: Your age? /My career goal is ____. How do you like
Rahul?
ii.Close Ended: I. Dichotomous – Yes/No, Middle Class/Upper Class (answer
have 2 options)
iii.MCQ
iv.Scales: Discussed previously

5. Criteria for Design

Clearly specify issue – Avoid Ambiguity – Avoid Leading Questions (favoured
towards one answer) – Avoid sensitive questions

6. Structure – Intro
Acknowledgement (to respondent)
Classification Information
Opening Questions
Study Questions
Close

A sequential order must be maintained

Physically the questionnaire must be well spaced, right font etc.

It can then be pilot tested to correct irregularities, followed by administering the
questionnaire.
Recommended – To see multiple questionnaire and form a framework.

Unit 4
 Sampling
Sampling Techniques
i.Probability Sampling Designs
a. Simple Random Sampling: Here nos. from 1 to 10 would be written on
slips, and then a slip from the box (containing all the slips) is randomly selected.
It may be replaced to keep the probability of selection of any slip same that is
1/10, or not replaced to increase the probability of selection to 1/9.
b. Systematic Sampling: The entire population is arranged in a particular
order according to a design. (calendar dates, alphabetical order). A sampling
interval K = N/n is calculated, say 1000 slips (population) / 20 slips (sample
size) = 50. A random no. is selected from 1 to 1000. Say 200. The next element
to be selected would be 200+50 = 250, followed by 300, 350 and so on.
c. Stratified Random Sampling: It involves dividing the population into
various strata to increase the representativeness of the sampling. E.g. Dividing
employees on the basis of salary, low, medium, high.
d. Cluster Sampling – Elements within cluster are heterogenous but
homogeneity exist between clusters. E.g. Different floors of a building would
have Manager, VP, President of a particular dept.
Therefor several floors would be homogenous in terms of the hierarchy of
workforce, but each floor would be heterogenous in terms of the positions
that exist within it.

ii. Non-Probability Sampling Design

a. Convenience Sampling: - Selecting Sampling units as per the
convenience of the researcher. It is used in the pre-test phase of a research
study. E.g. People interviewed for their political opinion at market.
b. Judgemental Sampling – The judgement of an expert is used to identify a
representative sample. This approach empirically (by means of experience) may
not produce satisfactory results.
c. Snowball Sampling – It is used when difficulty arises to select members
of desired population. Each respondent after being interview is asked to find one
more in the field.
d. Quota Sampling – The sample is selected on the basis of certain
demographic characteristics such as age, gender, occupation etc.

Determination of Sampling Size

1. Researchers may arbitrarily decide the size of the sample without giving
any attention to its accuracy or the cost of sampling. This should be
avoided.
2. The total budget of the survey maybe used which when divided by the
cost of sampling one unit would give us the total no. of samples.
3. Confidence Interval Approach may be used.
● Higher is the variability of the population (measured by s.d.) larger
would be the size of the sample.
● Higher the confidence researcher wants for the study, larger will be
the sample size. (measured by the value of Z)
● Greater precision means lower allowable margin of error, thus
resulting in a larger sample size.
Factor Analysis
It is a multivariate statistical technique in which no distinction exists b/w
dependent or independent variables. All variables are analysed together to
extract the underlined factors. It is a Data reduction method.
A researcher might have collected data on more than 50 attributes of a product
which is difficult to analyse. Factor analysis can help reduce the data to a few
manageable factors.
(A factor is a linear combination of variables.)
Uses:
1. Scale Construction: The process of Scale Construction starts with
generating a large set of items. Factor analysis can reduce the set of
statements to a concise instrument that represents the critical aspects of
the construct being measured.
2. Establishing Broad Factors: Reducing multiple input variables into
grouped factors called Broad Factors.
3. Psychographic Profiling: For identifying personality types
 4. Segmentation Analysis: Different set of pizza consumers would give
different importance to the same set of factors.

Discriminant Analysis
It is used to identify the variables or statements that are discriminating and on
which people with diverse views will respond differently.
The objective here is to find a linear combination of variables that discriminate
between categories of dependent variables in the best possible manner.
It can be used in the following ways:
1. Scale Construction: If one wants to assess two different views of two
different groups on the same matter, then one may generate a no. of
statements and conduct a pilot study to select only those statements on
which the two groups differ significantly.
2. Segment Discrimination: Understanding what key variables on which two
or more groups differ from each other can be done by Discriminant
Analysis.
3. Perceptual mapping: To create attribute based spatial maps of the
respondent’s mental positioning of brands which helps to determine
which attribute is the USP of which brand and which attributes are valued
by respondent that no brand satisfies.

Practical Tests: ​
Z-test (Mean, Difference of Mean, Difference of Proportion)
T-Test (Mean), Paired T-test
Chi-Square Test
One-Way ANOVA.
Extra Information
Qualitative Research Techniques: Collects Qualitative Data Depth Interviews,
Focus Groups, Projective Techniques
Quantiative Research: Collects Measurable/Quantitative data via Online
Surveys, Paper Surveys etc., Observations, Secondary Sources
Type I Error: Rejecting a true null hypothesis. (measured by alpha) (false
negatives) - Can be controlled - Usually Smaller
Difference: Defn - Example - Measuring Technique - Alt. Name - Control
-Magn
Type 2 Error: Accepting a false null hypothesis (measured by beta) (false
positives) (can’t be controlled)(usually greater)
(False Positive and Negative can be understood if
H0: The drug does not have any effect
H1: The drug does have effect
Type 1: Rejecting H0 when true: Drug does have effect you believe in but
actually it does not
Type 2: Accepting H0 when false: So drug does not have any effect is what you
believe but it actually does)

Random Sampling/Probability Sampling - Every element in our population has

non zero probability of being selected as a part of the sample and we have
accurate knowledge of this probability
Non Random Sampling/Non-Probability Sampling - It is sometimes not possible
to have every element have non-zero probability of being selected. The elements
which can not be selected as a part of the sample are called excluded units.

Criteria For Measurement (of Scales)

1. Reliability: It is concerned with the consistency, accuracy and
predictability of a the scale.
- Test-Retest: Repeated measurements of the same person under similar
condition.
- Split-half method: All the items are divided into two halves measuring
same construct, and the correlation between two is obtained. (higher the
better)
2. Validity - Measuring what we intend to measure
 - Content Validity: Experts judgement is used to measure the aptness of a
construct.
- Concurrent Validity: Validity of new measuring technique by correlating
with established technique.
- Predictive Validity: Ability of a measured phenomenon at one point of
time to predict another phenomenon at a future point of time, and
correlation is drawn between the two.
3. Sensitivity: Ability to accurately measure the variability of a concept.

End.