Chapter 3

CHAPTER 3

RESEARCH METHODOLOGY

This chapter explains the research methodology which deals with the scientific
approach of solving research problems. The chapter presents a systematic way of
converting the research gap identified in literature review into a researchable problem
statement which ultimately leads to the objectives of the study. The chapter also
highlights the research questions and hypotheses of the study. The chapter describes the
research design, data collection methods, development of the scale, sampling plan and
data analysis plan used to achieve the research objectives. The chapter briefly elucidates
the limitations of the methodology.

3.1 PROBLEM STATEMENT

The effect of digital marketing communication on product categories like books,
music, fashion accessories, clothing, banking and online gaming and many others has
been well researched by the researchers but automobile industry despite being one of
the largest digital spenders has faced dearth of academic studies especially in India.
Moreover, few studies have thoroughly documented the effect of digital marketing
communication on each stage of consumer buying decision process. This gap in
existing body of knowledge was identified and this study was envisaged to fulfill the
gap. An empirical academic study emphasizing upon explaining the impact of digital
marketing communication on entire consumer buying decision process along with
discovering the most preferred digital channel, would not only help the car marketers
who are spending considerable amount of money on various digital media platforms,
but would also help the customers in taking an informed decision. There remains little
research about the preferred digital channels of communication while searching for
information about cars as majority studies have used the term digital media as a collective
term for representing all the content based digital platforms and device. So, a study that
could identify the specific digital channel used at particular stage of buying decision
process would ascertain the most appropriate digital channel to be used for each stage
of CDM for the marketers.
Although, many studies tried to examine the demographic details of the
consumers using digital media, yet there remained a need for a study that could
precisely analyze the demographic variables so that marketer might use them for
effective segmentation. Moreover, an important geographical variable namely the place
of living was not considered by majority of the studies. Integrating this important
geographic variable with demographic variable could help marketers fine-tuning their
targeting strategies which would ultimately result in better return on investment. The
current study tries to overcome this limitation by comprehensively understanding the

62
 Chapter 3

demographic profile of the consumers along with making an attempt to discover the
important customer segments for marketers to target.
It was observed that digital marketing communication offers several advantages
to customers over traditional mode of communication, yet there are issues of trust,
authenticity and technicality involved with the usage of digital marketing
communication. It was seemed appropriate to see the integrated effect of various
benefits and issues on the digital marketing communication‟s usage while making a
buying decision as majority of the studies have examined them separately. The study
identified this gap demanding due attention as results might detail the difficulties faced
by the customers while using the digital marketing communication which if solved by
the concerned marketers would lead to an effective and successful marketing campaign.
The findings could yield more sound practical implications for the marketers if they
know why at all people don‟t use digital marketing communication while buying a car.
Consumers‟ attitude which deals with a predisposition to behave in a favorable
or unfavorable way towards any object, person or phenomenon, was cited as one of the
most important reason leading towards increased usage of digital marketing
communication in buying decision process across product categories. However, studies
concerning attitude towards digital marketing communication in Indian passenger car
market were not sufficient enough to demonstrate the importance of this worthy
marketing construct. Any study that would highlight the general attitude of consumers
toward digital marketing communication would result in better understanding of
customers‟ needs by the marketer. It was apprehended from the past researches that
apart from attitude, influence of other people, belief in one‟s own capabilities and
technology related factors like cost, easiness, and trust were also found as motivating
factors leading towards usage of digital marketing communication in buying decisions.
A study that might see the combined effect of attitude and other people‟s influence
given the constraints of one‟s capability, resources and very nature of technology might
reveal the meaningful insights for the marketers.

3.2 OBJECTIVES OF THE STUDY

The main objective of the study was to understand the impact of digital
marketing communication on consumer buying decision making process in Indian Car
market. Apart from the main objective, several sub-objectives pertaining to
demographic profile of people using digital channels, most preferred digital channel,
consumers‟ attitude, trends of digital marketing in Indian passenger car market, general
difficulties faced by people and scope of leveraging any factor related to digital
marketing communication were conceptualized. Precisely, objectives of the study were
as follows:
 To know the demographic profile of the customers using different digital channels
while buying a car.
 To identify the preferred digital channels used by the consumers while buying a
car.
 To measure the consumers‟ attitude towards digital marketing communication in
Indian Car Market.
 To understand the impact of digital marketing communication on consumer buying
decision making process in Indian Car market.
 To study the factors if any, related to digital marketing communication which can
be leveraged by the marketers to alter the consumer behaviour.

63
 Chapter 3

Apart from the major objectives, two secondary objectives were also envisioned
to better understand the research area:
 To study the trends in digital marketing communication in Indian Passenger Car
Market.
 To study the general/technical difficulties if any, faced by consumers that might
prevent them using digital channels while buying a car.

3.2.1 Research Questions

Following research questions were set up for achieving the research objectives:
 What is the demographic profile of the customers using digital channels of
communication while buying a car?
 Is there any preferred channel of communication used by the customers while
buying a car?
 What is the consumers‟ attitude towards digital marketing communication in
Indian car market?
 What is the impact of digital marketing communication while buying car on
consumer buying decision process?
 Are there any factors related to digital marketing communication that could be
leveraged by the marketers to alter the consumer behaviour?

3.3 HYPOTHESES OF THE STUDY

Hypotheses of the study were as follows:
H01: There is no relationship between demographic profile of the consumer and type of
digital media channel used while buying a car.
H02: There is no preferred channel of communication used by the customers while
buying a car.
H03: Consumers have either indifferent or negative attitude towards digital marketing
communication while buying a car.
H04: There is no impact of digital marketing communication on consumer buying
decision process while buying a car.
H05: There is a no scope of leveraging the factors related to digital marketing
communication to alter the consumer behaviour.

3.4 RESEARCH DESIGN

The present study is descriptive in nature that aims to analyze the impact of
digital marketing communication on consumer buying decision process while buying a
car. The study included all the communication channels used by Indian passenger car
marketers namely websites, social networking sites, mobile phones, YouTube, digital
TV, digital outdoors and online communities. The study identified the key
characteristics of digital marketing communication by conducting an „elicitation‟ study
and a comprehensive literature review and phrased the questionnaire‟s statements in
tandem with the popularly believed characteristics of digital marketing communication.
During the study, one time contact with the respondents was made to collect the desired
information which made the study cross sectional in nature.

64
 Chapter 3

3.5 DATA COLLECTION

The present study is „empirical‟ in nature which aimed to understand the impact
of digital marketing communication on consumer buying decision process on the basis
of knowledge derived from actual experience rather than theoretical beliefs. Empirical
studies of such nature normally measure the observed phenomenon and bank heavily
upon the primary data to arrive at the conclusion. The theory was developed on the
basis of secondary data obtained from secondary sources like books, journals,
magazines and internet.

a) Primary Data
Primary data refers to the data collected by the researcher himself for a specific
purpose. Primary data was collected from the respondents (actual and potential car
buyers) with the help of research tool questionnaire. The questionnaire made use of
nominal and interval data. Both open-ended and close-ended questions were used in the
questionnaire. Survey method was used for collecting primary data. Physical and online
modes were used for collecting primary data.

b) Secondary Data
Secondary data refers to the data collected by someone else and used by
researcher for the problem in hand. The present study made use of secondary data to
have an in-depth understanding of the concept of digital marketing communication.
Secondary literature including journals, books, internet, newspapers and magazines,
was reviewed to understand the conceptual framework of the study.

3.6 SCALE DEVELOPMENT

A mixed research methodology was adopted for the study as mixed
methodology is considered better than the pure qualitative and quantitative researches
and normally used when researcher is interested in developing the theory about the
phenomenon and later on tests it (Greene, 2006). Mixed methodology results in more
balanced and informative research results (Johnson, Onwuegbuzie and Turner, 2007).
Initially, an elicitation study was conducted with small group of respondents
comprising of six colleagues, two superiors, one research supervisor, two car
enthusiasts, one car dealer, one employee in an IT company and ten management
students to know the likely effect that should be observed on consumer buying
decision process in response to digital marketing communication practices while
buying a car. Elicitation studies help researchers in identifying the true beliefs of
target population and help reducing the researcher‟s bias thereby aiding in the
replication of the theory irrespective of the context (Downs and Hausenblas, 2005).
Fishbein (2000) asserts that elicitation process makes the theories applicable to all cultures,
thereby increasing the breadth of the research results.
Review of available literature also revealed the important characteristics of
digital marketing communication which proposed digital communication as the most
preferred channel of communication both for the consumers and marketers. These
characteristics of digital marketing communication were also found affecting the
consumer behaviour as per reviewed literature. The characteristics were later on used

65
 Chapter 3

as variables in the scale developed for assessing the impact of digital marketing
communication on consumer buying decision process. Table 3.1 exhibits the important
characteristics of digital marketing communication along with their reference studies,
unraveled by literature review.
Table 3.1: Variables Used in Scale Development

Sr. Digital Marketing Reference Studies

No. Characteristics
1 Interactive Kierzkowski et al. (1996); Deighton (1996); Reza Kiani (1998); Becherer and
Halstead (2004); Prahalad and Ramaswamy (2004b); Bauer, Barnes, Reichardt,
and Neumann (2005); Corniani (2006); Bird (2007); Egan (2007); Chester and
Montgomery (2008); Charan (2011); Muller et al. (2011); Smith (2011);
Srivastava (2012); Verbaas (2013); Kaptyukhin et al. (2013); Tiago and
Veríssimo (2014); Aqsa and Kartini (2015); Edelman and Heller (2015); Jenyo
and Soyoye (2015); Yasmin et al. (2015)
2. Measurable Kierzkowski et al. (1996); Admin (2006); Merisavo et al. (2007); Reitzin
(2007); Ferrar (2010); Charan (2011); Kaptyukhin et al. (2013); Verbaas (2013);
Hemann and Burbary (2013); Yasmin et al. (2015).
3. Informative Kierzkowski et al. (1996); Reza Kiani (1998); Wind and Mahajan (2002);
Becherer and Halstead (2004); Koiso-Kanttila (2004); Bauer et al. (2005);
Bayo-Moriones and Lera-López (2007); Merisavo et al. (2007); Kink and Hess
(2008); Constantinides and Fountain (2008); Charan (2011); Muller et al.
(2011); Smith (2011); Srivastava (2012); Klososky (2012); Stokes (2013); Aqsa
and Kartini (2015).
4. Relevant Reza Kiani (1998); Egan (2007); Mulhern (2009); Charan (2011); Muller, et al
(2011); Jones et al. (2011); Reitzin (2007); Kirkpatrick (2012); Accenture Study
(2014); Syme (2015).
5. Personalization/ Kierzkowski et al. (1996); Allen (1999); McDonald and Wilson (1999); Wind
Customization/ and Mahajan (2002); Becherer and Halstead (2004); Bauer et al. (2005);
Individual Corniani (2006); Bird (2007); Reitzin (2007); Merisavo et al. (2007); Charan
(2011); Ferrar (2010); Smith (2011); Jones et al. (2011); Kaptyukhin et al.
(2013); Kaufman and Horton (2014); Hawks (2015).
6. Convenient/Easy/ Becherer and Halstead (2004); Koiso-Kanttila (2004); Swaminathan et al.
Compatible (2005); Bayo-Moriones and Lera-López (2007); Kink and Hess (2008);
Delafrooz et al. (2009); Muller et al. (2011); Hooda and Aggarwal (2012);
Khalil (2014); Nguyen (2014); Yasmin et al. (2015).
7. Cost effective Kierzkowski et al. (1996); Becherer and Halstead (2004); Koiso-Kanttila (2004);
Bayo-Moriones and Lera-López (2007); Merisavo et al. (2007); Reitzin (2007);
Muller et al. (2011); Klososky (2012); Charan and Jaiswal (2012) Yasmin et al.
(2015).
8. Customer Participation/ Deighton (1996); Novak and Hoffman (1997); Prahalad and Ramaswamy (2004b);
Content Creation/ Corniani (2006); Bayo-Moriones and Lera-López, (2007); Bird (2007); Charan
Feedback/Sharing (2011); Kaptyukhin et al. (2013); Stokes (2013); Hudson et al. (2012); Tiago and
Veríssimo (2014); Charan (2014); Yasmin et al. (2015).
9. Easy Availability/ Koiso-Kanttila (2004); Becherer and Halstead (2004); Kink and Hess (2008); Muller,
Accessibility et al (2011); Srivastava (2012); Punj (2012); Kaptyukhin et al. (2013); Yasmin et al.
(2015); Jenyo and Soyoye (2015); Aqsa and Kartini (2015).
10. Attractive/ Koiso-Kanttila (2004); Kierzkowski et al. (1996); McDonald and Wilson
Entertaining (1999); Mulhern (2009); Bauer et al. (2005); Srivastava (2012); Charan (2014);
Aqsa and Kartini (2015).
11. Quick/Time saving McDonald and Wilson (1999); Koiso-Kanttila (2004); Merisavo et al. (2007);
Mulhern, (2009); Smith (2011); Kink and Hess (2008); Hooda and Aggarwal
(2012).

Engel, Kollat and Blackwell Model (EKB) model was chosen for the study for
being simple, flexible, comprehensive and giving clear explanation of variables

66
 Chapter 3

affecting buying behaviour (Edison, 1993; Jisana, 2014; Schiffman, Kanuk and Kumar,
2010). Moreover, the model clearly appreciates the level of involvement in particular
buying situation and identifies and stresses upon the decision making process ranging
from need recognition, search for information, evaluation of alternatives, selection and
purchase and post purchase. The model is capable of handling new evidences about
consumer behaviour and has withstood the test of time since its introduction; which
makes it the most used model of consumer behaviour. The present study made use of
EKB Model to study the impact on consumer decision making process.
5-points Likert‟s Scale was developed after getting substantial input both from
the available literature and elicitation study. A pilot study was conducted with 80
respondents from Delhi and „Cronbach‟s Alpha‟ was used to check the reliability of the
scale. Content or face validity for the scale was ensured by phrasing the statements on
the basis of variables identified through literature. Table 3.2 represents the scale
developed for each stage after researching the literature along with the reliability
statistics for each stage.
Table 3.2: Scale for Consumer Decision Making Process

Sr. Stage Questions in the Scale Reliability of

No. the Scale
1 Need  Digital marketing communication makes me aware of my *
Recognition need of buying a car.
2 Search for  I have started using digital channels more than traditional .743
Information channels for getting information about car brands.
 Digital media provides huge quantity of information
about a car as compared to traditional marketing while
buying.
 I think Digital media reduces the time taken for collecting
necessary information while buying a car.
 Digital media helps reducing the visits to showroom
before buying a car.
 I feel digital media is more interactive as compared to
traditional media.
 I get more relevant information about cars over digital
media as compared to traditional media.
 I find digital media easy to use while buying car.
3 Evaluation of  It is easy to have expert advice about car over digital .797
Alternative channels as compared to traditional media.
 Customers‟ reviews available over digital channels help
me in evaluating the car brand.
 I feel positive towards that car brand that has maximum
positive reviews.
 I give more weightage to consensus recommendations while
selecting a car brand rather than manufacturer‟s claim.
 I can easily compare various car brands over digital media
as compared to traditional media.

(continued)

67
 Chapter 3

Table 3.2 (continued)

4 Selection and  I like the idea of using digital media channels as digital .791
Purchase channels help in easy selection of car dealer as compared
to traditional media.
 I can easily select the car's brand based on the expert's
advice over digital media as compared to traditional
media.
 I intend to join the consumer forums/blogs/SNS of car
brands.
 I would subscribe to the newsletter of car brands to avail
the important information.
 I participate in online discussions about car brands.
 I would buy car based upon the information available
over digital media.
 I would like to book a car online for purchasing.
 I find financing options more easily through digital media
sources as compared to traditional media.
 Digital marketing communication enables me to pre-pone
my car buying decision.
 I may change the brand considered for purchase after
searching online.
5 Post-Purchase  I would like to give feedback about my .733
satisfaction/dissatisfaction over digital channels.
 Using Digital Channels throughout the decision making
process while buying a car is good idea.
 I recommend my friends any exciting digital content
about cars that I come across.

* Measured with single statement only, so reliability statistic was not applicable.

It can be seen from the table 3.2 that reliability statistics for each stage of
consumer buying decision process was found more than .7, which showed that scale
used for the study was „reliable‟.
The study also made use of Theory of Planned Behaviour (Decomposed) model.
Attitude, subjective norms and perceived behavioural control were taken as key
constructs forming usage intentions which in turn affected the actual usage. Reliability
of the scale developed for the model was checked with the help of „Cronbach‟s Alpha‟.
Discriminant and convergent validity of the scale was checked with the help of AVE
(Average Variance Extracted), ASV (Average Shared Variance) and MSV (Maximum
Shared Variance). Scale development for DTPB has been explained in detail in chapter
7.

3.7 SAMPLING PLAN

Sampling plan for the study dealt with sample size determination, sampling
technique, sampling unit and sampling area. Details of sampling plan have been given
below:

68
 Chapter 3

(a) Sample Size Determination: Population Proportion Approach

As the study dealt with the usage of digital channels of communication
throughout the decision making process, „proportion approach‟ (Chawla and Sondhi,
2011) was used to determine the sample size. This conventional approach deals with the
degree of variability in the attributes being measured or distribution of the attributes in
the population. The approach is used when researcher is interested in knowing the
proportion rather than the mean in the population (Malhotra and Dash, 2012). Lwanga
and Lemeshow (1991) in their study also advocate the use of population proportion
approach especially when objectives of the study are clearly defined. The more
heterogeneous a population, larger the sample size required. As researcher aimed to see
the impact of digital marketing communication across rural, urban and semi-urban areas
across various income groups, different educational backgrounds and occupations, the
population for the study was heterogeneous in its composition. For a heterogeneous
population, the maximum variability of 50% or .5; was used to determine the sample
size. The formula for deriving the sample size has been given in Eq. 3.1.

(3.1)

„p‟ is the estimated proportion of an attribute that is present in the population,

and „q‟ is 1-p. „D‟ refers to the desired level of precision which was also known as
margin of error. „Z‟ refers to the desired confidence level.
For the current study, „z‟ was taken as 95% whose table value was +- 1.96 as for
most educational and social sciences studies, this was the most common „alpha‟ level
used (Ary, Jacobs and Razavieh, 1996). Margin of error (D) was taken as 3.5% or .035
as normal range for margin of error taken for educational researches are 3 to 5%
(Krejcie and Morgan, 1970). Moreover, researchers may increase these values when a
higher margin of error was acceptable or may decrease these values when a higher
degree of precision was needed (Bartlett, Kotrlik and Higgins, 2001).
„p‟ values (estimated proportion) could be taken as .5 indicating the maximum
variability in the population if exact proportion in population is not known (Malhotra
and Dash, 2012; Chawla and Sondhi, 2011).

(3.2)

(3.3)

A sample size of 784 (Eq. 3.2 and 3.3) was obtained using the proportion approach
which was rounded off to 800. Keeping the general non-response of people towards
surveys, questionnaire was administered to 1200 people. 835 filled questionnaires were
obtained by the researcher with a response rate of 69.6% or 70%. After data cleaning, 801
responses were deemed appropriate for further inclusion in the study.

b) Sampling Technique
Area wise proportionate cluster sampling (Malhotra and Dash, 2012) was used
to collect the data from the respondents. In this technique, target population is divided
into mutually exclusive and exhaustive sub-groups called „clusters‟. These clusters are
normally the geographic areas. Thereafter each cluster is sampled in the ratio of its

69
 Chapter 3

contribution towards the total population (Malhotra and Dash, 2012). Census (2011)
was considered as a frame of reference which divided Delhi into 9 districts.
Deliberately two districts namely „New Delhi‟ and „Central Delhi‟ were not included in
the sample as both were composed of urban population only whereas all other districts
were heterogeneous in terms of composition as they had urban, semi-urban and rural
population within them. Researcher conveniently chose areas in each district. The list of
areas chosen by the researcher is given in Table 3.3.

Table 3.3: List of Sampled Areas in Each District

District Name
North South West South West North East East North
West Delhi Delhi Delhi Delhi Delhi Delhi Delhi
Ashok Vihar Aali Bapraula Chhawla Shahdara Gandhi Burari
(U) (SU) (SU) (SU) (U) Nagar (SU)
(U)
Bawana Meetha Pur Hastal Delhi Cant Seemapuri Preet Rihar Civil lines
(SU) (R) (R) (U) (U) (U) (U)
Begumpur Molar Band Nangloi Dwarka Jaffrabad Shakar Kotwali
(SU) (SU) (SU) (U) (SU) Pur (U)
(R)
Chandpur Hauz Patel Kakrola Badarpur Shamas Jagatpur
(R) Khas Nagar (R) (R) Pur (R)
(U) (U) (R)
Darya Pur Kalkaji Punabji Kapashera Mandoli Vivek Sadar
(SU) (U) Bagh (SU) (R) Vihar Bazar
(U) (U) (U)
Jahangir Puri Defense Rajouri Mundka Gharoli
(SU) Colony Garden (R) (SU)
(U) (U)
Jaunti Saidabad Tilangur Najafgarh
(R) (SU) Kotla (U)
Vihar (SU)
Kair Uttam Raota
(R) Nagar (U) (R)
Kirari Jaffarpur Samalka
(SU) (R) (SU)
Madanpur Prashant
(R) Vihar (U)
Model Town
(U)
Nangal
Thakran (R)
Narela (U)
Rani Khera
(SU)
Rawta (R)
Rohini (U)
Roshan
Pura (R)
Saoda (R)
Saraswati
Vihar (U)
R = Rural; SU = Sub-urban; U = Urban

70
 Chapter 3

Samples were conveniently picked from the chosen areas. The proportion of
respondents from each district in the sample has been given in table 3.4. A total of 801
responses were collected from the study‟s sampling area Delhi. The proportion of
respondents in the sample was kept same as it was in the population for each district.
The researcher stopped accepting and collecting the responses when assigned number
for a district was achieved in order to maintain the desired proportion in the sample.
The details of sample have been given in table 3.4.

Table 3.4: Area Wise Proportionate Cluster Sampling

Sr.No. District Population % of total population Proportion in sample

1 North West Delhi 3,656,539 22.76 183
2 South Delhi 2,731,929 17.01 136
3 West Delhi 2,543,243 15.83 126
4 South West Delhi 2,292,958 14.27 114
5 North East Delhi 2,241,624 13.95 112
6 East Delhi 1,709,346 10.65 85
7 North Delhi 887,978 5.53 45
Sample Size = 801

c) Sampling Unit
A sampling unit refers to the minimum unit of observation for obtaining
information about the variable/phenomenon of interest. Car is considered a family
product in India involving significant investment and interest of all the family members.
So, for the current study any adult member of the family which either owned a car or
planned to buy a car in near future from the selected geographical area was considered
eligible to be the sampling unit.

d) Sampling Area
Delhi was chosen as a sampling area.

3.8 DATA ANALYSIS

Data was analyzed with the help of SPSS version 21.0. Descriptive and
inferential statistics were used to analyze the data. Data was represented with the help
of tables, cross tables, pie-charts and bar-charts. „One Sample Binomial Test‟,
„Wilcoxon Signed Rank Test‟ and „Chi-square‟ were the main non-parametric tests
used in the study. The study also used advanced statistical techniques like „Discriminant
Analysis‟ and „Cluster Analysis‟ to arrive at the findings. Amos version 20.0 was also
applied to see the model fit of DTPB for the study.

71
 Chapter 3

3.9 LIMITATIONS OF THE METHODOLOGY

The present study made use of quantitative data whereby it sought the desired
information from 801 respondents belonging to Delhi area. The study made use of cross
sectional descriptive research design whereby respondents were contacted once. The
proposed study can produce different results with longitudinal research design and
increased sample size. It is also quite possible to have different results if the study is
conducted in non-metro cities. The questionnaire statements were explained to the rural
respondents in Hindi by the researcher. It is possible to obtain different results in similar
studies if language barriers are ignored for rural respondents. Moreover restrictions of
time, money and other resources can‟t be avoided in such descriptive studies, which
held true for the present study as well. The present study collected the data revealing
important aspects of consumer behaviour. However, it is really difficult to accurately
predict the consumer behavior based on this data as it keeps on changing according to
the time and moods of the consumers.

3.10 CONCLUDING REMARKS

This chapter dealt with the research methodology adopted for the study.
Problem statement, research questions and hypotheses of the study were discussed in
the chapter. The chapter described that descriptive research design was used for the
study. Both primary and secondary data were used for the study. Structured
questionnaire was used as a tool of collecting primary data. The chapter also detailed
the scale development process. Proportion approach was used to decide the sample size
for the study. A total of 801 responses were collected from the respondents from Delhi
area using area wise proportionate sampling. The chapter also described the data
analysis and limitations of the methodology. The next chapter of the study explains the
classification variables used in the study along with the demographic profile of the
respondents using digital channels of communication.

72