Three approaches as pillars for interpretive Information Systems research:

development research, action research and grounded theory

M. R. DE VILLIERS
University of South Africa
________________________________________________________________________________________________
This paper addresses practical approaches and models, based on the paradigm of the ‘interpretivist school’, to operationalise research in information
systems. The study overviews research paradigms and some current issues in IS research, then describes, discusses and illustrates three approaches,
namely, development research, action research, and grounded theory, advocating them as proposed pillars for interpretive IS research. With the
present emphasis on user-centricity and empowerment of previously technologically-disenfranchised domains, inquiry processes emanating from the
social sciences and humanities are relevant to IS, particularly with relation to interactive systems to bridge the digital divide and for the design and
development of emerging technology. Each of the approaches suggested has an underlying theoretical framework and reflective methods, and can
serve as a model to guide the research process, offering a unifying thread, cohesion and internal consistency to a research study.

Categories and Subject Descriptors:

H.0 [Information Systems General] Research; H.5.1 [Multimedia Information Systems] Evaluation/methodology
H.5.2 [User Interfaces] Theory and methods, user-centred design;
J.4 [Social and Behavioural Sciences]; K.6.3 [Software Management] Software development
General Terms: Theory, Human factors
Additional Key Words and Phrases: Research philosophy, Research methodology, Metaresearch, Models, Reflection, Human-computer interaction
________________________________________________________________________________________________

1. INTRODUCTION
Research design, paradigms and methods are under the spotlight in the Computing discipline, particularly in the domain
of Information Systems (IS) [Baskerville, 1999; Baskerville & Wood-Harper, 1996; Cockton, 2004; du Plooy, 2004;
Glass, Ramesh & Vessey, 2004; Myers, 2004; Pather & Remenyi, 2004; Roode, 2003; 2004; Travis, 1999; Walsham
1995a; 1995b; Wood-Harper, 1985]. With the increase of ‘interpretivist’ research, a perspective emanating from the
social sciences, this paper addresses practical approaches and models to operationalise interpretive research in IS.
Three candidate approaches with methodological rigour, are proposed, namely: development research, action research,
and grounded theory.

2. BACKGROUND: RESEARCH PARADIGMS AND METHODS

Different research paradigms and models are based on varying philosophical foundations and conceptions of reality
[Cohen, Manion & Morrison, 2000; du Poy & Gitlin, 1998; Lincoln & Guba, 1985; Olivier 2004]. Each paradigm, in
turn, is implemented by associated methodological approaches and strategies.

2.1 Positivist and interpretivist paradigms

The positivist paradigm holds that knowledge is absolute and objective and that a single objective reality exists external
to human beings. Positivism is equated with the scientific method, whereby knowledge is discovered by controlled
empirical means, such as experiments. Positivist research is intended to produce an exact representation of reality,
unbiased and value-free. Research results should be reliable and consistent, free from perceptions and biases of the
researcher. Findings should be replicable by other researchers. Positivist research relies primarily on quantitative
methods, where data comprises mainly numbers and measurements and analysis is done using statistical methods.
Results can be used for prediction. Studies are frequently hypothesis-driven. These methods originate from the natural
sciences, but are also applied in the social sciences.
Interpretivism, by contrast, aims to find new interpretations or underlying meanings and adheres to the ontological
assumption of multiples realities, which are time-and context dependent. (Terminology is not absolute: An alternative
term for interpretivist is ‘naturalistic’ [Cohen, Manion & Morrison, 2000; Lincoln & Guba, 1985]. Walsham [1995a]
refers to the ‘interpretivist school’, yet as a general adjective uses ‘interpretive approaches, -methods, -studies’, etc.
Travis [1999] and Roode [2003; 2004] prefer ‘interpretivist’.) Interpretive research emanated from the social sciences
and is also used in educational research. It is now becoming accepted in IS [Roode, 2003; Walsham, 1995a; 1995b].
Inquiry is value-related as interpretivism leads to subjective findings which may differ between researchers. It is an
appropriate view for studies of complex human behaviour and social phenomena. Just as positivism is most naturally

________________________________________________________________________________________________
Author Addresses:
M.R. de Villiers, School of Computing, University of South Africa, P O Box 392, UNISA, 0003, South Africa; [email protected].
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that the copies are
not made or distributed for profit or commercial advantage, that the copies bear this notice and the full citation on the first page. Copyrights for
components of this work owned by others than SAICSIT or the ACM must be honoured. Abstracting with credit is permitted. To copy otherwise, to
republish, to post on servers, or to redistribute to lists, requires prior specific permission and/or a fee.
© 2005 SAICSIT

Proceedings of SAICSIT 2005, Pages –

operationalised using quantitative methods (yet not exclusively), so interpretivism lends itself mainly (not exclusively)
to qualitative studies. Where positivism tests hypotheses, interpretivism investigates research questions, focused on
understanding phenomena that occur in natural settings (ethnographic) and which use verbal data. Qualitative data
collection and analysis produce findings related to intricate details where values and human experiences are relevant. In
such contexts, the ability to interpret data is important and, in fact, ‘the researcher is an instrument’ [Leedy & Ormrod,
2001:147]. Reliability in qualitative research can be considered as fit between the findings recorded and occurrences in
the natural setting. Research methods are frequently triangulated by multiple data collection methods.

2.2 Research methods: qualitative and quantitative

Mertens [1998] describes qualitative research as a naturalistic interpretive science which is multi-method in focus. It
involves the use of methods such as case studies, interviews, observation and textual analysis, which provide insights
into cultural aspects, organizational practices and human interactions.
Qualitative and quantitative methods are not mutually exclusive. Many studies require eclectic inquiry methods to
cover the terrain and provide triangulation. Another role of qualitative research is as exploratory work, setting the
foundation for quantitative research. Qualitative findings from basic research in new areas can be used to formulate
hypotheses and questions for quantitative analysis, which then yields empirical results to test, verify and extend the
qualitative hypotheses. Figure 1 shows leading research methods situated on a Positivist ─ Interpretivist axis, tending
from the quantitative to the qualitative, yet with an overlap.

Quantitative
Mathematical modeling and

Qualitative

Surveys: Questionnaires
Controlled experiments

Quasi experiments
Field experiments
Theorem proving

Documents and
artefact studies
and interviews

Focus groups

Ethnography
Case studies
Observation
simulation

Testing

Positivist Interpretivist

Figure 1 Research methods/strategies

3. FOREGROUND: INFORMATION SYSTEMS RESEARCH

With a view of paradigms and methods as background, we move on to a discussion of information systems research. IS
is a multi-perspective discipline and, as Wood-Harper [1985] asserts, it can rightly engage a plurality of research
methods. Its domain incorporates scientific, technological, engineering, organizational, managerial, psychological and
societal aspects. The subdiscipline of human-computer interaction (HCI) is currently prominent, highlighting the role
of the end user, and adding more disciplines: sociology, philosophy, physiology, anthropology, linguistics, ergonomics,
graphical design, marketing and engineering. The increasing power and stability of technology has pushed the
application of information systems into multiple domains [Pather & Remenyi, 2004], making it appropriate territory for
practicing reflective methods and, as Pather and Remenyi propose, IS research needs corresponding reorientation.
Walsham [1995b] examined IS research journals from 1992 onwards, four each from the UK and USA, and noted
the advent of interpretive studies. In an extensive analysis of computing research, based on a study of selected papers
from leading journals between 1995 and 1999, Glass, Ramesh and Vessey [2004] coded 628 papers from Computer
Science (CS) journals and 369 from Software Engineering (SE), including the ACM and IEEE publications, as well as
other well-recognised journals. They also examined 488 Information Systems papers, mainly from ‘Management’ IS
literature. This metaresearch investigated the kind of studies conducted in these three disciplines, noting similarities
and differences. For the purpose of this paper, mainly information relating to IS is extracted, but there is brief reference
to the other two, where relevant. The IS topics focused heavily on organisational concepts (66%) with systems/software
management and systems/software concepts next, both at 6-7%. Of the CS papers investigated, systems/software
concepts comprised 19% and the SE focused on systems/software (55%) and systems/software management concepts
(12%). Each category was further divided into subcategories and it was found that 24% of the IS related to
usage/operation and 19% to technology transfer, with other IS papers focusing on the problem domain, e.g. decision
support systems and group support. With regard to research approaches, 67% of the IS papers used evaluative
approaches. IS research methods were field study (27%), laboratory experiment on humans (16%), conceptual analysis
(15%), and case study (13%). Finally Glass et al report on levels of analysis, stating that most of the IS work was
related to behavioural aspects: organizational behaviour (26%), individual behaviour (26%), and group issues (11%),
whereas behavioural analysis in CS and SE occurred in only 2% and 8% of the cases respectively.
 113

Walsham [1995b] and O’Donovan and Roode [2002] attach importance to the editorial policy shift in the journal
MIS Quarterly, which has expanded its range of research by explicitly calling for papers based on interpretive or
integrated, as well as positivist, approaches.
This tendency of IS to take cognizance of human behaviour and to use evaluative approaches is notable. There is a
current gravitation, particularly in the HCI subdiscipline, towards user-centric systems, usability support for personal
computing, and the empowerment of domains beyond business and management information systems. Systems are
being developed to bridge the digital divide, to offer accessible computer technology, as well as emerging technology
and development software for the previously technologically-disenfranchised.
Much of the IS research done, is for postgraduate study. Du Plooy [2004] and Roode [2004] held a seminar relating
to non-positivist IS research methodologies, with particular reference to doctoral studies. Du Plooy [2004] notes
concerns, describing IS research as a ‘maturing science’ and a ‘fragmented adhocracy’ lacking theory and methods,
which does not fit neatly into the positivist paradigm, especially when qualitative methods are used. IS places a major
focus on the unique qualities of information itself, as occurs during enquiry into design, management decisions, and
social processes. Du Plooy suggests that academic IS research is often geared to the target audience of the dissertation
examiner, rather than the IS community, and that results seldom reach IS practitioners. His consequent axiology of
relevancy posits that interpretive research is typically weak in clear proposals on how to improve practice and that most
IS research shies away from value-laden research issues. PhD students should address practical problems, using sound
interpretive and evaluative approaches based on methods such as field studies, semi-structured and unstructured
interviews, and ethnographic data. Analysis should be conducted using, for example, grounded theory, hermeneutics, or
critical social theory, all of which contribute to the production and refinement of theoretical frameworks [du Plooy,
2004]. In line with his notable editorial [Roode, 2003:1] proposing ‘acceptance of interpretivist research on a semi-
equal footing with positivist research’, Roode [2004], referring to Hirschheim and Klein [2000], addresses the
intellectual state of IS research, with its internal and external views. The internal view of the IS research community
indicates fragmentation, particularly due to the paradigm war between interpretivists and positivists. The external view
relates to the gap between IS research and industry expectations, where current research is insufficiently relevant to
practice and links between academia and industry are inadequate. Research outputs produce a plethora of ad hoc
findings, yet lack generality and pluralism nor do they broaden theoretical constructs. Roode promotes a broadened
notion of relevancy. While the accepted scientific method – entailing hypethetico-deductive studies – is relevant, non-
positivist forms of scholarly research should not be excluded.
Regarding postgraduate research in general, Mouton [2001] points out the multidisciplinary and heterogeneous
nature of current knowledge. He further cites methodological difficulties as a major factor in the non-completion of
postgraduate studies.
In this milieu, the contribution of the present paper is to suggest three practical, methodologically- and theoretically
sound approaches for conducting interpretive research in IS. Each emanated from the social sciences and humanities,
yet is applicable to research in computing disciplines, where the design of human-computer interaction is highly
relevant in a society and economy increasingly geared to user-centric values. The approaches are applicable to
postgraduate studies, as well as to basic, ad hoc and contract forms of IS research.

4. FOCUS: INTERPRETIVE INFORMATION SYSTEMS RESEARCH – THREE APPROACHES

Against the context set by Section 3, this section focuses on theoretical frameworks to underpin IS research. A
precedent occurred in SAICSIT 2004, where Pather and Remenyi [2004] proposed critical realism as an appropriate
approach to bridge the gap between the positivist and interpretivist paradigms, using both qualitative and quantitative
techniques. This present paper suggests further approaches, highlighting models and methods that primarily implement
the ethos of the interpretivist school. When conducting IS research, an explicit theoretical framework or conceptual
model should, ideally, support the study and provide internal continuity and cohesion in the reasoning process.
We first briefly address research terminology, with its plethora of overlapping, exclusive, interrelated and, at times,
confusing terms. To clarify matters, a taxonomy of terms follows, explaining their usage within this study (other
authors may well use them differently):
─ Paradigm – the primary philosophical point of departure: in this study, the interpretivist (rather than positivist)
stance.
─ Approach – the underlying model used to operationalise the study: the approaches suggested here being
development research, action research, and grounded theory.
─ Methods – practical means/strategies/techniques/instruments used for data collection: each approach has its own
set of methods, usually multiple methods and often hybrid methods.
─ Adjectives to describe the methods: some methods are qualitative and others quantitative.

With the acknowledgement that computing has human and sociological, as well as technological and computational
dimensions, particularly in Sub-Saharan Africa with its technologically-unempowered groups, research methods from
the interpretivist paradigm have a definititive role to play. Preece et al [2002] distinguish between the usability and user
experience of software systems. The latter, in particular, lends itself to interpretive and qualitative analysis.
 This section of the paper describes, discusses and illustrates development research, action research, and grounded
theory, and their application in interpretive IS research. Each of the three has associated methods and techniques to
operationalise it. The proposals culminate in a diagrammatic ‘Pillars Proposal’ for IS Research.

4.1 Development research

4.1.1 Definition and origins
Development research (DR), also called research with a development goal, has a dual focus. It:
1. develops practical and innovative ways of solving real problems, and
2. proposes general design principles to inform future decisions.
DR is not to be confused with ‘developmental research’, which relates mainly to sustainable development in the
context of transformation and community issues in developing countries. The approach [Reeves, 2000; van den Akker,
1999; 2002], which aims to make both practical and scientific contributions, originated in educational technology
research, giving graduate students and researchers support in pursuing development goals after decades of research with
empirical goals. It is also applied in curriculum research, where some of its key concepts emerged. It is not yet a
mainstream IS research approach. Other candidate approaches for development goals have drawbacks. Postmodern
perspectives and critical theory address inherent problems (particularly injustices), but do not necessarily improve
conditions. Action research (Section 4.2) is not always explicitly geared to producing new solutions and is not always
generalisable. Development research is problem-oriented, searching for new and innovative solutions, while also
seeking findings that are transferable, practical, and socially responsible. DR acknowledges the complex and dynamic
relationship between theory and application, and aims to provide a relevant foundation to guide practice by generating
design principles and methods that are both theoretically underpinned and empirically tested.

4.1.2 Research process and methods

Development research generates different kinds of research questions. A descriptive research question examines the
nature and extent of a problem, while a design/development question investigates an intervention or new product to
address the need. A principial question inquires into generalisable principles and guidelines for the application domain.
The process commences with the analysis, design and development of an artefact or intervention as a solution for a
real-world problem. This, in and of itself, is not yet research (though many postgraduate students wish it was!). It
becomes research when the design-and-develop project is conducted from the perspective of a researcher striving to
understand the issues of the application domain and its target users, such as the required characteristics of products and
artefacts. Such research is based on iterative analysis, design, development, implementation and formative evaluation
(ADDIE − an instructional technology design model), which feeds into redevelopment. DR is closely related to
evolutionary prototyping, and entails formative research during the development process of the intervention/product to
improve its quality. Van den Akker [2002] terms the process ‘successive approximation of the ideals’. Evaluations can
be done by one or more usability evaluation methods, e.g. formal usability testing in a laboratory, logging, surveys
among end users, observation, etc.
There are various models of the DR process. The model used by Plomp and his co-researcher, van den Akker [van
den Akker, 1999; 2002; Plomp 2002] refers to outcomes of an intervention. Immediate outcomes relate to results of
using an intervention or product within the cyclic process, and distant outcomes emerge when the immediate outcomes
lead to distant outcomes in the form of generalisable principles. Reeves’ [2000] model emphasises the iterative
interaction between researchers and practitioners to clarify the problems and refine potential solutions in a process of
evolutionary prototyping. Plomp’s and Reeves’ models both influence the representation in Figure 2.

4.1.3 Epistemology of DR
Development research has a pragmatic epistemology as it acknowledges collaborative shaping by researchers and
practitioners. Van den Akker [1999], describing the knowledge acquired from DR, distinguishes between:
─ Substantive design principles, relating to the generic characteristics of suitable interventions or products.
─ Methodological aspects, with a procedural emphasis, suggesting optimal development processes.
In formative research a great deal of such knowledge is inductively extracted from the experience of using and
evaluating the prototype developed for the study. This provides a link between the two branches of the dual
development focus, namely the developing solution to a specific problem and the evolution of generalisable design
principles. The experiential evidence obtained from studying the prototype in use in its various iterations, is enhanced
when integrated with theoretical arguments.

4.1.4 Application within IS

Many IS studies involve the generation of software artefacts or web-based applications. These vary, for example, from
simple prototypes through interactive web sites with backend databases through to virtual reality simulations. Design,
implementation and testing comprise the focus area of development but are not research. The introduction of
evaluation, where evaluation entails more than mere testing of functionality, constitutes a meaningful contribution to
the body of knowledge. However, dual-focused research producing both an effective solution and generalisable
principles for the application domain, enriches the process even more and is a useful approach for a variety of studies in
domains such as e-learning, e-commerce, e-health, etc. A product – often a prototype – can be custom-built as a
solution to a problem and iteratively evaluated and refined, as described by Conradie and de Villiers [2004]. DR has
 115

also been used for computer-based support materials, and is relevant to computer science and engineering as well for
generating hardware solutions and associated generic principles.
Figure 2 is a generic model of the DR process. Its iterative phases can be effectively used to structure an IS
research process, providing continuity and cohesiveness.

implementation
Solution
D evelopm ent
( im m ediate
D esign used as
outcom e )
solution , D evelop intervention
Problem
based on (prototype )
analysis
theoretical solution
fram ework Evaluate and General design
reflection
test principles ( distant
outcom es )

Refinement of problem , solution , methods

Figure 2 Development research model (synthesized by the author, influnced by Plomp [2002] and Reeves [2000])

4.2 Action research

4.2.1 Definition and origins
The action research approach [Baskerville, 1999; Baskerville & Wood-Harper, 1996; Cohen, Manion & Morrison,
2000; Zuber-Skerrit, 1992] emanates from the behavioural sciences and encompasses a variety of research and
intervention methods. Action research (AR) originated in action-based social psychology. Its founder in the 1940s,
Kurt Lewin of the University of Michigan, contended that complex real social events could not be investigated under
laboratory conditions [du Poy & Gitlin, 1998; Wood-Harper, 1985]. AR was used independently at the Tavistock
Clinic to study post-WW2 social disorders among veterans [Baskerville, 1999]. Zuber-Skerrit [1992] defines AR as
inquiry by higher-education academics into problems of students learning. Its participative, practitioner-researcher
approach lends itself to the domain of educational research, where an evolving intervention or product is investigated
over several cycles. It is appropriate for inquiry into educational technology and for investigating the introduction of
technologies into any organizations.
Aiming to bridge the gap between research and practice, AR encompasses action outcomes and research outcomes
[Dick, Passfield & Wildman, 1995]. Commencing with the identification of a problem or situation that calls for action,
AR functions as a liberating agent of change, and is [Baskerville, 1999; Dick, Passfield & Wildman, 1995; du Poy &
Gitlin, 1998]:
─ Cyclic: as iterative steps recur in a longitudinal time frame, generating knowledge to inform further action.
─ Participative: as clients, end users and researcher collaborate in partnership as co-researchers, or as practitioner-
researchers examine their own work. Where stakeholders are full participants in the research process or where
practitioners serve both as subject and researcher, one refers to participative action research.
─ Qualitative: operating more via verbal aspects than by numbers.
─ Reflective: since critical reflection on the process and outcomes is vital to each cycle, and is used in designing
subsequent steps and events.
─ Responsive: as it reacts and adapts flexibly to the findings from each previous cycle.
In a parallel from the professional disciplines, Schön [1987] defines reflective practice or reflection-in-action as the
professional artistry that occurs when skilled practitioners tackle work-related activities, going beyond rigid rules of
inquiry, and generating new rules in situations that are uncertain and unique. Furthermore, the reflective practitioner is
both a participant in the process and a critic who observes and analyses. Similarly, AR aims to improve practice and
advance knowledge.

4.2.2 Research processes and methods

Zuber-Skerrit [1992] terms the four repetitive processes undertaken in each cycle as plan, act, observe, and reflect. The
megaprocess comprises a series of cycles that feed into each other. AR is accordingly more of an ongoing process than
an event. Du Poy & Gitlin [1998] state that action research employs or integrates methods from both the experimental
and naturalistic (interpretivist) traditions, yet is consistent with naturalistic inquiry in that all research occurs within its
natural context. Kock et al [2000], emphasizing rigour in research, contend that effective use of an iterative approach
to AR can provide rigour of a standard acceptable even to positivists.

4.2.3 Epistemology and philosophy of AR

Its ethos is interpretivist, incorporating social enquiry based on the views and interpretations of the participants, all
regarded as equals, making it an emancipatory process, while also incorporating the researcher as participant. It is a
holistic, not a reductionist, approach, which includes ethnographic enquiry and works from an ideographic standpoint,
promoting the uniqueness of each setting [Baskerville, 1999]. At the time when AR originated, emphasis was placed on
the precise collection of quantitative data and there was a shortage of qualitative research skills. In due course, it was
recognised that AR operated under a different epistemology and, although it can be less rigorous in design and
methodology than other approaches, it came into its own as a human-related research process which generates reliable
knowledge.
Action research differs from development research in that first, it operates over a longitudinal time framework of
several cycles and, second, in the in-depth involvement of researcher as participant. Third, it may well focus on
refinement of existing processes or products and not necessarily on new product development.

4.2.4 Application within IS

AR is a valid research approach for applied fields [Myers, 2004]. Baskerville [1999] asserts that the AR approach
generates highly relevant research results due to its basis in practical action, aimed at explicit problem solving while
also informing theory. In the 1980s, AR techniques were applied by Peter Checkland in systems analysis, as he
developed soft systems methodology, i.e. use in development, rather than in research [Baskerville, 1999].
IS research has been characterised by lack of relevance [Keen, 1991; Westfall, 1999; both cited by Baskerville,
1999]. In the previous decade, however, Trevor Wood-Harper had already set out to address the tensions between
theory and practice and the confusion that existed between traditional scientific research and the more sociological
approaches, by introducing AR to the IS community as a purely research methodology with his landmark paper,
Research Methods in Information Systems: Using Action Research [Wood-Harper, 1985].
AR is increasingly used for scholarly research in IS. Baskerville [1999] advocates it to inquire into the complex and
multivariate nature of IS’s social setting, using:
1. Joint goals of solving practical computing problems and expanding scientific knowledge.
2. Collaborative performance, which enhances the competencies of all participants.
3. An emphasis on action and change orientation in social settings.
4. Systematic, iterative stages.
Baskerville lists forms of IS action research: prototyping, soft systems methodology, action science, participant
observation, fieldwork, and process consultation. The present author adds investigation of evolving solutions in their
context of use, e.g. e-learning applications and customised interfaces, where the designer conducts research into his/her
own product. AR for producing e-learning solutions is also advocated by Derntl & Motschnig-Pitrik [2004].
Action research can be graphically depicted as a spiral, but the model developed for Figure 3 is a series of cycles
which close in as a solution is attained. The researcher occupies a central, participative, and influential position. This
model forms a useful framework to guide and monitor the progress of a research project.

Observe and
evaluate
Observe and
evaluate

Reflect Act Reflect

Action

Researcher

Plan

Plan response

Diagnose

Figure 3 Action research model (synthesized by the author)

4.3 Grounded theory

4.3.1 Definition and origins
Grounded theory [Cockton 2004; Glaser & Strauss, 1967; Glaser, 1992; Leedy & Ormrod, 2001; Strauss & Corbin,
1990] is an approach in which theory and models are generated inductively from the analysis of contextual data.
Grounded theory (GT) involves the discovery of concepts and hypotheses as theory emerges from data. There is no
testing or replication of a-priori theory. Like action research, it has roots in social science, but specifically in sociology,
where attitudes to conditions (initially attitudes to death) were investigated by Glaser and Strauss. Strauss and Corbin
[1990] describe its extension to anthropology, education, nursing and the economic sciences. Cockton [2004] uses it in
the design of computing interactions. It provides a conceptual grasp of substantive areas, which evolves and is
modified to fit as findings emerge and new data occurs. GT should account for variation in domain behaviour by
defining categories, properties and relationships. The data may be quantitative, qualitative or a combination, but in the
 117

case of qualitative, it is essential that it be systematically collected, analysed and coded. Defined sets of methods
should be applied so that a grounded theory emerges systematically and inductively through covariant ongoing
collection and analysis. Lincoln & Guba [1985] state that a GT is adjusted, expanded, and refined via this ongoing
process.
As with AR, there is a parallel from the social professions. The notion of emerging patterns has an analogy in
Christopher Alexander’s patterns within architecture and town planning, which form practical architectural languages,
as physical and social relationships articulate themselves [Alexander, Ishikawa & Silverstein, 1977]. Within a pattern
language, it is also possible to integrate overlapping patterns in a densifying process, which provides added meaning.
The four criteria for a well-constructed grounded theory within a substantive area are:
Fit: its categories and their properties should fit the realities being studied.
Work: in order to work, it should explain variations in behaviour.
Relevance: this is achieved when a grounded theory both fits and works.
Modifiability: the emerging theory is open to adaptation as new data and variations are integrated.

4.3.2 Research processes and methods

Covariant data collection and analysis, which are methods from the social sciences, entail initial interviews and
fieldwork, which after transcription, coding and analysis delimit the field, prompting theoretical sampling methods and
densifying [Glaser, 1992]. Other methods [Strauss and Corbin, 1990] are observation, document analysis, historical
records, and videotapes, which reflect perspectives of the subjects of study. GT integrates quantitative and qualitative
perspectives [du Poy & Gitlin, 1998]. Back-and-forth mobility occurs between data collection and analysis, with
analysis driving collection. Patterns are identified and conceptualized as the researcher codes, compares, analyses and
records. Constant comparative coding [du Poy & Gitlin, 1998; Glaser & Strauss, 1967; Glaser, 1992] is the validation
process whereby observations and behaviours are compared/contrasted with core categories and properties, then coded
into categories. The developing conceptual model is modified as new data is explored and new concepts are integrated
into the emerging theory, reviewing and expanding where necessary. When multiple behaviours indicate similar
patterns and properties, saturation has occurred. When disconfirming evidence is found, revisions are required. In this
way the emerging theory is inductively discovered, bounded and confirmed. Lincoln and Guba [1985] suggest the
generation of alternative theories. As the researcher encounters confirming cases, negative cases and discrepant cases,
the theory with the most confirming cases and the least negatives and discrepants, emerges as the most robust. GT thus
has similarities to case study research and ethnography, since both the latter aim to detect and interpret patterns within
activities and events.

4.3.3 Epistemology of GT
The researcher’s bias and subjectivity may influence conceptualization and interpretations. However, grounded theory
has built-in mechanisms to prevent this, such as constant comparison, saturation and core relevance [Glaser, 1992].
Furthermore, data collection, analysis and presentation to peers should be linked at each step, adjusting one another to
the emergent theory and preventing forcing.
The Glaser model posits, contentiously, that so as not to force or make preconceptions, there is little initial need to
review literature. Once the emerging theory is sufficiently grounded in core variables, literature reviews in the
substantive field can commence and be related to the new work. Where new fields are opened the researcher, according
to Glaser, is a ‘pre-empting pioneer’, producing a new general theory to be integrated with other literature. Scholarship
starts and expands as the emerging grounded theory develops. Glaser claims that this approach also obviates the
problem of more and more to read and less time to do!

4.3.4 Application within IS

Orliowski’s [1993] award-winning paper in an issue of the MISQ describes a project in which a grounded theory
research approach was used to study organizational experience with the adoption and use of CASE tools. Findings were
used to develop a theoretical framework conceptualizing organisational change and social issues in such cases, where
installation involves not only new technologies, but also organisational change over time. GT was a relevant approach,
because of its emphasis on contextual elements, process management, and human actions.
Cockton [2004] discusses the applicability of grounded theory to computable interactions, and explains how, as
theories and themes emerge, corresponding models can be defined and implemented using an HCI contextual approach.
Such models could include personas, scenarios and sequence models, where the persona describes a stereotypical user
and the scenario a stereotypical usage. It is a rich context-centred approach, which takes account of users’ goals and
aspirations, aiming for high relevance. The models, in turn, are used in design to generate prototypes, following which
the fit between context of use and interaction surfaces can be tested and mediated. Thus grounded theory research in IS
investigates data, resulting in theory, which leads to models, which lead to innovative grounded designs or design
models, which satisfy ‘fit to context’ [Cockton, 2004].
The GT process can be applied in studies that focus on the extraction of design guidelines by analysis of practice or
best practice in substantive areas, synthesizing them into theoretical proposals, which are further refined, tested and
ratified by use. This is being done in certain South African postgraduate studies currently in progress. The associated
generation of design principles and evaluation criteria for software applications and web site development is a notable
role for GT in information systems research and development.
Glaser’s [1992] concept of pre-emption is appropriate in the emerging Southern African technological domains
where innovative work is underway. Particular examples are the design of non-standard interactive environments, such
as development software explicitly for the formerly disadvantaged, emergent information systems, culturally-sensitive
environments, and contextualized e-learning and training. Projects are being conducted which use technology in
orientation, training, and the provision of information in low-literate communities. Research on culturally-sensitive
interfaces has been extended beyond the workplace, to address accessibility for under-educated indigenous peoples.
Figure 4 graphically illustrates the processes and concepts of GT. This model can serve as an underlying framework
for IS research processes which investigate phenomena to determine their underlying theory and to derive principles.

Systematic
Context: data in collection and
substantive area Discovery refinement Saturation

Data, incidents, Confirmed patterns

Concepts and properties
observations
Categories

constant Emergent
Properties
inductive comparative testing theory
analysis coding for fit and
Hypothesis principles

Relationships

New data
mobility between
collection and
analysis

Figure 4 Model of grounded theory emergence (synthesized by the author)

4.4 The ‘Pillars Proposal’ as a framework

Although simplified representations can hold the risk of trivialization, Figure 5 summarizes and consolidates this paper
with a synthesized structure entitled Proposed Pillars for IS research. The three research approaches advocated are
depicted as pillars on an interpretivist platform. Within each pillar, key concepts and possible applications are
concisely listed. Supported by the pillars is a pediment of interpretivist IS research. These pillars are particularly
relevant for genres of IS research focused on systems intended for end users who are not computing professionals, and
for milieus where the IS researcher is not restricted to working in an organizational environment. Examples are the
investigation of software or web applications for personal computing, and systems for the empowerment of learners or
indigenous peoples.
This diagram is not intended to be exclusive: other interpretive approaches exist, other pillars could be included, and
there is also considerable scope for positivist IS research. Nor is it the intention that the three suggested pillar
approaches should be used simultaneously; they are presented as options for the researcher.
Walsham [1995a:76, citing Eisenhardt’s (1989) discussion on theory in organizational research] stresses the role of
theory in interpretive IS research. He identifies ‘three distinct uses of theory: (i) as an initial guide to design and data
collection; (ii) as part of an iterative process of data collection and analysis; and (iii) as a final product of the research’.
From Sections 4.1, 4.2 and 4.3, it is evident that each of the pillars, DR, AR and GT respectively, meets the first
criterion in terms of offering a basic research design. With regard to the second required use of theory, Figures 2, 3 & 4
portray iterative collection, evaluation/critical reflection and analysis as key tenets of the three approaches. And,
finally, each of the three aims to generate some form of new theory or generalisable principles as a product/byproduct
of research.
 119

Pedim ent :
Interpretivist I S Research

A pplication : A pplication : A pplication :

Software development e- Learning Pre -emption
e- Commerce Synthesis of theoretical
Web applications frameworks
Design principles New technologies
Design principles
Innovative solutions Participative studies
Evaluation criteria
Prototypes User -centric systems Grounded design
Contextual HCI
Emergent technologies
Development software

K eywords : K eywords : K eywords :

Pillars :
Solutions to problems Change -oriented action Induction : emergent theory
and principles
Generalisable findings Participative researcher
Theory _ Practice Iterative cycles Systematic data collection
and analysis
ADDIE Reflection _ Response
Conceptualisation :
Feedback Natural setting categories , patterns ,
Social inquiry properties , relationships
Constant comparative
coding
Saturation

Development Action Grounded

research research theory

Platform : Interpretivist research paradigm

Figure 5 Proposed Pillars for IS research

5 CONCLUSION
Certain research issues are best suited to interpretive and context-dependent approaches. For situations where human
performance and social inquiry are relevant, this paper suggests three approaches, or models, as proposed pillars to
operationalise the interpretivist research paradigm. Development research, action research and grounded theory all
provide theoretical frameworks and methodologies to guide a research project, providing internal continuity and
consistency as the study progresses. When one of these approaches is used as a model, the activities, relationships, and
reasoning entailed in the research process should be explicitly linked to the selected approach, so that its framework
dictates the steps. Regular textual and diagrammatic reference to the model in the written document can enhance the
study, providing a unifying thread, both for the researcher and the readers.
I have used (or am in the process of using) each of the three – personally or with students – after literature
encounters had indicated its suitability for the study in hand. Enriched by reflection and experiential knowledge, I
undertook further in-depth literature surveys, which confirmed the utility of the approaches. The literature also
mentions other interpretive means such as case studies, critical theory, descriptive studies, document analysis,
ethnography, field studies, focus groups, hermeneutics, phenomenology, role play and semiotics [du Plooy, 2004;
Olivier, 2004; Travis, 1999; Walsham, 1995a; 1995b]. However, action research and grounded theory are consistently
advocated. One of the sources studied was Michael Myers’ detailed, hyperlinked overview of Qualitative Research in
Information Systems [Myers, 2004], which converges closely with the present independent suggestions. He
recommends the approaches of (i) action research, (ii) case studies, (iii) ethnography and (iv) grounded theory. The
first and the fourth confirm two of the present pillars, while the both the second (case studies) and the third
(ethnography), relating to field-based and contextual studies, can be used as methods within the pillars proposed here.
However, the development research pillar appears to be unique to this present study, and holds high utility for future
application in information systems research.
The advent of personal computing and the emergence of the HCI subdiscipline are changing the nature of
information systems and hence the forms of IS research. The findings of academic research and real-world best practice
should inform each other. The social inquiry research approaches outlined in this paper have the potential to impact on
performance in interactive computing systems. Findings should contribute to development theories and real-world
systems, should generate design principles and evaluation criteria, should encourage sound interaction practices, and
can help to support bridges over the digital divide.
Note: For some who may be unfamiliar with terminology, a few words are explained in a glossary after the references.
6 REFERENCES
ALEXANDER, C., ISHIKAWA, S. and SILVERSTEIN, M. 1977. A Pattern Language. New York: Oxford University Press.
BASKERVILLE, R.L. 1999. Investigating Information Systems with Action Research. Communications of the Association for Information Systems,
2 (Article 19) http://cais.isworld.org/articles/2-19/
BASKERVILLE, R.L. and WOOD-HARPER, A.T. 1996. A Critical Perspective on Action Research as a Method for Information Systems Research.
Journal of Information Technology 11:235-246.
CONRADIE, M.M. and DE VILLIERS, M.R. 2004. Electronic assessment of free text: a development research initiative. South African Journal of
Higher Education 18(2): 172-188.
COCKTON, G. 2004. A tutorial: Grounded Design and HCI. September 2004. Pretoria: University of South Africa.
COHEN, L., MANION, L. and MORRISON, K. 2000. Research Methods in Education. (5th ed.). London: RoutledgeFalmer.
DERNTL, M. and MOTSCHNIG-PITRIK, R. 2004. A Pattern Approach to Person-Centered e-Learning Based on Theory-Guided Action Research.
In Proceedings of the Networked Learning Conference 2004.
DICK, B., PASSFIELD, R. & WILDMAN, P. 1995. A beginner’s guide to action research. {Online].
http://www.scu.edu.au/school/sawd/arr/guide.html
DU PLOOY, N. 2004. Information Systems Research. Presentation at Doctoral Seminar: Research Methodologies in Informatics. January, 2004.
University of Pretoria.
DU POY, E. and GITLIN, L.N. 1998. Introduction to Research: Understanding and Applying Multiple Strategies (2nd ed.). St Louis: Mosby Inc.
GLASER, B.G. 1992. Basics of Grounded Theory Analysis. Mill Valley CA: Sociology Press.
GLASER, B.G. and STRAUSS, A.L. 1967. The Discovery of Grounded Theory. Chicago: Aldine.
GLASS, R.L., RAMESH, V. and VESSEY, I. 2004. An analysis of research in computing disciplines. Communications of the ACM 47(6): 89-94.
KOCK, N.F., McQUEEN, R.J. & SCOTT, J.L. 2000. Can Action Research be made more Rigorous in a Positivist Sense? The Contribution of an
Iterative Approach. Action Research E-Reports, 9. [Online] Available: http://www.fhs.usyd.edu.au/arow/arer/009.htm
HIRSCHHEIM, R. & KLEIN, H.K. 2000. Information Systems Research at the Crossroads: External versus Internal Views. In: R. Baskerville, J.
Stage and J De Gross (Eds.). Proceedings of the IFIP TC8 WG 8.2 International Working Conference on the Social and Organizational
Perspective on Research and Practice in Information Technology, Boston, Mass: Kluwer Academic Publishers.
LEEDY, P.D. and ORMROD, J.E. 2001. Practical Research: Planning and Design. Upper Saddle River, NJ: Merrill Prentice Hall.
LINCOLN, Y.S. and GUBA, E.G. 1985. Naturalistic Inquiry. Newbury Park: Sage Publications.
MERTENS, D.M. 1998. Research Methods in Education and Psychology: Integrating Diversity with Quantitative and Qualitative Approaches.
Thousand Oaks: Sage Publications.
MOUTON, J. 2001. How to succeed in your Master’s and Doctoral Studies: A South African Guide and Resource Book. Pretoria: Van Schaik
Publishers.
MYERS, M.D. 2004. Qualitative Research in Information Systems. Original version: MIS Quarterly 21(2) 1997:241,242. Updated in MISQ
Discovery modified Sept 08, 2004. [Online] Available: http://www.qual.aouckland.ac.nz/
O’DONOVAN, B. and ROODE, D. 2002. A framework for understanding the emerging discipline of informations sytems. Information Technology
& People, 15(1): 26-41.
OLIVIER, M.S. 2004. Information Technology Research: A Practical Guide for Computer Science and Informatics. (2nd ed.) Pretoria: Van Schaik
Publishers.
ORLIKOWSKI, W.J. 1993. CASE Tools as Organizational Change: Investigating Incremental and Radical Changes in Systems Development.
Management Information Systems Quarterly. 17(3).
PATHER, S. and REMENYI, D. 2004. Some of the Philosophical Issues Underpinning Realism in Information Systems: From Positivsim to Critical
Realism. In G. Marsden, Kotzé, P. & Adessina-Ojo, A. (Eds). Fulfilling the Promise of ICT - Proceedings of SAICSIT 2004. Pretoria.
PLOMP, T. 2002. Some reflections on development research (DR). Speaking notes for a breakfast meeting at the Faculty of Education, University of
Pretoria. August 2002.
PREECE, J., ROGERS, Y. and SHARP, H. 2002. Interaction Design: Beyond Human-Computer Interaction. John Wiley & Sons, Inc.
REEVES, T.C. 2000. Socially Responsible Educational Technology Research. Educational Technology, 40(6): 19-28.
ROODE, D. 2003. Information Systems Research: A Matter of Choice? South African Computer Journal, 30: 1-2.
ROODE, D. 2004. The role of Empirical versus Conceptual Approaches. Presentation at Doctoral Seminar: Research Methodologies in Informatics.
January, 2004. University of Pretoria.
SCHöN, D.A. 1987. Educating the Reflective Practitioner. San Francisco:Jossey-Bass Publishers.
STRAUSS, A.L. and CORBIN, J. 1990. Basics of Qualitative Research: Grounded Theory Procedures and Techniques. Newbury Park, CA: Sage.
TRAVIS, J. 1999. Exploring the Constructs of Evaluative Criteria for Interpretivist Research. In Proceedings of the 10th Australasian Conference
on Information Systems 1999: 1037-1049.
VAN DEN AKKER, J. 1999. Principles and Methods of Development Research. In: J. van den Akker, R.M. Branch, K.L. Gustafson, N. Nieveen &
T. Plomp (Eds). Design Approaches and Tools in Education and Training. Dordrecht: Kluwer Academic Publishers.
VAN DEN AKKER, J. 2002. The Added Value of Development Research for Educational Development in Developing Countries. In: K. Osaki, W.
Ottevanger C. Uiso & J. van den Akker (Eds). Science Education Research and Teacher Development in Tanzania. Amsterdam: Vrije
Universiteit, International Cooperation Center.
WALSHAM, G. 1995a. Interpretive case studies in IS research: nature and method. European Journal of Information Systems, 4(2): 74-81.
WALSHAM, G. 1995b. The emergence of interpretivism in IS research. Information Systems Research, 6(4): 376-394.
WOOD-HARPER, T. 1985. Research Methods in Information Systems: Using Action Research. In E. Mumford et al (eds). Research Methods in
Information Systems. North-Holland: Elsevier Science Publishers B.V.
ZUBER-SKERRIT, O. 1992. Action Research in Higher Education. London: Kogan Page.

GLOSSARY
Empirical: Based on the results of experiments and/or observations, not based on theory.
Epistemology: Theory of the grounds of knowledge, how knowledge is produced, basis of claims to knowledge.
Hermeneutics: The science of interpreting the intention of the original author or researcher.
Methodology: A set of methods used in a process of inquiry.
Ontology: The science of the essence of being; closely related to one’s view of reality.
Paradigm: The underlying philosophy and assumptions that form the foundation to one’s approach and methodology.
Substantive: Having a separate and independent existence, not merely inferential or implicit.