Mod 3.1 (big)research Designs

1

Module 3.1 RESEARCH DESIGN (RD)

•

RD is the framework or blue print or plan for the collection, measurement & analysis of data, so as to obtain answers to research questions

to

solve

management problems • Research design lays down the method and procedure for

collection

of

required

2

information

and

measurement

&

its

analysis

with a view to arriving at certain

meaningful

conclusions for management decision making. •

So a Research Design:-

– is an activity and time based plan –

is

always

based

research question

on

the

3

– guides sources

the

selection

of

types

of

&

information. – is a framework for specifying relationships

among

the

study’s variables. – outlines

procedures

for

every research activity. – Provides

answers

for

questions such as (i)

The

methodology

that

research will

employed in the survey

be

4

(ii)

What techniques will

be used to gather data? (iii)What kind of sampling will be used? ( iv)

How will time & cost

constraints be overcome? Types of Research design – three types RD classified into – Explanatory RD –

Descriptive RD ) Conclusive RD

5

–

Causal RD

)

---do----Refer

earlier

figure

in

research process

1. EXPLORATORY RESEARCH (ER) Exploratory research is a study that seeks to develop initial

6

insights

and

to

provide

direction

for

any

further

research

needed

regarding

problems that are vague or lack clarity. •

Exploratory Research could be

used

for

any

of

the

followimg purposes:oDefine a problem more precisely oIdentify

alternative

courses of action. o

Develop Hypothesis*

7

o

key

Identify / Isolate variables

for

further

examination o

Gain

insights

for

developing an approach to the problem oEstablish

priorities

for

further research. oIncreasing familiarity

an

analyst’s

with

problem oClarifying concepts.

the

8

• Exploratory study is helpful in

breaking

problem

broad

vague

statements

into

smaller, more precise sub problem statements. This will be helpful to form specific hypothesis. • In

the

early

stages

of

research, we usually lack sufficient understanding of the problem to formulate a specific hypothesis due to several

tentative

9

explanations

for

a

given

marketing concept. E.g.

1. Sales are low

because our prices are too high; 2.Our dealers are not doing their job properly 3.Our representatives

are

sales not

following up well 4.Our advertising is weak etc.,

10

• Exploratory Research can be

used

to

establish

priorities in studying these competing

explanations.

Priorities

would

established

subsequently

because hypothesis the

a

be

particular

discovered

exploratory

in

study

appears to be promising. •

Exploratory

research

is,

therefore, characterized by flexibility and versatility

11

w.r.t

methods

because

formal

research

protocols

and

procedures

are

not

employed. •

E.R

rarely

structured large

involves

questionnaires, samples

and

probability sampling plans. •

Once a new idea or insight is discovered, exploration is redirected in that direction until

its

possibilities

are

12

exhausted or failing another direction is sought/found. •

So focus of investigation may shift constantly as new insights are discovered. creativity

Thus,

and

ingenuity of the researcher plays

a

major

role

in

exploratory research. *

Hypothesis

unproven

is

an

statement

of

research question

13

*

Hypothesis

statement

that

is

a

specifies

how two or more measurable variables are related

and

the

tests

relationship. Vauge Problem Exploratory Research Hypothesis New Ideas

Conclusive Research Decision

Types of Exploratory Studies

Literature Search (Secondary data) Experience survey (Opinion of experts) Focus Groups (Qualitative Research) Analysis of selected cases

that

14

Exploratory Research involves assessing the quality of iiry data :1.Secondary

sources

information

eg:-

of Trade

Journals, published statistics (external sources) Company’s

own

records.(internal sources) Both

are

analyzed

in

a

qualitative manner, including quantitative data

15

2.Experience Survey’s refer to the

informally

opinions

and

gathered insights

of

knowledgeable experts in the field. 3.Focus groups no

of

are a small

individuals

brought

together in a room to discuss some topic of interest to the focus of sponsor to gain new insights to the problem

16

4.Analysis of selected cases of similar problems from past records 5.Unstructured

interviewing

of concerned people eg:- if in marketing the sales people and distributors are interviewed. It is an in depth study of subjects of interest that will serve as inputs to understand the problem on hand.

17

Other

sources

can

be

customers, stores, websites, market areas etc., Eg:- Exploratory Res. Design:Take the case of a chocolate manufacturer whose product was found defective (infected with worms) at the retail end. Application of exploratory res can help identify the problem. With the batch no. of the product records

the

production

(secondary

data)

18

pertaining to that batch can be reviewed. Experience

survey

can

get

opinion of experts to find out where the production process could have gone wrong. This may help management to look at specific production records which may reveal a defect in the sterilization process of the production or lack of proper supervision or both. This will

19

help take necessary corrective action. Subsequently, a descriptive research study can help find the effectiveness of publicity campaign and image building exercise. EXPLORATORY

RESEARCH

MORE EXAMPLES 1.A chocolate manufacturer wants to identify the ten most important variables his

20

consumers use to decide in buying a brand of chocolate. The

results

exploratory

of

study

this could

provide him with inputs for a second study using factor analysis

techniques

to

reduce the ten variables into a smaller set of factors. 2.Focus

group

discussion

among

house

wives

to

debate

the

fortune

of

convenience foods in India.

21

It may be used to throw in new

ideas

products

about or

modifications

new

suggest to

existing

products, modifications in packaging etc., through an inhibition free discussion. 3.

Soft

drinks

manufacturer

faced with decreasing sales might exploratory generate

conduct

an

study

to

possible

explanations. The study may

22

establish the attributes that consumers seek in a soft drink to gain satisfaction. An

open

ended

question

may result in consumers freely opinion

expressing which

their

can

be

reduced to some ten key factors

(by

experts).

A

consulting Focus

group

discussion may reveal that the present soft drink meets most consumer expectations

23

and packing is satisfactory. The

focus

exploratory

of

further

research

may

shift towards study of sales force and distribution. 4.

3M

home

conducted research steel

products exploratory

on

wool

its

standard

pads

through

focus groups. The research revealed that standard steel wool

pads

expensive

scratched

cookware.

This

24

finding produced the idea for “Scotch Brite” scrub pads made of gentle fibers, reducing

the

odds

of

damaging cookware. 5.

A luxury car manufacturer may want to know the real reasons why consumers buy up market

limousines through in-depth interviews and utilize this message / information in

25

product

positioning and

creative ads. 2. DESCRIPTIVE RESEARCH •

It is a type of conclusive research that has its major objective

description

of

subjects

of

interest

in

such

as

various

fields

marketing,

finance,

HR

etc., • It is intended to generate data

describing

the

26

composition

and

characteristics of relevant groups such as consumers, sales people, organization and market areas. •

1.Descriptive conducted

research for

is

following

reasons. 1.

To

describe

characteristic

the of

relevant groups such as consumers sales people, organizations or market

27

areas eg: - Developing a profile

of

shoppers

the of

heavy

Shopper’s

Stop, Pantaloons Dept Stores. E.g.:-2. profile of an average user of

product

w.r.t Age, Income, Gender, Education level etc., 2.

To

estimate

the

percentage of units in a specified

population

28

exhibiting

a

certain

behaviour

eg:-

Estimating

the

percentage who

of

people heavy

are

shoppers of Shopper’s Stop

who

vacation

also

take

packages

of

Mahindra Holidays. 3.

To

determine

the

perceptions of product characteristics eg:- how do households perceive

29

supermarkets like Food World, Nilgiris, to Local Grocery Stores. Eg:-

how

do

people

perceive herbal products 4.

To determine the degree to

which

variables

marketing are

associated eg:- to what extent is shopping at super markets related to eating out

30

5.

To

make

predictions eg.,

specific what

will be the retail sales of up-market shirt brands in

Shoppers’

Stop

outlets in Bangalore

•

2. Descriptive studies are rigid > It needs to

specify how the study is to be conducted.

31

> It requires a clear specification of who, what, when, where, why and how, of the research eg:- who should be considered a regular customer of a

particular Dept Stores.

- What information is needed from respondents - How should the information be obtained - Where should the respondents be contacted - Why are we obtaining this information -Similar questions need be asked until information to be obtained is clearly defined. >Hence,descriptive research is preplanned and structured.

32

3. It is typically based on relatively

large

representative samples. •

A formal Research Design specifies the methods for selecting

the

sources

information and

of

collecting

data from those resources. •

Descriptive marked

research by

a

is clear

statement of the problem, specific

hypothesis

and

33

detailed information needs. l. Eg. 1. Market studies to describe size of market, either local or Overseas ; Gallup nationwide survey of China on profiles, attitudes and life style trends (1997) is an example of Des. Res. Study at national level.

34 •

Size

of

power

markets, of

Buying

consumers,

Distribution chain Consumer profile etc. Eg. 2. Market Share studies

Eg. 3. Sales Analysis

studies Eg. 4. Image studies Eg. 5. Product usage studies Eg. 6. Distribution Studies studies

Eg. 7. Pricing Eg 8. Ad. Studies. Etc.,

35

•

A

majority

of

research

studies in marketing

are

descriptive research which incorporates following major methods: i)

secondary data analyzed in a quantitative manner

ii) surveys iii)panels iv) Observational data.

and

other

36

Combination of exploratory research

and

research

descriptive

example

to

understand. The miller Brewing Co. in the US

was

the

pioneer

in

introducing light colour beer, light weight beer and offered it as a premium product with a higher price. The product was designed for the efficient, the occasional drinkers and women

and

sold

only

in

37

bottles to be drunk at pubs and homes. Over the years competition

increased

with

similar products miller lost its product distribution and also its market share and fall from top slot to eighth place. Millers investigated the beer market. On the basis of its market contacts, pub owners, bar tenders, established

the company through

exploratory research that blue

38

collar workers are the big market. On the basis of this the

company

could

hypothesize that “Blue collar workers

are

the

biggest

market segment for the beer in the US” The company commissioned a market

research

to

get

a

profile of the American beer drinker research

----

a

study

descriptive -----

that

revealed that the heavy users

39

of beer were young, male and highly concentrated in the blue

collar

occupations

confirming the hypothesis. The study also revealed that he drank at a bar, not at home, and he drank with his buddies from the job and not with

his

wife.

It

further

estimated that some 20% of the drinkers consume 90% of the beer and that some 60% of the adult population does

40

not drink beer at all. The study helped miller make certain product modification – a

stronger

beer

made

available on taps in pubs, with promotion to go with it – and regain the top slot second only to Budweiser.

Difference between Exploratory and Conclusive Research Sl Research Project no. Components

Exploratory Research

Conclusive Research

41 1.

Research purpose

General :- To generate insights about a situation

2. 3.

Data Needs Data sources

4.

Data Collection

Vague Secondary defined Open ended rough

5.

6. 7. 8.

Specific:- To verify insights and aid in selecting a course of action. Clear Well defined i ry

Usually structured and smooth Sample Relatively small: Relatively large: Subjectively objectively selected to selected to maximize permit generalization of generalization of insights findings. Data collection Flexible : no set Rigid: well laid procedure out procedure Data analysis Informal: typical Formal typical qualitative quantitative. Inferences/recommendations More tentative More final than than final tentative

42

43

CAUSAL /EXPERIMENTAL RESEARCH •

This is intended to generate the

type

necessary

of for

evidence confidently

making causal inferences / reasons

about

relationships

among

variables. •

Helps decide level

of

the

the optimum independent

variable in order to get the

44

desired

result

of

the

dependent variable. • Helps

establish

consumer

preference levels. Eg. Consider a consumer goods firm that wants to determine the impact of advertising on sales. To achieve this objective, the

firm

can

proceed

follows: How to do this?

as

45

1.Select a group of distinct market

areas

that

have

similar demographic, socio economic

and

competitor

characteristics. Eg

Jayanagar

/

Koramangala ; or Chennai / Bangalore 2.

Vary the level of advertising expenditure variable)

from

(independent market

to

market, keeping all other marketing variables, such

46

as

price

and

promotion

constant. 3.

Monitor

sales

(dependent

variable) over a sufficient length of time 4.

Analyze

the

data

to

see

whether or not the pattern of variation in sales across markets is consistent with the pattern of variation in advertising expenditure.

In

experimental design, it is called

47

Causal

Variable

=

Independent Variable Effective

variable

=

dependent variable •

To make causal inferences with confidence, we must manipulate variable and

the

(eg: study

variables

causal

advertising) the

called

other effect

variable, namely sales •

Another condition is that the

causal

variable

48

(advertising)

&

effect

variable (sales) must occur in

proper

the

time

sequence. To ascertain if advertising first

we

causes must

sales, change

advertising levels (without being

influenced

anticipated measure

the

sales) change

by and in

sales later •

A plan is developed for controlling

conditions

49

relevant to the experiment so that some test unit or experimental units can be exposed experimental

to or

the causal

variable and their results measured. •

The

researcher

hypothesis

that

has

on

if

an

experimental variable (egadvertising, shelf display, training) is applied to an experimental unit (eg. a

50

group

of

store,

consumers, some

a

sales

representatives etc) it will have a measurable effect (eg

the

No.

remembering

of

people

the

brand

name, no. of units sold, No. of calls made etc.,) What is experimentation It is a research process in which

one

or

more

variables are manipulated, so as to demonstratate the

51

cause

and

effect

relationship. Experimentation is done to find out the effects

one

factor/

the

variable

on

other facor/variable. Fig.

Different

Elements of Experiment Input

Extrane Outpu ous variable s

t

52

Experim ental

Experi

Variable mental 

test units

Variable OR Explanat Extrane ory

dent or

units or Effect

OR Indep.

Depen

ous

Variable variable s

Variabl e

53

Test Units: These are units on which the experiment is carried out. It is done with one or more independent variables controlled by

a researcher

to find out its effect on a dependent variable Eg.

consumers,

markets,

stores, product etc Explanatory/ Independent / Experimental variable:

54

These are variables whose effects the researcher wishes to

examine.Eg.advertising,

pricing, packaging, display, promotion etc Dependent Variable: This is the variable/ factor which is under study. Eg. Sales, consumer attitudes, brand loyalty Extraneous Variables These are also known as blocking variables and affect

55

the

results

experiments.

of

the

These

are

present in the environment of the experiment. They are of two types. Controllable

extraneous

variable This is under the physical control of the

researcher.

Eg Price of a product under experiment Uncontrollable extraneous variable

56

This may be factors in the macro environment not in the

control

researcher. condition

of

Eg. of

competition,

the

economic

the

market,

govt.

policies

etc., Eg.1

Suppose

a

toffee

manufacturer is making an attempt

to

measure

the

response of the buyers to two

different

types

of

packaging, at two different

57

locations.

For

the

two

groups of buyers, all factors of the product must remain same except the packaging. Suppose the manufacturer changes the price for one group,

the sales will not

indicate which packaging is preferred because the price has acted as an extraneous variable. physically hence

But price can be contolled

called

and

controllable

58

extraneous variable. Eg.2 Suppose a company is introducing a product in two different cities. It would like to

know

the

impact

of

advertising on sales. During the experiment, suppose a major

competitor

in

the

market has a strike in the factory and maintain

is unable to

supplies

in

the

market. Now the researcher cannot conclude that his

59

sales

in

that

increased

city due

has to

advertisement. In this case, strike

the

is

an

uncontrollable extraneous variable

or

variable as can

be

confounding no conclusion

reached

on

the

experiment. •

Other studies

experimental

60

- impact

of

change

retail

price

on

in

brand

market share - Effect

of

different

advertising themes upon sales of a product -

Impact

of

certain

promotional offers -

To

measure

sales

effectiveness of a new packaging display etc.

/

shelf

61

•A

research

manipulate

study on

may

or

more

such variables & examine the

effect

of

this

on

consumer action Special

types

of

Exptl,

Design used in MR 1.

After Only Design / Purely Post-Design This

design

measuring

the

consists

of

dependent

variable after exposing the

62

test unit (consumers) to the experimental / independent variable. Eg1:- Consumers are given a discount coupon along with the

newspaper

(independent

ad

variable)

to

buy the brand. The study measures

the

coupons

redeemed

coupons

extent

collected

at

of (ie the

retail level) and the resulting sales (dependent

variable)

63

from the sample covered ie., consumer response

to an

ad in a given market can be measured by relating it

to

sale of the product) Eg2:-

Suppose, HLL wants

to conduct

an

experiment

on the “ Impact of free sample on the sale

of toilet soaps.

A

small sample of toilet soaps are

mailed

customers

to

selected

64

2.Before – after Design In this design, each test unit is

measured

dependent

on

the

variable

once

“before” and again “after” exposure

to

experimental variable.

the The

difference between the two measures is treated as the

65

effect

of

exptl

/

independentg variable (ad + Coupon = ind variable) ie., sales before ad for a given period and after ad for the same period, the difference can be attributed to effect of the

advertisement

(ie.,

independentg variable). 3.Before

–

after

with

Control GP Design In this design the research study

includes

a

control

66

group in the experiment, but the

control

group

is

not

subject to the experimental variable. Sales/Mkt. Share

Sales/Mkt. Share

Measurement Group

Exptl.

Control Group

Before

E1

C1 After

E2

C2

67

The

idea

of

choosing

the

control group is to incorporate any change in the result due to same

extraneous

varables.

Then the effect of experimental variable is given by (E2-E1) – (C2-C1) E.g.: Consider the example in the

‘before

–after”

design.

Suppose the market share had increased

by

10%

(Say

E1=25%, E2=35%) between the advertising

campaigns.

68

However suppose in this period there was a general decrease in the market share by 4% (due to say dull season), i.e.,

C1

25%

assuming

similarity, C2 21% dull season effect. then

the

net

effect

of

the

advertising should be (E2-E1) – (C2-C1) because C2 > C1 10% - (-4%) because C2 > C1

69

10% + 4% = 14% season effect Note: In general, it is assumed that members of the control group are similar in nature to the experimental test units. Suppose the general market share increased by 4% owing to boom season then net effect would be: 10%-4%

(E2-E1) – (C2-C1) C125%

Net effect 6%C2 29%

70

4. “After-Only with control group” design. The

effect of experimental

variable

is

computing between

determine the

the

by

difference after

two

measurements. i.e E2-C2 = net effect In

test

marketing

experiments conducted

are to

such

frequently

examine

the

possible impact of the new

71

product in the market with appropriate

selection

experimental

and

of

control

markets. Eg:-

Suppose, HLL wants to

conduct an experiment on the “ Impact of free sample on the sale of toilet soaps. A sample of toilet soaps are mailed to selected locality.

customers

in

a

After a month, a

coupon of 50 paise off on one cake of soap is mailed to each

72

customer to whom free sample was sent earlier. An equal no of these coupons are mailed to people (control

in

another group)

neighborhood.

The

locality in

the

coupons

are coded to keep an account of the no of coupons redeemed from

each locality. Suppose

400 coupons were redeemed from exptl., group and 250 coupons were redeemed from the

control

group

.

The

73

difference 150 nos is the effect of free samples In this method, conclusion can be only drawn from the two “after” measurements.

Factorial Design (FD)

74

FD permits the

researcher to

test two or more variables at the

same

time

FD

helps

determine the effect of each of the variables

and measure

the interacting effect of many variables Eg., A huge dept stores wants to study the impact of price reduction for a product . There is promotional display

being

carried out in the stores a)

near the entrances B1;

75

b)

at

the

actual

sales

counters B2; c)no

promotional

displayB3 Now assume mgmt applies two price

levels

namely

regular

price A1 and reduced price A2. There

are

2

x

3

six

combinations Which of these combination is best suited?

76

There are 60 dept stores of the chain

divided into groups of

ten each Now

randomly

assign

ONE

combination to the ten stores as follows Combination Sales B1A1 S1 BIA2 S2

77

B2A1 S3 B2A2 S4 B3A1 S5 B3A2 S6 S1 TO S6 represents the sale resulting from each variable combination.

The

data

gathered will provide details on product sales on account of

78

two

independent

variables

price and promotion Outcome

of

the

experiment

could be as follows 1.Price redn with display at entrance 2.Price redn with diplay at actual sales counters 3.No display and regular price 4.Display at entrance with regular price Advantage

79

There is a possibility of testing every

combination

of

the

variables Disadvantage As the no of variables increase, the

no

of

combinations

increase exponentially, making the experiment cumbersome. In the above example if three types of packaging were to be included,

the

combinations

increase to 18 (2x3x3=18) Latin Square Design (LSD)

80

In

LSD,

the

experiment square.

It

layout

of

the

represents

a

is

classification blocks

a

that

the

effect

out by

extraneous the

randomization (effect

of

way

scheme

confounding factors

two

row

process

of

of

treatments

and

column

variables) within each cell. The researcher chooses to test three shelf arrangements A, B, C , in three stores. He would

81

like

to

observe

the

sales

generated in each of the stores at different time periods.

The

researcher must make sure that

one

type

of

shelf

arrangement is used in each store only once and ensures this is not repeated in any row or column If we choose three stores and three time periods the total no of combination (treatments) is

82

3x 3 = 9 as follows ( with random selection of A,B & C): Time

period

Stores 1 2

3

T1 C

B A

T2 A

C B

T3 B

A C

83

A single dependent variable SALES is tested in the above combination in each of the combination

above

and

its

siginificance is tested by

ANOVA Ex-Post Facto Design In EPFD,

the

researcher

tries to find out one or more causal variable that can give

84

plausible( likely ) explanations as to why variable

the dependent

reflected

adverse

values. Eg 1 Why the sales in a showroom measures

dropped. the

causes

It in

retrospective affect. Eg 2 A magazine publisher wants to ascertain the impact of an advertisement on frozen foods in

“Women’s Era”.

He

asks those who read and do not read the magazine about

85

frozen foods. The difference in response

indicates

the

effectiveness of the ad - after effects of the ad.