Exploring the role of informants in interpretive case study research in is
Journal of Information Technology (2011) 26, 32–45 & 2011 JIT Palgrave Macmillan All rights reserved 0268-3962/11 palgrave-journals.com/jit/Research articleExploring the role of informants ininterpretive case study research in ISBendik Bygstad1, Bjørn Erik Munkvold21Norwegian School of IT, Oslo, Norway;2University of Agder, Kristiansand, NorwayCorrespondence:B Bygstad, Norwegian School of IT, 0185 Oslo, Norway.Tel: þ 47 97658061;Fax: þ 47 22059960;E-mail email@example.comAbstractRecent information systems research calls for interaction between the researcher and theinformants in interpretive case study research. In line with Van de Ven’s call for engagedscholarship, we investigate how to involve the informants in case studies, not only forthe collection of facts, but also in the co-construction and interpretation of the casenarrative. The paper builds on a longitudinal case study, in which we explored an approachof extensive informant involvement. Using the ladder of analytical abstraction as ouranalytical tool, we discuss how an extended involvement of informants may enrich theinterpretive process in case study research, and increase the relevance of the findings.We discuss how and under what conditions this form of involvement may take place, andpotential challenges of this approach.Journal of Information Technology (2011) 26, 32–45. doi:10.1057/jit.2010.15;Published online 24 August 2010Keywords: informants’ involvement; informants’ feedback; respondent validation; member validation;interpretive case study researchIntroduction nterpretive case study research constitutes an important anthropological use of the informant is distinguished fromI and increasing part of the information systems (IS) knowledge base (Walsham, 1993; Myers, 1997; Pareand Elam, 1997; Walsham, 2006). Interpretive case studies ´ the social survey in that the respondents are selected not for their representativeness but rather on the bases of infor- medness and ability to communicate with the socialcan be distinguished from positivist case study research scientist’ (339). We argue that engaging the informants in ´ ´(Benbasat et al., 1987; Lee, 1989; Dube and Pare, 2003) by the a discourse on the concepts and patterns of explanationfocus on close interaction between researcher and partici- arrived at through the case study may provide an oppor-pants throughout the case study process, viewing the case tunity for mutual reflection and learning on the phenomenamembers as active participants in the construction of the studied. Further, it offers a way for the researcher to verifycase narrative (Boland, 1985; Guba and Lincoln, 1989; Kvale, that her interpretation of the phenomenon makes sense to2002). However, while the interpretivist perspective ascribes the informants. While this does not necessarily implyan active role to the case study informants, in practice the a shared interpretation, the informants should at least beextent of this involvement is normally confined to the data able to acknowledge how the researcher has arrived at thiscollection process and discussion of early versions of the interpretation (Schatzman and Strauss, 1973). If not, wecase narrative. In few cases is the involvement of the argue that there is a risk that the interpretation arrived atinformants reported to continue further to the final stages of by the researcher, however conceptually sophisticated, mayanalytical abstraction of the case study data, where the aim is be focusing on aspects of the case that are less relevant toto develop the overall patterns and explanations. the world of practice. Thus, we also argue that a closer In this paper we explore the question of how the involvement of informants in construction of the findingsinformants may be involved in the co-construction of is one way to meet the call for increasing practical relevancethe case narrative and theory building in interpretive of IS research (Benbasat and Zmud, 1999; Rosemann andresearch. The concept of informant is here understood in a Vessey, 2008). In that sense, this paper can also be viewedbroad sense, as a stakeholder that gives qualified informa- as a response to Van de Ven’s (2007) call for re-vitalizationtion or opinion on a case. As noted by Campbell (1955), ‘the of the relationship between research and practice through
Exploring the role of informants B Bygstad and BE Munkvold 33the concept of engaged scholarship, emphasizing research constructing a case narrative, we present experiences fromas a collective achievement rather than a solitary exercise: a longitudinal case study in the airline industry. During the‘Engagement means that scholars step outside of them- case study process spanning 18 months, the researcherselves to obtain and be informed by the interpretations of engaged the informants in different forms of reflection andothers in performing each step of the research process; feedback. Through this, both the informants and the resea-problem formulation, theory building, research design and rcher shifted perspectives on aspects of the project studied.problem solving’ (ibid, 10). We discuss the challenges experienced by the researcher in The question about the role of informants in co- this process, where the informants in several situationsconstruction of interpretive research entails issues about voiced strong opinions on both the case study report andthe epistemological status of the informants’ interpretations research publications from the case study.with respect to the researcher’s interpretations. Taken The next sections review former research related to theliterally, Van de Ven’s (2007) bold call for involving practi- involvement of informants in case study research, both intioners and other stakeholders in all steps of research social science research in general and in IS research. Wechallenges some of the foundations of the relationship bet- then present our case study, and describe how informantsween researcher and practitioner. As pointed out in much were involved in various phases of the study. Then, asmethodology literature, negotiating a case construction a lense for analyzing our findings, we draw on the ladder-agreed upon by the researcher and the informants is a risky of-abstraction framework (Carney, 1990). The final sectionsproject that may result in a ‘distortion’ of the case study discuss the findings and implications from the analysis.findings through mechanisms of post rationalization andunderlying power structures framing the informants’ views(Silverman, 1993; Klein and Myers, 1999; Walsham, 2006). The role of informants in case study researchAddressing these epistemological issues in full breadth is A review of the literature on qualitative research in the socialbeyond the scope of this paper, and the main focus of our sciences reveals differing perspectives on the role and natureinvestigation is on the methodological question of how to of involving informants in the process of constructing theinvolve informants in IS case study research. In accordance case narrative. Several terms are used to denote aspects ofwith the process of engaged scholarship (Van de Ven, 2007), this process: member validation (Yin, 1994; Bloor, 2001),we view the relationship between researchers and informants informants’ feedback (Miles and Huberman, 1994), respon-as involving discussions, mutual respect and collaboration. dent validation (Bloor, 1978; Silverman, 1985; Fielding andHowever, rather than ascribing equal epistemological status Fielding, 1986; Bryman, 1988), host verification (Schatzmanto the interpretations of the researcher and informants, we and Strauss, 1973), member verification (Morse et al., 2002),discuss how the latter may serve as important sources of member checks (Guba and Lincoln, 1989), backtalk (Lanzara,data that may inform the researcher’s interpretations. ´ ´ 1991) and project reviews (Dube and Pare, 2003). The research questions framing the discussion in this Key textbooks of qualitative research recommend thatpaper are thus: How may an extended role of informants getting feedback from the informants should be incorpo-contribute to enhance interpretivist case study research in rated as a part of the research design. Miles and HubermanIS, and what are the conditions where this type of approach (1994) refer to this as ‘one of the most logical sources ofis appropriate? To analyze these questions we draw upon corroboration’ yet also ‘a venerated, but not always exe-literature on qualitative research methods in social sciences cuted, practice in qualitative research’ (275). Differentand case study research in IS. Further, as an empirical basis techniques for obtaining such feedback include review offor our discussion, we present experiences from involving draft case study reports by the informants (Yin, 1994), andthe informants in a longitudinal case study conducted by evaluating the accuracy of ‘causal networks with higher-the first author. inference findings’ (Miles and Huberman, 1994). From a The literature review presents different perspectives on positivist perspective, the involvement of informantsthe role of informants in qualitative research, and discusses should be confined to verifying factual information andvarious forms of feedback from case study participants. assuring that the researcher’s understanding of the studiedOften referred to as member validation or respondent phenomenon as presented in the case report or similarvalidation (Silverman, 1985; Miles and Huberman, 1994; account is correct, in an objective sense, in order toYin, 1994) this may involve different activities conducted increase validity. As stated by Yin (1994): ‘The informantsthroughout the case study process, such as distribution of and participants may still disagree with an investigator’sinterview transcripts to informants for verification, pre- conclusions and interpretations, but these reviewers shouldsentation of a case study report or summary to key stake- not disagree over the actual facts of the case’ (145).holders for approval prior to publication and/or group According to this perspective, involving informants impliesmeetings with informants to discuss different interpreta- a potential risk for undue influence on the results andtions of the case material. The literature shows that while should thus be minimized (Morse et al., 2002).using this type of activities for verifying factual informa- Several authors present an extended perspective in whichtion is common practice, there is more debate on the issue feedback from the informants also serves to verify theof whether or how to also involve case informants in researcher’s understanding of the case study events. Forco-construction of case narratives (Kvale, 2002; Lincoln and example, using the term ‘host verification,’ Schatzman andGuba, 2003). Further, our review of case study literature in Strauss (1973) discuss how ‘credibility may be establishedIS shows a lack of attention to this debate in IS research. with some audiences by showing or simply stating that at To illustrate how an extended involvement of case least the major propositions were tested or checked againstinformants may play an important role in the process of the experiences and understandings of the hosts’ (134).
Exploring the role of informants B Bygstad and BE Munkvold34Their argument for how this practice may increase validity which enables a process of corrrection of impressions tocontinues: ‘If it was found that the propositions offered to take place. This should not exclude the possibility of thethe hosts did not empirically contradict their own under- interviewee doing some theorizing and some checking too.standings of their situation then the researcher may [y] under the right conditions ‘‘interviewees’’ can quiteconvince audiences that he has a measure of validity – easily turn themselves into co-researchers’ (249). Thepossibly a large measure. This mode of validating one’s potential role of informants in co-constructing the casework does not require that the hosts actually concur in the narrative is also echoed in Lanzara’s (1991) discussion ofpropositions themselves, but that they recognize rather the the concept of backtalk. This refers to the process of havingvalidity of the grounds (events) upon which the proposi- the informants ‘inspect and test’ the researcher’s initialtions rest’ (134). Similar, in Longitudinal Process Research interpretation of the data. Through an example from(LPR) (Pettigrew, 1985; Pettigrew, 1990), member valida- studies of educational practices in a music school, Lanzaration plays an important role for ensuring that the case describes how this backtalk provided new contextual datadescriptions are meaningful for the practitioners. that otherwise would not have been readily accessible, In general, phenomenological and interpretive research leading to a more complex picture of the phenomenonacknowledges case members as playing a more active role studied. This process also involved extensive discussions(Boland, 1985): ‘When the phenomenologist studies a with the researcher and the informants, leading toperson, she does not look at them, but with them in a dia- reinterpretation of past events and shifting perspectives oflogue searching for understanding. Understanding comes both parties. He also discusses how this process can involvestep by step, layer by layer, as preconceptions, prejudices, informants producing different descriptions of the sameand assumptions are recognized and seen through’ (343). event at different points in time, as the informant shifts hisAccording to this perspective, validity implies a process of or her perceptions on and cognitions of this event. Reflec-social construction (Kvale, 2002). Kvale treats validity as an ting on his role as researcher in this process, Lanzara (1991)expression of craftsmanship, emphasizing quality of resea- states: ‘As the actors talked back and told different storiesrch by checking, questioning and theorizing on the nature about past events, my job became one of accounting forof the phenomena investigated. Getting feedback from both the shifting stories and my own evolving under-informants constitutes one of the tactics for checking the standing of the process, accounts that I then discussed andcredibility of the findings. Further, Kvale discusses the tested with the actors themselves’ (291).concept of communicative validity, testing the validity of Summing up, our review of key readings on qualitativeknowledge claims through rational discourse between research in the social sciences has identified several per-researchers and a set of ‘legitimate partners.’ The latter spectives on the preferred and potential role of informantsmay include the subjects interviewed, the scientific com- in case study analysis. This ranges from verifying factualmunity and the general public: ‘valid knowledge claims information, through using informant feedback as a sourceemerge as conflicting interpretations and action possibi- of verifying the researcher’s interpretation, to viewing infor-lities are discussed and negotiated among the members of a mants as potential co-researchers. While there is generalcommunity’ (Kvale, 1996: 239). According to the construc- agreement about the importance of verifying factual infor-tivist perspective, Guba and Lincoln (1989) define member mation and about how informants’ feedback may repre-checks as ‘the process of testing hypotheses, data, preli- sent a source of additional data and insight (Fielding andminary categories, and interpretations with members of Fielding, 1986; Silverman, 1993), the literature review identi-stakeholder groups from whom the original construc- fied conflicting views on the value of involving informantstions were collected’ (238–239). They regard this as the for refining the analysis of the case study. Those expressingsingle most crucial technique for establishing credibility, caution against involving the informants in constructing thethe constructivist ‘parallel criterion’ to internal validity. case findings point to potential negative effects such asThis criterion focuses on establishing the match between censorship and defensive reactions from the informantsthe constructed realities of respondents and those reali- (Bryman, 1988). When reading the interview transcripts orties as represented by the researcher and attributed to case description the informants may want to justify theirvarious stakeholders. Guba and Lincoln (1989) discuss how actions, thus threatening the integrity of the initial datamember checks can be both formal and informal, and may collection (Bloor, 1978; McDonnell et al., 2000). Further, itoccur both during the data collection and analysis stage, has been argued that informants should not be considered asand when the case narrative is developed. To distinguish to have a privileged status as commentators on their actionsmember validation from triangulation, they argue that (Fielding and Fielding, 1986; Silverman, 1993).while triangulation is limited to cross-checking factual data, The discourse surrounding these modes of involvementmember-checking processes are concerned with verifying can be related both to aspects of validity as represented inthat the constructions collected are those that have been different research paradigms, and to practical relevance ofoffered by the respondents. the research. By focusing on the basic dichotomy between Some also argue for an ethical imperative in including positivist and interpretive research, we have deliberatelyinformants in the case analysis. For example, Reason and chosen not to engage in a more refined analysis of possibleRowan (1981) argue that refining the tentative results based variations between different forms of interpretivist caseon the informants’ reactions to these is a key characteristic studies (e.g., social constructionism, symbolic interaction-of good research ‘at the non-alienating end of the spectrum’ ism and phenomenology), as we do not see this would(249). They go on asserting that ‘instead of a ‘‘hit and run’’ add significantly to our main argumentation. The sameapproach which sucks subjects dry and leaves her by the applies with regard to the recent contributions of criticalwayside, there needs to be an involvement with the person realism in IS research (Smith, 2006), in which the basic
Exploring the role of informants B Bygstad and BE Munkvold 35epistemological position of critical realism is congruent approach, as the latter does not involve interaction withwith interpretivism (Sayer, 2000). the informants (or practitioners) related to the data ana- lysis stage. In empirical IS research, very few studies document inAn assessment of IS case study research depth the interaction between informants and researchersIn IS research we have found relatively few studies that taking place during the analysis stage of a case study. Forexplicitly address the mutual relationship between resea- example, in a survey of 183 positivist case studies fromrchers and informants in case study analysis. For example, seven major IS journals, only 15% of the studies explicitlyin Walsham’s (1995) ‘primer’ on conducting interpretive reported any form of feedback session with the casecase studies in IS, now considered a standard reading for IS ´ ´ informants (Dube and Pare, 2003). Building on Yin (1994)Ph.D. students, he discusses how the role of the researcher ´ ´ and Schatzman and Strauss (1973), Dube and Pare (2003)may vary between that of the outside observer and that of define project reviews as soliciting research subject orthe involved researcher. However, he does not address the participant views of the credibility of interpretations andpotential role of the informants in this type of research. findings. On the basis of the results of their survey theyIn his follow-up article on conducting interpretive research, argue for wider use of the project review strategy, ‘whetherWalsham (2006) discusses ethical issues and tensions in the under the form of a formal presentation to key actors orrelationship between the researcher and the case infor- a review of the case report itself (y), to corroborate themants and their organization, such as how to report bad evidence presented in the case report’ (625). Similarly, innews uncovered through the field work to the organiza- analyzing 22 qualitative studies from four top IS researchtional sponsors and whether the case organization ought to journals (MIS Quarterly, Information Systems Research,be disclosed or not. Yet, the issue of how informants may Journal of AIS and Information and Organization) in theplay a role in interpretation and construction of the case period from 2001–2005, Myers and Newman (2007) foundstory remains unaddressed. that only six of the studies reported any feedback offered to Several of the principles for conducting and evaluating the companies/subjects. In their discussion on the craft ofinterpretive field studies in IS defined by Klein and Myers qualitative interviewing in IS research, presenting findings(1999) relate to the issue of informants’ feedback, although and results to subjects and organizations is included underthey do not explicitly discuss this practice. For example, the the guideline termed ‘Ethics of Interviewing’: ‘it may beprinciple of interaction between the researchers and the advisable sometimes to provide early feedback to subjectssubjects calls for critical reflection on how the case story and organizations and to check with them about factualwas socially constructed through interaction with case matters if needed’ (23). The recommendation for thisactors. Klein and Myers (1999) argue that participants can practice is thus rather vague and limited to checking factualbe seen as interpreters and analysts just as much as the matters, not taking into account the potential resourceresearcher: ‘Participants are interpreters as they alter their implied in getting feedback from the members.horizons by the appropriation of concepts used by IS In the empirical studies where some form of feedbackresearchers, consultants, vendors and other parties inter- from the informants is reported, this tends to be onlyacting with them, and they are analysts in so far as their briefly mentioned in passing to document that standardactions are altered by their changed horizons’ (74). Further, methodological procedure has been followed but withoutthe principle of suspicion addresses the problem of ‘false explaining in detail how the involvement took place or howconsciousness’ on the part of the participants; underlying it possibly did influence on the research outcome. Similara seemingly unified account of something, the actors may observations have also been made for qualitative organiza-be strongly biased by structures of power. This may pro- tional studies in general (Locke and Velamuri, 2004). Theduce distorted pictures of reality, which the researcher following examples serve as illustrations of this:must see through. The issue of engaging the informants in a dialogue on A preliminary draft of the case, in a slightly longer,the understanding of the case findings also ties into the undisguised version, was circulated to the subjects. Thisongoing debate on how to increase the practical relevance version [the published article] incorporates their com-of IS research (Saunders, 1998; Benbasat and Zmud, 1999). ments. In addition, it is disguised at the request of theRosemann and Vessey (2008) argue for increasing relevance company and the subjects. (Goldstein, 1990: 259)through an approach they term applicability checks, definedas ‘evaluations by practice of the theories, models, frame- Finally, an important guard against an observer effect wasworks, processes, technical artifacts, or other theoretically the presentation of our findings to Bremerton managersbased IS artifacts that the academic community either uses in ways that the findings could be challenged or discon-or produces in its research’ (5). This also include research firmed. The findings presented in this paper incorporateobjects resulting from case study research. Through this the results of the challenges raised by BI employees.approach the aim is to ensure the importance of the (Levine and Rossmoore, 1993: 63)research to the needs of practice. By being conducted eitherat the beginning or the end of the research life cycle, theauthors state that the approach ‘leaves untouched the While these statements document that the researchersrigorous methods used to conduct the study, that is, it does have taken advantage of feedback from the case informants,not compromise traditional research methods’ (1). This the reader is not able to learn about the nature of thisimplies that the form of informants’ involvement focused in feedback or how it was incorporated. A somewhat moreour paper is complementary to the applicability checks elaborate description of the feedback process is offered
Exploring the role of informants B Bygstad and BE Munkvold36by Nicholson and Sahay (2004) in their study of offshore projects. The theoretical point of departure was socio-software development: technical research on IS innovation in organizations (Avgerou, 2002) and actor-network theory (Latour, 1987; A second step in the analysis process was the preparation Hanseth and Monteiro, 1996). This section provides a brief of management reports that we submitted to the senior overview of the case and the data collection and analysis, managers of both Sierra England and India. Keeping in as a basis for the detailed discussion of the informants’ mind the readership, this report was quite different in involvement process in the next section. A more in-depth structure and length from our summary document, which presentation of the case study is available in Bygstad (2006) was primarily for our internal purposes. In this manage- and Bygstad et al. (2010). ment report, we tried to concisely identify the proble- matic issues in the relationship, a description of why Case overview we thought it so and our suggestions for action. We The case organization was an international airline deve- submitted two such reports during the course of our loping an e-business solution. The researcher conducted research, which helped us not only to maintain our resea- two workshops and 20 interviews with central stakeholders rch access, but also served as a vehicle for obtaining and IS developers over a period of 18 months, and was multiple perspectives, that of the respondents, on our given extensive access to project documents. At the outset interpretations. Feedback received from the respondents it was agreed that the project managers (one from the on our report helped to clarify our own understanding of business side and one from IT) should read the draft publi- the issues that we raised, and also identify some other cations to approve whether the airline could be identified in issues that we may have missed out (340–341). the publications from the project. As will be explained in the next section, the informants’ involvement processSimilarly, Walsham and Sahay (1999) discuss how their resulted in such approval. However, in this paper we haverole as researchers during a 3-year longitudinal case study chosen to anonymize the case company, as the focus here isgradually evolved from that of independent observer on the interaction with the case informants rather than the(Walsham, 1995) to ‘action researcher,’ with impressions actual results of the case study.from the first months of field work fed back to the research In 2000, acknowledging the commercial potential ofsites. They reflect on how this had impact on the case, Internet booking, the airline decided to establish a web-arguing that direct involvement of the researcher is based marketing channel in all important markets, includ-inevitable in longitudinal interpretive case studies. How- ing Europe, Asia and the Americas. To support this newever, apart from a brief mentioning of themes being business process, a new content management and publish-developed and discussed with informants during the data ing solution was needed. A project was initiated, organizedanalysis, they do not report on the process of obtaining with two project managers (one business and one tech-feedback from the informants. Another example in the nical), marketing executives from different countries andborderline between interpretive research and action re- a team of software developers. The main aim of the projectsearch is the dialogical action research approach described was establishing the new marketing channel. Further goals ˚by Martensson and Lee (2004), in which the researcher of the project were to enable the marketing executives withengaged extensively with practitioners over a long period an easy tool to publish materials and campaigns, withoutof time in order to build a mutual understanding of the the need for using html coding, and to integrate this newcase organization. Similar to the applicability checks app- system with the airline booking systems.roach suggested by Rosemann and Vessey (2008), dialogical The development project was structured in five itera-action research aims at ‘speaking the practitioner’s lang- tions, building on the Rational Unified Process (RUP)uage’ in order to increase IS research relevance. (Jacobson et al., 1999). The project was run over 1 year, In summary, there is little IS research that explicitly starting in May 2001, with the system going live in thediscusses the practices focused in this paper. With a few summer 2002. After two disappointing iterations, in whichexceptions, the discussion on the role of informants in the the developers failed to convince the marketing executivesqualitative research literature is not echoed in the IS of the need for the system, and the external delivery ofliterature, and the nature of interaction with informants in central components was seriously delayed, the project wasIS case study research tends to be underreported. Given the in a state of crisis. In response to this, the project groupclose link between the IS discipline and its vocational concentrated successfully on internal technical issues,nature (Baskerville, 1996; Benbasat and Zmud, 1999), we postponing integration of the e-business solution and theargue that there is a need for more attention to the question business process. This was addressed in an extended lastof how involvement of informants in case study research iteration, in which the social and technical integrationmay contribute to enhance practical relevance. To illustrate challenges were solved by improvisation. After a hecticthis, we present a case study example where involvement of finish, the system was taken into use by the internationalthe informants contributed to increased mutual learning airline with relative success.between informants and the researcher. Data collection and analysis Data were collected in accordance with the principles ofCase study LPR (Pettigrew, 1985; Pettigrew, 1990; Ngwenyama, 1998):The empirical basis for this paper is a longitudinal casestudy conducted by the first author, focusing on the Engaging with the research site at several times duringchallenges of socio-technical integration in IS development the study, to collect data reflecting changes over time.
Exploring the role of informants B Bygstad and BE Munkvold 37Table 1 Data collection in the airline casePhase Activities DocumentationInitial activities Initial meeting with three managers Summary of business and project Document collection objectivesEarly project phase A workshop with IT project and business The primary artifact was a graphical project managers, to get the broad picture illustration of how stakeholders and Individual interviews components were included into the projectLate project phase Group interviews with project group Interview summaries Individual interviewsAfter system in production Group interviews Final project reports, user evaluations Individual interviews Participant observation, to understand the actors’ between the researcher and the stakeholders. One result of language and problem solving, and to make sense of the study was a set of socio-technical integration patterns in different situations. IS development projects: big bang, stakeholder integration, Collecting systematically different types of data, to secure technical integration and socio-technical integration validity. (Bygstad and Nielsen, 2003). The paper focused strongly on the practical and managerial implications of the findings.The informants were two sets of employees in the case While informant feedback was conducted in complianceorganization; the first group was key personnel from the IS with the longitudinal research method, the practical stepsdepartment (the developers of the technical solution) and were a combination of planned and improvised action,the second group consisted of marketing executives at depending on the practical opportunities and the research-different geographical locations (the future users of the er’s initiatives. The next section presents the activitiessolution). They were selected based on the following related to the informants’ feedback process in the casecriteria: they were key actors in the e-business project, project.and they represented both the IT and business side of theproject. The researcher was granted permission to interviewthem at planned intervals during the project, and the key Informants’ feedback in the airline case: correcting,informants (a methodology manager, two project managers co-constructing and reflectingand three marketing executives) expressed an initial As informant feedback was part of the chosen researchinterest in the research project, in order to give input to approach (LPR), the researcher discussed the practicalitiesthe internal improvement process. of conducting this with key informants at the start of the The case was researched in four phases, as illustrated in study. The key informants were all highly educated, withTable 1. Data collection included interviews, workshops, master or bachelor degrees, and many years’ experienceproject documentation, technical documents, software with large projects in several organizations. From the startdemos and participant observation from meetings. of the researcher’s engagement they expressed that they The data were coded in an Atlas database. Interview would be interested in the researcher’s findings, althoughsummaries, project documents and technical reports were not necessarily convinced that this research would contri-coded following the guidelines of Miles and Huberman bute knowledge new to them.(1994). Then a systematic search for relationships in the Informants’ feedback in the airline case was carried outdata was conducted, based on the following guidelines for in three steps (Table 2). First, the documented chronologydata analysis (Pettigrew, 1985; Ngwenyama, 1998): of events and the illustrated socio-technical network from the workshops were sent to the participants for comments Comprehensive analysis; to identify underlying struc- and corrections. Then, at the end of the case study there tures and patterns of the organizational process. was a long review session with technical and business Temporal analysis; to aid in contextualizing findings by stakeholders to discuss the final report. And lastly, the placing events and situations in a narrative structure. research papers that were published were sent to the two Informants’ feedback; to ensure that the case description project managers and two business line managers for and researcher’s interpretation were considered factually comments. correct and meaningful to the organizational actors. The first step was concerned with identifying the timeData collection and analysis were conducted in an iterative line, and actors and events. The next step was constructingmode; one observation would often trigger a new inter- the case description, focusing on relationships and themes.pretation, which again could lead to a new question and/or The third step focused on identifying socio-technicalpossibly a new stakeholder. The function of infor- process patterns and explanations, with research papersmants’ involvement in this case was not only to ensure from the case as the main documentation.the factual correctness of the case description, it alsocontributed to the social construction of the case study Phase 1: Verifying facts(Klein and Myers, 1999). The case description was built The case was gradually constructed over the whole periodgradually over time, in a process of learning and discussion of data collection. An early event was a half-day workshop
Exploring the role of informants B Bygstad and BE Munkvold38Table 2 Informants’ feedback during three phasesResearch project phase Documentation Function of informants’ feedbackData collection A graphical illustration of the actor Correcting errors network plus interview summariesCase construction Case description Co-constructing the case narrativeResearch publishing Research papers Discussing implicationswith one group of important stakeholders (the marketing happened. For example, there was a disagreement about themanager, the project manager and the three technical results of the first two workshops of the project. The projectdevelopers), with the aim of establishing a time line for the group had followed the plan, involved users and tried todevelopment project and a model of the socio-technical model the use cases graphically, and judged the workshopsnetwork. The time line included the activities that initi- to be useful. The marketing executives, on the other hand,ated the project, the actual project milestones and the had felt that the communication between the project anditerations. The actor-network model illustrated the step- the users was poor, and that the use cases were verywise enrollment of stakeholders and technology into the theoretical and not related to their work process.project. Both artifacts were updated throughout the study The written project documents supported the developers’over 18 months, and were used to structure the findings view; they had really done what the methodology called for.and analyze changes over time. This was discussed with the marketing executives in a The graphical representation was sent to the workshop follow-up interview, where they gave two explanations forparticipants for comments. A number of corrections were their negative perception of the workshops. First, they didgiven, including for example: not really have time to participate, and were not well pre- pared for the sessions. Second, they were used to another Corrections to the chronology of events, such as the dates technology (Frontpage), which gave them more freedom in for meetings and workshops. the design, and they regarded the new solution as a step Corrections to the persons involved, for example that backwards. Both accounts were documented in the revised a vendor was also participating in a meeting. case report, which was accepted by both sides. This conflict Amendments concerning the description of the technical of interpretation became input to a higher level of analysis: solution, for example that the number of components in enrolling stakeholders from the business process was not the fifth iteration was incorrect, and that a certain successful, in spite of it being done ‘by the book’, that is by component was left out. management approval and formal planning. The project Amendments to lists of relevant documents, for example group had failed to convince the marketing people of the that a methodology document was produced before the need for a new solution. project started. Another issue was a finding that important businessSummaries of workshops and interviews were commented needs were not included in the use case requirements.on much the same way, focusing on factual issues. A follow-up e-mail commented:In addition, there were suggestions for other stakeholdersto interview, and other documents to draw upon. The You write that important business needs were not part ofatmosphere of this interaction was generally quite relaxed. the requirements. This is not because we did not know,Although the graphical actor-network representation was but because of corporate priorities. The number of usesomewhat complex, the comments were all to the point, and cases was reduced after the 9/11 attacks.errors were easily corrected. The researcher responded by adding a sentence in the case report, to acknowledge this point:Phase 2: Co-constructing the case descriptionThe case was described in a final case report. It includedthe timeline and actor-network, but concentrated on inter- Of course, the fact that the number of use cases were reducedpretations of the case process. Prominent themes were the from 20 to 10 had consequences for the functionality of theinteraction between the development team and the market- system ying executives, and also the integration challenges in theproject. Integration issues included the gradual enrollment of both The final case report was discussed at a formal review stakeholders and technology into the project, and thesession (suggested by the researcher), in which several project managers’ response to the complexity of thestakeholders from both the project and the business process. An observation was that project managers tendedorganization participated. The discussion focused on inter- to respond with project encapsulation; that is to concentratepretations of concepts and events, and the review meeting on internal project activities instead of addressing externalresulted in a number of changes to the case report. Special problems. Two case stakeholders had strong objections tocare was taken to analyze the instances in which data this notion of project encapsulation, which they felt waswere contradictory, for example when developers and not an accurate description, and also that the notion madeuser representatives had different accounts of what had the project manager appear somewhat defensive in the
Exploring the role of informants B Bygstad and BE Munkvold 39situation. In an e-mail response they wrote that what had that the theoretical perspectives of the paper had reframedactually happened was that they were forced to postpone the members’ interpretations of their own experience. Thesome of the technical and stakeholder integration, because concepts of ‘socio-technical integration’ and ‘Big Bangof factors they did not control. The alternative would project’ had been introduced in the research papers, andhave been to stop the whole project. they were now used by the members in their arguments. The researcher took this into account. But how should it The objections put the researcher in an interesting butbe interpreted? The researcher argued that the event that difficult situation, which (it should be admitted) had notthe plan could not be followed should not be interpreted as been included in the research design. How should they bean accident or merely as bad luck, but rather as an indica- handled? In responding to them, the researcher wastion that something was not working properly. Would it be concerned that at this stage of the research the informants’fair to say that the dependency of too many actors forced feedback could not be solved by negotiation. Arguably,the project into a certain degree of encapsulation? No, assessing the implications of a case is a research step inthe stakeholders replied, because the decision was deli- which the researcher must trust his methodology and drawberate, and the risks were assessed. But was encapsulation on earlier contributions, in order to arrive at a validan unintended effect of that decision, then, the researcher conclusion. On the other hand, it would be contrary to theasked, and so on. notion of engaged scholarship to disqualify the objections The atmosphere in the review meeting and the following on principle. But which criteria should guide the response?e-mail exchanges was polite, but quite engaged. E-mails The researcher chose to treat the objections in the sametypically started with ‘I do not agree with your interpreta- way as he would with an academic reviewer, that is at a dis-tion of this event y .’ An interesting aspect was that the cursive level. This meant that he would assess the relevancemembers started to use the vocabulary introduced in the of the objection, and if considered relevant and reasonable,case report, as the example of project encapsulation try to address it. Two examples will illustrate this point. Indescribed above. response to objection #1 (above), the researcher acknowl- edged that the project group indeed had tried to enroll stakeholders early in the process, but that the numberPhase 3: Reflecting on the case study implications of dependencies had made it difficult. This interchangeThe third phase of analysis was concerned with the of views became an important input to discuss the forces ofimplications of the case. These implications were discussed the proposed integration patterns, and to propose how theyin several academic research papers that were written on could be managed.the basis of the case materials (Bygstad and Nielsen, 2003; The other objection was different, and expressed aBygstad, 2006). The papers were written in the usual aca- concern that researchers have a limited understanding ofdemic style, including research reviews and theoretical the pressures of large IT projects. There might be amplediscussions. They focused on the more general aspects of reasons for this belief. Often, university textbooks presentsocio-technical integration, and used the case as an relatively simple recipes for handling IT projects, makingexample of integration patterns. the (maybe unintended) impression that the main reason It had been agreed in advance that all papers should for failure is poor project management. In responding, thebe subject to approval by the airline (represented by the researcher chose to add a section in the paper, highlightingkey stakeholders) prior to publication, that is if not appro- the increased pressures of large and complex IT projects inved, the researcher would be asked to anonymize the a turbulent business context.company in research publications. The papers were sent The researcher chose to discuss these issues in longto the methodology manager, the two project managers emails, in order to invite to a thoughtful and balancedand one of the marketing executives. Rather unexpectedly, reflection. The final responses from the stakeholders to thisthe papers triggered considerable response from the infor- were without the tension of the past exchanges, and the newmants. After receiving the first draft paper, lists of new version of the paper was accepted for publication with fullissues were brought up by three central stakeholders: the disclosure of the company.two project managers and the methodology manager. Thefirst in a series of e-mails started: ‘You cannot be allowedto publish accounts on this company that are not true y .’The objections included, for example: Discussion Our point of departure is the idea that a case narrative is (1) My opinion is that we really did aim for socio- not only a chronology of events, but a medium for sense- technical integration. The reason this was not achieved making for both the researcher and the informants (Butler, was that the marketing executives did not prioritize the 1998; Wagner, 2003; Czarniawska, 2004). As illustrated workshops. This was not a Big Bang pattern project. by the airline case, the reflexive dialogue between research (2) You do not seem to realize that a project manager and informants is both subtle and complex. It blurs the continuously has to balance the risks of such a complex difference between data and interpretations, since the infor- project with many dependencies. mants may start to use the terms of the researcher, as interpreted within their own frames of understanding.In total, 12 different issues were raised. The researcher Also, it challenges the traditional role set of researcher andwondered why these issues had not been mentioned at the informants, where the informants – within some limitslong and quite engaged validation meeting and e-mail discussed below – actively influence on the interpretationexchanges some months earlier. One explanation may be of the case.
Exploring the role of informants B Bygstad and BE Munkvold40 The case study shows how the different steps in the 3. Identifying patterns and explanationsinformants’ feedback process involved different levels ofabstraction, and different forms of involvement of the 2. Identifying themes and trendsinformants. To frame the discussion of these levels, we 1. Summarizing interviews and technical documentsintroduce in the next section a framework by Carney (1990)that defines different levels in the process of analytical Figure 1 Ladder of analytical abstraction, after Carney (1990).abstraction in qualitative research. On the basis of this, wethen discuss the findings and implications from our studyas related to our initial research questions. Although wedo not claim that the case can serve as a methodological discussing informants’ feedback as an occasion for genera-model for informants’ feedback, we argue that it, through ting new data (Silverman, 1993; Locke and Velamuri, 2004).the lense of the framework, serves to highlight that infor- At the second level of abstraction (identifying themesmants’ feedback play different roles in the different stages and trends), informants’ feedback is important for con-of analysis. structing the case narrative. Focus for this phase is the case study report, which provides a case story, and whereFramework: The ladder of analytical abstraction the researcher introduces her key terms. These terms toAs illustrated in Figure 1, the framework depicts the con- a large extent decide the framing of the case and perspectivestruction of a case story as a ‘ladder of analytical of the interpretation, and the informants may, or may not,abstraction’ (Carney, 1990; Miles and Huberman, 1994) identify with these terms.Starting with the interviews and other textual material, the While informants’ feedback at the first level of abstrac-first level is concerned with summarizing and coding of the tion takes the form of factual corrections, the process atdata. The next level is focused on identifying themes and the second level may be one of co-construction. As thetrends in the data, identifying the important concepts and airline case illustrates the understanding of the case eventsvariables. At the third level, the researcher aims at delinea- was discussed through the text, working on terms andting the ‘deep structure’ of the case, identifying patterns and sentences, and discussing nuances and exceptions. This isbuilding explanations. also exemplified by Locke and Velamuri (2004), referring to Climbing this ladder of abstraction is a process of trans- a dissertation project in which the Vice President of anformation; raw data are transformed to concepts and Indian case company provided a 36-page response to a casevariables, which again are synthesized to larger explanatory study draft of 42 pages, thus initiating an extensive iterativeframeworks. For the researcher this is a demanding pro- process of drafts and detailed responses going back andcess, which involves explicating and reflecting upon the forth over a period of 2 months.views and biases of the different stakeholders and the For the researcher, the process of involving the infor-researcher herself. While the researcher may use estab- mants in constructing the case report is both interestinglished methodological principles for qualitative research and challenging. While some of the informants’ views will(Klein and Myers, 1999) and analytical techniques such provide more depth to the narrative, other views may justas forward-chaining and backward-chaining (Pettigrew, reflect the informants’ wish to justify and defend their own1985; Pettigrew, 1990), a heavy responsibility resides with actions. This problem of distortion is important, and wethe researcher. We will argue that these transformations will return to it in the next section. But it is a premise for allalso represent opportunities for a systematic feedback from qualitative researchers that any respondent will providethe informants. information from her point of view, whether her intention The ladder of abstraction should not be taken too is to provide as complete a version of the story as possible,literally. It may give the misleading impression that case or to justify her own views. Neither the researcher nor theconstruction is linear, sequential and algorithmic, while it is informant have access to the God’s Eye point of view. Whenusually non-linear, iterative and experimental. However, reading and responding to the case description, the respon-as will be demonstrated in the following sections, the ladder dent may reframe her interpretation of it, and may also giveof abstraction as an analytical tool serves to frame our new answers. Our assessment is that this new informationdiscussion of informants’ feedback practice in the case is valuable in a double sense; it provides more informationstudy example. on the narrative and it may also reflect the learning and sensemaking process of the informant. The job of the researcher remains the same; to collect the data as faithfullyHow may an extended role of informants contribute to enhance as possible, to document the time and context in which itinterpretivist case study research in IS? was produced, and to build on all available sources whenA key observation is that informants’ feedback plays an constructing the case narrative. Not all evidence may pointimportant, but quite different role, at the various levels in the same direction in a complex case. As noted by Van deof abstraction. At the lowest level of abstraction, infor- Ven (2007), triangulation does not always converge intomants’ feedback is important as a means to verify factual one coherent story.information. This typically includes chronology, stake- Thus, we disagree with those who claim that theholders and events. The role of case members here is to cor- informants’ feedback strategy invalidates the researchrect errors and give additional information. As experienced (McDonnell et al., 2000; Morse et al., 2002). Rather, wein our case study, the process may also give important argue that if the researcher cannot reach a basic agreementinput for data collection, for example by pointing to a new on the case description with key stakeholders, this calls forstakeholder. This finding is congruent with earlier studies further analysis and reflection by the researcher on her
Exploring the role of informants B Bygstad and BE Munkvold 41interpretation of the case. If we accept that stakeholders practices as described in this paper, the researchers estab-are co-constructors of the case study, it is unreasonable lish an arena for mutual knowledge sharing with practi-that their interpretations should not be taken into account tioners and for testing the perceived practical relevance ofas important second-order data. As pointed to by Miles and their findings. Thus, while informants’ feedback at level 2Huberman (1994) (citing Blumer, 1969), ‘an alert and contributes to increase internal validity, at the discursiveobservant actor in the setting is bound to know more than level it also contributes to increase external validity, thethe researcher ever will about the realities under investiga- transferability of findings.tion’ (275). For how long should this process go on? In Correspondingly, we propose that without informants’principle, it should go on until the researcher believes all feedback the researcher may run the risk of missingrelevant information is on the table; in practice it goes on important aspects of a case study. In the airline case studyas long as the key informants are willing to respond. We the case narrative and also the papers were significantlyregard this process as steps to increase the internal validity, improved through the process of informants’ feedback. Inthe credibility, of the case narrative. the data collection phase a number of factual corrections At the third and highest level of abstraction the focus is were supplied, such as chronology, additional actors andon the implications of the case study. The role of the relevant events outside the project. In the case constructionmembers is now more discursive. While the researcher phase competing interpretations were offered on severaldraws on related research to assess the possible general- points. In addition to providing more information, theseization from the case, the case members will draw on perspectives also challenged the researcher to look fortheir previous experience and on practitioner sources alternative explanations, and to strengthen his own argu- ˚(Martensson and Lee, 2004). The examples from the airline ments. For example, the identification of one of the genericcase study cited in the previous section illustrate this, as patterns of integration was done during these discussions.the stakeholders refer to the perceived general conditions In the paper publishing phase the objections of the infor-of projects and project managers when defending their mants made the researcher reflect more systematicallypositions. This contributes to make the discourse informed on competing patterns, and helped to analyze the manage-and balanced, as an opportunity for mutual learning. rial challenges of each pattern in more depth. All in all, the In the airline case study, the project members also used changes were substantial.terms from the draft papers, such as socio-technical inte- Thus, without informants’ feedback, the researcher maygration. This illustrates that in the process of establishing lose factual information in the data collection phase. In thea common ground, the project members were influenced by construction of the case study report the text may be biasedthe researcher’s framing of the case. This does not imply that in favor of the perspectives of the researcher, excludingthey use the term in the researcher’s sense, rather that the important alternative interpretations from the informants.new shared ground constitutes a new context, in which also And in the publishing phase a lack of feedback may lead tothe researcher’s terms are appropriated. This phenomenon is reduced practitioner relevance. These potential risks arealso described by Lanzara as ‘backtalk,’ as discussed in the vividly illustrated by Fitzgerald and Howcroft (1998) inreview section. A similar reflection is made by Walsham and their anecdotal tale of the fictitious researcher Ethna O’Sahay (1999) in their case study of a Geographical Infor- Graphy, who in her idealistic attempts to ‘always lookmation System use in Indian district-level administration, beyond superficial cause-effect relationships to consider thewhere they experienced that the terms of reference for a later deeper meaning underpinning all human activity’ (315)phase in the project studied drew heavily from material ends up with losing sight of the practical realities of theprovided by the researchers in the earlier stages. research situation. Although clearly representing an extre- Table 3 summarizes our argument regarding significance mist position, this caricature underpins well the importanceof informants’ feedback in the different project phases, and of validating the researcher’s interpretation against thethe role of case members in this process. informants or stakeholders to be able to produce research Overall, we argue that these research activities, used of at least some practical relevance. In this sense, our sugg-systematically, can also contribute to increase the practical ested approach also can be seen to extend the applicabi-relevance of IS case study research. In the debate on the lity check proposed by Rosemann and Vessey (2008). Whilelack of relevance of IS research, limited exposure to their approach involves discussion with practitionerspractical contexts (Benbasat and Zmud, 2003) and lack of ex-ante and/or ex-post of the case study itself, we argueknowledge transfer between academics and practitioners for engaging in a dialogue with the informants concerning(Moody, 2000) have been stated to represent parts of the relevance of the case findings as part of constructing thethe problem. By involving the case informants through case narrrative.Table 3 Summary of argumentsProject phase Level of abstraction Role of case member Significance of informants’ feedbackData collection Low Verifying facts Increasing factuality Generating new dataCase study report Medium Co-constructing the case narrative Increasing internal validityResearch publishing High Reflecting on implications Increasing relevance
Exploring the role of informants B Bygstad and BE Munkvold42What are the conditions in which this type of approach is the explicit aim of contributing to improve socio-technicalappropriate? practices in complex projects. He shared with the keyFrom what we have written earlier in this paper it is clear respondents a theoretical and practical knowledge on thethat our approach cannot be used as a general strategy in development methodology used (RUP), the technologyinterpretive research, but is applicable in certain condi- used (web and database technology) and also a moretions. We will focus on four conditions for this approach to general knowledge on the challenges of large projects, suchbe successful. as risk management. Thus, this case study is representative First, it requires a longitudinal case study design. A of a large body of IS research projects that focus onreasonable objection to the last suggestion in Table 3 is the contributing to improve organizational practice throughquestion of how the researcher can ensure that the member analyzing the development and use of particular IS andengagement in phase 3 really is informed discourse and not related routines and processes. Some examples of domainsimply post hoc rationalization. As shown by Weick (1995), areas in IS in which this type of research is prevalent aremanagers tend to make sense of their actions only when implementation of enterprise systems, software processthey realize the consequences of them. This implies that improvement and IS project management. We argue that ininformants’ feedback conducted at a late stage in the ana- this type of IS research projects, the distance between thelysis may be ‘contaminated’ by post hoc rationalization, and researcher and the informants is often smaller than in manythus actually threaten the credibility of the case. This other disciplines. In this aspect, our findings contrast ear-problem of distortion is an important issue, and our reply is lier findings of how different language styles of resea-that it must be addressed in the research design. The short rchers and practitioners made informed discourse difficultanswer is that ‘truth is the daughter of time.’ As pointed out ˚ (Hoffman, 2004; Martensson and Lee, 2004). At the sameby Van de Ven (2007), engaged scholarship requires multi- time, we of course acknowledge that IS research alsople iterations between researchers and stakeholders, where includes other types of projects in which the condition ofthey share knowledge, discuss implications and make shared basic objectives and terminology would not bechanges where needed. Our point is that the researcher present to a similar degree. Examples of this would includemust be in control of the process. Thus, the three different critical research studies on (lack) of end user involvement,roles of informants’ feedback illustrated in Table 3 cannot and research on power and politics related to IS develop-be chosen arbitrarily. They are mutually dependent parts of ment and use (Doolin, 1998).a longitudinal study, in which data collection is performed Fourth, the informants need to engage in reflection-throughout the study. When the researcher gradually con- ¨ in-action. As defined by Schon (1983), reflection-in-actionstructs and interprets the case, she must keep track of the involves thoughtfully considering one’s own experiences intemporal dimension, in particular the time-stamp of applying knowledge to practice: ‘When someone reflects-in-various data. Project members may describe events and action, he becomes a researcher in the practice context.issues quite differently during a long project (Lanzara, [y]. He reflects on the phenomena before him, and on the1991). When the researcher puts this puzzle together, the prior understandings which have been implicit in hisphenomenon of post hoc rationalization should be known ¨ behavior’ (Schon, 1983: 68). Through this, ‘the practitionerand addressed in the analysis. For example, in the may surface and criticize his initial understanding of thediscussions with the stakeholders in the third phase of phenomenon’ and ‘construct a new description of it’ (ibid,informants’ feedback (as reported in the previous section), 63). As demonstrated in the study reported in this paper,the researcher could lean on a substantial amount of materials the approach is dependent on informants that are willingto assess whether new objections should be addressed or not. and able to interpret and discuss the case study report and the Second, some of the informants must be knowledgeable on scientific papers of the researcher, as a basis for challenginghow the case connects to other structures. They should be both their own and the researcher’s understanding.selected in order to ensure that they cover as much of the The informants in our study were highly educated andcase and its environments as possible. In this requirement experienced, and interested in spending their work timewe agree with the critique that has been raised against (though limited) in reading, discussing and commenting onconstructionist approaches, in which the emphasis is the researcher’s artifacts. We acknowledge that this may yetmainly on local contexts and situated practice (Kallinikos, be unusual in many organizations, but recent developments2004). We think it is absolutely essential that – in the actual in education and corporate practices may contribute totype of case research – that some (but obviously not all) change this. One aspect is the increasing level of educationinformants are knowledgeable on not only user aspects, but of the IS practitioners, in which many IT managers andalso on power, organization and technological structures. specialists have been trained in reading scientific papers asWithout this, knowledge informants would not be able to part of their Master’s degree. Another aspect is thegive informed feedback on suggested patterns and struc- proliferation of improvement programs in the IT industry,tures, particularly at level 3. such as Capability Maturity Model (Paulk, 1998) and action Third, the researcher and the informants should share research (Mathiassen et al., 2001), which encouragessome basic objectives and terminology. As pointed out by systematic learning and reflection.those addressing the relevance problem of IS, one part ofthe solution lies in the ability of the researcher to ‘speak thelanguage of the practitioner’ (Benbasat and Zmud, 1999; Practical limitations to informant’s feedback practices ˚Hoffman, 2004; Martensson and Lee, 2004); to facilitate a In addition to the conditions described above, there aredialogue within a terminology familiar to the case infor- a number of more practical limitations to informants’mants. The airline case was conducted by a researcher with involvement.
Exploring the role of informants B Bygstad and BE Munkvold 43 There are practical limitations related to the time frame increasing relevance through mutual learning and reflec-and scope of the activities of informants’ involvement, and tion. This study illustrates the complexity involved inthe access to different informants and stakeholders. Most informants’ feedback, which requires careful fieldwork andoften, time is a scarce resource both for the researcher and analysis. We acknowledge that the full scope of thisthe case members. Time constraints may thus limit the approach is only appropriate under certain circumstances:possibility for the informants to do an in-depth review of the research approach should be a longitudinal case study;the material, and the feedback (if any) may thus be limited some of the informants must be knowledgeable on how theto correcting factual errors. As we have experienced in case connects to other structures, the researcher and infor-several of our other research projects, the engagement from mants should share some aims and related vocabulary,and interaction with case companies during informants’ and the informants need to display characteristics of reflec-feedback may therefore often be much less than in the tive practitioners. We argue that although these conditionsairline case study presented in this paper. are certainly not satisfied in all case settings, they are Also, the case company may decide that the informants’ increasingly common in IS research.feedback is to be handled by the primary contact person Further research is needed on the practices related tothus serving as a gatekeeper in the process of assessing the informants’ involvement in different forms of qualitativeresearcher’s interpretation against those of the informants. inquiry. This may take the form of both descriptive and/orIn such cases, the feedback may run the risk of being overly normative research, and may address in more detail theinfluenced by concerns for the organization’s reputation, possible variation in these practices among different qua-rather than the goal of discussing potentially differing litative research approaches such as ethnographic studies,perspectives on the case events (Silverman, 1985; Guba and action research and critical studies. In particular, we see theLincoln, 1989; Locke and Velamuri, 2004; Walsham, 2006). need to investigate whether other research designs thanThus, the opportunity for using informants’ feedback as the Longitudinal Process Research can benefit from oura source of additional data may also be lost. proposed approach to informants’ involvement.Conclusion AcknowledgementThe point of departure for this paper was the methodolo- We thank Professor Kallinikos and the anonymous reviewers forgical challenges related to the interaction between research- their constructive and in-depth comments that helped in framinger and informant in interpretive research. As documented our discussion in this paper.in our research review, informants’ feedback is a recom-mended practice in both social science and IS research.The practical research steps involved in this process are, Referenceshowever, not described in much detail. Following Van de Avgerou, C. (2002). New Socio-Technical Perspectives of IS innovation inVen’s (2007) call for engaged scholarship, our research Organizations, in C. Avgerou and R.L. LaRovere (eds.) ICT Innovation:objective was to explore how, and under which circum- Economic and organizational perspectives, Cheltenham: Edward Elgar,stances, the informants could be involved in constructing pp. 141–161.and interpreting the case narrative. Baskerville, R. (1996). Deferring Generalizability: Four classes of generalization On the basis of the ladder of analytical abstraction as in social enquiry, Scandinavian Journal of Information Systems 8(2):the conceptual framework, and empirical findings from a 5–28. Benbasat, I., Goldstein, D.K. and Mead, M. (1987). The Case Researchlongitudinal case study, we find that the role of informants’ Strategy in Studies of Information Systems, MIS Quarterly 11(3): 369–385.feedback varies, and should be treated differently, during Benbasat, I. and Zmud, R.W. (1999). Empirical Research in Informationthree steps of analysis. Our findings can be summarized as Systems: The practice of relevance, MIS Quarterly 21(1): 3–16.follows: Benbasat, I. and Zmud, R.W. (2003). The Identity Crisis within the IS Discipline: Defining and communicating the discipline’s core properties, In the data collection phase, informants’ feedback can MIS Quarterly 27(2): 183–194. increase the quality of the case documentation through Bloor, M. (1978). On the Analysis of Observational Data: A discussion of the factual verification and the generation of new data. This worth and uses of inductive techniques and respondent validation, Sociology corresponds to the ‘traditional’ perspective on infor- 12(3): 545–557. mants’ feedback. Bloor, M. (2001). Techniques of Validation in Qualitative Research: A critical commentary, in R. Emerson (ed.) Contemporary Field Research: Perspectives In the case study report, informants’ feedback may enrich and formulations, 2nd edn, Long Grove, IL: Waveland Press Inc, pp. 383–395. the case narrative and increase the internal validity of Blumer, H. (1969). Symbolic Interactionism: Perspective and Method. the study. In this phase, the role of the informant is to Englewood Cliffs, NJ: Prentice-Hall. co-construct the case narrative and its interpretation. Boland, R. (1985). Phenomenology: A preferred approach to research on In the dissemination phase, the role of the informants is information systems, in E. Mumford (ed.) Research Methods in Information more discursive, contributing to increase the relevance Systems, North-Holland: Amsterdam, pp. 181–190. Bryman, A. (1988). Quantity and Quality in Social Research, London: Unwin of findings, the external validity, by assessing the con- Hyman. clusions and implications of scientific publications in a Butler, T. (1998). Towards a Hermeneutic Method for Interpretive Research practitioner context. in Information Systems, Journal of Information Technology 13: 285–300. Bygstad, B. (2006). Managing Socio-Technical Integration in IterativeOur contribution is practical in nature, intended as an Information System Development Projects, International Journal ofextension to the interpretive case study methodology Technology and Human Interaction 2(4): 1–14.literature. We argue that informants’ involvement, when Bygstad, B. and Nielsen, P.A. (2003). The Meeting of Processes, Proceedings ofused systematically, can enhance IS case study research by IRIS 26 (Helsinki, Finland).