40

CHAPTER 3

RESEARCH METHODOLOGY

3.1 INTRODUCTION

The research methodology section presents a description of the

approach taken and the methodologies employed to accomplish the research
objectives described in the earlier chapter. The research process flow diagram
gives the framework of the entire study. A description of the data collection
methods shows how and from where the information was collected. The
population and the sampling procedure describe the criteria used in consumer
survey, following information regarding the method for data analysis and
ending with validity and reliability concerns regarding the study.

3.2 RESEARCH STRATEGY

The purpose of this research is to investigate cloud kitchen

attributes and the effects upon behavioural intention and actual use behaviour
towards cloud kitchen-based mobile food delivery apps. Therefore, a
descriptive research was chosen to add new knowledge on mobile food
delivery apps, with special reference to cloud kitchen-based mobile food
delivery apps. When examining the actual use behaviour towards cloud
kitchen-based mobile food delivery apps, it is significant to focus on user
perspective and intention to use these apps.
 41

3.3 RESEARCH PROCESS FLOW DIAGRAM

Introduction Objectives Review of Research Gap

Literature

Empirical
Study 1
support found
for actual use
Testing the
behaviour -Theoretical and model
towards cloud Sample size = empirical
kitchen-based 674 users of Background
mobile food cloud kitchen-
-Hypotheses
delivery apps based mobile
development
food delivery
apps -Research Design Study 2

Testing the
Empirical
model with
support found
moderation
for moderating
effect
effect of health
consciousness
towards cloud
kitchen-based
mobile food
delivery apps

Measuring actual use behaviour towards cloud kitchen-based mobile food delivery
apps in mobile food delivery apps

Discussion, Limitations & Scope for further research, Conclusions and References

Figure 3.1 Research Process Flow Diagram

3.4 RESEARCH PHILOSOPHY

Research philosophy is defined as the development of knowledge

and nature of knowledge (Saunders, Lewis & Thornhill, 2009). The research
strategy adopted by a researcher in any social science research is based on a
research paradigm. A research paradigm is a set of agreements and common
 42

beliefs that have been shared by scientists on how research problems must be
addressed and understood by them (Scotland, 2012).

A research paradigm according to Terre Blanche Durrheim &

Painter (2006) can be characterized through three factors namely ontology,
epistemology and methodology respectively. Hay (2007) defines ontology as
„what there is to know‟, epistemology as „what can we can know about it‟ and
methodology as „how to go about acquiring that knowledge‟.

Positivism is used to describe an approach to research based on the

assumption that knowledge may be gained by collecting data through
observation, measurement and analysis to establish truths (Somekh & Lewin,
2005). “Positivist studies are premised on the existence of a priori fixed
relationships within phenomena, which are typically investigated with
structured instrumentation. Such studies serve primarily to test theory, in an
attempt to increase predictive understanding of phenomena” (Baroudi &
Orlikowski, 1991). Also suggested that positivism can be used to test theory in
an attempt to increase the predictive understanding of phenomena, in this case
adoption of cloud kitchen-based mobile food delivery apps. The hypotheses
and a conceptual framework described in chapter two and the intention to
examine them in terms of cloud kitchen adoption it is believed make this
research a positivist one.

3.5 RESEARCH APPROACHES

Based on the descriptions of the aim and objectives, this research

develops knowledge based on behaviour studies and information systems
derived from existing studies and theories. This research intends to generalize
the cloud kitchen adoption phenomenon among mobile food delivery apps users.
A deductive approach is used in this work. A deductive approach involves
developing a theory and testing the developed theory (Saunders et al. 2009) in
 43

order to fulfil the aim of the research. According to Gratton & Jones, (2010),
the deductive approach “comprises the development of an idea, or hypothesis,
from prevailing theory which can then be tested through the gathering of
data”

A deductive approach is mostly used when laws can present,

enlighten and predict the phenomenon. A deductive approach is appropriate
for quantitative data with large sample sizes and can help in elucidation of the
relationships among variables. The deductive research approach is commonly
connected to positivist and quantitative research. Given the positivistic and
quantitative nature of this research, the deductive approach is best suited
for it.

3.6 RESEARCH STRATEGIES

A research strategy proposes a complete direction to the research,

including a process of how research should be conducted and assisted such
that researchers can logically perform the study (Remenyi, 1998). The
strategies can be deliberated as a general strategy for researchers to achieve
their research objectives and obtain answers to their research question
(Saunders et al. 2009). The elements that help the researchers to choose an
appropriate research strategy are research objectives, research questions,
prevailing information and other properties as well as the selected research
philosophy.

A survey is a research strategy while the questionnaire is a research

tool that uses questions to gather data. Surveys help in finding answers to
„who, what, where, how much and how many‟ questions. A survey is
frequently used because it can economically gather huge amounts of data
(Saunders et al. 2009).
 44

A survey comprises collection of data using questionnaires (Chen

& Hirschheim, 2004). It is also a common strategy for business and
management studies and is related to the deductive approach. The data collected
by a survey strategy can be analysed using both descriptive and inferential
statistics. This means that a survey is used to provide reasons for particular
relationships between variables to create models that illustrate a relationship and
to allow more control over the research process (Saunders et al. 2009).

A questionnaire can be defined as a set of carefully designed

questions related to the research topic of interest and administered in exactly
the same form to a group of people when collecting data (Jupp, 2006). A
questionnaire usually provides an inexpensive and effective way to obtain
data in a structured and manageable way.

The questionnaire was the selected research instrument for this

research. For this research, a self-administered questionnaire is used.

3.7 RESEARCH CHOICES

3.7.1 Quantitative and Qualitative Data

The terms quantitative and qualitative data are used to understand

the research data characteristics. Quantitative data is “Mainly used as a
synonym for any data collection technique such as, a questionnaire or data
analysis procedures such as, graphs or statistics that generate or use numerical
data” (Saunders et al. 2009:151). Consequently, quantitative data is
information about the data in the form of numbers (Punch, 2013). Thus,
researchers convert data into numbers that can be measured and analysed
(Punch, 2009).

In contrast, Qualitative data is “used predominantly as a synonym

for any data collection technique such as, an interview or data analysis
 45

procedure such as categorising data that generates or use non-numerical data”

(Saunders et al. 2009:151).

In researcher‟s study of consumer behaviour towards cloud kitchen-

based mobile food delivery apps, data was collected using surveys and were
then converted into numbers; therefore, quantitative data was considered to be
appropriate for this research.

This research applied a quantitative approach because data was

obtained in numerical format. Secondly, the selected strategy and research
philosophy conformed to the research aims. The aim to understand the actual
use behaviour towards cloud kitchen-based mobile food delivery apps
requires quantitative data, just like earlier research in this field
(Alalwan, 2020; Kapoor & Vij, 2018; Roh & Park, 2019).

3.7.2 Primary and Secondary Data

Primary data is data collected precisely for an undertaken research

study (Saunders et al. 2009). Primary data can be collected using methods
such as surveys or observations (Zikmund et al. 2009).

Secondary data are data that have previously been collected by

other researchers and for other cases. Secondary data is raw data and in the form
of, for example, published précises, research articles, white papers and census
taken, annual reports etc (Saunders et al. 2009). Secondary data can be acquired
faster and is less expensive than primary data (Cooper & Schindler, 2013).

Due to the aim and research questions of this study, this research
needed to acquire primary data. However, secondary data such as literature
reviews were also required to acquire an initial understanding and form the
conceptual framework. In this research, secondary data such as journal
 46

articles, conference publications, white papers and newspaper information‟s

were used for problem definition, literature review, conceptual development,
method development and discussion phases. Statistical documents and
secondary quantitative data sets also helped in developing the problem
definition and evaluation phases.

3.8 RESEARCH INSTRUMENT AND DEVELOPMENT

The survey instrument measures were evaluated based on the five-

point Likert scale ranging from „strongly disagree‟ (1) to „strongly agree‟ (5).
All the items were developed following leads from literature, with minor
modifications to capture the content-specific aspects of cloud kitchen-based
mobile food delivery apps. In the following section, the construct measures
are described along with operational definitions.

Research gaps in the chosen area of study were identified on the

basis of review of the relevant literature, described in the previous chapter.
The outcome variable behavioural intentions influenced by performance
expectancy, effort expectancy, social influence, facilitating conditions, online
review, online coupons, various food choices and delivery experience and
these were used as a direct antecedents of cloud kitchen-based mobile food
delivery apps. The framework is a combination of the drivers that influence
actual use behaviour towards cloud kitchen-based mobile food delivery apps.
Additionally, the moderating variable – health consciousness was used to and
their effects on actual use behaviour towards cloud kitchen-based mobile food
delivery apps were analysed.

Performance Expectancy

The present study suggested the construct – performance

expectancy and refers to Venkatesh et al. (2003) who defined performance
 47

expectancy as “the degree to which an individual believes that using the

system will help him or her to attain gains in job performance”. Thus,
performance expectancy was measured using the concept developed by
Venkatesh et al. (2003) and four items were used to measure the construct. A
sample item is “I find cloud kitchen-based mobile food delivery apps
useful in my daily life”. The reliability value for this scale in earlier work
was 0.910 (Alalwan 2020).

Effort Expectancy

Effort expectancy is defined as “the degree to which a person

believes that using a system would be free of effort”. The cloud kitchen-based
mobile food delivery apps are used based on the mobile food delivery
applications. The usage of such applications will depend on ease of use. Thus,
four items were used to measure the effort expectancy of mobile food delivery
apps, as suggested by Venkatesh et al. (2003). The sample item is “Learning
how to use cloud kitchen-based mobile food delivery apps is easy for me”.

Social Influence

Social influence is defined as “The degree to which an individual

perceives that important others believe he or she should use the new system”.
To measure this construct, four items were used as suggested by Venkatesh
et al. (2003). A sample item is “People who are important to me think that I
should use cloud kitchen-based mobile food delivery apps”.

Facilitating Conditions

The construct – facilitating conditions – is defined as “The degree

to which an individual believes that an organizational and technical
infrastructure exists to support use of the system”. To measure this construct,
 48

four items were used as suggested by Venkatesh et al. (2003). A sample item
is “I have the necessary resources to use cloud kitchen-based mobile food
delivery apps”.

Online Ratings

The construct, online ratings, is defined as another type of crowd

opinion and indicates reviewers‟ average evaluation of the different features
of product or service (Fileri, 2015). To measure this construct, three items
were adopted from Alalwan, (2020). A sample item is “Online ratings have
helped me to learn about the cloud kitchen-based mobile food delivery apps”.

Online coupons

Online coupons offer a product at a lower price and consumers

redeem coupons because of increase in value of the product owing to reduced
price being offered in electronic coupon codes. To measure this construct,
three items were used as suggested by Lichtenstein et al. (1990). A sample
item is “I enjoy offers in the form of coupons, cashback and discount on the
cloud kitchen-based mobile food delivery apps”.

Various Food Choices

Individual food choices are primarily influenced by such

considerations as taste, cost and convenience. A quality attribute of food
delivery apps is food choice. To measure this construct, three items as
suggested by Cho, Bonn, & Li, (2019) were used. A sample item is “I can
order food with a wide range of prices through cloud kitchen-based mobile
food delivery apps”.
 49

Delivery Experience

Delivery experience refers to the positive experience related to the

delivery of food when ordered through cloud kitchen-based mobile food
delivery apps. To measure this construct, three-scale items were used as
described by Ray, Dhir, Bala, & Kaur, (2019). A sample item is “I like the
cloud kitchen-based mobile food delivery apps provisions to order food at late
night”.

Behavioural Intention

Behavioural intentions defined as consumer readiness and

likelihood to use cloud kitchen-based mobile food delivery apps (Pena &
Brody, 2014). In this study, behavioural intention was measured using the
scale items suggested by Venkatesh et al. (2012). A sample item is “I intend
to use cloud kitchen-based mobile food delivery apps in the future”.

Use behaviour

The actual use behaviour is defined as physical and mental acts

involved in incorporating the information discovered into the person‟s
existing information base (venkatesh, 2003). The scale adopted from
(Tan, 2013, Venkatesh 2003), measured use behavior in terms of adoption and
attitude towards electronic applications. Therefore, the scale adopted and
modified according to the current study. A sample item is “I currently use
cloud kitchen-based mobile food delivery apps”.

Health Consciousness

Health consciousness refers to the boundary condition that

influences people‟s food choices. Three-scale items were used to measure the
construct adopted from Lee, Conklin, Cranage, & Lee (2014). A sample item
 50

is “I think of myself as the sort of person who is concerned about the long-
term health effects of my food choices”.

3.9 SAMPLING PLAN

Sampling is the process of choosing elements from a population to

represent the overall population (Cooper & Schindler, 2013). The reasons to
use sampling are lower costs compared to a census and rapid data collection
speed. The sampling process followed in this research included defining a
relevant population, choosing the sample type, picking a sampling technique,
classifying and evaluating sampling frames and drawing samples (Cooper &
Schindler, 2013). The following section discusses the sample used.

The sampling frame is a list of elements in the population from which

the sample is drawn (Cooper & Schindler, 2013). The list includes particulars
such as the names of employees in a company (Saunders et al. 2009). However,
for this research, it was impractical to access the overall list of cloud kitchen
users in Chennai and Bangalore and risked violation of privacy of the users of
food aggregator apps. Therefore, this research used the non-probability
sampling. Non-probability sampling is a sampling technique in which, the
odds of any member being selected for a sample cannot be calculated. The
major advantage of non-probability sampling method is cost- and time-
effectiveness. There are many types of non-probability sampling methods.
Descriptive research design was followed in this study. It was cross sectional
in nature and a survey-based approach was used to collect data from the
selected population. The snowball sampling technique was applied to collect
data from the sample respondents. This study adopted the snow ball sampling
to increase the number of respondents as mentioned in other studies
(Gilboa et al. 2019, Ahmad and Thyagaraj, 2017).
 51

3.9.1 Pilot Study

This research also used a pilot test. A pilot test is a trial collection of
data that is used to distinguish flaws in the design, instrumentation and
endowment of proxy data for the selection of a sample (Cooper & Schindler,
2013). Pilot studies can also help with assessing question validity and
reliability (Saunders et al. 2009). A Pilot not only enables a researcher to
review a questionnaire used for distribution, but it also allows the researcher
to test the analysis methods and the framework (Bell, 2005).

Pre-test was performed to select the stimuli for this study. Initially,
users of mobile food delivery applications were selected and scrutinized with
the initial questions on whether they were aware of cloud kitchen-based
mobile food delivery apps and used it. The study focused on users of cloud
kitchen-based mobile food delivery apps. The pilot study was carried out with
50 cloud kitchen users in Chennai and Bangalore through the distributed
questionnaire method. The results of the pre-test were evaluated using
Cronbach‟s alpha for the items of each construct and the results showed
acceptable level of internal consistency.

3.10 DATA COLLECTION

A broad survey instrument was developed for this study based on

the review of relevant literature support for testing the research model. The
selection of India as the context of conducting empirical part of the current study
was because online food delivery industry is growing rapidly in India compared
to its neighbouring countries. According to the study, Indian online food ordering
market is slated to grow at a CAGR (Compound Annual Growth Rate) of 16.2
per cent to USD 17.2 billion by 2023 (Business Today, 2019). The participants
were based in Chennai and Bangalore city in the southern part of India. This
 52

research chose respondents who using cloud kitchen-based mobile food

delivery apps from food aggregator platforms such as Swiggy, Zomato and
also direct cloud kitchen applications such as Box8, Freshmenu, etc. The
cloud kitchen-based mobile food delivery services that used their own apps
were enlisted in food aggregator platforms. The appendix 1 list the survey
instrument used for this research.

Since snowball sampling method was followed in this study, the

potential respondents of this study were users of cloud kitchen. Initially, a few
diverse individuals were asked to generate a list of initial customers of cloud
kitchen who influenced these individuals and from these samples, a second
wave of respondents were interviewed known as a snowball sample
(Hancock & Gile, 2011). The questionnaire was distributed and circulated
also through Google forms online, of the 800 responses received, 674 were
found valid with a response rate of 84.25% from April to June 2019. The
invitation accompanying a personal message was sent to potential participants
requesting them to participate in the survey. Further, personal
communications were directed to friends who helped in recruiting volunteers
from the city of Chennai and Bangalore.

3.11 DATA ANALYSIS

Once the data was collected, analysis was necessary, which was
dependent on the research questions and objectives. For this, the data used to
test the conceptual model were non-parametric. The normality test is the test
for normally distributed (or bell-shaped) data. “In statistical analysis,
parametric tests can be done only the data is in normally distributed”
(Jupp, 2006) However, since this research applied non-parametric tests, the
normality test was not required.
 53

The analysis was done through SPSS 20 (Statistical Product and

Service Solutions) and Analysis of a Moment Structure (AMOS) an added
SPSS module. Based on the type of data and measurement levels, appropriate
statistical techniques were used to accomplish the objectives of the present
study. The percentage analysis was performed to understand the profile of
cloud kitchen users in mobile food delivery apps. The mean and standard
deviation were calculated for the factors influencing behavioural intention and
actual use behaviour towards cloud kitchen-based mobile food delivery apps.
A confirmatory factor analysis was also done to ensure the number of factors
required in the study and to confirm the measured variables related to the
latent variables. The correlation analysis was done to study the relationship
between the factors influencing behavioural intention and actual use
behaviour towards cloud kitchen-based mobile food delivery apps.

The first objective was to investigate the role of demographic

profile on behavioural intention of Indian consumers who use the cloud
kitchen-based mobile food delivery apps. The independent t- test and
Analysis of Variance (ANOVA) test were employed to examine the
difference between profiles of cloud kitchen users. This was done to
understand the difference between the demographic details of the respondents
and the factors influencing Behavioural Intention, differences between the
sample respondent demographic and the behavioural intention and the
differences between the respondent demographic and actual use behaviour
towards cloud kitchen-based mobile food delivery apps.

The second objective was to identify and empirically examine the

factors predicting behavioural intention towards cloud kitchen-based mobile
food delivery apps in India. Multiple regression analysis was carried out to
examine the factors influencing behavioural intention and the actual use
 54

behaviour towards cloud kitchen-based mobile food delivery apps. Multiple

Regression is a statistical tool used to improve a self-weighting estimating
equation, which predicts values for a dependent variable from the values of
independent variables, controls variables to better evaluate the contribution of
other variables, tests and explains a causal theory (Cooper & Schindler, 2013).

The third objective was to explore the primary association between

behavioural intention and Actual use behaviour in cloud kitchen-based mobile
food delivery apps by Indian consumers. Regression analysis was used for
this purpose. The analysis was carried out to find out the relationship between
behavioural intention and actual use behaviour. Regression analysis is the
process of calculating a regression coefficient and regression equation using
one independent variable and one dependent variable (Saunders et al. 2009).
The analysis used regression equations to predict the values of dependent
variables. The regression coefficient is the result of the analysis that shows
the strength of the relationship between a dependent variable and independent
variable.

Objectives four focused on the role of health consciousness on the

Indian consumers‟ decision of reuse of cloud kitchen-based mobile food
delivery apps. Moderation analysis was done for this. In this study, moderation
analysis was tested between exogenous and endogenous constructs through
moderator - health consciousness. This helped in understanding the effect of
moderator on the relationship between the endogenous and exogenous
constructs (behavioural intention and actual use behaviour).

To prove the overall model fit, the structural equation model (SEM)
was applied, to examine the structural relationship between variables and
overall model fit. This technique allows researchers to explore a set of
interrelated research questions in a distinct, logical and wide-ranging analysis
 55

by modelling the relationships among multiple independent and dependent

constructs simultaneously (Gefen et al. 2000). Structural Equation Modelling
(SEM) has been used in literature since the 1980s (Hair et al. 2011)

3.11.1 Techniques for Validating the Research Findings

After questions were selected and developed from previous

research studies, the questionnaire was validated. Validity is a representation
of measurement and encompasses testing the extent to which a researcher
wants to measure and the variances found with a measurement tool. These
redirect the true alterations amongst participants drawn from a population
(Cooper & Schindler, 2013). Instrument validation is a dynamic step for
researchers to ensure the generation of scientifically valid knowledge. For this
purpose, the questionnaire was printed and instructed interviews were
conducted to seek responses and comments from an expert panel. Thereafter,
content validity was performed, where the recommendations and suggestions
from the expert panel were incorporated. Before distributing the content
validity questionnaire to respondents, approval was obtained from the
supervisor. Following validation, some improvements were made to the
questionnaire.

It is important to validate the findings done by a researcher after

conducting the statistical analysis to ensure that the results obtained from the
statistical analysis were reliable and valid. The legitimacy of the collected
instrument and information is measured by validity. Both internal and
external validities have been ensured by the researcher. The reliability for
various components of questionnaire was tested by computing Cronbach‟s
alpha and the results are presented in Table 3.1.
 56

Table 3.1 Cronbach’s alpha Measures

Constructs No. of Items Cronbach’s Alpha

Performance Expectancy 4 0.657
Effort Expectancy 4 0.801
Social Influence 3 0.812
Facilitating Conditions 4 0.824
Online coupons 3 0.818
Various Food Choices 3 0.750
Online Ratings 3 0.800
Delivery Experience 3 0.842
Behavioural Intention 3 0.796
Use Behaviour 3 0.760
Health Consciousness 3 0.822

3.12 CHAPTER SUMMARY

In this chapter, the conceptual model used in this research work

was described. This model was designed to understand intention to use and
actual use behaviour towards cloud kitchen-based mobile food delivery apps.
The research method and methodology were also described. From this
chapter, it can be seen that the researcher applied a deductive and quantitative
approach along with a Survey Strategy using distributed questionnaire. In
terms of data, both primary and secondary data were used. The chapter also
provided details about the sample size, sampling type and sampling technique.
For analysis SPSS and Structured Equation Modelling were used and the
reasoning for their use were provided.