WHEN I WRITE MY MASTERPIECE: THOUGHTS ON WHATMAKES A PAPER INTERESTING

STEPHEN R. BARLEYStanford University

Academic papers are a bit like rock and rollbands: whether an audience finds them interestingis a matter of perspective, if not taste. We all knowthere’s no accounting for taste. There’s no unanim-ity of taste, either. This is why readers seem todisagree vociferously with every list of the Top 50albums Rolling Stone publishes. For proof that aca-demia suffers similarly diverse tastes, one need gono further than the poll of AMJ review board mem-bers that motivated my writing this paper. Al-though the count of nominations for “most inter-esting paper” indicated that I had more papersnominated than anyone else, the honor was thinlywon. The margin of decision was a small handful ofvotes. In fact, if just one ballot had fallen victim toa cluttered inbox, this moment in history mighthave turned out differently. Although I am honoredby the results, the caveat is clear: You should beskeptical of any authority that might accrue to meas a result of the balloting. On this score, George W.Bush and I finally have something in common—although my election, however narrow, did actu-ally reflect a plebiscite.

There is a second similarity between academicpapers and rock and roll bands. One might thinkthat finding a band’s music interesting would besynonymous with liking it, but we all know better.For me, at least, King Crimson, Nirvana, and Pri-mus are quite interesting, and I appreciate whatthey are (or were) trying to do. But I do not likethem! I own one CD by each band; I’ve listened toeach CD just once; and I have no intention of ac-quiring more music by any of them. On the otherhand, I possess hundreds of recordings of GratefulDead concerts as well as every album and CD theDead ever cut (in some cases, multiple copies).Hardly a day goes by without my listening to atleast one of these recordings. My oldest son evenjokes, with good reason, that until he went to col-lege the Dead provided the soundtrack for his life.Yet I never think of the Dead as being interesting.The Dead and I are way beyond that! So it is withacademics. Whether our colleagues find our papersto be “interesting!” (I loved it!) or “interesting” (Iread it, appreciated it, and shelved it!) is signaledby an intonation that anyone who has heard com-

ments muttered in hallways after academic talkscan easily mimic.

Rock also teaches us that being interesting doesnot imply being important. Consider the strangefascination the music of Tiny Tim, Herman’s Her-mits, Napoleon XIV, and the Count Five once heldfor some people. If you’ve never heard of thesemusicians, consider yourself lucky. Your ignoranceis bliss and proves my case. Tiny Tim brings me tomy second caveat: when it comes to being interest-ing, you should be careful what you wish for.

Finally, rockers and academics share anothercharacteristic: a peculiar kind of cluelessness. Al-though many people can teach you how to playguitar, no one can teach you how to play guitar like,say, Jerry Garcia, including Garcia himself (even ifhe were still alive). Whatever it was that allowedGarcia to play like Garcia was tacit: A feel, ormaybe a sensibility. I don’t think Garcia or anyother musician (including Robert Fripp!) ever wokeup and said, “Today, I’m going to play somethinginteresting.” It just happened or it didn’t. Yet, witha little effort and the right vocabulary, all of uscould describe what it is about a virtuoso’s stylethat strikes us as interesting.

The same is true of academic writing. I know ofno scholars who can will themselves to write inter-esting papers. I also doubt that anyone can tell ushow to write a paper that others will find interest-ing, although many scholars can teach us to writewell. The most any scholar can do is describe thebroad attributes of the papers that he or she hasfound interesting and then provide examples. Thisis all that I can hope to do too. Although a sufficienthandful of people apparently think that I have writ-ten interesting papers, there is no way I can explainhow I might have done what they believe I did,because I don’t understand it myself. Thus, myfinal caveat: If you expect to learn how to write aninteresting paper by reading this essay, you shouldstop reading now and go listen to rock and roll.

Difference Is the Root of All Interest

The otherwise diverse papers that I have found tobe interesting over the years have one commondenominator: They differed in some significant and

! Academy of Management Journal2006, Vol. 49, No. 1, 16–20.

16

striking way from most of the other papers in aca-demic journals. For this reason they captured myattention, like scarlet begonias against a sea of gray.I might have loved these papers or I might have not.I might have agreed with their authors or I mighthave not. I might have thought the papers impor-tant or I might have not. But one thing is certain:because each of these works stood out like figure onground, I haven’t forgotten it.

Subject matter. Differences of several kindsseem to pique my interest in a paper. Some articlesinterest me simply because they address subjectsthat depart noticeably from the mainstream. Likedire wolves, researchers run in packs. Thus, thepapers that appear in journals during an era oftencluster around a relatively small set of topics andconversely, papers written on particular topicstend to cluster in time. Robert K. Merton (1973) andDerek De Sola Price (1986) argued that such clus-tering occurs in the physical sciences becausemembers of invisible colleges agree on which ques-tions and problems are currently important fortheir field’s further progress. The same thing mayalso be true of the social sciences. Or it may simplybe that topics fall in and out of fashion as research-ers first crowd into areas and then eventually be-come bored. Either way, papers on rogue topics arespoonfuls of gold.

Mauro Guillen’s (1997) “Scientific Manage-ment’s Lost Aesthetic” is a paper that interests mefor its subject matter, although it qualifies on othergrounds as well. (For instance, Guillen uses themethods of a historian, which is rare in organiza-tion studies.) Contemporary management scholarswho mention Frederick Taylor and “scientific man-agement,” for instance, usually go little beyond ac-knowledging that scientific management was thefirst “management theory.” A few others argue,along with Braverman (1973), that Taylor providedthe ideology that has justified deskilling labor.Guillen reminds us that scientific management wasmuch more: it was a worldview with influence farbeyond the shop floor.

Specifically, Guillen shows how scientific man-agement influenced the aesthetic of modernist ar-chitecture. He substantiates his claim not only bydrawing on the writings of well-known modernists,like Gropius and La Corbusier, but also by demon-strating that modernism only became influential incountries where architects trained beside (or as)engineers. Although scientific management and itsimitators were enthusiastically embraced in theUnited States (Taylor’s home) and in Great Britain,modernist architecture only appeared in thesecountries after it spread across the Channel and theAtlantic from Continental Europe. Guillen claims

that modernism followed this diffusion pattern be-cause American architects were trained in schoolsof architecture, where engineers could not contam-inate them, and because in Britain, most engineersat the time had no formal training (also see Whal-ley, 1986). Guillen’s paper makes me think thatinteresting academic papers in organization studiesmay need no clear relevance for management.Whether this is often the case is an empiricalmatter.

Methods. All too often in graduate classes wecelebrate papers for their methods. Sometimes, Ithink this is a bit like admiring Mount Rushmorebecause of the dynamite and the shiny black steeljackhammers that chipped the faces out of the cliff.But the fact is, some papers are interesting pre-cisely because their methods are so different fromthe ubiquitous secondary data sets, attitude sur-veys, and interviews of top managers that providemost of the grist for our field. Particularly interest-ing to me are methods that get close enough tobehavior to show how people wittingly or unwit-tingly build and maintain their social worlds.

A recent example is David Gibson’s (2005) “Tak-ing Turns and Talking Ties.” In what may be a first,Gibson combines the tools of conversation analysiswith the methods of network analysis. In otherwords, he combines the techniques of the most“micro” of all sociologies with the tools of some ofthe most “macro” of sociologists. To put it yet an-other way, Gibson demonstrates that mixing whatothers see as apples and oranges is not alwaysfruitless. Gibson devises a system for exhaustivelyclassifying the turn-taking sequences in a conver-sation, which enables him to categorize the partic-ipation shifts (“p-shifts”) that marked the in situdiscussions of ten groups of managers who rou-tinely worked together. Gibson shows that manag-ers’ p-shifts mapped to their positions in the net-works formed by their friendship, coworker, andreporting relationships. In short, Gibson demon-strates that preexisting social relationships are cor-related with—and may actually influence—the sit-uated dynamics of group decision making. Gibson’spaper interests me not so much because of hisconclusion, but because he sings a rare and differ-ent tune. He combines techniques typically used byresearchers with vastly different and in some waysopposing perspectives to speak empirically to theissue of how actions and structures are entwined.

Orlikowski and Yates’s (1994) “Genre Repertoire:The Structuring of Communication Practices in Or-ganizations” is another paper that I find method-ologically interesting. Orlikowski and Yates use theconcept of a genre, which they draw from literarycriticism, to study the emergence, stabilization, and

2006 17Barley

modification of the e-mail practices of the group ofcomputer scientists who designed Common LISP, acomputer language of considerable significance inthe artificial intelligence community. The raw datafor the study were the group’s e-mail archive ofover 2,000 messages sent between December 1981and December 1982. By analyzing these messages,Orlikowski and Yates were able to identify distinctgenres (or forms) of e-mail and the points in timewhen these genres emerged. By relating the genres’appearance to important events, challenges, andmilestones in the LISP community’s history, Or-likowski and Yates were able to show how thegenres contributed to this community’s organizingand development.

Gibson’s and Orlikowski and Yates’s papersshare an important trait: both employ the methodsof a field generally thought to be outside the bound-aries of organization studies (respectively, conver-sation analysis and literary criticism) to shine alight on organizing. I do not know whether integrat-ing tools from disparate disciplines is a hallmark ofall methodologically interesting papers, but I doknow that others have claimed that bringing to-gether ideas or objects from previously unrelateddomains is a hallmark of innovation (Basalla, 1988;Hargadon & Sutton, 1997).

Theory. Finally, some papers have interested mebecause they propose theories, or at least perspec-tives, that differ from what has gone before. Thesepapers tend to be less common than papers withinteresting methods or subjects, in part becausetheory building is an arena in which books excel. Apaper is usually too short to provide adequatespace for a full accounting of “why,” especially ifthe primitives, logic, corollaries, and implicationsof a theory are complex. Thus, rather than forgefull-fledged theories, interesting “theoretical” pa-pers generally propose new models or metaphorsthat let us either see what we didn’t see before orsee in a new light what we thought we alreadyunderstood.

Compared to other types of papers, an interestingtheoretical paper may have a better chance of be-coming famous, infamous, or both. I also suspectthat writing theory papers is probably the hardestway to be interesting. They require more than anose for a good problem or the proclivities of ajack-of-all-trades. Authors of this kind of papermust see things differently, systematize their vi-sions, and then communicate them in language thatis at once comprehensible and persuasive. Scholarswho succeed on this front are likely to attract dis-ciples as well as detractors, because paradigms,whether big or small, are usually at stake.

Hannan and Freeman’s (1977) first major paper,

“The Population Ecology of Organizations,” isprecisely such a treatise. It is worth observingthat even though Hannan and Freeman use sev-eral equations, their paper contains no data ordata analysis. Instead, the paper presents a logi-cal and rhetorical case for applying the perspec-tive, concepts, and tools of bioecology to organi-zation studies. Hannan and Freeman argue that itis both plausible and fruitful to think in terms ofpopulations of organizations and to approachsuch populations using such ecological conceptsas competition, niche, environmental capacity,fitness, generalizing, and specializing. In otherwords, Hannan and Freeman invite us to see or-ganizations in a strange and new way. Their lensis a new metaphor that draws attention awayfrom individual firms toward sets of organiza-tions in competition over resource spaces.Whether readers agree or disagree with whatHannan and Freeman call “the ecological per-spective” (1977: 929), one cannot come awayfrom this paper without at least contemplatingorganizations in a way one has never thought ofbefore.

James Barker’s (1993) “Tightening the Iron Cage”is another paper that I find interesting on theoreti-cal grounds. Barker employs observations of self-managed teams on a factory floor to develop agrounded theory of “concertive control.” At thetime he collected his data, self-managed teamswere a crucial and popular component of the totalquality management (TQM) movement that wassweeping industry. TQM portrays self-managedteams as essential for decentralizing decision mak-ing by bringing it to where problems occur and forgranting workers both autonomy and responsibilityto solve them (as had sociotechnical systems theoryand the quality of work life movement years ear-lier). Barker was one of the first researchers to studysuch teams in situ and over time. He discoveredthat teams replaced supervisory control with peercontrol and that peer control was subtler, moreeffective, and potentially more coercive than super-visory control, because workers now policed eachother in the service of their organization’s goals andobjectives. As any high school student can tell you,peer pressure is always harder to resist than thepressure of an authority. Readers came away fromBarker’s paper viewing self-managed teams in anentirely different light. Barker’s paper also requiredan extension, if not a modification, of traditionalMarxist thought on control in organizations. Thus,in one fell grounding, Barker managed to reframeboth mainstream and critical theories of control—not bad for a single paper.

18 FebruaryAcademy of Management Journal

Limits on Transgression

Although I’ve argued that interesting papers usu-ally transgress the status quo, there are limits onhow far transgression can go. Papers that break toomany substantive, methodological, or theoreticalrules are more likely to be called flaky or wrong-headed than interesting. At minimum, interestingpapers need to conform to genre constraints. Em-pirical papers need to flow from introduction toproblem statement to methods to data and then to adiscussion and conclusions. Theoretical papersneed to work though implications of propositionsand consider counterarguments. In both cases,readers expect authors to warrant their claims inways that scholars find legitimate: with logic,mathematical models, data, and counterfactuals,for example. Without such warrants, a paper tooclosely resembles opinion, and when it seems to bemere opinion, a paper is unlikely to survive aca-demic skepticism long enough to have a chance tobe considered interesting.

In my experience, failing to conform to acceptedcanons of warranted claims is the qualitative re-searcher’s and the theoretician’s Achilles’ heel.Methods for designing quantitative studies and an-alyzing quantitative data come with built-in safe-guards for warranting claims. Given that quantita-tive researchers have the edge on this score, I thinkit is intriguing that of the 17 papers that receivedmore than one vote in the AMJ poll, 11 (65%) reliedsubstantially on qualitative data. Given that somany academics admonish graduate students toavoid doing qualitative work, how can it be that somany of our interesting papers are qualitative?

There is certainly some wisdom to warning grad-uate students to eschew qualitative work. Qualita-tive studies seem to have a higher rate of failure (inthe sense of being rejected from journals) than doquantitative studies, in part because they are harderto execute and lack clear genre constraints. Never-theless, if a qualitative study is rigorously done, Isuspect that it is more likely to yield importantdiscoveries than a quantitative study, if for no otherreason than this: qualitative researchers often dis-cover something because they usually approachtopics with little clue as to what they’ll find.

Readers may also find qualitative research inter-esting because qualitative researchers, having al-ready departed from mainstream methods, haveless to lose by studying odd topics and taking the-oretical risks. As Nancy Reagan’s “just say no” cam-paign implied, once you step over the first line, it’seasier to step over the second. Then again, all thatmay be going on is a trick of memory: qualitativeresearch may simply yield vivid and involving

tales that readers are more likely to remember whenquestioners ask them to nominate interesting pa-pers. Although I cannot adjudicate among theseexplanations, I am reminded of an examinationquestion that James March reputedly once askedstudents at Stanford: “Name one paper that hasmade a substantial theoretical contribution to ourfield that also contained a regression equation.”

Some Final Conjectures

It is worth contemplating whether a paper’s ca-pacity to be interesting decays over time. Will stu-dents 20 years from now still find interesting thosepapers that interest us now? Suppose an interestingpaper draws hordes of researchers to a topic,method, or theory. Over time, the topic, method, ortheory will become, by definition, mainstream. Al-though the paper may remain famous, I suspect thatyoung scholars schooled in the new status quo willfind the paper far less interesting than did theirmentors, unless it also has attributes that transcendits content and time. For example, the paper’s au-thor might have been such a clever writer that herwords remain interesting long after her contribu-tion has become blase. For my money, the essays ofEverett C. Hughes (1958) are like this. Their contentis worn and holds few surprises for those who haveread later scholarship on work and the professions.Nevertheless, Hughes’s prose is so sprightly andwell crafted that his essays twinkle with an en-chantment that the work of few researchers attains.

Still, I suspect that Hughes is an exception thatproves the rule: Interesting papers that start asuccessful line of inquiry are likely to lose lusterfor all but those who were drawn to the workwhen it was fresh. Originally interesting but lesssuccessful papers might actually remain moreinteresting to young scholars should they happento come across them in some scholar’s cut-outbin. Academic papers and rock and roll bandsthus share one more similarity: most just fadeaway.

Finally, we should consider whether we wouldwant all papers published in our journals to beinteresting. To wish otherwise might at first seemfoolish. Wouldn’t it be nice to open up an issue ofAMJ, AMR, or ASQ to a random paper knowing thatour reading would soon transport us to some peakof illumination or discovery? I certainly would liketo be transported a little more often!

But what if after reading the essays in this edito-rial, most of our colleagues committed to writinginteresting papers and succeeded? If being interest-ing requires a paper to be different, before long thefield would be a mess. Every paper would take on a

2006 19Barley

new topic, devise a new method, or offer a new wayof seeing things. With all of us so busily striving forthe next interesting paper, no subjects would bestudied more than once, no methods would be re-fined, and no ideas would be worked though. Thedevelopment of knowledge, at least in any scien-tific sense, would all but cease. Worse yet, becausethere would be no status quo to provide a measureof which new papers were interesting, the fieldwould implode into humdrum. At that point onlyby taking the risk of sticking doggedly to a topic,method, or theory could scholars rescue us fromthe quicksand of being interesting. In the endmaybe we are quite lucky that interesting papersonly come along every so often and that no one cantell us how to write more interestingly. If the worldwere made of candy, there could never be a WillyWonka!

REFERENCES

Barker, J. R. 1993. Tightening the iron cage: Concertivecontrol in self-managing teams. Administrative Sci-ence Quarterly, 38: 408–437.

Basalla, G. 1988. The evolution of technology. Cam-bridge, U.K.: Cambridge University Press.

Braverman, H. 1973. Labor and monolpoly capital. NewYork: Monthly Labor Review Press.

Gibson, D. R. 2005. Taking turns and talking ties: Net-works and conversational interaction. AmericanJournal of Sociology, 110: 1561–1597.

Guillen, M. 1997. Scientific management’s lost aesthetic:Architecture, organization, and the taylorized beautyof the mechanical. Administrative Science Quar-terly, 42: 682–715.

Hannan, M. T., & Freeman, J. H. 1977. The populationecology of organizations. American Journal of So-ciology, 82: 929–964.

Hargadon, A., & Sutton, R. I. 1997. Technology brokeringand innovation in a product development firm. Ad-ministrative Science Quarterly, 42: 716–749.

Hughes, E. C. 1958. Men and their work. Glencoe, IL:Free Press.

Merton, R. K. 1973. The sociology of science: Theoreti-cal and empirical investigations. Chicago: Univer-sity of Chicago Press.

Orlikowski, W. J., & Yates, J. 1994. Genre repertoire: Thestructuring of communicative practices in organiza-tions. Administrative Science Quarterly, 39: 541–574.

Price, D. J. D. S. 1986. Little science, big science . . . andbeyond. New York: Columbia University Press.

Whalley, P. 1986. The social production of technicalwork. Albany: State University of New York Press.

Stephen R. Barley (sbarley@stanford.edu) is the CharlesM. Pigott Professor of Management Science and Engi-neering, the codirector of the Center for Work, Technol-ogy and Organization at Stanford’s School of Engineer-ing, and the codirector of the Stanford/General MotorsCollaborative Research Laboratory. He holds a Ph.D. inorganization studies from the Massachusetts Institute ofTechnology. In collaboration with Gideon Kunda of TelAviv University, Barley recently published a book oncontingent work among engineers and software develop-ers, entitled Gurus, Hired Guns and Warm Bodies: Itin-erant Experts in the Knowledge Economy.

20 FebruaryAcademy of Management Journal