5 - YA RA - Chapter 4 - Research Methodology PDF
5 - YA RA - Chapter 4 - Research Methodology PDF
5 - YA RA - Chapter 4 - Research Methodology PDF
Research Methodology
4.1 Introduction:This chapter discusses the methodology used to conduct the research study as well as
the issues related to the chosen research methodology when investigating the different
factors affecting the adoption of E-Marketing by industrial and trading UK SBEs, the
different forms, tools and levels of implementation of E-Marketing by these enterprises as
well as the relationship between E-Marketing adoption and marketing performance of these
SBEs. Through this discussion the steps that were taken to address the research design, the
data collection and analysis methods used to conduct the research study will be illustrated.
These issues are addressed in light of the basic research objectives and the relevant
research questions discussed in chapter one and the research framework discussed in
chapter seven.
Firstly it discusses some possible research designs and methodologies and provides a
brief review of the literature of research methodologies, research strategies and research
methods implemented in the fields of E-Marketing and SBEs. Based on this analysis, the
research methodology for this study is justified in terms of its appropriateness and
usefulness to achieve the basic research objectives. The research population, research
sample, data collection methods, stages of collecting the data from the field, the procedures
and problems encountered during each stage of the fieldwork as well as the actual data
collected and methods of analysis are also explained in this chapter.
4.2 Research Methodologies and Strategies:According to Sarantakos (1998) research methodology is the theory of methods
(Sarantakos, 1998; P: 465); it is the way in which one makes sense of the object of enquiry.
Robson (2002) defines it as: The theoretical, political and philosophical backgrounds to
social research and their implications for research practice and for the use of particular
research methods (Robson, 2002; P: 549). According to Morvaridi (2005), the most
prevalent methodologies in social sciences and humanities research methodology are
quantitative and qualitative research (Morvaridi 2005; P: 2). Moreover, Sarantakos (2005)
argues that social science can be conducted within a quantitative or qualitative context.
Chapter 4
Research Methodology
4.2.1 Quantitative research:Quantitative approach is one of the most used approaches to conduct social
research. It is rooted in a strong academic belief that puts significant and substantial trust in
numbers that represent different opinions and/or concepts. As a result, quantitative research
can be conducted through emphasising quantification in the collection and analysis of data.
It relies mainly on a hypothesis which is derived from theory deductively; the objective is
to test the theory by way of observation and data collection, the findings of which,
following analysis, would either confirm or reject the theory (Morvaridi 2005; P: 2).
According to Bryman (1989), quantitative approach is clamed to be infused with positivism
which is an approach to the study of people which commends the application of the
scientific method.
4.2.2 Qualitative research:In contrast, the qualitative approach concentrates mainly on words and observations
to express reality and tries to describe people and research phenomena in natural situations
(Amaratunga, 2002; P: 19). Although quantitative and qualitative methodological
approaches are different, it is wrong to oppose them as two competing methodologies
(Morvaridi 2005; P: 3). This research argues that qualitative research facilitates and
illustrates quantitative research and quantitative research does the same thing with both
approaches helping to add to the available accumulative knowledge. This is in line with the
arguments of Bryman (1989) and Clark (1998). Within this context, Bryman (1989) argue
that quantitative and qualitative researches are different ways of knowing (Bryman,
1989; P: 27). Moreover, Clark (1998) argues that philosophically the qualitative and
quantitative paradigms are not as diverse or mutually incompatible as is often conveyed
(Clark, 1998; P: 1243).
On the other hand, Lincoln and Denzin (2002) summarised different aspects related
to qualitative research as follows: Qualitative research is many things at the same time. It
is multi-paradigmatic in focus. Its practitioners are sensitive to the value of the multimethod approach. They are committed to the naturalistic perspective and to the
interpretive understanding of human experience. At the same time, the field is inherently
political
and
shaped
by
multiple
4-2
ethical
and
political
allegiances
Chapter 4
Research Methodology
4.2.3 Research Strategy:Saunders and Thornhill (2003) define research strategy as: A general plan of how
the researcher will go about answering the research questions. (Saunders and Thornhill,
2003; P: 90). They distinguish between eight research strategies, namely: experiments,
surveys, case studies, grounded theory, ethnography, action research, cross-sectional
studies and exploratory studies. However, the three main strategies used by most
researchers are: experiments, surveys and case studies because of great benefits associated
with using them (Robson, 2002).
Bryman (1989) defines survey strategy as: The collection of data on a number of
units with a view to collecting systematically a body of quantifiable data in respect of a
number of variables which are then examined to discern patterns of association (Bryman,
1989; P: 104). According to Saunders, this approach is a popular and common strategy in
management and business research (Saunders and Thornhill, 2003; P: 92).
On the other hand, while case study can be defined as: A strategy for doing
research which involves an empirical investigation of a particular contemporary
phenomenon within its real life context using multiple sources of evidence (Robson, 2002;
P: 178), experiment strategy can be defined as: The measurement of the effects of
manipulating one variable on another variable and it is generally related to the natural
sciences (Robson, 2002; P: 88).
4.2.4 Triangulation Strategy:Triangulation 1 is a valuable and widely used strategy (Robson, 2002; P: 174)
because it examines the research problem from more than one viewpoint so that the study
become more robust. It can be defined from different angles; while Denzin (1978) defines
it as: The combination of methodologies in the study of the same phenomenon (Denzin,
1978; P: 301), Saunders and Thornhill (2003) define it as: The use of different data
collection methods within one study (Saunders and Thornhill. 2003; P: 99).
Moreover, there are four basic types of triangulation identified by Denzin (1978)
which are: data triangulation (refers to the use of a mixture of data sources in a certain
study), investigator triangulation (refers to the use of different researchers), theory
1
The term Triangulation is taken from land surveying, knowing a single landmark only locates the person
somewhere along a line in direction from the landmark, whereas with two landmarks a person can take
bearings in two directions and locate himself at their intersection (Fielding and Fielding, 1986; P: 23)
4-3
Chapter 4
Research Methodology
triangulation (refers to the use of multiple theories) and methodological triangulation which
refer to the use multiple methodologies to study a single problem (Denzin, 1978; P: 301).
Clearly there are a number of different approaches and it is important to select an
appropriate means to investigate the research question.
4.3.1 Primary Research Assumptions:Hiles (1999) argues that it is important to note that what defines human science is
not its methodology but its paradigms. Guba and Lincoln (1994) defines a paradigm as: "A
set of basic beliefs that deals with ultimates or first principles which represent a worldview
that defines, for its holder, the nature of the world , the individuals place in it and the
range of possible relationships to that world and its parts Guba and Lincoln (1994, p:
107).
Moreover, Hiles (1999) argue that all scientific research follows a set of procedures
that must begin with a group of assumptions, a set of beliefs, and a paradigm. Nevertheless,
choosing a certain paradigm for a study does not establish directly what strategies and
types of data collection and analysis methods can be employed (Hiles, 1999, p: 1). On the
other hand, according to Guba and Lincoln (1994), a paradigm is a basic belief systems
based on ontological, epistemological and methodological assumptions (Guba and Lincoln,
1994, p: 107). The fundamental beliefs that define a paradigm can be determined by the
responses given by proponents of a certain paradigm to three fundamental questions which
are: the ontological question, the epistemological question and the methodological
question, which are defined by Guba and Lincoln (1994; P: 108) as follows:-
The ontology question: What is the form and the nature of reality? And, therefore,
what is there that can be known about it?
4-4
Chapter 4
-
Research Methodology
The epistemology question: What is the nature of the relationship between the
knower, or the would-be knower, and what can be known? It is concerned with how
we know the world and what is the relationship between the inquirer and the
known.
The methodology question: How can the inquirer (the would-be knower) go about
finding out whatever he/she believes can be known? It focuses on how we obtain
knowledge about the world and indicates which research techniques are considered
appropriate for collecting valid empirical evidence. (Guba and Lincoln, 1994; P:
108)
According to Guba and Lincoln (1994), there are four major paradigms that
structure and organise social science research namely: positivism, post-positivism, critical
theory and constructivism or interpretivism. These paradigms, their associated strategies,
approaches and methods as well as the differences among them are illustrated in table 4-1.
Positivist and constructivism (interpretive) paradigms represent two contrasting
constellations of beliefs about how valid and applicable knowledge may be generated
(Denzin, 2000). Traditionally, there has been a deep emphasis on quantification in most
aspects of science (Guba and Lincoln, 1994). This led to a large scale adoption of the
positivist approach by most researchers in social science. The main reasons for this include:
the common belief that only quantitative data are highly valid and provide high level of
research quality (Sechrest, 1992), the positivist approach enables a researcher to replicate
the findings of one study in a different study or in a different context (Winfield, 1990) and
the availability, reliability and ease of use of a solid collection of statistical and
mathematical tools and models that can be used within the approach (Guba & Lincoln,
1994). However, there are some criticisms that face the positivist approach. Within this
respect, one of the most common criticisms about the positivist approach arise from the
argument that positivism is a poor and misleading approach to conducting social science
research since it assumes an objective external reality upon which inquiry can converge
(Hirschheim, 1992). Moreover, Guba and Lincoln (1994) illustrate a variety of problems
associated with the positivist approach which include context stripping, exclusion of
meaning and purpose, disjunction of grand theories with local contexts, inapplicability of
general data to individual cases and exclusion of the discovery (Guba and Lincoln, 1994, p:
106).
4-5
Chapter 4
Research Methodology
Table 4.1: Basic assumptions and practical issues of different research paradigms
Item
Positivism
Post-Positivism
Ontology
Naive Realism
Critical realism
Real
reality
(apprehendable reality)
is assumed to exist.
Knowledge
is
summarised in the form
of time and context
free
generalisations
which take the form of
cause-effect laws.
Epistemology Dualism/
Objectivism:
- The investigator and
the investigated object
are assumed to be
independent
entities
and can not affect each
other.
- When influence in
either direction (threats
to
validity)
is
recognised
or
suspected
various
strategies are followed
to reduce or eliminate
it.
- Findings are true
Methodology Experimental and
Critical theory
Historical realism:
Constructivism / interpretive
Relativist
Reality is assumed to be
apprehendable that was
once plastic and was
shaped by a congeries of
social, political, cultural,
economic, ethnic, and
gender factors.
Modified Dualism/
Objectivism:
- It is possible to approximate reality
but it is never fully known.
- Special emphasis is placed on
external guardians of objectivity
such as critical traditions and the
critical community.
- Findings are probably true
Transactional/
Subjectivist:
- The investigator and the
investigated object are
assumed
to
be
interactively linked, with
the
values
of
the
investigator
inevitably
influencing the inquiry.
- Therefore findings are
value mediated
Transactional/
Subjectivist:
- The investigator and the
object of investigation are
assumed to be interactively
linked so that the findings are
literally
created
as
the
investigation proceeds.
- Therefore findings are created
by the investigator.
Modified experimental/
Hermeneutical/
4-6
Chapter 4
Research Methodology
manipulative
Questions
and/or
hypotheses are stated in
propositional form and
subjected to empirical
test to verify them;
possible confounding
conditions must be
carefully controlled to
prevent outcomes from
being
improperly
influenced.
manipulative
- Emphasis is placed on "critical
multiplism" (a refurbished version of
triangulation) as a way of falsifying
hypotheses,
- The methodology aims to redress
some of the research by doing
inquiry in more natural settings,
collecting
more
situational
information,
and
reintroducing
discovery as an element in inquiry
- All these aims are accomplished
largely through the increased
utilisation of qualitative techniques.
Explanation, prediction and control
dialectical
- The variable and personal
nature of social constructions
suggests
that
individual
constructions can be elicited
and refined only through
interaction between and among
investigator and respondents.
- The final aim is to distil a
consensus construction that is
more
informed
and
sophisticated than any of the
predecessor constructions.
Critique
and
trans Understanding; reconstruction
formation; restitution
and emancipation
Verified
hypotheses Nonfalsified hypotheses that are Structural/historical
Individual reconstructions
Nature of
established as facts or probable facts or laws.
insights.
coalescing around
knowledge
laws.
Consensus.
Knowledge Accretion - "building clocks" adding to "edifice of Historical revisionism; More informed and
generalisation
by Sophisticated reconstructions;
accumulation knowledge"; generalisations and cause-effect linkages.
similarity.
vicarious experience.
situatedness; trustworthiness and authenticity
Goodness or conventional benchmarks of "rigor": internal and external historical
validity, reliability and objectivity
erosion of ignorance
quality
criteria
Excludedinfluence denied
Includedformative
Values
Source: Adapted from (Guba & Lincoln, 1994)
Inquiry aim
4-7
Chapter 4
Research Methodology
4.3.2 Selecting the Post-Positivism Research Approach:From this research point of view, there is no optimal approach that can be used
for any type of investigation regardless of the nature of enquiry. Moreover, it is
4-8
Chapter 4
Research Methodology
believed that the research approach depends on the problem that the research is trying
to solve and the questions that the research is trying to answer. Within this context,
this research aims to investigate the different factors that might have an impact on the
adoption of E-Marketing by SBEs as well as the effect of this adoption on the
marketing performance of these enterprises. In other words, the main aim of the
research is to test and investigate the causal relationship between E-Marketing
adoption and the SBEs marketing performance.
4-9
Chapter 4
Research Methodology
4-10
Chapter 4
Research Methodology
4.4.1 Research question:The study attempts to answer the following questions:1- What are the different factors affecting the adoption of E-Marketing by
industrial and trading SBEs?
2- What are the different forms, tools and levels of implementation of EMarketing by industrial and trading SBEs?
3- What is the relationship between E-Marketing adoption and the marketing
performance of industrial and trading small business enterprises?
However, the main objective of the research is to analyse the different factors
affecting the adoption of E-Marketing used by UK SBEs and the contribution of this
adoption to the SBEs marketing performance to develop a theoretical model to
understand and interpret the use of E-Marketing by SBEs in the UK. Accordingly, the
most appropriate research methodology to this research study will be that
methodology that provides the researcher with the right tools to answer the above
research questions and achieve the research aims.
4-11
Chapter 4
Research Methodology
in the last two decades; it appears that only from the late 1980s research on EMarketing began to appear in the literature. These research studies investigated and
covered a broad rang of E-Marketing areas, such as Internet Marketing, Mobile
Marketing, E-Mail Marketing, Intranet Marketing and Extranet Marketing.
To examine the different methodologies and methods used in contemporary EMarketing research, the literature related to the fields of E-Marketing and SBEs has
been investigated from a methodological context in the period from 1993 to 2009. The
literature time period started from 1993 because of two reasons; firstly, The World
Wide Web (WWW), which is the main E-Marketing element, started in 1993 and
secondly, studies related to E-Marketing began to appear in the literature from the late
1980s to mid 1990s.
The following online databases were searched to provide a comprehensive
bibliography of the E-Marketing methodological literature: Emerald (MCB) Database,
EBSCO Electronic Database, Proquest Direct Database, Science Direct (Elsevier)
Database, ABI/Inform database and University of Bradford catalogue. The review of
the literature yielded 387 studies. Although it was planned initially (because of the
limited resources regarding time and effort) to construct a representative unprobability sample from the literature in order to come up with an accurate and fair
representation of the methodological literature characteristics, and to investigate the
methodologies and methods used in contemporary E-Marketing research, it was
decided to take all the studies resulting from the literature into consideration because
of the relatively limited number of studies yielded.
By reviewing these literatures, it was found that the majority of studies within
the literature were conceptual with a percentage of 41 % of the total number of
studies. It was also found that the majority of empirical studies depended on
qualitative methodology with a percentage of 28 % of the total number of studies
followed by a 24 % of the total number of studies depending on quantitative
methodologies. On the other hand, triangulation methodology was implemented in 6
% of the total number of studies. These findings are illustrated in the following figure.
4-12
Chapter 4
Research Methodology
Chapter 4
Research Methodology
Chapter 4
Research Methodology
from 78% of the total number of studies from 1993 1997 to 46 % of the total
number of studies from 1998 2002 and then declined more to reach 24% of the total
number of studies from 2003 2009. This reflects not only growing attention from
the researchers in the field of E-Marketing towards conducting empirical studies as a
result of achieving mature conceptualisation of the major concepts within the field,
but its also suggests that the field is in its infancy in terms of research development. It
is also noticed that the majority of the empirical studies conducted from 1993 2009
depended on qualitative methodology with a percentage of 28 % of the total number
of studies and the number of studies using this approach increased across the literature
period. Within this respect, qualitative studies increased from 11 % of the total
number of studies from 1993 1997 to 28 % of the total number of studies from 1998
2002 and then increased more to reach 34 % of the total number of studies from
2003 2009. Moreover, only 24 % of the total number of studies conducted from
1993 2009 depended on quantitative methodology and the number of studies
depending on this approach increased across the literature period. Within this respect,
quantitative studies increased from 11 % of the total number of studies from 1993
1997 to 19 % of the total number of studies from 1998 2002 and then increased
more to reach 32 % of the total number of studies from 2003 2009.
Based on the previous findings, although there is a great increase in the
number of empirical studies, most of the studies conducted from 1993 2009 were
either conceptual or qualitative with a percentage of 69 % of the total number of
studies, which reflects the need to conduct more quantitative studies in the field of EMarketing.
On the other hand, it is also noticed that the number of studies conducted
based on triangulation methodology increased across the literature period. Within this
respect, these studies increased from zero studies from 1993 1997 to 8 studies with a
percentage of 6 % of the total number of studies from 1998 2002 and continued
increasing to reach 17 studies with a percentage of 9% of the total number of studies
from 2003 2009. This not only reflect the growing importance of triangulation as a
research methodology in the field of E-Marketing but also reflects academic
researchers attitudes towards triangulation methodology as a suitable methodology to
be employed in investigating different issues within the field. This research argues
4-15
Chapter 4
Research Methodology
that the increase in triangulation usage is due to the benefits associated with its usage
and that the emergence of triangulation studies implies that there is more need to
validate earlier studies towards theory building within the field.
With regards to research strategies employed across the literature period, as
can be seen from table 4-4, it is noticed that not only are survey and case study
strategies the most commonly used research strategies from 1993 2009 in the field
of E-Marketing but also the number of studies conducted depending on these two
strategies increased dramatically from 1993 2009. Within this respect, studies
conducted depending on survey strategy increased from 7 studies with a percentage of
11% of the total number of studies from 1993 1997 to 34 studies with a percentage
of 25 % of the total number of studies from 1998 2002 and continued increasing to
reach 68 studies with a percentage of 36 % of the total number of studies from 2003
2009. On the other hand, studies conducted depending on case study strategies
increased from 5 studies with a percentage of 8 % of the total number of studies from
1993 1997 to 24 studies with a percentage of 18 % of the total number of studies
from 1998 2002 and continued increasing to reach 48 studies with a percentage of
26 % of the total number of studies from 2003 2009. This reflects academic
scholars attitudes towards survey and case study strategies as suitable research
methods to be employed in investigating different issues within the field of EMarketing due to the benefits associated with the usage of both of them.
Table 4-4: Research strategies employed by researchers in the
and SBEs from 1993 to 2009
N Research
Number of studies
% of the total
strategies 1993 1998 2003 1993 1998 2003
1997 2002 2009 1997 2002 2009
1
Survey
7
34
68
11 % 25 % 36 %
2 Case study
5
24
48
8 % 18 % 26 %
3 Exploratory
1
11
20
2 % 8 % 11 %
4 Experiment
1
3
4
2% 2% 2%
5 Conceptual
51
63
45
77 % 46 % 24 %
6
Literature
0
1
1
0% 1% 1%
review
Total
65
136
186
100
100
100
4-16
fields of E-Marketing
% of empirical
1993 1998 2003
1997 2002 2009
50 % 47 % 49 %
35 % 34 % 34 %
7.5 % 15 % 14 %
7.5 % 4 % 3 %
100
100
100
Chapter 4
Research Methodology
With regards to research methods employed by researchers in the fields of EMarketing and SBEs from 1993 to 2009, as can be seen from table 4-5, it is noticed
that not only are questionnaires and interviews the most commonly used research
methods from 1993 2009 in the field of E-Marketing but also the number of studies
conducted depending on these two research methods increased dramatically from
1993 2009. Within this respect, studies conducted using questionnaires increased
from 7 studies with a percentage of 11% of the total number of studies from 1993
1997 to 34 studies with a percentage of 25 % of the total number of studies from 1998
2002 and continued increasing to reach 67 studies with a percentage of 36 % of the
total number of studies from 2003 2009. On the other hand, studies conducted using
interviews increased from 5 studies with a percentage of 8 % of the total number of
studies from 1993 1997 to 25 studies with a percentage of 18 % of the total number
of studies from 1998 2002 and continued increasing to reach 61 studies with a
percentage of 33 % of the total number of studies from 2003 2009. This reflects
academic scholars attitudes towards questionnaires and interviews as suitable
research methods to be employed in investigating different issues related to EMarketing due to the benefits associated with the usage of both of them.
Table 4-5: Research methods employed by researchers in the fields of E-Marketing
and SBEs from 1993 to 2009
Number of studies
% of the total
% of empirical
N
Research
1993 1998 2003 1993 1998 2003 1993 1998 2003
Methods
1997 2002 2009 1997 2002 2009 1997 2002 2009
1 Questionnaire
7
34
67
11 % 25 % 36 % 50 % 47 % 48 %
2
Interviews
5
25
61
8 % 18 % 33 % 36 % 35 % 44 %
3 Observation
2
9
6
3 % 7 % 3 % 14 % 12 % 4 %
4 Focus groups
0
4
6
0% 3% 3% 0% 6% 4%
5
Conceptual
51
63
45
78 % 46 % 24 %
6
Literature
0
1
1
0% 1% 1%
review
Total
65 136
186
100
100 100
100 100
100
Based on the previous discussion and the review of the methodological
literature for the research methodologies, research strategies and research methods
employed by researchers in the fields of E-Marketing and SBEs from 1993 to 2009,
this research can argue that there is a need to conduct more quantitative studies to
increase the degree of generaliseability of the findings of such studies. Moreover, this
4-17
Chapter 4
Research Methodology
research suggests that when conducting such studies a triangulation methodology can
be used depending on survey and case study research strategies by employing
questionnaires and interviews as research methods. Appendix 12 illustrates a
classification of the literature according to the methodology employed.
4.5 The methodology and methods of this study:E-Marketing studies have utilised a combination of qualitative and
quantitative approaches. It is noticed that qualitative approaches have been used
mainly where profound understandings were needed in respect of particular EMarketing phenomena, whilst quantitative approaches have been useful in cases
where the data was available. On the other hand, this research argues that commonly
there is no optimal research methodology or technique since every methodology,
method or technique has some drawbacks or limitations. A researcher must employ
the most appropriate research methodology and method to his/her research.
Based on that, and the previous discussions, the general principle is that the
research methodology, strategy, and methods or techniques employed must be
appropriate for the questions that the researcher wants to answer. It is believed that
triangulation approach is the most suitable methodology to be used in this research.
Data, method and methodological triangulation will take place, in which quantitative
and qualitative data will be collected based on a combined research strategy (survey
and case study strategies) using questionnaires and interviews to address different
levels of the study. The survey strategy allows the answer of the 'what' question of the
research, such as: what is the relationship between using E-Marketing by the SBEs
and its marketing performance? On the other hand, the semi-structured interviews
strategy gives in-depth information and answers the 'how' and why questions (for
example: why the relationship is taking one direction instead of another?).
4-18
Chapter 4
Research Methodology
4-19
Chapter 4
Research Methodology
Find out new insight and identify general patterns (Robson, 2002).
4-20
Chapter 4
-
Research Methodology
Provide control over the environment, order of the questions, time, date and
place.
As a result, this research argues that interviews are better than any other data
collection method to collect the data for case studies. Beside that Kaplan and Duchon
(1988) believe in the appropriateness of the interviews approach for studies in which
research is in its early stages, and where the context and respondents are of particular
importance to the study (just like this research). This research argues that the
complexity of the context being investigated and the diversity of the issues related to
E-Marketing sustainability make the case study interviews approach both fruitful and
particularly useful.
4.6 The research data:4.6.1 Secondary data:Secondary data is: the data already collected in some other context than the
present study (Robson, 2002; P: 552). It provides the necessary background
information, builds credibility for the research report and helps to clarify the problem
during the exploratory research process. In order to collect the secondary data and to
get a richer picture of the level of importance of the elements that constitute EMarketing adoption, implementation and the factors that contribute to the success of
this implementation, a literature review were conducted. This included published
literature, reports, secondary case studies, published vendor success stories and
official, governmental and SBEs web sites on the Internet.
4.6.2 Primary data:Primary data is: the data collected specifically for the research project being
undertaken (Saunders and Thornhill. 2003: 486). It has been collected based on two
research strategies, namely, survey and case studies from Egypt and the UK. While
survey strategy provides several data collection methods like questionnaire, interview
and participant observation, the case study strategy involves various methods, such as
documentary analysis, interview, participant observation and focus group. For the
purpose of conducting this research, questionnaires will be used to conduct survey
strategy and interviews and focus group to conduct the qualitative part of the research.
4-21
Chapter 4
Research Methodology
4.7 Implementation of data collection methods:4.7.1 Sample design:As explained in chapter one the original intention of this research was to
conduct a comparative study between two countries (Egypt as a developing country
and the UK as a developed one). But from the findings of the first exploratory study
conducted in Egypt and due to the more elementary level of E-Marketing practices by
Egyptian SBEs, it was more appropriate to focus this research exclusively on UK
SBEs since this is more likely to generate more reliable and generaliseable results (the
Egyptian exploratory study will be discussed in detail in chapter five). Consequently,
this research will apply the research framework to UK industrial and trading SBEs
sectors, to investigate the potential for managerial insights. Therefore, as it is
extremely difficult to reach an entire population for any research, for empirical
reasons, this research will examine a sample of the population. As research
investigations usually involves several hundred or thousands of elements, it would not
be possible to collect data from, or to test, or examine, every element in terms of time,
cost, and other human resources (Sekaran, 2000). Other scholars in the fields of social
science like Bryman and Bell (2003), Aaker and Day (1986) and Churchill (1979)
agree that the use of samples is an appropriate and sufficient alternative if the research
population is large in size or when the cost or time associated with the collection of
data from the population is high. But it is recognised that, from an empirical
viewpoint, supported generalisations are usually based on partial information because
the object of any enquiries is a population consisting of a finite number of individuals,
each with some measure that is observable (Branett, 2002).
In this respect, the sample design will be a process that involves a number of
tasks and decisions in sampling. The main purpose of this process is to construct a
subset of the research population which is adequate and sufficient to represent the
population under investigation. Krathwohl (1997) defines sampling procedures as:
The ways of selecting a small number of units from a population to enable
researchers to make reliable inferences about the nature of that population
(Krathwohl, 1997; P: 160). Moreover, Aaker et al (2004) defines sampling as: The
4-22
Chapter 4
Research Methodology
process of surveying only a sample of the whole population to make inferences about
the population (Aaker et al., 2004; P: 760).
The sample design within this research will include the following steps:1. Definition of the research population
2. Definition of the population frame
3. Determination of the sampling technique (sample type)
4. Determination of the sample size
These different steps are discussed in the following part of the chapter and the
research sample design process is illustrated in figure 4-4.
Complete Sampling
Handling the
non-response
problems
4.7.1.1 The research population and population frame:The first step in determining the research sample is to define the research
population of interest both clearly and accurately (Schofield, 1996). According to
Kumar (2000), the population is the set of all objects that have some common set of
4-23
Chapter 4
Research Methodology
Chapter 4
Research Methodology
3- They provide the ability to evaluate the registered enterprises to make sure
that it can satisfy the essential requirement to be considered as SBEs
(number of employees and annual turnover).
4- They are very common databases and business directories among small
business owners (based on the results of the research pilot study within the
UK).
Based on the search conducted in the nine previously mentioned databases and
business directories, a population frame was constructed containing detailed subindustry classification information for 2124 SBEs within the UK. This information
included the enterprise name and web address. These 2124 enterprise fulfilled the first
three characteristics to be included in the population frame. To ensure that these
enterprises met the fourth criteria (using E-Marketing to conduct marketing activities)
and to confirm that it really met the first three characteristics, each single web address
of the yielded enterprises was visited by the researcher (totalling 2124 addresses) to
make sure that these enterprises are small in size, uses E-Marketing to conduct its
marketing activities and to complete any missing data about any of these enterprises.
This scanning process was conducted for two months over the period of time from
15/8/2008- 15/10/2008. Based on this investigation and scanning, 171 enterprises
were excluded from the population frame for the following reasons:1. Some sites were terminated and not working any more.
2. Some enterprises had more than one link in the databases and business
directories and as a result of that counted more than one time.
3. Some enterprises were classified by the databases and business directories
as small businesses while in reality they were not small in size with
regards to the number of employees and annual turnover.
Investigating and determining whether each of these enterprises is considered
to be a small business or not was the most difficult phase in the scanning process. It
was conducted through investigating the information provided in the enterprise web
site. Some enterprise mentioned clearly that they are small, but for most of the
enterprises it was a must to visit some icons (sub-web pages) like About Us or
Company Profile to find out whether they are small in size or not. This scanning
procedure was essential and determining the population frame without conducting it
4-25
Chapter 4
Research Methodology
may have lead to spurious results and obliterated the random base for the research
sample. Since in such a situation it would not meet the criteria for a probability
sample, where each person or unit in the population has an equal probability of being
selected (De Vaus, 1996). Consequently, a modified population frame database was
constructed from the enterprises that met the relevant applicable four population
criteria discussed earlier. This modified population frame database contained
complete detailed information about the enterprises in it, which included the
enterprise name, address, the enterprise web address, e-mail address (or the URL for
the contact us web page in the enterprise website in case of not finding an e-mail
address for the enterprise), marketing department contacts, phone numbers and fax
number.
On the other hand, as mentioned earlier, the study planned to obtain responses
from different industries so that generalisation of the findings could be established.
The modified population comprises SBEs in eleven specific industries, namely,
automotive, agriculture, chemical and allied products, constructions, computers and
IT, engineering, food and drink, healthcare, leisure, publishing and textiles.
Distribution of the population frame according to industry is illustrated in table 4-7.
Meanwhile, table 4-8 illustrate the Distribution of research population by location
Table 4-7: The distribution of the research population according to industry
N
1
2
3
4
5
6
7
8
9
10
11
Industry
Automotive
Agriculture
Chemical and Allied Products
Constructions
Computers and IT
Engineering
Food and drink
Healthcare
Leisure
Publishing
Textile
Total
4-26
Frequency
223
201
197
30
422
115
368
97
35
15
250
1953
Percent
11 %
10 %
10 %
2%
21 %
6%
19 %
5%
2%
1%
13 %
100.00%
Chapter 4
Research Methodology
4.7.1.2 Sample type:According to Sekaran (2000) a research sample can be defined as: A subset of
the population (Sekaran, 2000; P: 266). On the other hand, Schofield (1996) define
research sample as: A set of elements selected in some way from a population
(Schofield, 1996; P: 25). Moreover, Bryman and Bell (2003) define it as: The
segment of the population that is selected for investigation, it is a subset of the
population (Bryman and Bell, 2003; P: 93).
Generally, there are two main types of research samples: probability and nonprobability samples (De Vaus, 1996; Schofield, 1996; Bryman and Bell, 2003 and
Sekaran, 2000). A probability sample is a sample in which each element within the
population has an equal, or at least a known, probability of being selected within the
sample. Bryman and Bell (2003) define probability sample as: The sample that has
been selected using random selection so that each unit in the population has a known
chance of being selected (Bryman and Bell, 2003; P: 93).
In probability sampling, as a result of the fact that all the units within the
population have the same probability of being included in the sample, Bryman and
4-27
Chapter 4
Research Methodology
Bell (2003) and Bryman and Cramer (1994) argue that it is generally assumed that
probability sample will be a more representative sample of the population and that the
main aim of using it is to reduce the sampling error and to keep it to a minimum. On
the other hand, in order to have a random probability sample certain procedures to
ensure that the different units within the population have equal probabilities of being
chosen can be used (e.g., systematic random sampling, simple random sampling,
cluster random sampling, and stratified random sampling). In contrast, a nonprobability sampling contains some procedures that do not include random sampling
at some stage in the process (Krathwohl, 1997: P: 171). Within the same line, Bryman
and Bell (2003) define probability sample as: A sample that has not been selected
using random selection methodthis implies that some units in the population are
more likely to be selected than others (Bryman and Bell, 2003; P: 93).
According to Sekaran (2000) non-probability samples can be divided into
convenience sampling, quota sampling and purposive (judgmental) sampling.
Convenience sampling is used when the researcher selects sampling units that are
conveniently available. In purposive sampling, a researcher selects sampling units for
a certain purpose (De Vaus, 1996). Figure 4-5 illustrates the classification of sampling
techniques.
Sampling techniques
Probability samples
Un-probability samples
Simple Random
Convenience
Stratified
Judgmental Sampling
Systematic Sampling
Quota Sampling
Cluster Sampling
Figure 4-5: Sampling techniques.
Source: Adopted from Sekaran (2000).
However, for the purposes of conducting this research study, probability
samples are both preferable and desirable because (as discussed earlier) these type of
samples are more likely to produce a representative sample, reduce the sampling error
and keep it minimum and enable estimates of the sample's accuracy. Accordingly, the
4-28
Chapter 4
Research Methodology
ideal arrangement would be a probability sample taken from the research population
frame.
When determining the most suitable type of probability samples to be used
within this research, the efficiency of the different probability samples must be taken
into consideration. As the efficiency of the sample is measured by the size of its
sampling errors relative to other samples of equal cost, as illustrated by Sudman
(1976), stratified sampling is intended to provide the smallest sampling error. This is
also confirmed by Sekaran (2000) who illustrated that stratified random sampling is
the most efficient among all probability designs Sekaran (2000; P: 281).
Consequently, for the purposes of conducting this research and in order for the
eleven main industries to have good representation, stratified sampling is selected. As
discussed earlier, this study planned to obtain responses from different industries so
that generalisation of the finding can be made.
4.7.1.3 Sample size:Because of the limited resources (regarding time and effort) of the researcher
and to come up with an accurate and fair representation of the population
characteristics, the researcher depended on a stratified research sample which has
been selected randomly from the population of 1953 SBEs resulting from the
sampling frame. It was planned to determine the sample size according to the Aaker
and Day (1986) sample size equation which is highly accepted by social science
researchers since it takes into account the degree of required confidence, the sample
error, ratio of population characteristics available in the sample (50% in social
sciences) and population size. According to Aaker and Day (1986) the sample size
can be determined depending on the following equation:-
S=Z
p (1- p )
Nn
n
N-1
Where: Z = Degree of required confidence (95 %)
S = Sample error (5%)
P = Ratio of population characteristics available in the sample (50%)
N = Population size
n = Sample size
4-29
Chapter 4
Research Methodology
Many scholars like Bryman and Cramer (1998), De Vaus (1996), Sekaran
(2000) and Bryman and Bell (2003) illustrate that a large and adequate sample size is
the main method to ensure that the data collected would provide a reliable basis for
drawing inferences, making recommendations and supporting decisions. Within this
respect, a large and adequate sample size would remove bias and meet the criteria
required by the analytical methods used within the research. However, Bryman and
Cramer (1998) highlight that the sample size has to be related to the size of the
population. Moreover, Malhotra (2004) highlight that the required sample size
depends on factors such as the proposed data analysis techniques used to analyse the
data. On the other hand, according to De Vaus (1996) the required sample size
depends on two key factors, namely, the degree of accuracy the researcher require for
the sample, and the extent to which there is variation in the population in regard to the
key characteristics of the study (De Vaus, 1996; P: 70).
Based on the argument of Malhotra (2004), a researcher has to consider data
analysis techniques used within the study when determining the study sample size.
Within this respect, the most demanding proposed data analysis technique for this
study is Structural Equation Modelling (SEM) which is sensitive to sample size and
less stable when estimated from small samples (Garson, 2009; Tabachnick and Fidell,
2001). By reviewing the literature it was found that there are no generally accepted
criteria for determining a specific sample size for using structural equation modelling
(Hair et al, 1998; Garson, 2009; MacCallum et al., 1996; Chin, 1998; Mitchell, 1992
and Khong, 2005).
However, there are some general guidelines that have been proposed by some
researchers with regards to the suitable sample size to be used when using structural
equation modelling in data analysis. Within this respect, Hair et al (1998) suggest that
a sample with a size of less than 100 is considered to be a small sample. They also
suggest that a medium sample size is between 100 and 200, and a large sample size in
more than 200. On the other hand, Garson (2009) suggest that a sample size has to be
more than 100. Moreover, many researchers have used a sample size of around 100 to
conduct research using structural equation modelling (e.g. Khong, 2005; Graham,
2005; Eid 2003 and Battor 2008). Based on that, it is generally regarded that a sample
size of 100 is the practical acceptable size for using structural equation modelling.
4-30
Chapter 4
Research Methodology
Considering this sample size as a guide, the sample size in different degrees of
required confidence was calculated using the Aaker and Day (1986) sample size
equation. As illustrated in table 4-9, for achieving a 90 % degree of confidence the
sample size needs to be 21 and a sample size of 87 SBE will generate a degree of
confidence of 95%.
Table 4.9: Implementing Aaker and Day sample size equation in different degrees of
required confidence
N
Degree of confidence
Sample size (SBE)
1
90 %
21 SBE
2
95 %
87 SBE
Meanwhile, Aaker et al (2004) illustrate that a common approach in
determining the sample size is to find similar studies and use their sample size as a
guide (Aaker et al, 2004; P: 403). In light of the argument of Aaker et al (2004), many
scholars within the fields of social sciences like Michael and Beck (1995) argue that
simple random sampling (SRS) yields a sampling fraction of 1/10 (Michael and Beck,
1995; P: 3). In line with that, De Vaus (1996) considers that having a population of 50
using the sample of 10 is sufficient and that the sampling fraction would be 1/5 (De
Vaus, 1996; P: 64). Within this respect, a sample size of twenty percent of the total
population is accepted by most researchers within the field. Eid (2003) and Michael
and Beck (1995) are examples of researchers who used a sample size of twenty
percent of the total research population.
To this point, as far as this study is concerned, the research sample was chosen
to represent twenty percent of the research population. The information from 391
SBEs from the intended population frame was collected. The sample size was chosen
to represent twenty percent of the research population because:1- This sample size is expected to fulfil the requirements of all the statistical
techniques used within the study.
2- This sample size is appropriate to justify the cost and time limitations of
the researcher.
3- This sample size exceeds the sample size required to achieve 95% degree
of confidence according to the Aaker and Day (1986) sample size equation
which is highly recognised among researches in the field.
4-31
Chapter 4
Research Methodology
Sample Size
43
39
39
8
82
23
74
20
8
4
51
391
4.7.1.4 Unit of analysis:The unit of analysis is the unit from which information is obtained (De Vaus,
1991). As the main aim of this research is to clarify the different factors affecting the
adoption of E-Marketing by UK industrial and trading SBEs, as well as the impact of
this adoption on SBEs marketing performance, the unit of analysis is conducted at the
organisational level of analysis. Therefore, the SBE owner, marketing manager and/or
sales manager perceptions of E-Marketing adoption and the impact of this adoption
on performance are measured. They are regarded as the main source of the
information because they are the key decision-makers in the SBE and directly
responsible for planning, adopting and implementing E-Marketing activities within
the small enterprise.
4.7.2 Construction of the Questionnaire:Questionnaire construction is a very important aspect in conducting any
research. In order to achieve effective questionnaire design, research aims and
objectives should be formulated before the questionnaire is designed (Sudman and
Bradburn, 1982; P: 261), because a good questionnaire is one that accomplishes the
4-32
Chapter 4
Research Methodology
insure developing a good questionnaire the Tull and Hawknis (1980) model for
questionnaire design was adopted (see appendix four). The questionnaire was divided
into six different parts, each one of them addressing one particular area of interest.
Part one contained questions that give background information about the SBE, part
two measured the different factors affecting the adoption of E-Marketing by SBEs,
part three measured the main independent variable of the study (E-Marketing
adoption), and part four measured the different E-Marketing tools, forms and levels
of implementation within the small businesses. Part five looked at the current and
future marketing performance of the SBEs, and finally part six contained some
questions related to the individual respondents.
In designing the questionnaire great attention has been given to the questions,
it was short, direct, clear and comfortable in discussing to ensure getting the right data
and to avoid any harm to the participants. Beside that the questions that may have two
meanings or may lead to specific answers were avoided. On the other hand, the
questionnaire depended on several kinds of questions, which were developed by the
researcher. Within this context, the questionnaire included open questions, yes and no,
category, ranking, scale and quantity questions. Different groups of questions have
been used to cover each item and clear instructions were given to the participants to
avoid any confusion.
Because the research depended on SPSS (the Statistical Package for Social
Sciences V16) for analysing the collected data, a coding scheme was designed prior
to the process of collecting the data. Moreover, the order and flow of the questions in
the questionnaire was considered to get a logical flow of questions to help in
collecting the data. Gaining consent from the participants enterprises has been done
on a fully informed and freely given bases; the questionnaire was introduced carefully
4-33
Chapter 4
Research Methodology
to the respondents to ensure having a high response depending on a cover letter which
declared that participation is voluntary. Also, it was used to give the participants full
information about the research which included: the title and purpose of the research/
the research team (the researcher and the supervisors)/ the research sample and who is
being asked to participate/ the kinds of data required/ assurances about participants
privacy, confidentiality and anonymity/ assurances about data security and that it is
only for the purpose of, and will be used only for scientific research /..etc).
variables
which
can
not
be
measured
directly.
Accordingly,
the
operationalisation of these variables was done through generating multiple scale items
to measure these variables in quantitative terms.
All the measures of the research constructs has been developed and tested by
the researcher except for the measures related to the TAM and IDT factors which was
4-34
Chapter 4
Research Methodology
adopted from the measurements used by Moore and Benbasat (1991) to measure
perceived ease of use, perceive relative advantage (usefulness) and perceived
compatibility (see chapter three section 3.3). This was mainly due to the lack of
existing well-established measures to measure the research constructs. All these
research measures has been developed based on the in depth literature review as well
as the results of the research exploratory studies and have been found valid and
reliable based on the results of the pilot study as well as different reliability tests
(namely item-to-total correlation and Cronbachs Alpha). As discussed in detail in
chapter nine (section 9.2), the values of item-to-total correlation and Cronbachs
Alpha were considerably and significantly higher than the reliability acceptable levels
suggested by Edgett (1991) and Nunnaly (1978). More details about the pilot study as
well as the validity of the research measures are discussed in sections 4.8.1 and 4.9 of
this chapter. The following table present the research measures as well as its source.
Table 4-12: Research measures
N
1
Constructs
Items
Source
Internal Factors:
SBE owner skills
I find it easy to use E-Marketing tools for
conducting my business.
I find it easy to interact with E-Marketing tools
Interacting with E-Marketing tools require a little
mental effort by me.
I think that the Internet and other E-Marketing tools
are very important to conduct business.
There is a sufficient support from the top
management for the adoption of E-Marketing .
E-Marketing
adoption
SBE resources
We have good, qualified and skilled marketing staff
in our enterprise.
We have good technological infrastructure in our
enterprise.
We have sufficient financial resources in our
enterprise for adopting E-Marketing .
4-35
The researcher
based on the in
depth literature
review and the
results of the
research
exploratory
studies
Chapter 4
Research Methodology
We conduct E-Marketing because we have sufficient
financial resources.
We will implement E-Marketing even if we did not
have sufficient financial resources.
We will implement E-Marketing even if we did not
have skilled and qualified staff.
We can not conduct E-Marketing without good and
sufficient technical resources.
Type of products
One of the factors that influenced our decision of
adopting E-Marketing is the types of products
produced by our enterprise.
We would implement E-Marketing regardless of the
types of products produced by our enterprise.
The type of products produced by our enterprise did
not affect our decision of adopting E-Marketing .
If we changed the types of products that we
produced, we might choose not to adopt E-Marketing
.
SBE size
We are too small to adopt E-Marketing .
The size of our enterprise did affect our decision to
adopt E-Marketing .
We will adopt E-Marketing when we become a
bigger enterprise.
We adopted E-Marketing
regardless of our
enterprise size.
Using E-Marketing
performance.
improves my job
Compatibility
Using E-Marketing is compatible with all aspects of
my work.
4-36
Adopted from
Moore and
Benbasat
(1991)
Chapter 4
Research Methodology
Using E-Marketing is completely compatible with
my current situation
I think that using E-Marketing fits well with the way
I like to work.
Governmental influence
We adopted E-Marketing because of the incentives
provided by the government.
We adopted E-Marketing because of the protection
provided by the government.
We adopted E-Marketing because of government
influences.
There was no influence of the government on our
decision of adopting E-Marketing .
Our customers do not like purchasing through the
Internet.
Our customers do not trust E-Marketing tools
There is a lack of trust between enterprises
conducting E-Marketing activities.
Our customers prefer to pay in cash instead of
electronic payment methods
The customers usually do not trust E-Marketing
tools because of security issues.
The customers usually do not trust E-Marketing
tools because of privacy issues.
The customers usually do not trust E-Marketing
tools because they distrust the companies that
provide products using these tools.
4-37
The researcher
based on the in
depth literature
review and the
results of the
research
exploratory
studies
Chapter 4
Research Methodology
The majority of our customers are able to utilize
technology.
E-Marketing Implementation
2
E-Marketing
Forms
3
E-Marketing
tools
4-38
The researcher
based on the in
depth literature
review and the
results of the
research
exploratory
studies
The researcher
based on the in
depth literature
review and the
results of the
research
exploratory
studies
Chapter 4
E-Marketing
implementati
on level
E-Marketing
impact on
marketing
performance
Research Methodology
to conduct up to --- % of our marketing activities
In our enterprise we depend on E-Mail Marketing to
conduct up to --- % of our marketing activities
In our enterprise we depend on Mobile Marketing to
conduct up to --- % of our marketing activities
In our enterprise we depend on Intranet Marketing to
conduct up to --- % of our marketing activities
In our enterprise we depend on Extranet Marketing
to conduct up to --- % of our marketing activities
In our enterprise we depend on Business to Business
(B2B) to conduct up to --- % of our marketing
activities
In our enterprise we depend on Business to
Consumer (B2C) to conduct up to --- % of our
marketing activities
In our enterprise we depend on Business to
Government (B2G) to conduct up to --- % of our
marketing activities
In my enterprise, implementing E-Marketing leaded
to:
- New Sales
- New Customers
- Increased Profits
- Good Customer Relationships
- Reduction of sales costs
- Faster discovery of customer needs
- Greater customisation of products
- New markets
- Fast communication with customers
- Increased customer satisfaction
- Developing new products
- Faster adaptability of customer needs
- Providing better service quality
- Increased market share
- Increased brand equity
The researcher
based on the in
depth literature
review and the
results of the
research
exploratory
studies
The researcher
based on the in
depth literature
review and the
results of the
research
exploratory
studies
4.8 Data preparation (reliability and validity):Reliability and validity are tools of an essentially positivist epistemology
(Watling, as cited in Winter, 2000; P. 7)
To gain meaningful results from the data analyse stage, some steps were
conducted to make sure that the data was well prepared and to assess the reliability
and validity of the measurements used in the survey. These steps are discussed in the
following sections of the chapter.
Howell et al (2005) illustrate that validity refers to the degree to which a study
accurately reflects or assesses the specific concept that the researcher is attempting to
4-39
Chapter 4
Research Methodology
With regards to validitys definition, Joppe (2000) sees validity as a tool to:
determine whether the research truly measures that which it was intended to
measure or how truthful the research results are. In other words, does the research
instrument allow you to hit "the bulls eye" of your research object? Researchers
generally determine validity by asking a series of questions, and will often look for the
answers in the research of others (Joppe, 2000; P. 1). On the other hand, Bell (1999)
defines validity as: Regarding an item measures or describes what it is supposed to
measure or describe (Bell, 1999; P: 104). Moreover, while validity is concerned
with the study's success at measuring what the researchers set out to measure,
reliability is concerned with the accuracy of the actual measuring instrument or
procedure (Howell et al, 2005).
Joppe (2000) defines reliability as: The extent to which results are consistent
over time and an accurate representation of the total population under study and if
the results of a study can be reproduced under a similar methodology, then the
research instrument is considered to be reliable (Joppe, 2000; p: 1). In addition
Robson (2002) defines reliability as: The extent to which a measuring device, or a
whole research project, would produce the same results if used on different occasions
with the same object of study (Robson, 2002; P: 551). On the other hand, Kirk and
Miller (1986) identify three types of reliability which relates to:(1) The degree to which a certain measurement remains the same.
(2) The stability of a measurement over time.
(3) The similarity of measurements within a given time period (Kirk and
Miller, 1986, P: 41 and 42).
4-40
Chapter 4
Research Methodology
4.8.1 Content validity:The process of assessing the research measurements reliability and validity
started with ensuring the content validity of the research instrument. According to
Carmines and Zeller (1991) content validity can be defined as: The extent to which a
measurement reflects the specific intended domain of content (Carmines & Zeller,
1991, p.20). On the other hand, The USA Environmental Protection Agency - USEPA
(2009) defines it as: The ability of the items in a measuring instrument or test to
adequately measure or represent the content of the property that the investigator
wishes to measure (The United States Environmental Protection Agency, 2009, P: 1).
Based on these definitions and the arguments of researchers (Carmines and
Zeller, 1991; Nunnally and Bernstein, 1994 and Haron 2002), a measure has content
validity if there is a general agreement among the subjects and researchers that the
instruments has measurement items that cover all the content domain of the variables
being measured. Moreover, according to McDaniel and Gates (1996), a measurement
must satisfy some certain criteria before it can be applied in an empirical fieldwork.
These criteria are:1. Carefully defining what is to be measured.
2. Conducting a careful literature review and interviews with the target
population.
3. Let the scales be checked by experts.
4. The scale has to be pre-tested.
For the purpose of ensuring the content validity of this research instrument,
the criteria proposed by McDaniel and Gates (1996) was adopted. Within this respect,
the proposed research variables were developed and defined carefully through a
deductive process from the literature review as well as an inductive process through
semi-structured interviews with the target population (SBEs with the UK). Moreover,
the research questionnaire and scales has been checked, reviewed and examined by
4-41
Chapter 4
Research Methodology
the researcher as well as academic research experts from Bradford University School
of Management, Strathclyde Business School, Wolverhampton Business School, St
Andrews Business School and Cairo University Business School. Within this
examination process, the dimensionality of the research scales and it relevance to the
target population was examined. Afterwards, the suitability and acceptability of the
research questionnaire were tested through a pilot study of small business
entrepreneurs and small business marketing managers within West Yorkshire County
as well as academic staff and doctoral researchers within Bradford University School
of Management and across the UK. Based on that, the research scales within this
study are considered to possess content validity since if has gone through all the
needed actions and procedures to gain this validity.
4-42
Chapter 4
Research Methodology
that the values of the data have been entered into the computer software correctly. The
review process was conducted by the researcher and one other researcher within
Bradford University School of Management through having the researcher assistant to
read the codes directly from the coded questionnaires while the researcher is listening
and making sure that the values are correct in the SPSS data file. By finishing this
stage, the data became ready for analysis.
4.9 Purification of measures:After completing the data entry and reviewing processes, all the measures used
within the research were then tested and purified by assessing their reliability and
validity prior to any future data analysis. The main aim of this process is to make sure
that all the measures used within the research is sufficiently and significantly
reflecting the underlying variables that it is attempting to measure. This will not only
provide support for triangulation of the research results but will also provide a more
significantly meaningful explanation of the phenomena that the research is
investigating. On the other hand, a reliable and valid measuring instrument enhances
the methodological rigour of the research (Eid, 2003).
Purification of the measures within this study has been done depending on the
internal consistency approach through a two stages process based on item-to-total
correlation and coefficient alpha (Cronbach alpha). Item-to-total correlation and the
Cronbach alpha coefficient is considered to be a very popular reliability index in the
field of social science research (Haron, 2002; Eid 2003; Shoobridge, 2004, Churchill,
1979 and Mat-Saad, 2001).
4.9.1 Item-to-total correlation test:Item-to-total correlation is one of the most widely accepted reliability
measures among social science researchers. It is mainly based on the determination of
the relationship of a particular item to the rest of the items in a specific dimension
(Haron, 2002). According to Churchill (1979), using item-to-total correlation ensures
that the items making up a specific dimension share a common core. When using
item-to-total correlation for the purification of measures (which can be calculated by
SPSS), according to Edgett (1991), only those items with an item-to-total correlation
of 0.3 or more can be accepted and should be retained for further analysis because
4-43
Chapter 4
Research Methodology
these are the only items that are considered to have high reliability. On the other hand,
any items with an item-to-total correlation less than 0.3 should be removed and
excluded because it has low correlation unless they represent an additional domain of
interest.
4.9.2 Coefficient alpha (Cronbach alpha):According to Nunnally (1978), the estimation of reliability is based on the
average correlation between items within a dimension which is concerned with
internal consistency. The statistical technique used in determining the reliability based
on this internal consistency is called Coefficient Alpha and known among
researchers as Cronbach alpha.
Cronbachs alpha is a coefficient of reliability (or consistency) first named as
alpha by the American educational psychologist Lee J. Cronbach (1916 - 2001)
(Kupermintz, 2003). Cronbach alpha technique is based on the calculation of the
mean reliability coefficient for all possible ways of splitting a set of items into two
halves (Haron, 2002). Cronbach's alpha can be calculated as a function of the number
of test items and the average inter-correlation among these items according to the
following formula:-
Chapter 4
Research Methodology
indicates that the scale is more reliable. Moreover, Van de Ven and Ferry (1980)
contended that a value of 0.7 or higher is sufficient. For the purpose of this study, the
Magal et al., (1988) and Nunnally (1978) point of view will be adopted within this
research.
Within this study Cronbach alpha was used to assess the reliability of the
research measurement scales for the following reasons:1- Cronbach alpha coefficient is considered to be a very reliable, accepted and
widely used reliability index in the field of social science research (Haron,
2002; Eid 2003; Shoobridge, 2004 and Mat-Saad, 2001).
2- Cronbach alpha is applicable to this study since it requires the measurement
scale to include equal intervals (Peter, 1979) and the study used Likert scale
(equal intervals scale) to measure the research variables.
3- Depending on Cronbach alpha will ensure the reliability of the research
measurement scales which in turn will lead to have a more meaningful
explanation of the E-Marketing phenomena investigated within this study.
Cronbach alpha were calculated for all the measuring instruments in the
research questionnaire namely; E-Marketing adoption, compatibility, ease of use,
relative advantage, owner skills, organisational culture, SBE resources, SBE size, type
of product, international orientation, competitive pressure, government influence,
cultural orientation, implementation level, impact on future performance and finally
impact on current performance. Any item with a Cronbach alpha value less than 0.6
will be excluded.
4.9.3 Reliability Analysis Results:The statistical package for social science was used to calculate the item-tototal correlation as well as the coefficient alpha (Cronbach alpha) to perform the
reliability analysis process for the research measures. The process started with the
calculation of item-to-total correlation and coefficient alpha to obtain the total
dimension of the constructs. Within this respect, Cronbach alpha was calculated first
for all the research measures to determine its level of reliability. Based on calculations
Cronbach alpha coefficient ranged from 0.649 to 0.924, which was considerably
higher than the reliability acceptable level of 0.60 suggested by Magal et al., (1988)
4-45
Chapter 4
Research Methodology
and Nunnally (1978). Consequently, based on coefficient alpha results the research
measures are satisfactory acceptable for conducting further data analysis through
inferential statistics to test the research hypothesis.
To confirm the coefficient alpha results, item-to-total correlation was
calculated for all the items within each domain (112 items in total) to determine the
degree of correlation for each item which was then used to determine the low item-tototal correlation items. Out of the research dimensions, eleven items in three
dimensions were found to have a low item-to-total correlation that is below the
acceptable limit of 0.3. These items include three items in internal factors group, one
item in the external factors group and seven items in the E-Marketing adoption
group. Based on the results of the item-to-total correlation, although coefficient alpha
results were acceptable, the items with low item-to-total correlation were removed in
order to improve the reliability of the research scales. Moreover, both item-to-total
correlation and coefficient alpha (Cronbach alpha) was recalculated.
After removing all the eleven items and recalculating the item-to-total
correlation, the item-to-total correlation values for all the items were found to have a
high value which is above the acceptable limit of 0.3, and the correlation ranged from
0.314 to 0.919. Afterwards, Cronbach alpha was recalculated and it was found that
not only all the research measures were found to have a value of coefficient alpha
which is significantly above the acceptable level of 0.60 and ranging from 0.728 to
0.924, but also the values of the coefficient alpha recorded an improved reliability.
The results of the item-to-total correlation and coefficient alpha (Cronbach alpha) first
and second analysis are presented and discussed in chapter nine in more detail. These
results (as summarised in table 4 12) confirm that the research instrument and scales
used within this research possesses a high level of reliability and is satisfactorily
acceptable for conducting further data analysis through inferential statistics to test the
research hypothesis.
Table 4-12: Reliability Analysis for the Research Variables
Item
Code
A
A1
Item
A. Internal factors:
Entrepreneur (owner) skills
4-46
Cronbachs
Alpha
Total
Number of
Items
.802
Chapter 4
A2
A3
A4
A5
A6
A7
A8
A9
B
B1
B2
B3
C
H1
H2
Research Methodology
Organisational culture
The SBE resources
Type of the product
International orientation
Size of the firm
Perceived ease of use
Perceive relative advantage
Perceived compatibility
B. External Factors:Competitive pressure
Government influences
Cultural orientation towards E-Marketing
C. E-Marketing adoption
The current effect of implementing EMarketing on the marketing performance
The future effect of implementing EMarketing on the marketing performance
.762
.807
.728
.828
.783
.924
.858
.846
5
4
4
4
4
7
4
4
.889
.913
.836
.847
6
4
8
10
.736
16
.862
15
4.9.4 Data collection of the questionnaires:After the pilot test and checking the validity and reliability, the questionnaire
was sent by mail to be returned after completion to the researcher by prepaid
unmarked envelope which provides confidentiality for the participants, avoid any
harm to them, and gives them the chance to choose a suitable time to complete the
questionnaire. Additionally a link for an electronic version of it was sent to the
research participants via e-mail. The completion of the questionnaire was followed by
e-mails and telephone calls (reminders) to the participants, but that was done for a
suitable and reasonable number of times and in appropriate timing to avoid any harm
to the participants. Finally, after receiving the completed questionnaires, the answers
of each respondent was organised, coded and entered into SPSS. It was then analysed
by using the appropriate techniques.
The survey questionnaire targeted a sample of 391 SBEs within the UK that
were selected randomly from a population of 1953 SBEs within the same region. Of
the 391 SBEs, a total of 153 questionnaires were returned. This included 14 returned
(not completed) with a label stating Gone Away or Not in this address any more.
A further 19 questionnaires were returned with a covering letter explaining why they
had not been completed. For the most part of the responses, the participants indicated
that it was not the enterprise guiding principles to participate in surveys or that they
4-47
Chapter 4
Research Methodology
did not participate in the research as a result of lack of time due to work pressures.
Also 3 questionnaires were returned without completing all the needed questions
within the questionnaires and another 3 questionnaires were complete but the annual
sales of these enterprises exceeded 8.5 millions pounds and were excluded from
participating in the study as these enterprises can not be considered as SBEs
according to the definition used within this study. Therefore, the number of returned
completed questionnaires was 114. Table 4-13 provides a summary of the responses
distribution and rate. The response rate scored 32.38% (the useable response rate
scored 29.16%) and was calculated depending on the technique developed by De
Vaus (1991; P: 99) and Bryman and Bell (2003; P: 104). Based on these methods the
response rate is calculated according to the following equation:Number of usable questionnaires
X 100
Total sample unusable or un-contactable member of the sample
391
114
14
19
6
32.38 %
4.10 Semi-structured interviews:On the basis of the findings of the Egyptian primary survey, as well as the
quantitative findings of the UK primary survey, in-depth interviews have been applied
to collect the research data. Yin (1994) argues that semi-structured interviews is a
suitable choice of research method when researchers have small control over the
environment; when the phenomena under investigation is contemporary; and when the
context of the research is important. Since case studies would enable the researcher to
gain an in-depth understanding of why and how the SBEs under investigation were
using E-Marketing, it was adopted. Moreover, Bryman (1989) argues that case studies
provide one of the chief arenas in which quantitative and qualitative research can be
combined (Bryman, 1989; P: 175). He further adds that semi-structured interviews is
4-48
Chapter 4
Research Methodology
4-49
Chapter 4
Research Methodology
the semi-structured interviews participants for accuracy. The case studies provided a
more in-depth investigation of the motivation and experience that could not be
obtained by the survey. The semi-structured interviews results provided a level of
validation of the information gathered through the survey.
4.11 Chapter summary:Within this chapter the methodology used to conduct the research study as
well as the issues related to the chosen research methodology were discussed. This
discussion was built on the outcomes of chapters two and three and through it the
steps that were taken to address the research design, the data collection and analysis
methods used to conduct the research study were illustrated. These issues were
addressed in light of the basic research objectives and the relevant research questions.
The chapter started by providing a basic background about the different research
methodologies and strategies and the importance of selecting an appropriate research
approach, research methodology and strategy. Afterwards, the basic assumptions and
practical issues of different research paradigms were illustrated and discussed. Based
on this discussion, the chapter proved that there are sufficient philosophical and
practical reasons for depending on the post-positivist approach as the research
approach for this research. Moreover, the research methodology and stages were
discussed with reference to the adopted research philosophy.
Subsequently to identify the most appropriate research methodology and
methods to conduct this research, a review of the literature on the available research
methodologies, research strategies and research methods implemented by scholars
within the fields of E-Marketing and SBEs from 1993 to 2009 were discussed and
illustrated. Based on this review as well as the evaluation of this literature, it was
found that most of the researchers who contributed to the field of E-Marketing
depended on a single methodology or method when conducting research studies and
that most of the studies conducted have some critique points regarding the
methodologies and methods used in it. The discussion within the chapter illustrated
that there is no optimal research methodology or method since each methodology or
method has some drawbacks or limitations, but a researcher must employ the most
appropriate research methodology and method to his research. Based on that it was
illustrated that this research will depend on a data, method and methodological
4-50
Chapter 4
Research Methodology
4-51