A Conceptual Model of the Research Methodology Domain

With a Focus on Computing Fields of Study

Colin Pilkington

and Laurette Pretorius

School of Computing, University of South Africa, Florida Park, South Africa

College of Graduate Studies, University of South Africa, Pretoria, South Africa

Keywords:

Research Methodology, Conceptual Model.

Abstract:

Recognising the need for the development of research capacity and changing learning paradigms that include

online and collaborative approaches, an ontology of research methodology needs to be developed to allow for

the shared creation of knowledge in this domain. An ontology engineering approach is followed in developing

a conceptual model of the domain using UML, with a focus on studies in the computing disciplines. A

research scheme that is made up of a philosophical world view, a research design, and research methods is

proposed. Appropriate relations between these are identiﬁed, as well as attributes of the various concepts in

the conceptual model. A focus group consisting of senior researchers in the ﬁeld of computing was utilised to

validate the model.

1 INTRODUCTION

Developing research capacity is vital if there is to

be the creation of new knowledge and innovation, as

well as having economic beneﬁts (Crossouard, 2008;

Green Paper, 2012). Postgraduate enrolments have

been increasing globally (Engebretson et al., 2008),

and the inclusion of online and collaborative learning

approaches in educational paradigms (Johnson et al.,

2014) should be harnessed to support the increase in

postgraduate output, particularly as it relates to the

supervision of these postgraduate students. Supervi-

sors are faced with a range of dilemmas in the context

of increasing student numbers, with increasing stu-

dent diversity, and often with additional problems as-

sociated with distance education (Wisker et al., 2007).

As the Internet has expanded and online technologies

have become more popular, the way students access

and build knowledge has also become more diverse

(H¨akkinen and H¨am¨al¨ainen, 2012). However, use of

technology is not just about getting content to stu-

dents (where intelligent search engines can locate ma-

terial) – students also need to be brought into learning

networks (Chatti et al., 2007).

It has been noted that research methodology is of-

ten a problem for masters and doctoral students (Hof-

stee, 2006). Thus, to support postgraduate students

in their collaborative, online learning of research

methodologies, an ontology of the domain will pro-

vide the necessary communication support through

the development of a common vocabulary, as well as

allowing for accessing, sharing, integrating, and an-

notating such semantically driven knowledge (Bera

et al., 2010; DiGiuseppe et al., 2014; Motik et al.,

2002; Noy and McGuinness, 2001). Central to this

task is the creation of a conceptual model of the re-

search methodology domain, and the purpose of this

paper is to propose a model of the domain that would

be suitable for use by postgraduate students study-

ing in computing disciplines. The research question

is thus: What are the main concepts and relations

that make up a research methodology that will guide

and support students in their understanding of the do-

main?

Recognising that the provision of software tools in

the support of postgraduate supervision may be inad-

equate (Yew et al., 2011), such a conceptual frame-

work could provide the basis for learning environ-

ments that may be of value in both online as well as

blended approaches to postgraduate research and su-

pervision. Such support can be offered to supervisors

and students separate from a particular model of post-

graduate supervision or research education, in both

physical and virtual learning environments. However,

the model will also be limited to the support of the

discovery and learning of research methodologies and

will not cover the whole scope of research education

such as developing research questions, doing a litera-

Pilkington, C. and Pretorius, L..

A Conceptual Model of the Research Methodology Domain - With a Focus on Computing Fields of Study.

In Proceedings of the 7th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2015) - Volume 2: KEOD, pages 96-107

ISBN: 978-989-758-158-8

ture review, or writing up the results.

A review of the ﬁeld of ontologies and ontology

engineering (section 2) will provide the basis for the

development of a conceptual model. The research

methodology domain and its conceptualisation will

then be described (section 3). Finally, conclusions

will be drawn and pointers to future work outlined

(section 4).

2 ONTOLOGIES AND

ONTOLOGY ENGINEERING

The concept of the Semantic Web, dating back to

2001, refers to a “new form of Web content that is

meaningful to computers” (Berners-Lee et al., 2001),

and it aims at allowing for the interconnection of facts

or data (Gonzalez, 2013) in diverse locations and for-

mats. This term captures two important concepts

(Feigenbaum, 2013a):

1. “Semantic” refers to the meaning of the data

that is explicitly represented, and this meaning is

transferred along with the data.

2. The individual facts or pieces of data are linked in

a network of information.

The two point to the centrality of data in the Se-

mantic Web and that reasoning about it should not

become the primary objective (Feigenbaum, 2013b).

It is believed that it is in the potential behind the

linking of data with consistent interpretations across

a variety of sources that the power of the Semantic

Web lies (Berners-Lee et al., 2006). Furthermore, an

educational Semantic Web that seeks to link educa-

tional knowledge will occur in tandem with develop-

ments in the wider semantic web technology sphere

(Carmichael and Jordan, 2012).

2.1 Ontologies in Education

One approach to the Semantic Web proposes the use

of ontology-based formal annotations (Dzbor et al.,

2007), which can enhance an educational social net-

working approach (such as semantic wikis, for exam-

ple) (Limpens et al., 2008). This formal approach

has been used in e-learning, where semantic technolo-

gies have been used in the tagging of learning mate-

rials (Cu´ellar et al., 2011) – efforts that can be seen

in EdNA, Educational Modelling Language (EML),

SCORM, and IMS Learning Design speciﬁcations

(Cope and Kalantzis, 2011; Dzbor et al., 2007). This

approach has led to a move away from creating con-

tent to the collecting of current distributed learning

objects from multiple sources into a personalised as-

semblage of learning material, using semantic meta-

data to construct individualised courses (Wen et al.,

2012). The effect of the Semantic Web on e-learning

has, thus, not just been in the semantic tagging of

learning material, but also in the non-trivial provi-

sion of brokerage tools and services that match stu-

dents to learning materials (Dzbor et al., 2007; Wen

et al., 2012), and these tools and services could be

used in research methodology studies for postgradu-

ate students. The purpose of using semantic meta-

data in these situations is primarily to open up the

way to resource discovery – a primary purpose of the

Dublin Core standard (Cope and Kalantzis, 2011; Dz-

bor et al., 2007).

In this work of making resource discovery more

accurate and unambiguous via semantic annotation,

formal ontologies play an important role (Cope and

Kalantzis, 2011; Limpens et al., 2008). An ontol-

ogy provides structure to data by providing a shared,

formal, and explicit conceptual model of a domain

(Berners-Lee et al., 2006; Charlton et al., 2012;

Naeve, 2005) and allows a common understanding of

data by providing a controlled and limited vocabulary

that deﬁnes the concepts of a particular domain rig-

orously as well as deﬁning the relationships among

them (Berners-Lee et al., 2006; Cope and Kalantzis,

2011). Although such ontologies are often built by

an expert, this is not a requirement of an ontology

per se, and there are approaches (such as CGMean-

ing) that allow a community of users to extend such

an ontology and establish a dialogue with the commu-

nity on the particular expression of reality in the on-

tology (Cope and Kalantzis, 2011). The critical value

of such a shared domain model where co-construction

of knowledge is contemplated and a common vocabu-

lary is necessary has been noted previously (Charlton

et al., 2012), as tools based on an ontology are able

to interpret what the user is intending, as the seman-

tics, or meaning, of words is formally represented,

and, as such, are an improvement over simple char-

acter strings and mark-up languages such as metadata

tags and XML (Charlton et al., 2012).

2.2 An Ontology Engineering Approach

It has been argued that there is no single, correct on-

tology engineering methodology, as well as there not

being only one correct way to model a domain (Brusa

et al., 2006; Di Maio, 2011; Nagyp´al, 2007; Noy and

McGuinness, 2001). Various approaches have been

proposed and used (Corcho et al., 2003; De Leenheer

and Christiaens, 2007; Di Maio, 2011; Keet et al.,

2013; Nagyp´al, 2007; Noy and McGuinness, 2001).

A Conceptual Model of the Research Methodology Domain - With a Focus on Computing Fields of Study

Yet such a methodology, or ontological engineering

process, is useful, if not crucial, as it facilitates a prag-

matic and systematic path through the complex pro-

cesses and tasks of building an ontology (Devedzic,

2002; Di Maio, 2011; Nagyp´al, 2007). The ontology

engineering process used in this research is based on

the work of

• Nagyp´al – METHONTOLOGY, which has been

suggested to be the most mature of the various ap-

proaches (Corcho et al., 2003; Nagyp´al, 2007),

• Noy and McGuinness – Ontology Development

101 (Noy and McGuinness, 2001),

• Di Maio – JEOE (Just Enough Ontology Engi-

neering), although the author proposed this an

approach to ontology engineering rather than a

methodology (Di Maio, 2011), and

• Brusca, Caliusco, and Chiotti (Brusa et al., 2006),

all of which are largely application independent

methodologies.

The strategy that will be applied in the case of the

research methodology ontology will be a middle-out

approach (Nagyp´al, 2007). This will allow a form of

brain-storming at the start of the process to ﬁnd the

most salient concepts in the research methodologydo-

main. Once this is done, the model can be reﬁned, and

the ontology can grow both upwards (to more abstract

concepts) and downwards (to more speciﬁc concepts).

Ontology development methodologies suggest

separating the two main tasks of the process – speciﬁ-

cation, and implementation (Di Maio, 2011) – and in

this paper the focus will be on speciﬁcation. How-

ever, the conceptualisation of the research method-

ology model is seen as separate from its speciﬁca-

tion (Brusa et al., 2006; Nagyp´al, 2007), and this

was added as a separate step; speciﬁcation and con-

ceptualisation can thus be taken as the knowledge

acquisition part of the ontology engineering process

(Devedzic, 2002). Figure 1 lays this process out in

summary form. The approach followed allowed the

ontology to be expressed in a conceptual model ﬁrst

before being implemented in an ontology representa-

tion language.

2.2.1 Speciﬁcation

The point of this step is to gain knowledge about the

domain, and what needs to be achieved.

Stakeholders: The stakeholders would include this

researcher, postgraduate students (and their supervi-

sors) in the ﬁelds of computer science and informa-

tion systems, as well as researchers more generally.

Purpose, goals, requirements: The purpose of the

Ontology development

process

Specification

Conceptualisation

Implementation

Identify stakeholders

Define purpose, goals,

and requirements

Outline knowledge

sources

Delimit scope and

granularity

Plan quality assurance

Propose competence test

Generate vocabulary

Formulate concepts

Define properties,

relations, axioms

Figure 1: An ontology development process map.

ontology is to be a content- and communication-

oriented ontology that describes the research method-

ology ﬁeld/domain and allows for sharing of knowl-

edge, with the goal being to allow the ontology to

be used to tag data in research publications, as well

as set up and consolidate a knowledge base about re-

search methods that can be used by stakeholders; se-

mantic searching and reasoning over the content of

the ontology should also be possible. The ontology

should be usable by both masters and doctoral stu-

dents (novices) and supervisors (experts).

Knowledge sources: These would be current, expe-

rienced supervisors/researchers in the domain under

consideration, as well as books that have been written

to support the postgraduate student in the learning of

research methodologies.

Scope: This would be a domain ontology that will

cover all major research methodologies used in post-

graduate studies in computer science and information

systems.

Quality plan: The quality was initially ensured via a

focus group discussion based on the conceptualmodel

of the domain being modelled, and will be further

tested in the next phase of the research.

Competence/knowledge boundary tests: The research

methodology ontology should be able to answer ques-

tions similar to the following.

• Which methods are used by a particular design?

• What are a particular design’s assump-

tions/theoretical base?

• With a particular theoretical base, which designs

could be used?

• Which designs can speciﬁc methodsbe used with?

• Which is the best design to achieve a certain out-

come?

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

• What are the characteristics of a particular design?

2.2.2 Conceptualisation

This is one of the most complex tasks in ontology de-

velopment (Nagyp´al, 2007), and aims to produce a

model of the research methodology domain in a form

that will allow communication with domain experts

who may not be fully conversant with ontology lan-

guages (Nagyp´al, 2007) – UML in this case. This

phase has been referred to as “developing the arte-

facts” (Di Maio, 2011, 7), and here the use of se-

mantic networks, brainstorming, and mind maps are

recommended for the early stages of the development

process (Nagyp´al, 2007). The following steps were

not attempted linearly, recognising that many of these

ideas are closely interlinked (Noy and McGuinness,

2001), and thus they were accomplished as a group of

tasks rather than completed individually.

Vocabulary: The important terms in the research

methodology domain were listed and organised in an

informal way as a start to the process of building the

ontology.

Concepts: A concept could be described as “funda-

mental to our ability to think, express, represent and

communicate, ... Concepts can correspond to things,

but also to fuzzy clouds of ideas and notions identi-

ﬁed by words and related to a certain thing or sub-

ject. ... cognitive artifacts that support categorization

and communication, and are necessary to support hu-

man and artiﬁcial thinking and reasoning.” (Di Maio,

2011, 7). Concepts in a research methodology that

were considered important in this domain related to

the types of designs of a research endeavour as well

as the methods that would be used. The grounding in

some philosophical position was also considered im-

portant.

Properties: The properties, or attributes, of the iden-

tiﬁed concepts were established.

Relations: Relations are the “semantic interdepen-

dence” (Di Maio, 2011, 8) between concepts, and

these were identiﬁed. It is in this task that a concept or

taxonomic hierarchy is developed (Noy and McGuin-

ness, 2001).

2.2.3 Implementation

Transferring the conceptual model to a formal no-

tation such as ﬁrst order logic or description logics

is sometimes included as a formal part of the ontol-

ogy engineering process (Nagyp´al, 2007), and this ap-

proach will also be followed here in work in develop-

ing the ontology further into description logics and,

ﬁnally, OWL. As work has been done on transferring

a UML conceptual model directly to OWL (Zedlitz

and Luttenberger, 2014), this can be used to check the

OWL implementation developed from the description

logics phase. The implementation activity thus in-

volves taking the conceptual model and representing

it in some ontology language.

3 A CONCEPTUAL MODEL OF

THE RESEARCH

METHODOLOGY DOMAIN

There are several books that discuss the research

methodology domain for postgraduate students that

will be referenced in this section, as well as the

Sage Research Methods Online (SRMO) website

(http://srmo.sagepub.com), and although these cover

much the same material, there are differences in the

way the research methodologies are structured and

discussed, what topics are included, and the termi-

nology used. It is hoped that in the construction of

a conceptual model of the domain, a single model can

be designed that would bring these various disparate

ideas together in a form that will ultimately beneﬁt

the postgraduate student. Similar work has been done

in the Ontology of Clinical Research (OCRe) where

such a model serves as a common semantics for hu-

man clinical studies supporting data description, data

sharing, and knowledge-based decision support (Sim

et al., 2014).

Conceptual modelling can be seen as “identify-

ing, analyzing and describing the essential concepts

and constraints of a domain with the help of a (di-

agrammatic) modeling language” (Guizzardi et al.,

2002, 69). There is, furthermore, basic agreement that

a conceptualisation is a “formal structure” (Guarino,

2005, 7), or representation, about some aspect of the

real world with the aim of allowing that domain to be

better understood and communicated – that is, pro-

viding clear and unambiguous semantics is an essen-

tial objective (Bera et al., 2010; Partridge et al., 2013;

Wand et al., 1999; Weber, 2003).

The conceptual model will be built using UML,

and so a discussion of this will precede the presen-

tation of the model. The model will then be pre-

sented following the main components of the con-

ceptual model. Note that only some representative

classes will be included in the diagrams of the model

to give an idea of the structure of the model, and that

the data types that are given in the class diagrams

will be listed in a table along with possible instances.

Note that the string data type used in the diagrams al-

lows for the entry of any string that would describe

the data variable, and that bool(ean) allows for yes/no

A Conceptual Model of the Research Methodology Domain - With a Focus on Computing Fields of Study

or true/false type values. The validation of the model

will close the discussion of the model.

3.1 Modelling using UML

Conceptual models are often built using some mod-

elling grammar (Bera et al., 2010; Weber, 2003),

and as conceptual models largely subscribe to object-

oriented views of the world (Borgida and Brachman,

2003), it was decided to use UML as the modelling

grammar for this conceptual model – a choice that

has often been been made (Zedlitz and Luttenberger,

2014). Further reasons for using UML include its

open standard and that it has a strong history in con-

ceptual modelling (Burek, 2003).

It has been noted that there are problems with us-

ing UML as a conceptual modelling grammar as it is

not based on sound ontological theory (Burek, 2003;

Weber, 2003). Such problems relate to questions

about creating one single, all-encompassing concep-

tual modelling grammar as against several specialised

grammars, although no deﬁnitive answer is arrived at

(Weber, 2003). However, it is conceded that UML

is the current standard conceptual modelling gram-

mar, and acknowledged that it does offer power-

ful conceptual modelling constructs (Weber, 2003).

UML also models specialisation/generalisation rela-

tions, and uses composition/aggregation to handle

part-whole relations(Keet and Artale, 2008). It is on

this basis that the choice of UML has been main-

tained.

3.2 Research Scheme

A conceptual model is easier to understand and man-

age if it has a root concept or class (Motik et al.,

2002), and thus the model presented here is rooted

in the concept of a research scheme. It is necessary

to make it clear what is meant by the term research

methodology as against the term research scheme (as

these terms are used in this paper). Research method-

ology has been used in various ways in the literature,

from being a synonym for a research design, to being

the research process, to being the speciﬁc implemen-

tation of the methods (Balian, 2011; Hammond and

Wellington, 2013; Mouton, 2001; van Wyk, sa). Here

it is taken to mean the overall justiﬁcation, rationale,

or logic for undertaking the research in terms of lo-

cating it within the larger body of scientiﬁc enquiry,

explaining which, how, and why particular research

designs should be applied in the research, and decid-

ing and describing which appropriate methods will be

employed (Clough and Nutbrown, 2007; Grix, 2010;

Hofstee, 2006; Kothari, 1985). A research scheme,

on the other hand, is a structure that describes the con-

cepts that are included in a research methodology, and

covers the choice of particular approaches and meth-

ods (as opposed to methodologies) to meet the needs

of the proposed overall methodology.

A research scheme is thus made up of a philo-

sophical world view which underpins the research, a

research design which provides the structure of the

research, and research methods that are used in a de-

sign. A detailed description of all the various options

for these three elements will not be provided, and this

can found elsewhere (Biggam, 2011; Creswell, 2014;

Grix, 2010; Hammond and Wellington, 2013; Hofs-

tee, 2006; Mouton, 2001; Oates, 2006; Olivier, 1999;

Wisker, 2001). Figure 2 shows that this root class (the

ResearchScheme) is

• underpinned by a single PhilosophicalWorldview,

and

• has one or more ResearchDesigns – thus allowing

for mixed-method type studies where more than

one method is used in the research methodology.

The inverses of these two relationships indicate that

a PhilosophicalWorldview may underpin, and a Re-

searchDesigns may be used in, none or many Re-

searchSchemes. The ethicalClearance attribute en-

sures that this important part of any research scheme

is included in the model.

Note that the relationship of ResearchScheme

to PhilosophicalWorldview and ResearchDesign is

a part-whole relationship, where the philosophical

world view and research design are part of the re-

search scheme. This is a meronymic, component–

complex/integral object type relation, where the cate-

gory of the parts differfrom the category of the whole.

The two components provide the philosophical basis

and guiding design structure employed in a research

scheme.

Note also that PhilosophicalWorldviews and Re-

searchDesigns need to be consistent with each other,

and that a research scheme would not simply be un-

derpinned by any world view, and use any designs.

Further, each ResearchDesign may have one or

more ResearchMethods, and each ResearchMethod

may be used in none or many ResearchDesigns.

Again, the method, or methods, employed may be

seen as a part of the research design in a similar part-

whole relationship as that that exists between the re-

search scheme and a research design.

Each of these three main components of a research

scheme will be presented separately below.

3.3 Philosophical World View

All research is based on some (albeit sometimes

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

100

ResearchScheme

+ ethicalClearance: bool

PhilosophicalWorldview

+ statusOfTruth: TruthType

+ statusOfReality: RealityType

+ epistemology: EpistemologyType

+ roleOfResearcher: RoleType

+isUnderpinnedBy

0..*

underpins

ResearchDesign

+ researchApproach: ResearchApproachType

1..*

+hasDesign

0..*

+isUsedInScheme

+hasConsistentDesign

+hasConsistentWorldview

ResearchMethod

+ dataCollectionProcess: string

+ dataSource: DataSourceType

+ levelOfControl: LevelOfControlType

+ dataFormat: DataFormatType

1..*

+hasMethod

0..*

+isUsedForDesign

Figure 2: The high level classes that make up a research

scheme.

undeclared and implicit) philosophical world view

(Wahyuni, 2012). This is also sometimes known as

a (meta-theoretical) paradigm, tradition, perspective,

or meta-science (Creswell, 2014; Ebersohn, 2011;

Grix, 2010; Hammond and Wellington, 2013; Mou-

ton, 2001; Wisker, 2001). Essentially, though, this

provides the underlying fundamental beliefs, or basis,

for statements about the status of truth and reality in

the research endeavour, as well as the grounding for

what is to be studied, how it is to be studied, and the

nature of the knowledge found or created by the re-

search (Creswell, 2014; Hammond and Wellington,

2013; Wisker, 2001).

Common examples of philosophical world views

include post-positivism, constructivism, transforma-

tive/critical theory, interpretivism, and pragmatism.

It does need to be noted that there is also refer-

ence to qualitative and quantitative paradigms in the

literature on research methodologies (Clarke, 2005),

although these are also referred to as approaches

(Creswell, 2014).

The PhilosopicalWorldview class is related to

the ﬁve classes that represent common philosophical

world views in computer science and information

systems research via an inheritance relation (see

Figure 3). The relationship is one of subsumption

where each of the world views is in an is-a relation

with the superclass, PhilosopicalWorldview, and

could be seen as subclasses of this class (Welty,

2002).

PhilosophicalWorldview

+ statusOfTruth: TruthType

+ statusOfReality: RealityType

+ epistemology:EpistemologyType

+ roleOfResearcher: RoleType

PostPositivist

+ hypothesis: string

Constructivist

+ approach: string

TransformativeCriticalTheory

+ praxisContext: string

Pragmatist

+ application: string

Interpretivist

+ setting: string

Figure 3: Subclasses of the PhilosophicalWorldview class.

See Table 1 for possible values/instances for the

PhilosopicalWorldview class.

3.4 Research Design

A research design (or strategy, method, or method-

ology, as it is also sometimes known (Biggam,

2011; Creswell, 2014; Hofstee, 2006)) identiﬁes

and describes the overall kind of study that will

be done (Mouton, 2001) and provides a struc-

ture/framework/blueprint/plan for the whole process

(Hammond and Wellington, 2013; van Wyk, sa). Dif-

ferent research designs are best suited to answering

certain types of questions (Mouton, 2001), and the de-

cision of which design to use is usually guided by the

research questions and the kind of information that

the researcher is wanting (Hammond and Wellington,

2013; Wahyuni, 2012; Wisker, 2001). The design

should also be consistent with the philosophical world

view which underpins the research scheme.

A Conceptual Model of the Research Methodology Domain - With a Focus on Computing Fields of Study

101

Table 1: PhilosopicalWorldview data types.

Data type Possible values

TruthType Absolute

Relative/CriticalRealism

MultipleTruths

RealityType External/Independent

Objective

Rational

Subjective

EpistemologyType Empiricism

Constructivism

LogicalPositivism

Realism

RoleType ObjectiveRole

SubjectiveRole

ParticipantRole

AdvocateRole

Research designs are often divided into two

main groups, or types, of designs (Mouton, 2001;

Remenyi and Money, 2004): empirical and non-

empirical/theoretical designs. However, this is not a

universal approach, and some authors prefer to divide

designs by the basic approach: qualitative or quan-

titative (Kothari, 1985). It has been argued that it

is a mistake to group research designs in this way

as designs are not intrinsically qualitative or quan-

titative (Biggam, 2011). Furthermore, it has been

pointed out that the empirical/non-empirical division

neatly follows the types of questions that designs can

be used to answer (Mouton, 2001). Also, a partic-

ular design could use both qualitative and quantita-

tive research methods with it being rare that only

one such approach will be used (Biggam, 2011), and

so the empirical/non-empirical approach will be used

here. However, the alternative qualitative/quantitative

structure, or even a basic/applied structure, can also

be accommodated in the ontology. A observa-

tional/interventional classiﬁcation has also been pro-

posed (Sim et al., 2014).

A research scheme may use more than one re-

search design (either concurrently or sequentially) in

answering the research questions, and the study be-

comes what is known as a mixed methods one that

often integrates both qualitative and quantitative ap-

proaches (Balian, 2011; Creswell, 2014; Hofstee,

2006); this is sometimes used as a deliberate attempt

to triangulate, and so reach a deeper understanding of

the research topic (Hammond and Wellington, 2013).

Also, it is possible that a variation of a standard de-

sign can be used (Hofstee, 2006), although such adap-

tations will not be addressed in the ontology; rather,

the basic design will be covered, allowing the individ-

ual researcher to adapt as necessary, and after further

Table 2: ResearchDesign and sub-class data types.

Data type Possible values

ResearchApproachType Qualitative

Quantitative

Hybrid/MixedMethod

CaseStudyDesignType SingleCase

MultipleCase

CaseStudyFocusType CriticalCase

UniqueCase

RevelatoryCase

ExploratoryCase

DescriptiveCase

ExplanatoryCase

SurveyType Longitudinal

CrossSectional

Panel

Cohort

SampleType-Random SimpleRandom

Stratiﬁed

Cluster

Systematic

SampleType-NonRandom Convenience

Quota

Accidental

Theoretical

Purposive

SystematicMatching

Snowball

ExperimentalSettingType Laboratory

Field

ExperimentalGoalType Explore

Test

Prove

PresentationType Graphical

Mathematical

ModelTheoryType ReducedScale

Abstraction

reading about the chosen design.

Common examples of research designs in com-

puter science and information systems include case

studies, surveys, algorithm development, model or

theory building, and experiments.

Thus, the ResearchDesign class has been sub-

classed into two basic research design types: Empir-

icalResearchDesign and NonEmpiricalResearchDe-

sign – see Figure 4. These two sub-classes are again

sub-classed into common research design types ac-

cording to the two main categories.

See Table 2 for possible values/instances for the

ResearchDesign class and its sub-classes.

3.5 Research Methods

Research methods are the tools or instruments that

will be used to gather data, and a research design may

use several different research methods (Balian, 2011;

Hammond and Wellington, 2013; Hofstee, 2006).

Commonly, methods may be categorised as either

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

102

ResearchDesign

+ approach: ApproachType

EmpiricalResearchDesign

+ contextDescription: string

NonEmpiricalResearchDesign

CaseStudy

+ caseStudyDesign: CaseStudyDesignType

+ caseStudyFocus: CaseStudyFocusType

SurveyDesign

+ surveyType: SurveyType

+ surveySample: SampleType

Experimental

+ experimentalSetting: ExperimentalSettingType

+ experimentalGoal: ExperimentalGoalType

+ nullHypothesis: string

Algorithms

+ algorithmNotation: PresentationType

ModelTheoryBuilding

+ modelType: ModelTheoryType

+ modelNotation: PresentationType

Figure 4: Subclasses of the ResearchDesign class.

quantitativeor qualitative (Grix, 2010; Hammond and

Wellington, 2013). However, it has been argued that

researchers need to be aware that the methods them-

selves are atheoretical and do not have philosophical

or methodological assumptions (and can thus be used

in various research designs), and that it is why they

are chosen, and how they are applied, that is linked to

such assumptions (Grix, 2010; Wahyuni, 2012).

The collected data must be analysed so that it can

be interpreted and used as evidence in the research ar-

gument to generate research ﬁndings (Biggam, 2011;

Creswell, 2014; Hofstee, 2006), again using speciﬁc

approaches and tools. Note that the analysis will not

be dealt with in detail in this research scheme, and is

an avenue for developing the ontology further.

Common examples of research methods in com-

puter science and information systems include inter-

views, observations, measurements, argumentation,

and survey questionnaires.

The ResearchMethod class has also been sub-

classed to provide for qualitative, quantitative and

theoretical research methods - see Figure 5. This

largely follows the usual categorisation, although a

sub-class has been added for the methods that can be

used in the more theoretical research designs.

See Table 3 for possible values/instances for the

ResearchMethod class and its sub-classes.

3.6 Validating the Model

A focus group was organised (with the necessary eth-

ical clearance) to validate the conceptual model of

the research methodology domain. This was an ex-

ploratory focus group which has been recommended

for reﬁning the design of artefacts such as this model

(Hennink, 2014; Tremblay et al., 2010). The focus

group allowed the participants to spontaneously inter-

act with each other in reviewing the initial concep-

tual model that was discussed, permitting the moder-

ator to use the group synergy and dynamic to gain an

understanding of the perceptions and impressions of

the participants about the model (Stewart et al., 2007;

Krueger and Casey, 2009).

The focus group was made up of ten senior re-

searchers (six professors, three associate professors,

and a senior lecturer who teaches a module on re-

search methodologies; ﬁve from computer science

and ﬁve from information systems backgrounds) who

supervise postgraduate students in a distance educa-

tion environment. These researchers knew each other,

which allowed the group to relate to each other eas-

A Conceptual Model of the Research Methodology Domain - With a Focus on Computing Fields of Study

103

ResearchMethod

+ dataCollectionProcess: string

+ dataSource: DataSourceType

+ levelOfControl: LevelOfControlType

+ dataFormat: DataFormatType

QualitativeResearchMethod

+ analysisMethod: QualAnalysisType

QuantitativeResearchMethod

+ analysisMethod: QuanAnalysisType

TheoreticalResearchMethod

Interview

+ interviewType: InterviewType

+ interviewMode: ModeType

+recorded: bool

Observation

+ participant: bool

+recorded: bool

QualitativeQuestionnaire

+ questionnaireMode: ModeType

Measurement

+ measurementDesign: MeasurementDesignType

QuantitativeQuestionnaire

+ questionnaireMode: ModeType

Argumentation

+ ArgumentationType: ArgTypeType

+ ArgumentationNotation: ArgNotationType

Figure 5: Subclasses of the ResearchMethod class.

ily, but were all senior enough to respect the ideas of

others in the group and not to be threatened by such

views. There was also no dominant member in the

group, which could have limited its usefulness (Hen-

nink, 2014). Group participants were provided with a

copy of the model, as well as the proposed questions,

before the group met.

The focus group was moderated by the devel-

oper of the conceptual model, realising that this role

needed to balance the aspects of having someone

present the model who knew it well but at the same

time avoiding introducing any bias into its presenta-

tion (Tremblay et al., 2010). However, this approach

allowed the model developer to interact directly with

the researchers, enhancing opportunities to probe and

follow-up on comments made in the interactions.

The one factor that was problematic relates to the

expected outcome of the focus group. Generally, the

aim of a focus group is not to come to a consensus

(Hennink, 2014; Krueger and Casey, 2009; Stewart

et al., 2007). However, the point of an ontology is a

shared conception of a domain. So the focus group’s

expected outcome was to understand to what extent

the model could be accepted by the group partici-

pants, rather than to reach ﬁnal consensus, so that de-

cisions could be made concerning the model.

The focus group led to reﬁnements to the model

in the following ways:

• The use of some terminology had to be modiﬁed

to reduce misunderstandings. For example, the

research method class named Measurement was

originally named Experimental, and it was felt

that Measurement better encapsulated the task of

the method.

• A relationship indicating a consistency be-

tween PhilosphicalWorldviews and ResearchDe-

signs was added to improve the structure of the

hierarchy of concepts.

• Some speciﬁc research designs and research

methods were added, although these are not part

of the fragment that is presented here.

• Alternative categorisations (or sub-classing) of

the research designs were proposed such as

the qualitative/quantitative and basic/applied cat-

egorisations mentioned before (see 3.4).

• Some additional concept attributes were added to

classes. For example, it was felt that introduc-

ing the ethicalClearance attributein the Research-

Scheme class would highlight the importance of

this aspect of research.

• The scope of the model was reﬁned in terms of

what should be included and what is seen as ex-

ternal to the model. The conceptual model thus

focusses only on the philosophical world views,

research designs, and research methods that are

common in computing ﬁelds of research.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

104

Table 3: ResearchMethod and sub-class data types.

Data type Possible values

DataSourceType Observed

Self-reported

Archival

Physical

LevelOfControlType Low

Medium

High

DataFormatType Video

Image

Audio

Text

Numeric

Hybrid

QualAnalysisType Thematic

GroundedTheory

QuanAnalysisType DescriptiveStatistics

InferentialStatistics

NonParametricStatistics

InterviewType StructuredInterview

UnstructuredInterview

SemiStructuredInterview

ModeType FaceToFace

Telephonic

Postal

Electronic

PaperBased

MeasurementDesignType SingleGroup

Blind

Experimental/Control

ArgTypeType Inductive

Deductive

ArgNotationType MathematicalNotation

Symbolic

Textual

4 CONCLUDING REMARKS

Ontology development, and thus its conceptual mod-

elling, is an iterative process (Noy and McGuinness,

2001), and this paper is just one step in the pro-

cess of modelling the research methodology domain.

Three key concepts were identiﬁed (the philosophi-

cal world view, the research design, and the research

methods), modelled around a central research scheme

(which encapsulates the decisions being made about

a research methodology). The relationships linking

these major components were explained, and depicted

in a UML class diagram. Moreover, each key con-

cept, modelled as a UML class, was extensively ex-

panded into its subclasses, as they occur in the re-

search methodology literature. This comprehensive

model of these key concepts constitutes a novel con-

tribution to the formalisation of the research method-

ology for computingdomain towards the development

of an ontology. The development of such an ontology

in OWL forms part of future work.

While there may be little controversy surround-

ing the three key components of a research scheme,

it is at the level of the research designs and methods

that ﬁnding consensus may be more problematic, even

though the proposed model has been largely accepted

by a focus group of senior supervisors in computing

research. A workable and shared conceptual model

has to be agreed upon to allow an OWL ontology of

the domain to be developed. It does need to be made

clear, however, that this model will not (as with others

(Devedzic, 2002)) fully represent the whole research

methodology domain, but only that part that will al-

low computer science and information systems post-

graduate students to explore, and contribute to, an un-

derstanding of this domain.

The conceptual model could be expanded to in-

clude a wider variety of research designs and meth-

ods (particularly for study domains that may use other

research designs and methods from those commonly

used in computing). Further, an ontology (once built)

could be integrated into a semantic wiki to allow

for the learning of, and collaborative generation of

knowledge about, research methodologies. It could

also allow research articles to be parsed, drawing data

relevant to the ontology into a searchable form to

further assist students in learning how such research

schemes have been implemented by other researchers.

REFERENCES

Balian, E. S. (2011). The Graduate Research Guidebook:

A Practical Approach to Doctoral/Masters Research.

Silver Sky Publishing, Encinitas, CA, 4th edition.

Bera, P., Krasnoperova, A., and Wand, Y. (2010). Using

ontology languages for conceptual modeling. Journal

of Database Management, 21(1):1–28.

Berners-Lee, T., Hall, W., Hendler, J. A., O’Hara, K., Shad-

bolt, N., and Weitzner, D. J. (2006). A framework for

web science. Foundations and Trends in Web Science,

1(1):1–130.

Berners-Lee, T., Hendler, J., and Lassila, O. (2001). The

Semantic Web. Scientiﬁc American, 284(5):34–43.

Biggam, J. (2011). Succeeding With Your Master’s Disser-

tation: A Step-By-Step Handbook. Open University

Press, Maidenhead, 2nd edition.

Borgida, A. and Brachman, R. J. (2003). Conceptual mod-

eling with description logics. In Baader, F., McGuin-

ness, D. L., Nardi, D., and Patel-Schneider, P. F.,

editors, The Description Logic Handbook: Theory,

Implementation, and Applications, chapter 10, pages

359–381. Cambridge University Press, Cambridge.

Brusa, G., Caliusco, M. L., and Chiotti, O. (2006). A pro-

cess for building a domain ontology: An experience

in developing a government budgetary ontology. In

A Conceptual Model of the Research Methodology Domain - With a Focus on Computing Fields of Study

105

Proceedings of the Australasian Ontology Workshop

– Conferences in Research and Practice In Informa-

tion Technology, Volume 72, pages 7–15. Australian

Computer Society, Inc.

Burek, P. (2003). Conceptual modeling based on upper-

level ontologies and meta-ontological foundations.

Content-und Wissensmanagement, pages 117–122.

Carmichael, P. and Jordan, K. (2012). Semantic web tech-

nologies for education – time for a turn to practice?

Technology, Pedagogy and Education, 21(2):153–169.

Charlton, P., Magoulas, G., and Laurillard, D. (2012).

Enabling creative learning design through semantic

technologies. Technology, Pedagogy and Education,

21(2):231–253.

Chatti, M., Jarke, M., and Frosch-Wilke, D. (2007). The

future of e-learning: A shift to knowledge networking

and social software. International Journal of Knowl-

edge and Learning, 3(4/5):404–420.

Clarke, R. J. (2005). Research models and methodologies.

HDR Seminar Series, Faculty of Commerce, Spring

Session.

Clough, P. and Nutbrown, C. (2007). A Student’s Guide to

Methodology. Sage, London.

Cope, B. and Kalantzis, M. (2011). Creating an inter-

language of the social web. In Cope, B., Kalantzis,

M., and Magee, L., editors, Towards a Semantic Web:

Connecting Knowledge in Academic Research, pages

371–427. Oxford: Chandos.

Corcho, O., Fern´andez-L´opez, M., and G´omez-P´erez, A.

(2003). Methodologies, tools and languages for build-

ing ontologies. Where is their meeting point? Data &

Knowledge Engineering, 46(1):41–64.

Creswell, J. W. (2014). Reseach Design: Qualitative,

Quantitative, and Mixed Methods Approaches. SAGE,

Thousand Oaks, 4th edition.

Crossouard, B. (2008). Developing alternative models

of doctoral supervision with online formative assess-

ment. Studies in Continuing Education, 30(1):51–67.

Cu´ellar, M., Delgado, M., and Pegalajar, M. (2011). Im-

proving learning management through semantic web

and social networks in e-learning environments. Ex-

pert Systems with Applications, 38(4):4181–4189.

De Leenheer, P. and Christiaens, S. (2007). Mind the gap!:

Transcending the tunnel view on ontology engineer-

ing. In Proceedings of the 2nd International Confer-

ence on the Pragmatic Web, pages 75–82. ACM.

Devedzic, V. (2002). Understanding ontological engineer-

ing. Communications of the ACM, 45(4):136–144.

Di Maio, P. (2011). ’Just enough’ ontology engineering.

In Proceedings of the International Conference on

Web Intelligence, Mining and Semantics, pages 1–10.

ACM.

DiGiuseppe, N., Pouchard, L. C., and Noy, N. F. (2014).

Sweet ontology coverage for earth system sciences.

Earth Science Informatics, 7(4):249–264.

Dzbor, M., Stutt, A., Motta, E., and Collins, T. (2007).

Representations for semantic learning webs: Seman-

tic web technology in learning support. Journal of

Computer Assisted Learning, 23(1):69–82.

Ebersohn, L. (2011). Postgraduate supervision. Workshop

presented at the University of South Africa, 9 July

2011.

Engebretson, K., Smith, K., McLaughlin, D., Seibold, C.,

Terrett, G., and Ryan, E. (2008). The changing reality

of research education in Australia and implications for

supervision: A review of the literature. Teaching in

Higher Education, 13(1):1–15.

Feigenbaum, L. (2013a). The many names for the Se-

mantic Web, Semantic University, Cambridge Seman-

tics. http://www.cambridgesemantics.com/semantic-

university/many-names-for-the-semantic-web.

Feigenbaum, L. (2013b). Semantic Web miscon-

ceptions, Semantic University, Cambridge Seman-

tics. http://www.cambridgesemantics.com/semantic-

university/semantic-web-misconceptions.

Gonzalez, R. (2013). Introduction to the Seman-

tic Web, Semantic University, Cambridge Seman-

tics. http://www.cambridgesemantics.com/semantic-

university/introduction-to-the-semantic-web

Green Paper (2012). Green paper for post-school educa-

tion and training. Department of Higher Education

and Training, Republic of South Africa.

Grix, J. (2010). The Foundations of Research. Palgrave,

Houndmills, Basingstoke, 2nd edition.

Guarino, N. (2005). Ontology-driven conceptual mod-

elling. 9th East-European Conference on Advances in

Databases and Information Systems (ADBIS’2005).

Guizzardi, G., Herre, H., and Wagner, G. (2002). On the

general ontological foundations of conceptual mod-

eling. In Proceedings of 21st International Confer-

ence on Conceptual Modelling, ER 2002, LNCS 2503,

pages 65–78, Tampere, Finland.

H¨akkinen, P. and H¨am¨al¨ainen, R. (2012). Shared and per-

sonal learning spaces: Challenges for pedagogical de-

sign. Internet and Higher Education, 15:231–236.

Hammond, M. and Wellington, J. (2013). Research Meth-

ods: The Key Concepts. Routledge, Abingdon.

Hennink, M. M. (2014). Focus Group Discussions. Series

in Understanding Qualitative Research. Oxford.

Hofstee, E. (2006). Constructing a Good Dissertation.

EPE, Sandton.

Johnson, L., Adams Becker, S., Estrada, V., and Freeman,

A. (2014). NMC Horizon Report: 2014 Higher Ed-

ucation Edition. Technical report, The New Media

Consortium, Austin, Texas.

Keet, C. M. and Artale, A. (2008). Representing and reason-

ing over a taxonomy of part–whole relations. Applied

Ontology, 3(1):91–110.

Keet, C. M., Khan, M. T., and Ghidini, C. (2013). Ontol-

ogy authoring with FORZA. In Proceedings of the

22nd ACM International Conference on Information

& Knowledge Management, pages 569–578. ACM.

Kothari, C. R. (1985). Research Methodology: Methods

and Techniques. Wiley Eastern, New Delhi.

Krueger, R. E. and Casey, M. A. (2009). Focus Groups. A

Practical Guide for Applied Research. Sage, Thou-

sand Oaks, 4th edition.

KEOD 2015 - 7th International Conference on Knowledge Engineering and Ontology Development

106

Limpens, F., Gandon, F., and Buffa, M. (2008). Bridg-

ing ontologies and folksonomies to leverage knowl-

edge sharing on the social web: A brief survey. In

23rd IEEE/ACM International Conference on Auto-

mated Software Engineering-Workshops, 2008, pages

13–18. IEEE.

Motik, B., Maedche, A., and Volz, R. (2002). A con-

ceptual modeling approach for semantics-driven en-

terprise applications. In On the Move to Meaningful

Internet Systems 2002: CoopIS, DOA, and ODBASE,

LNCS 2519, pages 1082–1099.

Mouton, J. (2001). How to Succeed in your Master’s and

Doctoral Studies: A South African Guide and Re-

source Book. Van Schaik, Pretoria.

Naeve, A. (2005). The human semantic web shifting from

knowledge push to knowledge pull. International

Journal on Semantic Web and Information Systems,

1(3):1–30.

Nagyp´al, G. (2007). Ontology development: Methodolo-

gies for ontology engineering. In Studer, R., Grimm,

S., and Abecker, A., editors, Semantic Web Services:

Concepts, Technologies and Applications, chapter 4,

pages 107–134. Springer, Berlin.

Noy, N. F. and McGuinness, D. L. (2001). Ontology de-

velopment 101: A guide to creating your ﬁrst on-

tology. Technical report, Stanford Knowledge Sys-

tems Laboratory Technical Report KSL-01-05 and

Stanford Medical Informatics Technical Report SMI-

2001-0880.

Oates, B. J. (2006). Researching Information Systems and

Computing. Sage, London.

Olivier, M. S. (1999). Information Systems Research: A

Practical Guide. Published by author.

Partridge, C., Gonzalez-Perez, C., and Henderson-Sellers,

B. (2013). Are conceptual models concept models? In

Proceedings of ER 2013, Lecture Notes in Computer

Science 8217, pages 96–105.

Remenyi, D. and Money, A. (2004). Research Supervision

for Suprevisors and their Students. Academic Publish-

ing International, Reading.

Sim, I., Tu, S. W., Carini, S., Lehmann, H. P., Pollock,

B. H., Peleg, M., and Wittkowski, K. M. (2014). The

ontology of clinical research (OCRe): An informatics

foundation for the science of clinical research. Jour-

nal of Biomedical Informatics, 52:78–91.

Stewart, D. W., Shamdasani, P. N., and Rook, D. W. (2007).

Focus Groups. Sage, Thousand Oaks.

Tremblay, M. C., Hevner, A. R., and Berndt, D. J. (2010).

Focus groups for artifact reﬁnement and evaluation in

design research. Communications of the Association

for Information Systems, 26(Article 27):599–618.

van Wyk, B. (s.a.). Research design and methods, Part

I. http://www.uwc.ac.za/Students/Postgraduate/Docu-

ments/Research

and Design I.pdf.

Wahyuni, D. (2012). The research design maze: Under-

standing paradigms, cases, methods and methodolo-

gies. Journal of Applied Management Accounting Re-

search, 10(1):69–80.

Wand, Y., Storey, V. C., and Weber, R. (1999). An ontologi-

cal analysis of the relationship construct in conceptual

modeling. ACM Transactions on Database Systems,

24(4):494–528.

Weber, R. (2003). Conceptual modelling and ontology:

Possibilities and pitfalls. Journal of Database Man-

agement, 14(3):1–20.

Welty, C. (2002). Ontology-driven conceptual modeling.

Slide presentation at CAISE-02.

Wen, L., Brayshaw, M., and Gordon, N. (2012). Person-

alized content provision for virtual learning environ-

ments via the semantic web. ITALICS Innovations in

Teaching and Learning in Information and Computer

Sciences, 11(1):14–26.

Wisker, G. (2001). The Postgraduate Research Handbook.

Succeed with your MA, MPhil, EdD and PhD, volume

Palgrave Study Guides. Palgrave, Houndmills, Bas-

ingstoke.

Wisker, G., Robinson, G., and Shacham, M. (2007). Post-

graduate research success: Communities of practice

involving cohorts, guardian supervisors and online

communities. Innovations in Education and Teaching

International, 44(3):301–320.

Yew, K.-T., Ahmad, W., and Jaafar, J. (2011). A frame-

work for designing postgraduate research supervision

knowledge management systems. In National Post-

graduate Conference (NPC), pages 1–6.

Zedlitz, J. and Luttenberger, N. (2014). Conceptual mod-

elling in UML and OWL-2. International Journal on

Advances in Software, 7(1 and 2):182–196.

A Conceptual Model of the Research Methodology Domain - With a Focus on Computing Fields of Study

107