ASET OF RINCIPLES FOR ONDUCTING AND … been called the interpretive paradigm. Klein & Myers/Evaluating Interpretive Field Studies. and. Principles. 4. 4. Studies,

Klein & Myers/Evaluating Interpretive Field Studies

A SET OF PRINCIPLES FOR CONDUCTING AND

EVALUATING INTERPRETIVE FIELD STUDIES IN

INFORMATION SYSTEMS1

SPE C I A L IS S U E

By: Heinz K. KleinSchool of ManagementState University of New YorkBinghamton, New York [email protected]

Michael D. MyersDepartment of Management Science

and Information SystemsUniversity of AucklandPrivate Bag 92019AucklandNEW [email protected]

AbstractThis article discusses the conduct and evaluationof interpretive research in information systems.While the conventions for evaluating informationsystems case studies conducted according to thenatural science model of social science are nowwidely accepted, this is not the case for interpre-tive field studies. A set of principles for the con-duct and evaluation of interpretive field researchin information systems is proposed, along withtheir philosophical rationale. The usefulness ofthe principles is illustrated by evaluating threepublished interpretive field studies drawn fromthe IS research literature. The intention of the

paper is to further reflection and debate on theimportant subject of grounding interpretiveresearch methodology.

Keywords: IS research methodologies, interpre-tivist perspective, critical perspective, casestudy, field study, ethnography, hermeneutics

ISRL Categories: IB01, AI0802, AI0803, AI0102,AI0108, AI0112, AI0116

IntroductionIn recent years, interpretive research has emergedas an important strand in information systemsresearch (Walsham 1995b). Interpretive researchcan help IS researchers to understand humanthought and action in social and organizationalcontexts; it has the potential to produce deepinsights into information systems phenomenaincluding the management of information systemsand information systems development. As theinterest in interpretive research has increased,however, researchers, reviewers, and editors haveraised questions about how interpretive fieldresearch should be conducted and how its qualitycan be assessed. This article is our response tosome of these questions and suggests a set of prin-ciples for the conduct and evaluation of interpre-tive field research in information systems.

Purpose and MotivationThe conventions for evaluating information sys-tems case studies conducted according to the

MIS Quarterly Vol. 23 No. 1, pp. 67–94/March 1999 67

1Allen Lee was the accepting senior editor for thispaper.

Janice DeGross

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.


natural science model of social science are nowwidely accepted. One of the key contributions ofthe research methods stream in IS research hasbeen the formulation of a set of methodologicalprinciples for case studies that were consistentwith the conventions of positivism (Benbasat etal. 1987; Lee 1989; Yin 1994). As a result, casestudy research is now accepted as a validresearch strategy within the IS research commu-nity. The principles proposed in this stream ofwork have become the de facto standard againstwhich most case study research in IS is evalu-ated. However, while the criteria are useful inevaluating case study research conductedaccording to the natural science model of socialscience, the positivist criteria suggested are inap-propriate for interpretive research. Several con-ferences (Cash and Lawrence 1989; Lee et al.1997; Mumford et al. 1985; Nissen et al 1991),workshops (Baskerville et al. 1994; Davies et al.1993; Lee at al. 1992; Orlikowski et al. 1991),journal contributions (Harvey and Myers 1995;Lee 1991; Walsham 1995a, 1995b), and a mono-graph (Walsham 1993) have addressed them-selves to qualitative research and shown how thenature and purpose of interpretive research dif-fers from positivist research (see also Orlikowskiand Baroudi 1991). This paper takes this workfurther, and a set of principles are proposed forthe conduct and evaluation of interpretive fieldresearch in IS.

This paper can be seen as a response to the call“to discuss explicitly the criteria for judging qual-itative, case and interpretive research in informa-tion systems” (Lee et al. 1995, p. 367). Therefore,just as principles and guidelines for case studieswere provided by analyzing them from the philo-sophical perspective of positivism (Lee 1989), sothis paper will do the same for interpretive fieldresearch, but from the philosophical perspectiveof hermeneutics. Also, just as suggestions weremade for researchers who wished to undertakeresearch employing the case research strategyand offered “criteria for the evaluation of casestudy research” (Benbasat et al. 1987, p. 369), sothis paper does the same, except that we focuson interpretive field research.

The set of principles proposed in this paper isderived primarily from anthropology, phenome-nology, and hermeneutics. We readily acknowl-

edge that there are many other forms of interpre-tivism that are not necessarily hermeneutic (suchas postmodernism or deconstructionism). Wecaution, therefore, that the particular set of prin-ciples suggested here applies mostly to the con-duct and evaluation of interpretive research of ahermeneutic nature. We leave open the possibil-ity that other authors may suggest additional setsof principles; indeed, other voices representingthe many different forms of interpretivism arewelcome and needed.

Some readers may feel that, in proposing a set ofprinciples for conducting and evaluating inter-pretive field studies, we are going too farbecause we are violating the emergent nature ofinterpretive research, while others may think justthe opposite. In this debate, we have adopted amiddle position. While we agree that interpre-tive research does not subscribe to the idea thata pre-determined set of criteria can be appliedin a mechanistic way, it does not follow thatthere are no standards at all by which interpre-tive research can be judged. Our claim is simplythat we believe our proposed principles are con-sistent with a considerable part of the philo-sophical base of literature on interpretivism andhence an improvement over the status quo. Webelieve that it is better to have some principlesthan none at all, since the absence of any crite-ria increases the risk that interpretive work willcontinue to be judged inappropriately.Ultimately the quality (and status) of interpretiveresearch within IS will benefit from a livelydebate about its standards. Therefore anotherimportant purpose in publishing these principlesalong with their philosophical rationale is to fur-ther debate on the important subject of ground-ing interpretive research methodology. Wediscuss our use of the word “principles” in moredetail below.

This paper has two audiences. First of all, itshould be of interest to all those who are directlyinvolved with interpretive research, i.e.,researchers, reviewers, and editors conducting,evaluating, or justifying interpretive research ininformation systems. Hopefully the principlesoffered in this paper will contribute to the raisingof the aspirations for individual research projects.Second, many readers, while not doing interpre-tive research themselves, have become aware of

68 MIS Quarterly Vol. 23 No. 1/March 1999


its importance and wish to better understand itsmethodological foundations and potential. Ourpaper provides one possible and concise answerto some of their questions.

The Nature of Interpretive ResearchIt is important to explicitly define what we meanby interpretive research. This is especially sogiven that often no clear distinction is madebetween “qualitative” and “interpretive” re-search. However, the word interpretive as usedhere is not a synonym for qualitative—qualitativeresearch may or may not be interpretive, depend-ing upon the underlying philosophical assump-tions of the researcher (Myers 1997). Forexample, if one follows Chua’s (1986) classifica-tion of research epistemologies into positivist,interpretive, and critical, qualitative research canbe done with a positivist, interpretive, or criticalstance. This implies that case study research canbe positivist (Yin 1994), interpretive (Walsham1993), or critical, just as action research can bepositivist (Clark 1972), interpretive (Elden andChisholm 1993) or critical (Carr and Kemmis1986). A brief description of the three researchphilosophies is given below, although weacknowledge that the distinctions themselves areoften contentious, and some research papersmay be more difficult to classify than others(Walsham 1995b).

Generally speaking, IS research can be classifiedas positivist if there is evidence of formal propo-sitions, quantifiable measures of variables,hypothesis testing, and the drawing of inferencesabout a phenomenon from a representative sam-ple to a stated population (Orlikowski andBaroudi 1991). Examples of a positivist approachto qualitative research include Yin’s (1994) andBenbasat et al.’s (1987) work on case studyresearch.

IS research can be classified as critical if the maintask is seen as being one of social critique,whereby the restrictive and alienating conditionsof the status quo are brought to light. Criticalresearch seeks to be emancipatory in that it aimsto help eliminate the causes of unwarrantedalienation and domination and thereby enhancethe opportunities for realizing human potential(Alvesson and Wilmott 1992b; Hirschheim and

Klein 1994). To make this possible, critical theo-rists assume that people can consciously act tochange their social and economic conditions.They do, however, recognize that human abilityto improve their conditions is constrained by var-ious forms of social, cultural, and political dom-ination as well as natural laws and resourcelimitations. Examples of a critical approach toqualitative research include Forester’s (1992) andNgwenyama and Lee’s (1997) work.

IS research can be classified as interpretive if it isassumed that our knowledge of reality is gainedonly through social constructions such a lan-guage, consciousness, shared meanings, docu-ments, tools, and other artifacts. Interpretiveresearch does not predefine dependent and inde-pendent variables, but focuses on the complexityof human sense making as the situation emerges(Kaplan and Maxwell 1994); it attempts to under-stand phenomena through the meanings thatpeople assign to them (Boland 1985, 1991;Deetz 1996; Orlikowski and Baroudi 1991).Interpretive methods of research in IS are “aimedat producing an understanding of the context ofthe information system, and the process wherebythe information system influences and is influ-enced by the context” (Walsham 1993, pp. 4-5).Examples of an interpretive approach to qualita-tive research include Boland’s (1991) andWalsham’s (1993) work.

The Scope and Approach of This PaperKeeping the above definition of interpretiveresearch in mind, the scope of this paper is lim-ited to addressing the quality standards of onlyone type of interpretive research, namely, theinterpretive field study. Field studies include in-depth case studies (Walsham 1993) and ethno-graphies (Suchman 1987; Wynn 1979, 1991;Zuboff 1988). Although there is no hard and fastdistinction between the two, their principle dif-ferences are the length of time that the investiga-tor is required to spend in the field and the extentto which the researcher immerses himself or her-self in the life of the social group under study. AsYin (1994, pp. 10-11) describes, “Ethnographiesusually require long periods of time in the ‘field’and emphasize detailed, observational evidence. . . . In contrast, case studies are a form ofenquiry that does not depend solely on ethno-

MIS Quarterly Vol. 23 No. 1/March 1999 69


graphic or participant-observer data.” This paperis therefore concerned with addressing the eval-uation criteria for both case studies and ethno-graphies, as long as the underlying philosophy isinterpretive.

The approach used in this paper was to analyzethe literature on the philosophy of interpretivismfor insights that could ground a methodology forthe conduct and evaluation of interpretive fieldstudies. We have summarized the key results ofour analysis of this philosophical base in sevenprinciples, which are discussed in the next sec-tion. We decided to concentrate on thehermeneutic philosophers, especially Gadamerand Ricoeur, for a few reasons. First, the com-plete literature of interpretive philosophy com-prises so many varied philosophical positionsthat it is unlikely to yield one consistent set ofprinciples for doing interpretive research. Whathas been called the interpretive paradigm(Burrell and Morgan 1979) is in fact a family ofparadigms too large and diverse to tackle for thisproject. However, since hermeneutics is a majorbranch of interpretive philosophy with Gadamerand Ricoeur arguably being its most well knownexponents, it made sense to focus primarily onthem—as well as pay attention to phenomenol-ogy as suggested by Boland (1985). Second,hermeneutics has a relatively settled philosophi-cal base and therefore lends itself to being usedas a “bridgehead” for making a contribution tointerpretive research methodology. Third, bothauthors are well versed in interpretive philoso-phy and hermeneutics in particular. One of theauthors was trained as an anthropologist and hasused hermeneutics (and especially criticalhermeneutics) in his research work for over 20years; the other has acquired a varied philosoph-ical background in positivism, critical social the-ory, and hermeneutics. It was therefore natural tofocus on a topic that was centred on the sharedcore of knowledge among the co-authors andallowed them to capitalize on their rather differ-ent formal education and life experiences.

Given the above, it should be clear that this is notan interpretive paper as such (let alone an inter-pretive field study). Rather, it is a conceptualpaper drawing on work in anthropology andhermeneutics. Our paper discusses the nature ofinterpretive research and the philosophical ratio-

nale for how such research, which takes ahermeneutic stance, can be evaluated. Conse-quently, not all of the principles suggested belowcan apply to this paper by definition. For exam-ple, it is quite obvious that there are no “living”participants in this project.

Organization of This PaperThe paper is organized as follows. The next sec-tion suggests a set of principles for the evaluationof interpretive field research. The following sec-tion applies these principles to three publishedexamples of interpretive field research to demon-strate their usefulness for authors and evaluators(reviewers and editors). The final section con-cludes that, while not all of the principles mayapply in every situation, their systematic consid-eration is likely to improve the quality of futureinterpretive field research in information systems(especially that of a hermeneutic nature). Webelieve there is a need to stimulate further debateconcerning whether and how the quality of inter-pretive research can be assessed.

A Set of Principles forInterpretive Field ResearchIn this section we propose a set of principles forthe evaluation of interpretive field research ininformation systems. There are two sources forthese principles: the practice of anthropologicalresearch and our understanding of the underlyingphilosophy of phenomenology and hermeneu-tics. As mentioned earlier, we wish to emphasizethat the particular set of principles suggested hereapplies mostly to the conduct and evaluation ofinterpretive research of a hermeneutic nature.Interpretive research takes many different forms,not all of which are hermeneutic in orientation.We hope that our proposal will encourage othersto advance additional sets of principles for alter-native forms of interpretivism.

Our choice of the word “principles” requiressome explanation. By the use of this word, wewant to emphasize that the general principles wepropose are offered in the spirit of being funda-mental ideas that may be helpful to authors andreviewers. The ideas are “fundamental” because



they are derived from certain philosophical writ-ings (as referenced below) that are consideredclassical contributions to interpretivism; theprinciples should be helpful because they sum-marize important insights in interpretivism whichare (as yet) not embedded in the practice of inter-pretive field research. However, the principlesare not like bureaucratic rules of conduct,because the application of one or more of themstill requires considerable creative thought. Thisfollows in part from the idiographic nature ofinterpretive field studies and in part from theirphilosophical grounding. Our use of the word“principles,” therefore, guards against the ideathat their use is mandatory; rather, it is incumbentupon authors, reviewers, and editors to exercisetheir judgement and discretion in decidingwhether, how, and which of the principlesshould be applied and appropriated in any givenresearch project. However, this does not meanthat we advocate arbitrarily selecting some prin-ciples while ignoring others, since at the end ofthis section we shall argue that the principles are,to some extent at least, interdependent.

Table 1 summarizes the seven principles that wehave identified together with illustrative examplesfrom the IS research literature. These seven princi-ples are discussed in more detail below. We beginwith theassumption that theprinciples for theeval-uation of interpretive research are not independentof the guidelines for its conduct. Therefore, ourapproach is to derive a set of principles for the con-duct and reporting of interpretive research andthen illustrate that the same principles can also beused for post hoc evaluation.

The Fundamental Principle of theHermeneutic CircleThe most fundamental principle of hermeneuticsis that of the hermeneutic circle. This principle isfoundational to all interpretive work of ahermeneutic nature and is in effect a meta-prin-ciple upon which the following six principlesexpand. The idea of the hermeneutic circle sug-gests that we come to understand a complexwhole from preconceptions about the meaningsof its parts and their interrelationships. To clarifyits generality it is best to relate to Gadamer’s(1976b) example of how we are to translate themeaning of a sentence into a foreign language.

As a case in question, consider the sentence“they are playing football.” In order to under-stand the individual parts of the sentence (i.e.,whether football is a round ball, an egg-shapedball, or no ball at all), we must attempt to under-stand the meaning of the sentence as a whole.The process of interpretation moves from a pre-cursory understanding of the parts to the wholeand from a global understanding of the wholecontext back to an improved understanding ofeach part, i.e., the meanings of the words. Thesentence as a whole in turn is a part of somelarger context. If from this context it is clear thatnobody is engaged in sport at all, then we canconclude that the meaning of “they are playingfootball” must be metaphorical. To apply themetaphor, one needs to interpret “football” as anissue which is contested which in turn involves anew understanding of the meaning of the term“playing” as involving something abstract whichis being “thrown or kicked around.” Also, “play-ing” no longer means physical movement on agrassy field.

Thus the movement of understanding is con-stantly from the whole to the part and backto the whole. Our task is to extend in con-centric circles the unity of the understoodmeaning. The harmony of all the details withthe whole is the criterion of correct under-standing. The failure to achieve this har-mony means that understanding has failed(Gadamer 1976b, p. 117).

In Gadamer’s description of the “hermeneutic cir-cle,” the terms“parts” and“whole” shouldbegivenabroadand liberal interpretation.Theycanbepartsofahistorical story,and then thewhole is theproperperspective of the historical context. This interpre-tation should be used to apply the principle of con-textualization discussed below. Alternatively, theparts can be the interpretive researchers’ and theparticipants’ preliminary understandings (i.e., pre-understandings) in the study. The whole consists ofthe shared meanings that emerge from the interac-tions between them. Note that participants appro-priate (i.e., make their own) ideas from theresearcher and vice versa. Hence, in a number ofiterations of the hermeneutic circle, a complexwhole of shared meanings emerges. This interpre-tation should be used in applying the principle ofinteraction between the researchers and subjects.




Table 1. Summary of Principles for Interpretive Field Research

1. The Fundamental Principle of the Hermeneutic CircleThis principle suggests that all human understanding is achieved by iterating betweenconsidering the interdependent meaning of parts and the whole that they form. This principleof human understanding is fundamental to all the other principles.

Example: Lee’s (1994) study of information richness in e-mail communications. It iteratesbetween the separate message fragments of individual e-mail participants as parts and theglobal context that determines the full meanings of the separate messages to interpret themessage exchange as a whole.

2. The Principle of ContextualizationRequires critical reflection of the social and historical background of the research setting, sothat the intended audience can see how the current situation under investigation emerged.

Example: After discussing the historical forces that led to Fiat establishing a new assemblyplant, Ciborra et al. (1996) show how old Fordist production concepts still had a significantinfluence despite radical changes in work organization and operations.

3. The Principle of Interaction Between the Researchers and the SubjectsRequires critical reflection on how the research materials (or “data”) were socially constructedthrough the interaction between the researchers and participants.

Example: Trauth (1997) explains how her understanding improved as she became self-conscious and started to question her own assumptions.

4. The Principle of Abstraction and GeneralizationRequires relating the idiographic details revealed by the data interpretation through theapplication of principles one and two to theoretical, general concepts that describe the natureof human understanding and social action.

Example: Monteiro and Hanseth’s (1996) findings are discussed in relation to Latour’s actor-network theory.

5. The Principle of Dialogical ReasoningRequires sensitivity to possible contradictions between the theoretical preconceptions guidingthe research design and actual findings (“the story which the data tell”) with subsequentcycles of revision.

Example: Lee (1991) describes how Nardulli (1978) came to revise his preconceptions of therole of case load pressure as a central concept in the study of criminal courts several times.

6. The Principle of Multiple InterpretationsRequires sensitivity to possible differences in interpretations among the participants as aretypically expressed in multiple narratives or stories of the same sequence of events understudy. Similar to multiple witness accounts even if all tell it as they saw it.

Example: Levine and Rossmore’s (1993) account of the conflicting expectations for theThreshold system in the Bremerton Inc. case.

7. The Principle of SuspicionRequires sensitivity to possible “biases” and systematic “distortions” in the narrativescollected from the participants.

Example: Forester (1992) looks at the facetious figures of speech used by city planning staffto negotiate the problem of data acquisition.


During repeated cycles of the hermeneutic circle,all of the suggested principles can be applied itera-tively, forming a complex web of interpretations.

In the IS literature, an example of how thehermeneutic circle can help broaden the under-standing of IT is given by Lee’s (1994) study ofinformation richness in conventional e-mailcommunications. Lee’s study iterates betweenthe separate message fragments of individual e-mail participants as parts and the global contextthat determines the full meanings of the separatemessages to interpret the message exchange as awhole. In so far as conventional e-mail is limitedto the exchange of textual information, our pre-understanding is that “electronic mail is a leanmedium that does not readily support the level ofcommunication richness associated with, forinstance, a face-to-face meeting” (Lee 1994, p.143). However, if this partial understanding isrelated to the larger whole of the literature on theconceptual and empirical weaknesses in infor-mation richness theory, contradictions arise,because these studies include empirical findingsin which e-mail supports a level of richness com-parable to some “rich” media (for detailed refer-ences see Lee 1994). If we merely concentrate onthe contradiction between information richnesstheory and the empirical findings, this also fitsthe pattern of the principle of dialogical reason-ing. However, this becomes the starting point fora second hermeneutic circle focusing on theinterpretive processes that impart meaning to e-mail communications. By guiding the collectionof detailed records with concepts from thehermeneutic theory of Ricoeur (1981), Lee canshow that a complex world of social construc-tions may be evoked through e-mail communi-cations in a way that is not unlike what happensthrough face-to-face meetings with subsequentphone calls, memos, etc.

The Principle of ContextualizationThe principle of contextualization is based onGadamer’s insight that there is an inevitable dif-ference in understanding between the interpreterand the author of a text that is created by the his-torical distance between them. The hermeneutictask consists, not in covering up the tensionbetween the text and the present, but in con-sciously bringing it out (Gadamer 1976b, p. 133).

In interpretive research in IS, therefore, one of thekey tasks becomes one of seeking meaning incontext. Various contexts can be explored (e.g.,those associated with the informants or theresearchers), the choice largely depending uponthe audience and the story the author wants totell. For some research purposes it may be fruit-ful to extend Gadamer’s viewpoint with a clarifi-cation of the power bases which over time haveprivileged certain beliefs as legitimate truths. Thiskind of analysis can be based on Foucault’s(1972) archaeology of knowledge.

The contextualization principle requires that thesubject matter be set in its social and historicalcontext so that the intended audience can seehow the current situation under investigationemerged. The spirit in which this is done differsfrom a positivist account of history. Positivistresearchers also study the way the organizationhas been in the past, but then presume that pat-terns observed in the past will repeat themselvesin the future. Positivist researchers tend to ignorethe fact that people think and act, that people areactive makers of their physical and social reality(Orlikowski and Baroudi 1991). Positivist studies

are premised on the existence of a priorifixed relationships within phenomena . . . .Such studies serve primarily to test theory, inan attempt to increase predictive under-standing of phenomena (Orlikowski andBaroudi 1991, p. 5).

In distinction to this, interpretive researchersinsist that any observable organizational patternsare constantly changing because, as Parmenidesobserved, “you cannot swim in the same rivertwice.” Interpretivists argue that organizationsare not static and that the relationships betweenpeople, organizations, and technology are notfixed but constantly changing. As a conse-quence, interpretive research seeks to understanda moving target. In so far as each instance istreated as a unique historical occurrence, inter-pretive research is idiographic. From this it hassometimes mistakenly been concluded that inter-pretive research cannot generalize—this claimwill be taken up shortly under the principle ofabstraction. When the researcher does the fieldresearch, the results of his or her work are influ-enced by the total history of the organization andthe research itself becomes a part of the organi-



zation’s future history. The principle of contextu-alization requires that this be explicitly reflected.The researcher needs to see people as the pro-ducers and not just as products of history and thedescription of the historical context shouldreflect this in the write-up of the research study.

An IS example illustrating the principle of con-textualization is Ciborra et al.’s (1996) study ofFiat’s new assembly plant where the Punto, 1995“European Car of the Year,” is manufactured. Theauthors discuss the historical forces that led toFiat establishing the new assembly plant as an“integrated factory.” At the beginning of the1980s, Fiat

embarked resolutely into the so-called high-tech factory concept, which privilegedintense automation to yield high productiv-ity and quality, while reducing the role ofhuman work drastically (Ciborra et al. 1996,p. 405).

The results, however, were disappointing, withmanagement learning that quality cannot be theoutcome of sophisticated technology only, withlittle involvement of the workforce. The newassembly plant represented the first example ofthe new production concept adopted by Fiat inthe 1990s: the integrated factory. Despite radi-cal changes in work organization and opera-tions in the new factory, however, the authorsshow how Fiat’s earlier history of applying oldFordist production concepts still had a signifi-cant influence.

The Principle of Interaction Between theResearcher(s) and the SubjectsWhereas the principle of contextualizationplaces the object of study in context, this princi-ple requires the researcher to place himself orherself and the subjects into a historical perspec-tive. In social research, the “data” are not just sit-ting there waiting to be gathered, like rocks onthe seashore. Rather, interpretivism suggests thatthe facts are produced as part and parcel of thesocial interaction of the researchers with theparticipants.

Within the discipline of anthropology, this prin-ciple has become more widely recognized inrecent years. For example, Kahn argues for a

reflexive anthropology where it is recognizedthat the interpretation of culture(s) “is in fact partof a process of construction” and says thatanthropologists themselves “are similarly part ofa broader socio-historical process” (Kahn 1989,p. 22).

To put the argument in another way, whenI, or any other anthropologist, produces intext an account of another culture, what Iam in fact doing is engaging in a processwith a history. That history is the culturalproduct of a longstanding relationbetween “us” and “them” within which Iand my “informants” are embedded. Atthe same time, the knowledge which I/weproduce out of that relation—in my casefor example the knowledge I might chooseto term “Minangkabau culture,” or “themeaning of mosque symbolism in a WestSumatran village,” or whatever—is newknowledge in the sense that it does notpre-exist in West Sumatra, hardwired as itwere in the brains of the Minangkabaufrom time immemorial. It exists, and canonly exist, in the relationship betweenMinangkabau and the West (Kahn 1989, p.16).

It follows from this that interpretive researchersmust recognize that the participants, just as muchas the researcher, can be seen as interpreters andanalysts. Participants are interpreters as they altertheir horizons by the appropriation of conceptsused by IS researchers, consultants, vendors, andother parties interacting with them, and they areanalysts in so far as their actions are altered bytheir changed horizons. This effect is lessened ifthe researcher is not interacting with the partici-pants, i.e., relies solely on historical secondarydata or a concealed one-way window. Even inthis case, however, the researcher’s preconcep-tions about the participants still affect the con-struction, documentation, and organization ofthe material.

Read’s (1965) account of his fieldwork in Papua,New Guinea was one of the first anthropologicalstudies to fully reflect this principle. Read suc-ceeds in giving the reader an extraordinarilyvivid insight into the life and character of thepeople. But the story is told “as it appearedthrough my own eyes, filtered through my own



background, my likes and dislikes, qualified bymy own strengths and weaknesses” (Read 1965,p. 247).

In information systems, an example of this prin-ciple is Trauth’s (1997) discussion of the lessonsshe learned over the course of conducting threefield studies. The three research projects wereconcerned with the education of IS profession-als, the societal influences and impact of theemerging information economy in Ireland, andthe influence of societal factors on the diffusionof information technology in the Netherlands.She explains how her understanding of all threesituations improved as she became self-con-scious and started to question her own assump-tions. While making an effort to “embrace thecontexts” in which she was conducting herresearch, she also attempted to “bring into con-sciousness the emotional and intellectual reac-tions to experiences and observations” (Trauth1997, p. 241).

The Principle of Abstraction andGeneralizationThe two previous principles emphasize thosefeatures unique to the particular situation understudy. Whereas it is true that interpretive researchvalues the documentation of unique circum-stances and is highly suspicious of any claim thathuman affairs are governed by natural laws thatare culturally independent, this is not the wholestory. One outcome of the extensive debates inphilosophy is that there is a philosophical basisfor abstraction and generalization in interpretivefield studies.

Interpretive philosophers such as Heidegger(1962) and Husserl (1970, 1982) attempted toarticulate the essence (the most basic character-istics) of the human condition in terms of a num-ber of elementary categories. Examples of suchcategories which have been used in informationsystems research are “embodiment,” or “break-down” (Madsen 1989). What is important here isthe recognition that these types of concepts wereextracted from common, everyday experiencessuch as hammering, or misunderstandings ineveryday language (breakdowns). Therefore,intrinsic to interpretive research is the attempt torelate particulars as may be described under the

principle of contextualization to very abstractcategories; unique instances can be related toideas and concepts that apply to multiple situa-tions.2 This does not mean that it is appropriate totest theory in any simple or direct manner (Deetz1996), as is suggested for positivist case studyresearch (Benbasat et al. 1987). Nevertheless, itis important that theoretical abstractions andgeneralizations should be carefully related to thefield study details as they were experiencedand/or collected by the researcher. This is soreaders can follow how the researcher arrived athis or her theoretical insights.

The principle of abstraction with its philosophi-cal backing in the works of Heidegger andHusserl supports Walsham’s argument that thevalidity of the inferences drawn from one ormore cases does not depend on the representa-tiveness of cases in a statistical sense, “but on theplausibility and cogency of the logical reasoningused in describing the results from the cases, andin drawing conclusions from them” (Walsham1993, p. 15; see also Lee 1989). Walsham arguesthat there are four types of generalizations frominterpretive case studies: the development ofconcepts, the generation of theory, the drawingof specific implications, and the contribution ofrich insight (Walsham 1995a). The key point hereis that theory plays a crucial role in interpretiveresearch, and clearly distinguishes it from justanecdotes. However, theory is used in a differentway than is common in positivist research; inter-pretive researchers are not so interested in “falsi-fying” theories as in using theory more as a“sensitizing device” to view the world in a cer-tain way. Interpretive researchers in informationsystems tend not to generalize to philosophicallyabstract categories but to social theories such asstructuration theory or actor network theory.

An example of this principle in information sys-tems is Monteiro and Hanseth’s (1996) article,which focuses on general theoretical relation-ships between technological and organizationalissues. Drawing mostly on Latour’s actor-net-work theory, they discuss the development of an


2Not all anthropologists agree with this position, advo-cating cultural relativism instead. However, even cul-tural relativists use some abstract general concepts,e.g., the concepts of culture and enculturation.


information infrastructure for the Norwegianhealthcare sector. The project involved definingthe Norwegian standard for drug prescriptionexchange, the main task being the definition ofan EDI message representing a prescription. Theauthors show how the identifiers for prescriptiondrugs were defined by the pharmacies, withoutconsidering the needs of GPs. What is usuallyconsidered a mere “technical detail” (the defin-ing of identifiers) in fact represented the politicalinterest of the pharmacies.

In this article, the idiographic details revealed bythe data interpretation were thus related to theo-retical, general concepts drawn from actor-net-work theory. According to this theory, humansand non-humans are linked together into actor-networks. The theory assumes that actors pursueinterests, and that these interests can becomeinscribed in technical or social arrangements. InMonteiro and Hanseth’s case, their findings areused to show how organizational behavior isinscribed into “technical” details of standardsand how adoption and diffusion of a standardinvolves making it irreversible.

The Principle of Dialogical ReasoningThis principle requires the researcher to confronthis or her preconceptions (prejudices) thatguided the original research design (i.e., the orig-inal lenses) with the data that emerge through theresearch process. The most fundamental point isthat the researcher should make the historicalintellectual basis of the research (i.e., its funda-mental philosophical assumptions) as transparentas possible to the reader and himself or herself.As a minimum, the researcher should identifywhat type of interpretivism s/he prefers, identifyits philosophical roots, and relate the particularstrengths and weaknesses of the preferred philo-sophical direction to the purpose of the work.

The intellectual basis of the research design pro-vides the lenses through which field data are con-strued, documented, and organized. It could bethat the research findings do not support thesepreconceptions. Therefore, they may have to bemodified or abandoned altogether. This process isone instance of the hermeneutic rule that “preju-dice,” prejudgement, or prior knowledge plays animportant part in our understanding. In positivist

social science, “prejudice” or prejudgement isseen as a source of bias and therefore a hindranceto true knowledge; objectivity, according to posi-tivism, is best attained if a social scientist adopts avalue-free position and does not let biases inter-fere with his or her analysis. By contrast,hermeneutics recognizes that prejudice is thenecessary starting point of our understanding. Thecritical task of hermeneutics then becomes one ofdistinguishing between “true prejudices, bywhich we understand, from the false ones bywhich we misunderstand” (Gadamer 1976b, p.124). Of course, the suspension of our prejudicesis necessary if we are to begin to understand a textor text-analogue. But as Gadamer points out, thisdoes not mean that we simply set aside our preju-dices. Rather, it means that we, as researchers,must become aware of our own historicity(Gadamer 1976b, p. 125).3

The above rule can be applied several times insequence so that the improved understanding ofone stage becomes the prejudice for the next. Anexample of this is the six-stage process summa-rized by Lee (1991, p. 356). Lee describes howNardulli (1978) came to revise his preconcep-tions of the role of caseload pressure as a centralconcept in the study of criminal courts severaltimes. Beginning with the “legal man” notion thatthe court system is a due process system whereimpartial judges seek to serve justice based ontruth finding through adversarial debate, severalrevisions of this basic notion are confronted withconventional empirical and participant observa-tional records. Eventually this rationalistic “legalman model” of the court has to be dismissed. Thecourt system functions more like a clubhouse inwhich the collective efforts of the court roomelite (judges, prosecutors, and defense lawyers)are guided by the common interest to processcases expeditiously (Lee 1991, p. 359).

Interestingly enough, in this particular examplesome of the preconceptions came from positivistresearch. It shows that hermeneutic theory can


3This awareness of the dialogue between the text and theinterpreter is peculiar to contemporary hermeneutics.The earlier hermeneutic philosophers such as Diltheyignored this dialogical relationship between the textand the interpreter and attempted to understand theobjective meaning of a text in its own right.


embrace meanings that originate from both posi-tivist and interpretivist types of research.

The Principle of Multiple InterpretationsWhereas it is possible to apply the previous prin-ciples to texts from only one source, this wouldignore that human actions are conditioned by asocial context involving multiple agents. Theprinciple of multiple interpretations requires theresearcher to examine the influences that thesocial context has upon the actions under studyby seeking out and documenting multiple view-points along with the reasons for them. The analy-sis of reasons may include seeking to understandconflicts related to power, economics, or values.Moreover the researcher should confront the con-tradictions potentially inherent in the multipleviewpoints with each other, and revise his or herunderstanding accordingly. This follows fromRicoeur’s (1974) work, The Conflict ofInterpretations. These revisions are similar to theapplication of the hermeneutic rule referred to inthe principle of dialogical reasoning, except thatit is not a confrontation of the researcher’s pre-conceptions and the data, but a confrontation ofconflicting interpretations of the participants inthe field. In either case, revisions of theresearcher’s preconceptions may be the outcome.

Examples of the fruitfulness of this principle in ISare most common in ethnographic studies of sys-tem requirements formulation. One example isthe requirements for the Threshold system in thestudy of Bremerton Inc. (BI), an integrated finan-cial services company (Levine and Rossmore1993). Threshold was an IT initiative that shouldhave made BI more competitive, by not onlyreplacing its bread and butter core applications,but also by expanding into new products andmarket services to place BI at a competitiveadvantage. However, as the planning forThreshold moved into the second year, conflict-ing interpretations among the participantsbecame apparent. There were disagreements asto whether Threshold was supposed to be “a sys-tems project with major business implications”or a “business project with major systems impli-cations” (Levine and Rossmore 1993, p. 64). Athird group argued that “work on Threshold waspremature since the overall strategic vision forBremerton was still evolving” (p. 65). The ethno-

graphic study documented the factors thatallowed this situation to arise.

Of course, it is not mandatory that such conflict-ing interpretations be present. This is exemplifiedby the ethnographically informed systems designfor air traffic control as reported by Bentley et al.(1992). The ethnographic record documented indetail the consensual teamwork character ofcooperation in flight control among various pro-fessional specializations (one chief controller,two radar controllers, two assistants with militaryliaison officers) via flight strips. However, wewould argue that the possibility of conflictinginterpretations among the participants them-selves of what an air control system should bedoing was dismissed too easily in this study, asthe reader is left wondering why no such con-flicting interpretations exist in this case. Even ifeventually none are found, the principle of mul-tiple interpretations is of heuristic value becauseit leads to probing beneath the surface. Suchprobing is strengthened further with the principleof suspicion discussed next.

The Principle of SuspicionEven though the above principles alreadyencourage various forms of critical thinking, onthe whole they are more concerned with theinterpretation of meanings than with the discov-ery of “false preconceptions.” We therefore adaptthe principle of suspicion from Ricoeur (1976). Inwhat he describes as a “hermeneutics of suspi-cion,” Ricoeur argues that it is possible in certaincircumstances to see consciousness as “false”consciousness (Ricoeur 1976, pp. 194-203).Ricoeur illustrates the operation of the principleof suspicion with examples of critical analysisfrom Marx and Freud. The idea is to reveal theeffects of socially created distortions (Marx) andpsychopathological delusions (Freud). Fromm(1955) and Adorno et al. (1950) provide manyexamples for the operation of false consciousnessand its recognition.

In subsequent literature, Critical Social Theoryhas pursued this idea more vigorously than inter-pretivism. As Deetz describes,

Either explicit or implicit in critical work is agoal to demonstrate and critique forms ofdomination, asymmetry, and distorted



communication through showing howsocial constructions of reality can favor cer-tain interests and alternative constructionscan be obscured and misrecognized (Deetz1996, p. 202).

It is therefore not surprising that examples of inter-pretive field studies implementing the principle ofsuspicion tend to be influenced by the writings ofcritical theorists, especially Habermas andFoucault (Alvesson and Wilmott 1992a, 1992b;Forester 1992; Lyytinen 1992; Mingers 1981;Myers and Young 1997; Ngwenyama 1991).

A possible method to reveal distortions in con-versations is presented by Forester (1992) in hisbrief study of an interaction between city plan-ning staff. After brainstorming a long list of infor-mation needed to create a developmental planfor a small city, one of the planners raises thetricky question of how this information is to beobtained without having the resources to pay forit. By carefully analyzing the figures of speechemployed (like “that’s a minor detail”), it isshown how the five people involved negotiatedthe difficult situation by making obviously falseclaims and facetiously consenting to them. Whatis at stake here is not the truth or untruth of theclaims, but the world of social relations betweenthe planning staff and the other departments ofcity government. By questioning the surfacemeaning of what is being said in a systematicway, Forester can

explore a four-layered practical structure ofsocial and political interactions shaping(more or less true) beliefs, (more or lessappropriate) consent, (more or less deserved)trust, and (more or less aptly focused) atten-tion. In so doing, we can identify subtle, yetpowerful pragmatic moves of social actorswho seek ends instrumentally and yet con-tinually reproduce social and political rela-tions too (Forester 1992, pp. 61–62).

This kind of approach clearly goes beyondunderstanding the meaning of the data because itpoints the researcher to “read” the social worldbehind the words of the actors, a social worldthat is characterized by power structures, vestedinterests, and limited resources to meet the goalsof various actors who construct and enact thissocial world.

The application of the principle of suspicionappears to be one of the least developed in the ISresearch literature. However, since there is con-siderable disagreement among interpretiveresearchers concerning the extent to whichsocial research can (or should be) critical (Deetz1996), we leave open the possibility that someinterpretive researchers may choose not to followthis principle in their work.

The Interdependence of the SevenPrinciples

We began this section with the suggestion thatthe principle of the hermeneutic circle is theoverarching principle upon which the other sixprinciples expand. Having considered each ofthe principles separately, we now wish to comeback full circle and consider the interdepen-dence of the principles and the “whole” that theycreate; this whole needs to guide the researcher’sjudgement with regard to the application of eachprinciple individually.

For instance, a researcher’s deciding on what rel-evant context(s) should be explored (principletwo) depends upon the following: how theresearcher “creates data” in interaction with thesubjects (principle three); the theory or conceptsto which the researcher will be abstracting andgeneralizing (principle four); the researcher’sown intellectual history (principle five); the dif-ferent versions of “the story” the researchunearths (principle six); and the aspects of the“reality presented” that he or she questions criti-cally (principle seven). It is obviously infeasibleand distracting to describe every aspect of thecontext. The researcher has to choose what tosay depending upon the audience and the storythat he or she wants to tell. Clearly, the whole(the final published story) affects the parts (howeach individual principle is applied), and theapplication of each part affects the whole.

This means that, while we believe that none of ourprinciples should be left out arbitrarily, researchersneed to work out themselves how (and which of)the principles apply in any particular situation. Onthe other hand, we do not believe that the princi-ples can be used a la carte, since the importanceand relevance of each principle is derived in partfrom the manner in which the others are applied to



the collection and interpretation of the field mate-rials. We argue that qualified readers can judge ifany of the principles have been left out arbitrarilyby finding “holes” in the researcher’s story. Thereason for this is that neglect of the interdepen-dence among the applicable principles would ren-der the account less cogent and trustworthy. Thebasic rationale for this claim seems to have beenintuitively recognized by Prasad (1997), when shesuggests that an ethnography is ultimately judgedon its ability to offer convincing explanations ofaction in a particular culture or subculture—i.e., ithas to be “plausible.”

Plausible accounts refer to ethnographicwritings that are convincing not onlybecause they pay attention to detail, butbecause the overall narrative incorporatesthe viewpoints of multiple actors and tiesthese together in a culturally coherent andarticulate fashion . . . . Many features cancontribute to the plausibility of the researchnarrative including the development of astrong story line, evidence of theresearcher’s involvement in the field, a senseof historical context and a coherent weavingof disparate events within the field (Prasad1997, p. 108).

We would like to think that our seven principlesare a more detailed and useful description of the“features” to which Prasad refers that make aresearch story plausible and convincing to its tar-get audience. We therefore conclude this sectionby reiterating the point that the whole storyresulting from the application of the individualprinciples is greater than the sum of the parts,i.e., the separate application of each principle.Interpretive researchers need to write an accountthat is not only interesting, but also plausible andconvincing. Our suggested set of principles isdesigned to help interpretive researchers improvethe plausibility and cogency of their accounts.

Three Examples of InterpretiveField Research in InformationSystemsThree published examples of interpretive fieldresearch from the IS research literature are evalu-

ated in this section. The articles are evaluated inthe light of our proposed set of principles forinterpretive field research in order to demon-strate how authors, reviewers, and editors canapply the principles. The three examples are asfollows: Orlikowski’s (1991) ethnographic studyof the changes in forms of control and forms oforganizing in a large, multinational software con-sulting firm; Walsham and Waema’s (1994)analysis of the IS strategy formation and imple-mentation process in a medium-sized UK build-ing society; and Myers’ (1994) examination of thefailed implementation of a centralized payrollsystem by the New Zealand EducationDepartment. Whereas the ethnographic researchmethod was used in Orlikowski’s study, the casestudy method was used in the other two studies.The three articles are summarized in Table 2.

We selected these three particular papersbecause we consider them to be very goodexamples of interpretive field research; theyexemplify most of our suggested principles ofinterpretive field research, even though someprinciples are better exemplified than others.Also, they were among the first interpretivepapers to be published in information systemsjournals, which meant that they were available tous for analysis at the early stages of our project.

The Fundamental Principle of theHermeneutic CircleThe first principle, the hermeneutic circle, is fun-damental in the sense that it should guide theapplication of the other six to the interpretationof the field study material. Each of the followingsix principles will guide the researcher to revealspecific aspects of the case, none of which, whentaken by themselves, are necessarily complete.Hence, each principle may help the researcher todiscover or better understand a significant part ofthe case that contributes to an understanding ofthe field study as a whole. At the same time, asthe understanding of the parts becomes clearer,they themselves help to codetermine the mean-ing of the whole (Gadamer 1976b, p. 117). Thisprocess is not unlike putting the pieces of a puz-zle together, except that the pieces are not allgiven but have to be partially fashioned andadjusted to each other.



In all three articles, the principle of thehermeneutic circle is implied, but only Myersgives it brief, explicit recognition (Myers 1994,pp. 188–9).5 We suggest this could be becausethe application of the first principle depends onthe application of the other six. Given that theother principles were only applied to a varyingdegree, it is perhaps to be expected that this firstprinciple cannot be clearly identified in thesearticles. Therefore, the value of this principle liesmore in guiding future work than in evaluatingwork already published.

Having said this, however, we can suggest howthe principle of the hermeneutic circle could beapplied in future. A researcher could use theprinciple to move back and forth between thedifferent interpretations of the field study mater-ial to which he or she was led in the application

of the remaining six principles. This activitywould be guided by the goal to fill in any gaps orunresolved contradictions in the field studymaterial that may persist, and would continueuntil the “pieces of the puzzle” fit together (atleast in the researcher’s mind).

The Principle of ContextualizationAll three articles set the subject matter in its his-torical, political, and economic context,although the manner and degree in which this isdone varies. One of the articles (Myers) placesthe contextualization toward the end of theanalysis section, which is somewhat unusual.This is done as a rhetorical device to explain tothe reader certain surprising aspects of the casepresented earlier. For example, one of the sur-prising aspects is the government’s decision “toscrap the system despite the fact that it was work-ing correctly” (Myers 1994, p. 194). Myers’description of the sequence of events that led tothe government’s decision provides the explana-tion of this puzzle. He shows how all decisions


Table 2. Summary of Three Interpretive Field Studies in Information Systems 4

1. Orlikowski (1991) 2. Walsham and Waema (1994) 3. Myers (1994)

Research Ethnography Case study Case studyMethod

Research A large, multinational A medium-sized UK A New ZealandSite software consulting firm building society government

department

Theoretical Implications of IT for IS strategy formation and Implementation of Focus forms of control and implementation process information systems

forms of organizing

Key Findings Instead of facilitating a The IS strategy formation Most theories of IS more flexible organiza- and implementation implementation aretion, IT enabled existing process is a dynamic one, too narrow and forms of control to be involving time-varying mechanistic; IS intensified and fused relationships, multilevel implementation can

contexts, and cultural only be understoodand political aspects as part of the

broader social andorganizationalcontext

4The abstract of each of the three papers is included inthe Appendix.

5Walsham discusses the way in which his research wasinformed by hermeneutics elsewhere (Walsham 1993).


regarding the centralized payroll system makesense in the light of political changes surround-ing government strategy to restructure educa-tional administration in New Zealand.

The other two papers place the context earlier inthe presentation of the case, which is more com-mon. Of the two, Walsham and Waema’s paperis the most comprehensive, providing a discus-sion of the multi-level analysis of the Britishfinancial services sector and an overview of theevents and actions over the whole period in theUK building society from 1981 to 1989.Orlikowski’s realization of the principle of con-textualization is the weakest of the three papers.She discusses the increasing formalization andstandardization of the software firm’s disciplinarypractice since the early 1960s, along with thefirm’s current political structure and overallobjectives. The reader’s understanding couldbenefit, however, by a better description of thesocial and historical context that would explainwhy the management of the consulting companycommitted to CASE tools “to streamline as muchof the software development process as possible”(Orlikowski 1991, p.14). It is not clear to thereader why the company chose to limit its rangeof options this way.

All three articles contribute to an overall under-standing of the historical forces affecting theorganizations under study. At the same time, theysee the subjects of their research as the produc-ers and not just as products of history. The sub-jects are portrayed as actors and not simply aspassive respondents to a situation over whichthey have no control.

For instance, Walsham and Waema discuss thedifferences in management style between twochief executives. One of the chief executives (Mr.Brown) is described as “the key actor attemptingto manage strategic change over this period” (p.163). Through specific quotes from Brown andthose working with him, the authors are able todocument Brown’s conviction that he can makea difference through his commitment andactions. They give numerous examples for this,one of which involved Brown manipulating theorganization to apply prototyping in the imple-mentation of an information system. In Brown’sown words, “So what I did in Sky . . . was tointroduce a half-baked system which was then

modified by a lot of feedback (from branchstaff)—trying it on and polishing it until it fits thebusiness” (p. 164). The article explores in somedepth this chief executive’s background and hisrelationship with senior management and otherinterest groups. The authors are therefore able toexplore the social and historical context withinwhich Brown and the organization as a wholeoperated, while at the same time showing thatBrown was an actor who himself significantlyinfluenced the organization.

In a similar way, Orlikowski describes the soft-ware consultants as “agents.” She says that theexercise of control is never one-sided; on occa-sion SCC consultants “are able to refuse to con-form to the tools and the culture, and do thingsthe way they see fit” (p. 32). While the consul-tants are thus subject to institutional controls(and significantly influenced by the tools, the cul-ture, and the organizational context withinwhich they work), they are nevertheless por-trayed as actors and not simply as passiverespondents to a situation over which they haveno control.

The Principle of Interaction Between theResearcher and the SubjectsIn all three articles, this principle is followed toa much lesser extent than any of the other prin-ciples. The authors tend to focus on the organi-zations as if their account is the only possibleone and gloss over the social interactionbetween the participants and the researcherthrough which the data were constructed andrecorded. For example, Walsham and Waemaacknowledge that their particular perspective onorganizational culture developed at a fairly latestage, after analyzing participants’ perspectives,and after seeing the importance attached tosymbolic elements. To some extent, this indi-cates how the researchers’ interpretationsemerged as their understanding of the caseimproved. A critical reviewer, however, couldpoint out that if the principle of interactionbetween the researcher and subjects had beenapplied more forcefully, they could have takenthis further by more fully describing the ways inwhich data collection and interpretation activi-ties affected each other.



With regard to this principle, Myers’ article is theweakest of the three, since only a brief summaryof the research methodology is given (Myers1994, p. 186). Orlikowski improves upon this byhaving a slightly more extensive discussion ofresearch methods and the ways in which datawere collected. However, how the data were cre-ated through the interaction between theresearcher and participants is not revealed(Orlikowski 1991, p. 14).

If the authors of the three papers had acknowl-edged this principle in conceiving and conduct-ing their studies, they might also have givenmore attention to the effects that their researchproduces on the participants. Clearly, informalcontacts, interviews, requests for specific docu-ments, and conversations will affect how the sub-jects view their own affairs and how they presentthat to the researcher. This in turn will have aneffect on the kind of data that the researcherobtains. In none of the papers are these kinds ofeffects of the researcher on the participants andvice versa acknowledged or analyzed.

The Principle of Abstraction andGeneralizationAll three articles generalize their findings to the-oretical constructions of interest to the widerinformation systems research community.Orlikowski draws on Giddens’ structuration the-ory as a vehicle to derive abstractions and gener-alizations from her findings. She focuses on theforms of control adopted in organizations andprovides an understanding of the possible controland organizing consequences of deploying infor-mation technology in work processes. Walshamand Waema generalize their findings using con-cepts from contextualism, a type of organiza-tional change framework proposed by Pettigrew(1987) and his colleagues. The Pettigrew frame-work focuses the researcher on the concepts ofcontent, context, and process (the interconnect-edness of actions and events over time). Thisframework is used by the authors to developsome general implications about the process of ISstrategy formation and implementation.

An additional dimension of abstraction is pro-vided by Myers in that his case description isused for two purposes: to reflect on some general

principles of research methodology (i.e., criticalhermeneutics) and to generalize to key conceptsin IS implementation theory such as IS successand power. The broader social and historicalcontext is introduced as a cause for implementa-tion failure (p. 197) and discussed in the contextof different IS implementation models. Myerssuggests that the lack of progress in IS imple-mentation research may be due to

an underlying mechanistic view of the rela-tionship between information technologyand organizational change. That systemsdevelopment is conducted within a com-plex, intertwined set of social and politicalinteractions is generally ignored (p. 188).

Using the critical hermeneutics of Gadamer andRicoeur, the case analysis then proceeds to showin some detail how concepts such as IS “success”and “user satisfaction” are matters of interpreta-tion. The idiographic details revealed by the datainterpretation are thus related to theoretical, gen-eral concepts; Myers uses the case to generalizeto concepts from IS implementation theory.

The Principle of Dialogical ReasoningAs explained earlier, this principle requires theresearcher to confront his or her preconceptions(prejudices) which guided the original researchdesign (i.e., the original lenses) with the data thatemerge through the research process. The mostfundamental point is that the researcher shouldmake the historical intellectual basis of theresearch as transparent as possible to the reader.In all three papers, the intellectual basis of theresearch is made clear. Orlikowski acknowledgesthat her research was informed by assumptionsabout organizations and information technologywhich draw on Giddens’ theory of structuration.Contrary to the theoretical position of much of theIS research literature, which assumes that infor-mation technology will transform existing bureau-cratic organizational forms and social relations,Orlikowski shows how the use of new informa-tion technology led to the existing forms of controlbeing intensified and fused. Walsham andWaema propose a content, context, and processframework derived from a number of social sci-ence and IS researchers, especially Pettigrew.Myers proposes the critical hermeneutics of



Gadamer and Ricoeur as an integrative frame-work for researching IS implementation.

However, the dialogical aspect of the research isnot discussed explicitly in any of the three arti-cles to be little awareness of the researchers’ ownhistoricity; we are presented with a complete,polished theoretical framework with little infor-mation provided regarding the way in whichthese theoretical perspectives emerged. Why thismay have been the case is discussed later.

The Principle of Multiple Interpretations

Although Orlikowski’s (1991) article mostlyemphasizes the uniformity of the software firm’smethodology and culture, alternative viewpointsare presented, including the views of those whooccasionally undermine the workings of the con-trol mechanisms. For example, Orlikowski saysthat some consultants deliberately bypassed theCASE tools they were supposed to use if theybelieved that the technology was slowing theirwork. Walsham and Waema (1994) also offeralternative perspectives within the organizationunder study. Although the focus of their article ison senior management and in particular the dif-ferences in management style between two chiefexecutives, they do document the views of otherinterest groups.

Of all three articles, Myers (1994) has the mostextensive discussion and analysis of the view-points of the various stakeholders. He showshow the centralized payroll project was charac-terized by conflicting interpretations among theparticipants about what happened, who was toblame, and how successful the project was. Forexample, while the system received very badpublicity in the media and was seen as a failureby teachers,

the Director of Management Services pro-claimed some 4 months after implementationthat the system was successful in that it wasnow on target to meet its main financial goalof saving the government millions of dollars ininterest payments (Myers 1994, p. 196).

Despite this apparent “success,” Myers explainsthe continued opposition of the teachers’ unionto the payroll system and the government’s deci-sion to scrap it just a few months later.

The Principle of SuspicionAll of the authors adopt a critical perspective anddo not take their informants’ views at face value.The articles provide informative examples ofhow this principle can be implemented in differ-ent ways. In Orlikowski’s case, the firm isdescribed as a “quintessential ‘knowledge-based’firm,” where all employees are referred to as“professionals.” It is our expectation that “profes-sional practitioners apply significant discretion”over their own work. However, Orlikowskishows how the firm exercised a number of sys-tematic and personal forms of control, and howIT enabled the existing forms of control to beintensified and fused. She says that the “objectiv-ity,” “integrity,” and “independence” expected ofprofessionals “would appear to be underminedby the short-term focus on profits evident in SCCand its emphasis on standards, routines, method-ologies, and tools that shape the reality they pur-port to serve” (p. 38).

Walsham and Waema’s critique emerges mostlythough the juxtaposition of alternative view-points of the participants, and by adopting a cul-tural and political perspective on the IS strategyand implementation process. For example, theynote that the dominance of one chief executiveofficer over other members of the senior man-agement group “was a conscious political act onhis part, at least if we accept his own testimonyafter the event” (p. 164). He believed thisapproach was necessary to get things done.Rather than merely accept the words of the CEO,however, Walsham and Waema exhibit suspicionby considering the views of other staff. “Lesscharitable” interpretations were put forward bythem, with some staff claiming that he “only tookactions likely to produce improved financialresults in the relatively short term, but not neces-sarily best for Sky in the long term” (p. 165).

In Myers’ case, a critical perspective is taken byexamining the statements and actions of the var-ious stakeholders in terms of the political andeconomic interests of the actors. For example,the payroll system itself is seen as “one of themeans by which the government intended torestructure educational administration in NewZealand” (p. 197). The opposition of the teach-ers’ union and others to the system is explained(at least in part) by noting that “the system



became in effect a symbol of the government’sresolve to push through sweeping changes toeducation administration” (p. 197). Many ofthem opposed this government initiative.

The Interdependence of the SevenPrinciplesSince we have argued that all seven principlesdepend upon each other and form an interde-pendent whole, at least to some extent, we nowneed to return full circle to discuss each paper asa whole in relation to the principles as a whole.From our overall assessment of the three papersin Table 3, we can see that all three articles areremarkably similar with respect to the way inwhich they exemplify our suggested principles.Four of the principles are clearly illustrated,while one of the principles (that of thehermeneutic circle) is explicitly recognized inone paper and implied in the others.

Our overall assessment, therefore, is that all threepapers can be considered exemplars of interpre-tive field research in information systems. Theyare plausible and convincing, and clearly illus-trate the principles of contextualization, ofabstraction and generalization, of multiple inter-pretations, and of suspicion.

It is striking, however, that all three papers areespecially weak in relation to the third and fifthprinciples of interpretive research (of interactionbetween the researcher(s) and subjects, and ofdialogical reasoning). We are given little under-standing of how the researchers’ analysis devel-oped over the course of the project. As it stands,we are presented with a finished piece of inter-pretive research with few indications given of itsemergent nature.

A possible explanation for this is that all threeauthors may have been conforming to normativepressures for writing up their research as if theresearcher was an unobtrusive, objectiveresearcher. However, by (consciously or uncon-sciously) conforming to positivist criteria for validresearch, the quality of all three papers wasaffected from an interpretive perspective. Thishighlights the importance of explicit recognitionby the IS research community of the differencesbetween the evaluation criteria for positivist andinterpretive research. Our proposed principles

for interpretive research should help in thisregard. If some consensus emerges that our pro-posed principles are useful, interpretiveresearchers will no longer feel the need to justifytheir work by the use of inappropriate criteria.

With regard to the interdependence among thoseprinciples that are actually applied in the threepapers, we were surprised to discover the pivotalinfluence exerted by principle four. In the case ofMyers (1994), the principal purpose is to demon-strate the usefulness of critical hermeneutics as atheoretical frame of reference for implementationresearch. Critical hermeneutics provides thepaper’s constructs for abstraction and generaliza-tion. This particular perspective guides theresearcher to investigate the roles of differentstakeholders and their multiple interpretationswith the distortions in their viewpoints resultingfrom social conflicts. It also encourages theresearcher’s “suspicion” not to take anything atface value. It is, therefore, not surprising thatMyers focuses on the wider social and historicalcontext to help explain the multiple viewpointsand conflicting social interests that he saw as dri-ving the sequence of events that occurred.

Since we have access to the field notes and rec-ollections of the original researcher, we can con-firm that in the Myers’ (1994) article, principlefour had an overwhelming influence in givingpriority to principles six and seven, which then inturn drove the application of principle two (con-textualization). This suggests that if the theoreti-cal ideas used for generalization and abstractionemphasize a critical viewpoint, then principle sixis naturally connected to principle seven (if thereare multiple interpretations, none can be taken atface value). The connection between principlessix and seven provides the key point for Myers’conclusions. In trying to plausibly explain thereasons why the implementation of payroll sys-tem failed, he writes:

On the surface there are many reasons whythe implementation of this system was sofraught with problems. While some of themcan be easily discounted (e.g., a journalistsuggesting that the problem was a “com-puter jinx”), many of the reasons suggestedby the participants are perfectly legitimate.However, none of these reasons on theirown really explain what happened; the




Tabl

e 3.

Ove

rall

Ass

essm

ent o

f the

Thr

ee In

terp

retiv

e F

ield

Stu

dies

1.O

rlik

owsk

i (19

91)

2.W

alsh

am a

nd W

aem

a (1

994)

3.M

yers

(19

94)

1.Th

e Fu

ndam

enta

l Pri

ncip

leIm

plie

d, b

ut n

o ex

plic

itIm

plie

d, b

ut n

o ex

plic

itA

pp

lied

and

brie

f rec

ogni

tion

of t

he H

erm

eneu

tic

Cir

cle

reco

gni

tion

giv

en to

itre

cog

nitio

n g

iven

to it

giv

en to

it

2.Th

e Pr

inci

ple

ofD

iscu

sses

the

firm

’s c

urre

ntP

rovi

des

a m

ulti-

leve

l ana

lysi

sU

ses

the

pol

itica

l con

text

toC

onte

xtua

lizat

ion

pol

itica

l str

uctu

re a

nd o

vera

llof

the

Brit

ish

finan

cial

ser

vice

sex

pla

in s

ome

surp

risin

g a

spec

tsob

ject

ives

sect

or a

nd a

n ov

ervi

ew o

f key

of th

e ca

sehi

stor

ical

eve

nts

3.Th

e Pr

inci

ple

of I

nter

acti

onIg

nore

s th

e so

cial

inte

ract

ion

Igno

res

the

soci

al in

tera

ctio

nIg

nore

s th

e so

cial

inte

ract

ion

betw

een

the

Res

earc

hers

and

bet

wee

n th

e p

artic

ipan

tsb

etw

een

the

par

ticip

ants

bet

wee

n th

e p

artic

ipan

tsth

e Su

bjec

tsan

d th

e re

sear

cher

and

the

rese

arch

eran

d th

e re

sear

cher

4.Th

e Pr

inci

ple

ofU

ses

Gid

den

s’ s

truc

tura

tion

Use

s P

ettig

rew

’s c

onte

nt,

Use

s th

e cr

itica

l her

men

eutic

sA

bstr

acti

on a

ndth

eory

and

focu

ses

on fo

rms

cont

ext,

and

pro

cess

of G

adam

er a

nd R

icoe

ur a

ndG

ener

aliz

atio

nof

con

trol i

n or

gan

izat

ions

fram

ewor

k an

d fo

cuse

s on

focu

ses

on IS

imp

lem

enta

tion

IS s

trat

egy

form

atio

n an

dim

ple

men

tatio

n

5.Th

e Pr

inci

ple

ofTh

e in

telle

ctua

l bas

is o

f the

The

inte

llect

ual b

asis

of t

heTh

e in

telle

ctua

l bas

is o

f the

Dia

logi

cal R

easo

ning

rese

arch

is m

ade

clea

r, b

utre

sear

ch is

mad

e cl

ear,

but

rese

arch

is m

ade

clea

r, b

utth

e d

ialo

gic

al a

spec

t is

not

the

dia

log

ical

asp

ect i

s no

tth

e d

ialo

gic

al a

spec

t is

not

dis

cuss

edd

iscu

ssed

dis

cuss

ed

6. T

he P

rinc

iple

of

Alte

rnat

ive

view

poi

nts

are

Focu

ses

mos

tly o

n se

nior

Has

an

exte

nsiv

e d

iscu

ssio

nM

ulti

ple

Inte

rpre

tati

ons

pre

sent

ed in

clud

ing

the

view

sm

anag

emen

t, b

ut th

e vi

ews

of th

e vi

ewp

oint

s of

of th

ose

who

occ

asio

nally

un-

of o

ther

inte

rest

gro

ups

are

vario

us s

take

hold

ers

der

min

e th

e co

ntro

l pro

cess

esal

so p

rese

nted

7.Th

e Pr

inci

ple

ofC

ritic

izes

the

shor

t-te

rm fo

cus

Juxt

apos

es a

ltern

ativ

e vi

ew-

Exa

min

es th

e vi

ews

and

Susp

icio

non

pro

fits

at S

CC

and

dem

on-

poi

nts

and

sug

ges

ts th

at o

neac

tions

of v

ario

us s

take

hold

ers

stra

tes

the

way

in w

hich

IT h

asC

EO

faile

d to

take

a lo

nger

in te

rms

of th

eir

pol

itica

l and

enab

led

exi

stin

g fo

rms

of c

on-

term

vie

wec

onom

ic in

tere

sts

trol t

o b

e in

tens

ified

and

fuse

d


scrapping of this system by the governmentonly makes sense if we come to understandthe broader social context within which theimplementation of this system took place(Myers 1994, p. 197).

The principle of suspicion expressed in this pas-sage (“on the surface,” “none of these reasons ontheir own”) clearly motivates the selection ofitems to be included in the characterization ofthe “broader social and historical context.”Hence principle seven also had a strong influ-ence. The key interdependencies among theprinciples exemplified by Myers (1994) are sum-marized in Figure 1.

In the Myers (1994) example, it is relatively easyfor us to reconstruct the interdependenciesamong the principles, because we have access tothe original research notes. In the case of theother two examples we can only guess.However, we may expect some similarities, asthe other two papers also place principle fourhigh on their research priorities and, like Myers,ignore principles three and five.

Orlikowski’s research is concerned with “theextent to which information technology deployedin work processes facilitates changes in forms ofcontrol and forms of organizing” (p. 9). While the

relationships between forms of control and organi-zational forms could be explored on the basis ofmany theories, Orlikowski chose structurationtheory for that purpose. As far as we can tell fromthe paper, key constructs from structuration theoryhave driven the application of principle two, i.e.,the selection of material that explains the firm’scurrent political structure and overall objectives.Furthermore, the pivotal theoretical distinctionsintroduced in the paper deal with systemic formsof control and their changes. These distinctionsappear to be related to key constructs of structura-tion theory, called mediation of structure throughhuman action, which draw on the rules andresources provided by the organization. If theserules and resources are embodied within IT, thenthe use of IT will not change the existing forms ofcontrol, but reaffirm their importance (unlessexceptional circumstances upset the whole orga-nizational structure). Orlikowski illustrates theseideas in her telling of the story. Hence it appears tous as if the main influence is directly from principlefour to two, with some consideration of six at thesame time. In comparison to Myers (1994), princi-ple seven receives much less attention. It is usedprimarily to explain why the principle actors haverelied on IT to intensify the existing forms of con-trol (i.e., to meet short-term profit goals).


Principle Four

Principle Six

Principle Two

Principle Seven

Figure 1. Interdependencies Among the Principles Exemplified in Myers (1994)


The research goal in Walsham and Waema’spaper is to provide “an example of a moredetailed perspective on processes in IS strategyand implementation then [sic] typically avail-able in the literature” (p. 150). To this end,they draw on Pettigrew’s contextualism as atheoretical foundation. When applied to ISstrategy, contextualism suggests tracing thedynamic interlinking among the followingcomponents of IS strategy: content, context,and process. Through their interpretation of thecase, Walsham and Waema explore these ideasas they relate to the case details. Hence, onceagain, we see a clear dominant influence fromprinciple four to two and six in order to depictthe evolving story. At the core of their inter-pretation are the views and actions of seniormanagement to deal with IS strategy issues inthe building society and the reactions of othersto these actions. As in Orlikowski’s research,principles six and seven receive less attentionthan in Myers (1994), but are still clearlyvisible.

In summarizing the interdependence of theprinciples in all three papers, it is interesting tonote that all three are strong theoretically. Itappears that principle four played a dominantrole in them, and that this drove the applicationof the others. The three papers that we ana-lyzed would therefore support the conjecturethat principle four assumes a leading role in allinterpretive field studies of a hermeneuticnature. This is so because hermeneutics makesthe philosophical case that no understanding ofnew “data” is possible without relying on somepre-understanding. However, this raises thequestion of whether more attention to some ofthe other principles (especially principles threeand five, which were neglected in these papers)by other researchers in the future will affect theinterrelationships among the principles. At thispoint, we leave open the possibility that manyother combinations of the principles are possi-ble; these new combinations might well lead toother types of interpretive field studies with pat-terns different from those exhibited in our threeexamples. However, it is difficult for us to seehow an interpretive field study could be mean-ingfully conducted without some sort of theo-retical filter.

Discussion and ConclusionsInterpretive research has emerged as a valid andimportant approach to information systemsresearch. Most mainstream IS journals now wel-come interpretive research and significant groupsof authors are working within the interpretive tra-dition (Walsham 1995a). With this growing inter-est in interpretive research, however, questionshave been raised about how its quality can beassessed. This article has suggested a set of prin-ciples for the conduct and evaluation of interpre-tive field research in information systems derivedprimarily from anthropology, phenomenology,and hermeneutics.

Earlier we emphasized the point that our sug-gested set of principles for interpretive fieldresearch is just one among many plausible anduseful sets of principles for interpretive research.We readily acknowledge that not all interpretivework is hermeneutic in orientation and thereforewe leave the door open for other interpretiveresearchers to suggest a different set (or sets) ofprinciples. We do believe, however, that explicitarticulation of the principles is a contribution toimproving interpretive field research methodol-ogy in information systems in the following ways.

First, we have managed to crystallize a diffuse lit-erature into a manageable set of principles forthose who accept philosophical hermeneutics asa foundation for interpretive research. Thisshould help those interpretive researchers whowish to concentrate on fieldwork, as opposed tothe study of philosophical foundations, to designtheir investigations more systematically. Withoutthe principles proposed in this paper, each andevery interpretive researcher would have tospend considerable time deriving the theoreticalfoundations for their research from diverse litera-ture sources. This would be true even if theywished to limit themselves to the same subset onwhich this paper is based, i.e., key ideas inanthropology, phenomenology, and hermeneu-tics. Second, researchers can now defend theirwork by appealing to principles that are firmlygrounded in at least one major direction of inter-pretive philosophy. Authors no longer need tojustify their work by the use of inappropriate(positivist) criteria. In fact, authors may find ituseful to refer to the principles when their work



is submitted for peer review. Third, the introduc-tion of a set of principles encourages researchersto consider each one of the principles systemati-cally and ensures that none has been left outarbitrarily. It may well be that researchers wouldhave otherwise neglected an important principlein their work. The corollary of this is that theprinciples can also help reviewers and readers tocheck the selective judgements of researchers.Reviewers and editors can ask for clarificationsand explanations when some of the principlesappear appropriate to them but were not appliedby the researcher.

We caution, however, that our proposed set ofprinciples cannot be applied mechanistically. Itis incumbent upon interpretive scholars to appro-priate them and use their own judgement as totheir specific application. We do not absolveauthors, reviewers, and editors of the effort ofworking out whether, how, and which of theprinciples should be applied in any givenresearch project.6

As well as being a contribution to improvinginterpretive field research methodology in IS, webelieve our paper is a contribution in otherways. First, the introduction of this set of princi-ples should help those readers who do not con-duct interpretive research themselves to betterappreciate its nature. We hope that this paperwill lead to interpretive research being betterunderstood and more widely accepted, eventhough the underlying philosophical assump-tions are distinctly different from those of posi-tivism. Second, our proposed principles providean important stimulus for the advancement ofinterpretive research approaches. The introduc-tion of this set of principles potentially creates avisible target and thereby a challenge for othersto articulate their disagreement in a constructiveway. Without continuing debate on mattersmethodological, research methods and stan-dards will stagnate. Compared to the historicaldebate on the core assumptions of positivistresearch methods, the advocates of interpre-

tivism have been relatively silent.7 We hope thatour paper will stimulate further reflection anddebate concerning whether and how the qualityof interpretive research can be assessed.

AcknowledgementsWe are grateful for the critical and constructivecomments of the senior editor,Allen S. Lee, and theanonymous reviewers. We would also like tothank Justo Diaz, head of the Department ofManagement Science and Information Systems attheUniversityofAuckland, forproviding thefinan-cial support that made this research possible. H. K.Klein gratefully acknowledges the release time forthis project from the School of Management, StateUniversity of New York at Binghamton.

ReferencesAdorno, T. W., Frenkel-Brunswick, E., Levinson,

D. J., and Nevitt Sanford, R. The AuthoritarianPersonality, Harper, New York, 1950.

Alvesson, M., and Willmott, H. (eds.). CriticalManagement Studies, Sage Publications,London, 1992a.

Alvesson, M., and Willmott, H. “On the Idea ofEmancipation in Management and OrganizationStudies,” Academy of Management Review(17:3), 1992b, pp. 432–464.


6In fact, a particular study could illustrate all of our sug-gested principles and still not come up with interestingresults. However, a discussion of appropriate and“interesting” substantive research topics is outside thescope of this paper.

7Interpretive researchers should take to heart an impor-tant lesson from the history of positivist research meth-ods. The current conventions regarding methods andstandards of empirical research as codified in the mod-ern handbooks of empirical research emerged out ofheated debate over three centuries. In the 16th cen-tury, Bacon and Galileo started this debate with theircontroversial claims on the nature of truth and humaninquiry followed by Locke and Hume in the 17th and18th centuries. This debate continued well into the20th century, with the polemics between the ViennaCircle and Popper and the post-Popperian revisionists(i.e., Kuhn and Lakatos). It seems to us that interpre-tivism needs a similar debate about its core assump-tions if it is to prosper. Any major school of thoughtcan only advance through continuing debate.Compared to the heated controversy about the natureof the scientific method in positivism and its claim tosupremacy (in the Popper-Adorno and Habermas-Popper controversies), the philosophical advocates ofinterpretivism have been relatively silent.


Baskerville, R., Pentland, B. T., and Walsham, G.“A Workshop on Two Techniques forQualitative Analysis: Interviewing andEvaluation,” in Proceedings of the FifteenthInternational Conference on InformationSystems, J. I. DeGross, S. L. Huff, and M. C.Munro (eds.), Vancouver, BC, Canada,December 14–17, 1994, pp. 503–4.

Benbasat, I., Goldstein, D. K., and Mead, M.“The Case Research Strategy in Studies ofInformation Systems,” MIS Quarterly (11:3),1987, pp. 369–386.

Bentley, R., Rodden, T., Sawyer, P., Sommerville,I., Hughes, J., Randall, R., and Shapiro, D.“Ethnographically-informed Systems Designfor Air Traffic Control,” ACM 1992 Conferenceon Computer-Supported Cooperative Work:Sharing Perspectives, ACM Press, New York,1992, pp. 123–129.

Boland, R. J. Jr. “Phenomenology: A PreferredApproach to Research in InformationSystems,” in Research Methods in InformationSystems, E. Mumford, R. A. Hirschheim, G.Fitzgerald, and A. T. Wood-Harper (eds.),North-Holland, Amsterdam, 1985, pp.193–201.

Boland, R. J. Jr. “Information System Use as aHermeneutic Process,” in Information SystemsResearch: Contemporary Approaches andEmergent Traditions, H-E. Nissen, H. K. Klein,and R. A. Hirschheim (eds.), North-Holland,Amsterdam, 1991, pp. 439–464.

Burrell, G., and Morgan, G. SociologicalParadigms and Organisational Analysis,Heinemann, London, 1979.

Carr, W., and Kemmis, S. Becoming Critical:Education, Knowledge and Action Research,Falmer Press, London, 1986.

Cash, J. I., and Lawrence, P. (eds.). TheInformation Systems Research Challenge:Qualitative Research Methods, HarvardBusiness School Research Colloqium, Boston,1989.

Chua, W. F. “Radical Developments inAccounting Thought,” The Accounting Review(61), 1986, pp. 601–632.

Ciborra, C. U., Patriotta, G., and Erlicher, L.“Disassembling Frames on the Assembly Line:The Theory and Practice of the New Divisionof Learning in Advanced Manufacturing,” inInformation Technology and Changes in

Organizational Work, W. J. Orlikowski, G.Walsham, M. R. Jones, and J. I. DeGross (eds.),Chapman and Hall, London, 1996, pp.397–418.

Clark, P. A. Action Research and OrganizationChange, Harper and Row, London, 1972.

Davies, L., Newman, M., and Kaplan, B. “AWorkshop on Two Techniques for QualitativeAnalysis: Interviewing and Evaluation,” inProceedings of the Fourteenth InternationalConference on Information Systems, J. I.DeGross, R. P. Bostrom, and D. Robey (eds.),Orlando, FL, December 5–8, 1993, p. 399.

Deetz, S. “Describing Differences in Approachesto Organization Science: Rethinking Burrelland Morgan and their Legacy,” OrganizationScience (7:2), 1996, pp. 191–207.

Elden, M., and Chisholm, R. F. “EmergingVarieties of Action Research: Introduction tothe Special Issue,” Human Relations (46:2),1993, pp. 121–142.

Forester, J. “Critical Ethnography: On Field Workin an Habermasian Way,” in CriticalManagement Studies, M. Alvesson, and H.Willmott (eds.), Sage Publications, London,1992, pp. 46–65.

Foucault, M. The Archaeology of Knowledge,Tavistock, London, 1972.

Fromm, E. The Sane Society, Rhinehart, NewYork, 1955.

Gadamer, H-G. Truth and Method, TheContinuing Publishing Corporation, New York,1975.

Gadamer, H-G. Philosophical Hermeneutics,University of California Press, Berkeley, CA,1976a.

Gadamer, H-G. “The Historicity ofUnderstanding,” in Critical Sociology, SelectedReadings, P. Connerton (ed.), Penguin BooksLtd, Harmondsworth, UK, 1976b, pp.117–133.

Giddens, A. The Constitution of Society: Outlineof the Theory of Structure, University ofCalifornia Press, Berkeley, CA, 1984.

Harvey, L., and Myers, M. D. “Scholarship andPractice: The Contribution of EthnographicResearch Methods to Bridging the Gap,”Information Technology & People (8:3), 1995,pp. 13–27.

Heidegger, M. Being and Time, Basil Blackwell,Oxford, 1962.



Hirschheim, R., and Klein, H. K. “Four Paradigmsof Information Systems Development,”Communications of the ACM (32:10), 1989,pp. 1199–1216.

Hirschheim, R., and Klein, H. K. “RealizingEmancipatory Principles in InformationSystems Development: The Case for ETHICS,”MIS Quarterly (18:1), 1994, pp. 83–109.

Husserl, E. Logical Investigations, Routledge andKegan Paul, London, 1970.

Husserl, E. Ideas Pertaining to a PurePhenomenology and to a PhenomenologicalPhilosophy, Kluwer, Boston, 1982.

Kahn, J. S. “Culture: Demise or Resurrection?”Critique of Anthropology (9:2), 1989, pp.5–25.

Kaplan, B., and Maxwell, J. A. “QualitativeResearch Methods for Evaluating ComputerInformation Systems,” in Evaluating HealthCare Information Systems: Methods andApplications, J. G. Anderson, C. E. Aydin, andS. J. Jay (eds.), Sage, Thousand Oaks, CA,1994, pp. 45–68.

Lee, A. S. “A Scientific Methodology for MISCase Studies,” MIS Quarterly (13:1), 1989, pp.33–52.

Lee, A. S. “Integrating Positivist and InterpretiveApproaches to Organizational Research,”Organization Science (2:4), 1991, pp.342–365.

Lee, A. S. “Electronic Mail as a Medium for RichCommunication: An Empirical InvestigationUsing Hermeneutic Interpretation,” MISQuarterly (18:2), 1994, pp. 143–157.

Lee, A. S., Baskerville, R. L., and Davies, L. “AWorkshop on Two Techniques for QualitativeData Analysis: Action Research andEthnography,” in Proceedings of the ThirteenthInternational Conference on InformationSystems, J. I. DeGross, J. D. Becker, and J. J.Elam (eds.), Dallas, TX, December 13–16,1992, pp. 305–306.

Lee, A. S., Baskerville, R. L., Liebenau, J., andMyers, M. D. “Judging Qualitative Research inInformation Systems: Criteria for Acceptingand Rejecting Manuscripts,” in Proceedings ofthe Sixteenth International Conference onInformation Systems, J. I. DeGross, G. Ariav, C.Beath, R. Hoyer, and C. Kemerer (eds.),Amsterdam, December 10–13, 1995, p. 367.

Lee, A. S., Liebenau, J., and DeGross, J. I. (eds.).Information Systems and Qualitative Research,Chapman and Hall, London, 1997.

Levine, H. G., and Rossmore, D. “Diagnosing theHuman Threats to Information TechnologyImplementation: A Missing Factor in SystemsAnalysis Illustrated in a Case Study,” Journal ofManagement Information Systems (10:2),1993, pp. 55–73.

Lyytinen, K. “Information Systems and CriticalTheory,” in Critical Management Studies, M.Alvesson and H. Willmott (eds.), SagePublications, London, 1992, pp. 159–180.

Madsen, K. H. “Breakthrough by Breakdown,” inInformation Systems Development for HumanProgress in Organizations, H. K. Klein, and K.Kumar (eds.), North-Holland, Amsterdam,1989, pp. 41–53.

Mingers. J. C. “Towards An Appropriate SocialTheory for Applied Systems Thinking: CriticalSocial Theory and Soft Systems Methodology,”Journal of Applied Systems Analysis (7), 1981,pp. 41–49.

Monteiro, E., and Hanseth, O. “Social Shapingof Information Infrastructure: On BeingSpecific about the Technology,” inInformation Technology and Changes inOrganizational Work, W. J. Orlikowski, G.Walsham, M. R. Jones and J. I. DeGross (eds.),Chapman and Hall, London, 1996, pp.325–343.

Mumford E., Hirschheim, R. A., Fitzgerald, G.,and Wood-Harper, A. T. (eds.). ResearchMethods in Information Systems, North-Holland, New York, 1985.

Myers, M. D. “A Disaster for Everyone to See: AnInterpretive Analysis of a Failed IS Project,”Accounting, Management and InformationTechnologies (4:4), 1994, pp. 185–201.

Myers, M. D. “Qualitative Research in InformationSystems,” MIS Quarterly (21:2), June 1997, pp.241–242. MISQ Discovery, archival version,June 1997, http://www.misq.org/misqd961/isworld. MISQ Discovery, updated version,February 22, 1999, http://www.auckland.ac.nz/msis/isworld/.

Myers, M. D., and Young, L. W. “HiddenAgendas, Power, and Managerial Assumptionsin Information Systems Development: AnEthnographic Study,” Information Technologyand People (10:3), 1997, pp. 224–240.



Nardulli, P. F. The Courtroom Elite: AnOrganizational Perspective on CriminalJustice, Ballinger Press, Cambridge, MA, 1978.

Ngwenyama, O. K. “The Critical Social TheoryApproach to Information Systems: Problemsand Challenges,” in Information SystemsResearch: Contemporary Approaches andEmergent Traditions, H-E. Nissen, H. K. Klein,and R. A. Hirschheim (eds.), North-Holland,Amsterdam, 1991, pp. 267–280.

Ngwenyama, O., and Lee, A. S. “Com-munication Richness in Electronic Mail:Critical Social Theory and the Contexuality ofMeaning,” MIS Quarterly (21:2), 1997, pp.145–167.

Nissen, H-E., Klein, H. K., and Hirschheim, R. A.(eds.). Information Systems Research:Contemporary Approaches and EmergentTraditions, North-Holland, Amsterdam, 1991.

Orlikowski, W. J. “Integrated InformationEnvironment or Matrix of Control? TheContradictory Implications of InformationTechnology,” Accounting, Management andInformation Technologies (1:1), 1991, pp.9–42.

Orlikowski, W. J., and Baroudi, J. J. “StudyingInformation Technology in Organizations:Research Approaches and Assumptions,”Information Systems Research (2:1), 1991, pp.1–28.

Orlikowski, W. J., Markus, M. L., and Lee, A. S.“A Workshop on Two Techniques forQualitative Data Analysis: Analytic Inductionand Hermeneutics,” in Proceedings of theTwelfth International Conference onInformation Systems, J. I. DeGross, I. Benbasat,G. DeSanctis, and C. M. Beath (eds.), NewYork, 1991, pp. 390–1.

Pettigrew, A. “Context and Action in theTransformation of the Firm,” Journal ofManagement Studies (24:6), 1987, pp.649–670.

Prasad, P. “Systems of Meaning: Ethnography asa Methodology for the Study of InformationTechnologies,” in Information Systems andQualitative Research, A. S. Lee, J. Liebenau,and J. I. DeGross (eds.), Chapman and Hall,London, 1997, pp. 101–118.

Read, K. E. The High Valley, Charles Scribner’sSons, New York, 1965.

Ricoeur, P. The Conflict of Interpretations: Essaysin Hermeneutics, Northwestern UniversityPress, Evanston, IL, 1974.

Ricoeur, P. “Hermeneutics: Restoration ofMeaning or Reduction of Illusion?” in CriticalSociology, Selected Readings, P. Connerton(ed.), Penguin Books, Harmondsworth, UK,1976, pp. 194–203.

Ricoeur, P. Hermeneutics and the HumanSciences, Cambridge University Press,Cambridge, UK, 1981.

Suchman, L. Plans and Situated Actions: TheProblem of Human-Machine Communication,Cambridge University Press, Cambridge, UK,1987.

Trauth, E. M. “Achieving the Research Goal withQualitative Methods: Lessons Learned alongthe Way,” in Information Systems andQualitative Research, A. S. Lee, J. Liebenau,and J. I. DeGross (eds.), Chapman and Hall,London, 1997, pp. 225–245.

Walsham, G. Interpreting Information Systems inOrganizations, Wiley, Chichester, UK, 1993.

Walsham, G. “Interpretive Case Studies in IsResearch: Nature and Method,” EuropeanJournal of Information Systems (4:2), 1995a,pp. 74–81.

Walsham, G. “The Emergence of Interpretivismin IS Research,” Information Systems Research(6:4), 1995b, pp. 376–394.

Walsham, G., and Waema, T. “InformationSystems Strategy and Implementation: A CaseStudy of a Building Society,” ACMTransactions on Information Systems (12:2),April 1994, pp. 150–173.

Wynn, E. Office Conversation as an InformationMedium, unpublished Ph.D. dissertation,University of California, Berkeley, 1979.

Wynn, E. “Taking Practice Seriously,” in Designat Work, J. Greenbaum and M. Kyng (eds.),Lawrence Erlbaum, Hillsdale, NJ, 1991.

Yin, R. K. Case Study Research, Design andMethods, 2nd ed., Sage Publications,Newbury Park, CA, 1994.

Zuboff, S. In the Age of the Smart Machine, BasicBooks, New York, 1988.

About the AuthorsHeinz K. Klein is associate professor of informationsystems at the School of Management at the StateUniversity of New York at Binghamton and leads



the information systems research group there. Amajor focus of his research has been the analysis ofalternative paradigms of information systemsdevelopment and the application of critical socialtheory to rationality, user participation, and con-flict management in ISD. Dr. Klein has publishedarticles on decision support systems, office infor-mation systems, information systems researchmethodology, and alternative approaches to infor-mation systems development in Communicationsof the ACM, Advances in Computers, InformationSystems Research, MIS Quarterly, and Office:Technology and People. He is a member of the edi-torial boards of the Information Systems Journal,Information, Technology and People, and theWiley Series in information systems. He is co-edi-tor of Systems Development for Human Progressand Information Systems Research: ContemporaryApproaches and Emergent Traditions. He is alsoco-author of Information Systems and DataModeling: Conceptual and PhisosophicalFoundations.

Michael D. Myers is associate professor in theDepartment of Management Science andInformation Systems at the University of

Auckland, New Zealand. His research interestsare in the areas of information systems develop-ment, qualitative research methods in informa-tion systems, and the social and organizationalaspects of information technology. His papershave been published in Accounting, Manage-ment and Information Technologies, Commun-ications of the ACM, Information SystemsJournal, Information Technology and People,Journal of International Information Manage-ment, Journal of Management InformationSystems, MISQ Discovery, New Zealand Journalof Computing and New Zealand Journal ofBusiness. He is co-author of two books, includ-ing New Zealand Cases in Information Systems.Dr. Myers is an associate editor of theInformation Systems Journal, an associate editorof MIS Quarterly, and on the editorial boards ofCommunications of the AIS, InformationTechnology and People, and Journal of Systemsand Technology. In 1992, he was a visiting fac-ulty member in Information Science at theClaremont Graduate School, California, and in1996 a visiting faculty member in theDepartment of Management, University ofSouthampton, UK.


APPENDIX

Abstracts of the Three Interpretive Field Studies Discussed in the Third Section

Orlikowski (1991). “Integrated Information Environment or Matrix of Control: The ContradictoryImplications of Information Technology.”

Abstract: This paper examines the extent to which information technology deployed in work processesfacilitates changes in forms of control and forms of organizing. A field study of a single organization thatimplemented information technology in its production process is presented as an empirical investigationof these issues. The findings indicate that information technology reinforced established forms of orga-nizing and facilitated an intensification and fusion of existing mechanisms of control. While debunkingthe technological imperative once again, the results also provide a number of insights into the contra-dictory implications of computer-based work and control in organizations. In particular, this paper showsthat when information technology mediates work processes, it creates an information environment,which, while it may facilitate integrated and flexible operations, may also enable a disciplinary matrix ofknowledge and power. These findings and their implications for forms of control, forms of organizing, andprofessional practice are discussed.


Walsham and Waema (1994). “Information Systems Strategyand Implementation: A Case Study of a Building Society.”

Abstract: The formation and implementation of strategy with respect to computer-based information sys-tems (IS) are important issues in many contemporary organizations, including those in the financial ser-vices sector. This paper describes and analyzes an in-depth case study of the strategy formation andimplementation process in one such organization, a medium-sized UK building society, and relates theprocess to its organizational and broader contexts; the organization is examined over a period of severalyears and under the contrasting leadership of two different chief executives. The case study is used todevelop some general implications on IS strategy and implementation, which can be taken as themes fordebate in any new situation. The paper provides an example of a more detailed perspective on processesin IS strategy and implementation than typically available in the literature. In addition, a new frameworkfor further research in this area is developed, which directs the researcher toward exploring the dynamicinterplay of strategic content, multilevel contexts, and cultural and political perspectives on the processof change.

Myers (1994). “A Disaster for Everyone to See: An InterpretiveAnalysis of a Failed IS Project.”

Abstract: The New Zealand Education Department attempted to implement a centralized payroll systemin 1989. The difficulties that the department experienced were broadcast on national radio and televisionand publicized on the front page of The New Zealand Herald. In the end, the centralized payroll systemwas scrapped by the government. This paper examines this case study using the critical hermeneutics ofGadamer and Ricoeur. Critical hermeneutics, as an integrative theoretical framework, combines bothinterpretive and critical elements, and addresses those social and organizational issues which are key tothe successful implementation of information systems. This paper suggests critical hermeneutics as a con-ceptual foundation for information systems implementation research.


Documents

ASET OF RINCIPLES FOR ONDUCTING AND … been called the interpretive paradigm. Klein & Myers/Evaluating Interpretive Field Studies. and. Principles. 4. 4. Studies,