This article was downloaded by: [131.151.244.7] On: 06 November 2014, At: 22:27Publisher: Institute for Operations Research and the Management Sciences (INFORMS)INFORMS is located in Maryland, USA

Organization Science

Publication details, including instructions for authors and subscription information:http://pubsonline.informs.org

Coordinating Expertise Among Emergent GroupsResponding to DisastersAnn Majchrzak, Sirkka L. Jarvenpaa, Andrea B. Hollingshead,

To cite this article:Ann Majchrzak, Sirkka L. Jarvenpaa, Andrea B. Hollingshead, (2007) Coordinating Expertise Among Emergent GroupsResponding to Disasters. Organization Science 18(1):147-161. http://dx.doi.org/10.1287/orsc.1060.0228

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial useor systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisherapproval, unless otherwise noted. For more information, contact permissions@informs.org.

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitnessfor a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, orinclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, orsupport of claims made of that product, publication, or service.

Copyright © 2007, INFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, managementscience, and analytics.For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org

OrganizationScienceVol. 18, No. 1, January–February 2007, pp. 147–161issn 1047-7039 �eissn 1526-5455 �07 �1801 �0147

informs ®

doi 10.1287/orsc.1060.0228©2007 INFORMS

Coordinating Expertise Among EmergentGroups Responding to Disasters

Ann MajchrzakMarshall School of Business, University of Southern California, Los Angeles, California 90089, majchrza@usc.edu

Sirkka L. JarvenpaaCenter for Business, Technology and Law, McCombs School of Business, University of Texas at Austin, Austin, Texas 78712,

sirkka.jarvenpaa@mccombs.utexas.edu

Andrea B. HollingsheadAnnenberg School for Communication, University of Southern California, Los Angeles, California 90089, aholling@usc.edu

In the aftermath of catastrophic events, when plans for organized and timely response break down, impromptu groupsoften emerge to provide disaster relief. Much remains to be learned about the internal dynamics of these emergentresponse groups whose representatives may include members from organizations with relief missions; private sector organi-zations offering resources; and private citizens with the information, relationships, or physical and mental stamina to help.Organizational theories have the potential to contribute to a better understanding of emergent response groups and howthey efficiently coordinate knowledge, people, resources, tasks, and technology, thereby substantially improving disasterresponse for future catastrophes. We apply one organization science theory toward better understanding of these groups—transactive memory systems theory—which is a theory about knowledge coordination in groups. Our application of thistheory to emergent response groups requires extending the theory in three ways: the role of expertise in task assignment,how groups function when credibility in member expertise cannot be validated, and how expertise is coordinated. Bydemonstrating how transactive memory systems theory can be extended to the unique operating conditions of emergentresponse groups, we hope to inspire organization science researchers to accept the challenge of adapting their theories tostudy this important problem of our time.

Key words : emergence; tranactive memory systems; coordinating catastrophes

1. IntroductionIn the immediate aftermath of recent large-scale disasterssuch as Hurricanes Katrina and Rita, the SoutheasternPacific tsunami, and the Pakistani earthquake, govern-ment emergency response plans failed to meet the urgentneeds of those affected (see Exhibit 1). Hurricane Kat-rina, a disaster nearly 20 times larger than any previousnatural emergency in the United States, led to break-downs in emergency response at all levels of government(GAO 2005, Time 2005). In disasters of large scale andscope, formal plans break down in unexpected ways asthe disaster unfolds. Authority structures and communi-ties react in unforeseen ways. Planned communicationlinks break down. Information about the disaster arrivesat a pace, level of detail, level of credibility and con-nectedness, and across a variety of sources that rapidly

make any planned response too slow, disconnected, andinadequate for the task. In such operating environments,disaster researchers have recognized the importance ofemergent response groups, i.e., groups with no preexist-ing structures such as group membership, tasks, roles, orexpertise that can be specified ex ante (e.g., Drabek andMcEntire 2003, Tierney et al. 2001, Tierney and Trainor2004). These groups are distinctly different from disasterresponse groups that operate with preexisting structuresand that have experience working together on a varietyof tasks, such as police or firefighters (e.g., Bigley andRoberts 2001).Emergent response groups are characterized by a sense

of great urgency and high levels of interdependence,operating in environments that are constantly changingas new information arrives about needs for victims and

147

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: Perspective148 Organization Science 18(1), pp. 147–161, © 2007 INFORMS

Exhibit 1 An Unanswered Cry for Help

“On the morning of August 29th, I received a call that I will neverforget, and once I tell you about it, I hope you will never forgetit either. My friend and colleague, former appointee to the SocialSecurity Administration, Susan Daniels called me to enlist my helpbecause her sister-in-law, Benilda Caixetta, a New Orleans residentwho was quadriplegic, paralyzed from the shoulders down, hadbeen unsuccessfully trying to evacuate to the Superdome for twodays. Despite repeated requests to be evacuated, in her powerwheelchair, which is a vital tool for mobility and independence, theparatransit system that serves the transportation needs of peoplewith disabilities never showed up. In my naiveté I thought a fewphone calls to the ‘right’ people would help, and I was sure I knewwho to call. I was wrong. After many calls to the ‘right’ people, itwas clear that Benny was NOT being evacuated. I stayed on thephone with Benny for most of the day, assuring her that I was doingall I could to make sure help would be coming as soon as possible.I was on the phone with her that afternoon when she told me, withpanic in her voice, “the water is rushing in” and then her phonewent dead. We learned five days later that she had been found inher apartment, dead, floating next to her wheelchair.”

Note. Statement of Marcie Roth, Testimony Before the Subcommit-tee on Oversight, House Committee on Ways and Means, Decem-ber 13, 2005.

resources (Drabek and McEntire 2003). This volatilityand the need to adapt create unstable task definitions,flexible task assignments, fleeting membership, and pur-suit of multiple simultaneous, possibly conflicting pur-poses. Group members come and go as they have volitionand resources to help, making membership in the groupfleeting and often unclear, and resembling swarms ratherthan traditional groups. Members are likely to representa diverse set of perspectives (e.g., firefighting, home-ownership, military command, independent boat own-ership, animal rights activism) that is ever changing.Members are unlikely to know one another before theemergency, and may never see one another again after-ward. Members may operate remotely from one chas-ing distributed resources and focusing the resources onthe needs of the victims regardless of where the victimsare located. Members may represent not only their ownaltruism and self-interest in helping, but also the interestsof their institutions, which may sometimes conflict withthe interests of other institutions during these momentsof great need.It is impossible to predict which organizations will

and will not engage in disaster response; what tasks,people, and knowledge are needed; and how expertisewill be coordinated in an emergent group. However, wecan learn much from examples of emergent responsegroups that efficiently coordinated expertise in the after-math of Hurricane Katrina. As Exhibit 2 illustrates,social networks were tapped to locate an electronic pay-ment service provider that not only had capabilities (asdid the other parties already contacted), but also the voli-tion to perform under the most resource-scarce condi-tions. Exhibit 3 illustrates that, within hours of Hurri-cane Katrina’s landfall in Louisiana and Mississippi, a

Exhibit 2 Issuing Money in Emergent Situations

The Red Cross had been using LocalTexasBank to administerdebit-card-based relief for Katrina and Rita disaster victims. Theprocess was slow, involving the Red Cross first issuing a check tothe victim and then the victim presenting the check to LocalTexas-Bank to obtain a debit card. As the scope of the Katrina disasterbecame known and the town was flooded by evacuees, LocalTex-asBank became increasingly unable to process and deliver thedebit cards in a timely fashion. “The bank was capable of issuing100 cards a day, when we needed about a 1,000 a day,” remarkeda Red Cross representative. Sharing his frustration to an officerat another financial institution, the officer was unable to help butreferred the Red Cross to a long-term customer, EPAY, a small pri-vately held electronic payment organization. The officer thoughtEPAY might be able to help because it issued debit cards for theprivate sector and had an aggressive strategy to expand nation-wide. EPAY’s CEO had personally lived through a hurricane severalyears earlier and wanted to help, so he offered EPAY’s servicesat no cost. EPAY and the Red Cross had no prior working experi-ence, nor had EPAY any experience with large-scale disaster reliefefforts. Nevertheless, on a handshake, the relationship between theRed Cross and EPAY was formed. The process was streamlined inaction. Instead of waiting to get a check from the Red Cross andthen presenting the check to LocalTexasBank to receive the debitcard, the Red Cross provided information electronically on quali-fied victims to EPAY and EPAY issued a debit card to the victims.EPAY provided e-mails to the Red Cross with daily issuance totals.Both the Red Cross and EPAY reported results that went beyondtheir expectations.

Notes. LocalTexasBank and EPAY are pseudonyms. Compiled frominterviews of Steven S. Eastland with Red Cross on 12/20/2005 and1/07/2006; a case study of Disaster Strikes: Aftermath of HurricaneRelief. University of Texas McCombs School of Business.

KatrinaHelp Wiki emerged. With the independent effortsof hundreds of people across many continents, the Wikiprovided lists of shelters, government resources, animalrescue resources, the latest health and safety informa-tion, and a people-finder service that helped to coordi-nate rescue, recovery, and relief efforts. Exhibit 4 illus-trates how the U.S. Coast Guard collaborated with civil-ian boat operators, the National Oceanic and Atmo-sphere Administration (NOAA), the U.S. Army Corpsof Engineers, and the U.S. Maritime administration inunexpected ways to rescue an estimated 22,000 peoplein the urban areas affected by the hurricane, as well asto reopen waterways to maritime traffic.Efficient coordination of expertise has been studied

in many organizational contexts such as software devel-opment (Faraj and Sproull 2000) and emergency med-ical response units (Faraj and Xiao 2006) where, inthese contexts, the group membership, tasks, roles, andknowledge can be specified ex ante. Although the con-ventional indicators of efficient coordination—expertisespecialization, credibility in expertise, and coordinationof expertise—are relevant in disaster response, disasterspresent a unique operational environment. Disasters are“events, observable in time and space, in which societiesor their subunits (e.g., communities, regions) incur phys-ical damages and losses and/or disruption of their routine

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: PerspectiveOrganization Science 18(1), pp. 147–161, © 2007 INFORMS 149

Exhibit 3 Katrina Wiki

Within hours of Katrina’s landfall, the KatrinaHelp Wiki was created.Internet-based Wiki technology allows anyone to add content in theform of web pages and reorganize existing content on those webpages from their web browser. The notion of a KatrinaHelp Wikisite was simultaneously in many different people’s minds, but gotits start as Rob Kline’s initial e-mail was quickly echoed by oth-ers, many of whom were members of the TsunamiHelp Blog andTsunamiHelp Wiki team, started after the December 2004 IndianOcean tsunami disaster. Rudi Citibrasi, a student in Amsterdamstarted hosting the KatrinaWiki server space. An emergent groupof people started adding any knowledge they had that might helpothers: names of people they knew were missing, names of peoplerecently found, software that searched and matched missing andfound names and notified people, addresses of shelters, directionsto the shelters, job opportunities for displaced residents, types ofjobs that victims needed or skills they had for jobs they wanted.The site quickly contained lists of shelters, government resources,animal rescue resources, the latest health and safety information,a people finder service, lists of job opportunities for displacedresidents, activities for children in the affected areas, ham radioresources, fundraising events, and even a life and death sec-tion for immediate assistance—all added by an emergent continu-ously changing set of participants. Only a few people consistentlyremained involved in the site over time, with most of the emergentgroup contributing to the site joining and leaving the group as siteand personal needs dictated. Four days after Katrina hit land, thewebsite received a million hits a day, overloading Rudi’s server. SiteMeter joined the emergent group agreeing to host the site. Whenit became clear that people without Internet access needed theinformation offered by the site. Skype joined the emergent group tostaff a Skype help line connected to the site.

Note. Compiled from http://katrinahelp.info/wikiarchives.katrina.asiaquake.org and www.skypejournal.com.

functioning” (Kreps 1984, p. 312). Disasters have wideimplications for expertise coordination because the pre-conditions known to facilitate expertise coordination arelimited or nonexistent in disaster response. Such precon-ditions include but are not limited to, a shared goal; aclear reward structure; known group membership, exper-tise, and skills to accomplish the task; and time to sharewho knows what. Despite extensive disaster researchon emergent response groups, Drabek, a noted disasterresearcher pioneering the notion and value of emergentresponse groups, in fact lamented that “we still lackmuch insight into the internal dynamics of these emer-gent organizations” (1986, p. 161). The integration ofresponders and resources in an adaptive manner is gen-erally not well understood (Trainor 2004).Organization science theories have contributed much to

our understanding of the dynamics of organizations in gen-eral, and of groups in particular. However, groups gen-erally studied by organizational scientists rarely face thelevel of unpredictability, urgency, and reconfigurabilityrequired by emergent response groups. Hence, muchopportunity exists to apply and extend organizationaltheories to this unique setting. In addition to contributing

Exhibit 4 Waterways

Twenty-two thousand people in a hazardous and constantly chang-ing urban disaster environment needed to be rescued by waterin the immediate aftermath of Katrina. The Coast Guard operatingoutside of its normal scope of work, in conjunction with an emer-gent and ephemeral flotilla of civilian boat operators convergingon the heavily damaged areas, both on their own initiative and inresponse to a call for assistance by political leaders, did the job.The ability of Coast Guard operational commanders to act rela-tively autonomously in the field, and the development of a sharedvision of what was necessary by both Coast Guard and civilianboat operators facilitated the ability to improvise at a multiorganiza-tional level. Not only did people need to be rescued from the water-ways, but the waterways needed to be reopened. With few pre-planning documents in place, the Coast Guard, U.S. Army Corpsof Engineers, National Oceanic and Atmospheric Administration[NOAA], and U.S. Maritime Administration collaborated in unprece-dented ways to get waterways reopened to maritime traffic. NOAA’snavigation response teams performed emergency data collection,conducted sonar surveys to update government navigation charts,coordinated navy dive teams to check for hazardous obstructionsin the waterways and provided mapping support. Working with theCoast Guard, NOAA repositioned, repaired, and replaced naviga-tion aids such as signal buoys and channel markers. For example,NOAA performed a survey that identified a channel obstruction inMobile, AL that prevented a shipload of coal to a fuel-starved powerplant in Southeast Mississippi. NOAA and the Coast Guard workedtogether to help Drummond load and navigate its coal vessel insuch a way to avoid the obstruction, ensuring electricity for thecustomers in one of the hardest hit areas of the Gulf.

Note. Compiled from Nagle (2005) and Wachtendorf and Kendra(2005).

to an important problem of our time, theoretical exten-sions can provide value to organizational theories morebroadly by pushing the definitional limits and theoreticalboundaries beyond traditional organizations.In this article, we focus on emergent response

groups and how they coordinate expertise. We firstbriefly review the disaster research literature on emer-gent response groups to underscore the need for suchgroups—despite federal preparedness plans—and theneed to understand the internal dynamics of emer-gent response groups, specifically expertise coordina-tion. Transactive memory systems (TMS) theory is oneorganization science theory focused on understandinghow expertise is coordinated in groups. We apply TMStheory to expertise coordination in emergent responsegroups and find that extensions to the theory are needed.Our extensions on group-level expertise coordinationhave implications for research topics in organization sci-ences in general.

2. The Need for Emergent GroupsResponding to Disasters

Consistent with other large crises (e.g., Dynes 1983,Dynes and Tierney 1994, Neal and Phillips 1995, Wach-tendorf 2004), Hurricane Katrina demonstrated how for-mal systems fail to respond and how the consequences

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: Perspective150 Organization Science 18(1), pp. 147–161, © 2007 INFORMS

of failed plans can be severe. In the United States, theNational Preparedness System, consisting of the NationalResponse Plan, the National Incident Management Sys-tem, and the National Preparedness Goal, establishesstructures to bring federal, state, local, and nonprofitorganizations together in the event of a disaster. TheNational Preparedness System specifies that an incidentcommand manager and an emergency management officeplay central roles in assigning tasks, managing resources,and making decisions that are disseminated through achain of command (Anderson et al. 2004, DHS 2004).The formal system assumes that responding organiza-tions are known ahead of time, have trained together,and are available when needed (Bigley and Roberts2001, Trainor 2004). Despite the existence of these for-mal plans, extensive training, and bureaucratic structures,when the authority structure breaks down, as occurredduring Katrina, so do the formal plans (GAO 2005).In Hurricane Katrina’s aftermath, some emergency

workers in the affected areas fled to care for their crisis-stricken families. Volunteers without knowledge of theCity of New Orleans stepped in to fill the vacuum, butoften were without specific experience in emergencyresponse or knowledge about the streets and neighbor-hoods of New Orleans. People trapped in the LouisianaSuperdome, the city’s pronounced final safe haven, hadextensive knowledge of the city and desire to help theirhometown but had no ability to make the knowledge andtheir volition known to volunteers. As authority struc-tures and plans crumbled, trucks idled at their destina-tions waiting for official approval to unload ice (Wach-tendorf and Kendra 2005). Warehouses full of beddingthat had been donated by the private sector sat unusedawaiting approved transport, while evacuees nearby slepton the floor (U.S. Congress 2005). Personnel and sup-plies were turned away by some officials, while otherofficials called for more of the same resources (Fore-man et al. 2005). Communication capabilities in areasaffected by Hurricane Katrina were so poor that officialshad to use human couriers to transmit messages (NewYork Times 2005).In large-scale crises, the scope and preponderance of

uncertainties give rise to a critical role for emergentresponse groups (Drabek 1986, Drabek and McEntire2003, Dynes 1983, Neal and Phillips 1995, Tierney et al.2001, Wachtendorf 2004). Disaster researchers havedefined emergent response groups as collectives of indi-viduals who use nonroutine resources and activities toapply to nonroutine domains and tasks, using nonroutineorganizational arrangements (Bigley and Roberts 2001,Drabek et al. 1981, Drabek 1986, Drabek and McEntire2003, Kreps 1984, Tierney et al. 2001). As our examplesof emergent response groups in Tables 2–4 illustrate,these groups include representatives from relief organi-zations; private sector organizations with volition andrelevant resources; and private citizens with information,

relationships, or physical and mental stamina to help. Inmany cases, these emergent response groups operate out-side the formal authority structures and response plans.For example, emergency shelters were run even beforethe shelters were approved and managed by the RedCross (U.S. Congress 2005). Other emergent responsegroups rescued victims from flooded buildings, orga-nized food drives, delivered drinking water, providedfirst aid, and transported victims to shelters. Given thevast range of uncertainties and circumstances, none ofthe emergent response groups could rely on preexistingstructures such as group membership, tasks, roles, andexpertise that could be specified and relied on ex ante.Whereas emergent response groups were initially

viewed by federal agencies as an aberration that neededto be stopped, recent disaster research concludes thatsuch groups are not aberrations at all, but can be ob-served in all large-scale disasters; that emergent behaviorcannot be stopped; and that emergent activity fills a voidthat cannot be filled by command and control approachesto disaster response (Tierney et al. 2001). Emergentresponse groups not only provide physical labor, butthey also serve the additional valuable functions of tol-erating learning and fostering flexibility and innovation,thereby minimizing ritualistic behavior commonly seenin bureaucratic responses to disasters (Drabek and McEn-tire 2003).Much of the sociological disaster research has focused

on documenting the existence, value, and characteris-tics of emergent response groups (e.g., Clarke 1989,Drabek et al. 1981, Drabek 1986, Drabek and McEntire2003, Dynes 1983, Kreps 1984, Neal and Phillips 1995,Quarantelli 1998, Tierney et al. 2001, Trainor 2004,Wachtendorf 2004). This research has found that emer-gent response groups have unclear and fluid bound-aries; fleeting and unclear membership; unclear, fluid,and dispersed leadership; highly unstable task definitionsand assignments as environmental conditions continu-ously change; and geographic dispersion that makes com-munication difficult. These characteristics require thatemergent response groups adopt specific approaches forknowledge coordination. One such approach commonlydocumented in studies of such groups is their use ofa learn-by-doing (versus decision making) action-basedmodel of coordinated problem solving, in which sensemaking and improvisation are the norm rather than theexception. For example, Moynihan (2005) studied theefforts of an emergent response group that successfullycontained a highly contagious poultry disease amongchicken farms in the western United States. The groupsucceeded only after it dropped initial plans and engagedin learning by doing, facilitated by intensive communi-cation through cell phones and a centralized informationsystem to track farms and actions. Clarke (1989) foundthat an interorganizational emergent group responding toPCB contamination at a New York City office building

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: PerspectiveOrganization Science 18(1), pp. 147–161, © 2007 INFORMS 151

learned by doing rather than spent time deliberating onfuture consequences of actions or choosing the course ofaction that offered the best expected outcome. Tierneyand Trainor’s (2004) analysis of the emergent recoverygroups in the aftermath of the World Trade Center disas-ter in 2001, and Hale et al. (2005) offer other evidence oflearning by doing as a common approach to knowledgecoordination during disaster response.As emergent response groups learn by doing, coor-

dination becomes opportunistic, with emergent leaders,emergent norms, emergent coordination principles, andemergent coordination channels. The Katrina disasterfound ham radio operators, for example, serving not justin an information transmittal role in the immediate after-math, but as 911 operators, ambulance dispatchers, andrescue coordinators. The urgency of the situation meansthat the objective of coordination is to achieve minimallyacceptable and timely action, even when more effec-tive responses may be feasible—but would take longerand use more resources. Finally, coordination focusedon action suggests that expertise coordination in emer-gent response groups is done by coordinating action.Knowledge conveyed needs to be tied to possible actionsthat can be implemented by emergent parties. Despitethe extensive sociological research on emergent responsegroups, there is much that is not understood about theinternal dynamics of these groups (Drabek 1986, Drabekand McEntire 2003). Thus, a challenge of theorizingabout emergent response groups is to understand howgroups coordinate their knowledge and actions whenresponding to large-scale disasters.

3. Transactive Memory Systems TheoryOne organization science theory of knowledge coordi-nation among groups—applied generally to more stablegroups than to emergent response groups—is transactivememory systems (TMS) theory. TMS theory, a theoryof group-level cognition, explains how people in collec-tives learn, store, use, and coordinate their knowledge toaccomplish individual, group, and organizational goals.It is a theory about how people in relationships, groups,and organizations learn who knows what, and use thatknowledge to decide who will do what, resulting in moreefficient and effective individual and collective perfor-mance.A TMS is a shared system for encoding, storing,

retrieving, and communicating information that developsnaturally in relationships and in groups (Hollingshead1998a, Wegner 1987). TMS theory borrows heavily fromwhat is known about the memory processes of individ-uals and applies it to groups. A TMS can be thoughtof as a network of interconnected individual memorysystems and the transfer of knowledge among them(Wegner 1995). Individuals who are part of a TMSassume responsibility for different knowledge domains,

and rely on one another to access each other’s expertiseacross domains. Expertise is defined in the TMS litera-ture to broadly include the know-what, know-how, andknow-why of a knowledge domain (Quinn et al. 1996),what Blackler (1995) refers to as embodied competen-cies. Expertise specialization, then, reduces the cognitiveload of each individual and the amount of redundantknowledge in the group, while collectively providingthe dyad or group access to a larger pool of knowl-edge. What makes transactive memory transactive arethe communications (called transactions) among individ-uals that make possible the codifying, storing, retrieving,and updating of information from individual memorysystems. For transactive memory to function effectively,individuals must have a shared conceptualization of whoknows what in the group. Evidence of a TMS has beenfound in a variety of relationships and groups, includ-ing married couples, dating couples, families, friends,coworkers, and project teams in both organizational andlaboratory settings (e.g., Argote 1999; Argote et al. 2000;Hollingshead 1998a, b, 2000, 2001; Faraj and Sproull2000; Lewis et al. 2005; Liang et al. 1995; Moreland andMyaskowsky 2000).Research on TMS has identified three indicators of the

level of development of a TMS (Lewis 2003, Morelandand Argote 2003):

1. Memory (or expertise) specialization: the tendencyfor groups to delegate responsibility and to specialize indifferent aspects of the task;2. Credibility: beliefs about the reliability of mem-

bers’ expertise; and3. Task (or expertise) coordination: the ability of team

members to coordinate their work efficiently based ontheir knowledge of who knows what in the group.

The greater the presence of each indicator, the more de-veloped the TMS and the more valuable the TMS is forefficiently coordinating the actions of group members.As a theory of knowledge coordination, TMS has

been used to describe and explain knowledge shar-ing in close relationships, work groups, and organiza-tions when three conditions are met: group member-ship is known, members perceive cognitive interdepen-dence, and members have shared goals (Brandon andHollingshead 2004). TMS was not developed to predictor explain the emergence and behavior of interorganiza-tional emergent response groups, in which some, if notall, of these conditions are violated.When applied to emergent response groups, the

assumptions about how specific behaviors (e.g., coordi-nation) occur and the resulting effects of TMS on groupresponse might be different from stable groups. Dynamicsituations require that groups be flexible and able torespond quickly because delays could result in moredamage or more lives lost (cf. Moreland and Argote2003). Emergent response groups form to accomplish

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: Perspective152 Organization Science 18(1), pp. 147–161, © 2007 INFORMS

specific tasks and often then disband when the tasksare accomplished, when the group realizes that mem-bers will not be able to accomplish the task, or whena more pressing situation is presented to the group.These groups are likely to be composed of diverse mem-bers, many of whom have not worked together before,and may not work together again. Moreland and Argote(2003) suggest that the dynamic conditions under whichthese groups form and work together are likely to havenegative effects on the development of transactive mem-ory. Constant changes in team membership threatentransactive memory. Members do not have significantincentives to learn about what others know since they areunlikely to work together again, and it is risky for peopleto rely on one another’s knowledge without familiarityor accountability. The high levels of stress that emergentresponse groups encounter are likely to have negativeeffects on individual and group information processing(cf. Ellis 2006, Moreland and Argote 2003).Thus, we believe challenges occur in all three indica-

tors of the level of development of a TMS—expertisespecialization, credibility, and expertise coordination—requiring a need to consider extending theorizing abouteach indicator for emergent response groups.

Need for Extension 1: Reconceptualizing the Role ofExpertise Specialization as a Basis for Task Assignment.The relative expertise of members in the group servesas the primary basis for deciding who will do what intraditional TMS theory. The processes through whichmembers identify and communicate expertise have beena central theme in previous iterations and more recentexpansions of TMS theory (Brandon and Hollingshead2004, Lewis et al. 2005). Current TMS theory suggeststhat task performance will be highest in groups whentasks are assigned to members on the basis of their rel-ative expertise, when members share the same mentalmodels regarding members’ expertise, and when recog-nized experts are willing and able to contribute theirexpertise to the group (Brandon and Hollingshead 2004).Emergent response groups present a new operating

environment in which groups must interact and behave—an environment in which it is difficult to identify expertsand assign tasks based on expertise. Unlike many tradi-tional work groups whose members have the requisiteknowledge, skills, and tools to accomplish the grouptask, there may not be any members in an emergentresponse group with the specific expertise or previousexperience to perform the task at hand. In the debit cardexample (Exhibit 2), neither LocalTexasBank’s deep dis-aster payment expertise nor the Red Cross’ expertise indisaster recovery was able to cope with the enormityof the situation; instead, the willingness and ability ofEPAY’s CEO and the Red Cross team to flexibly adapttheir knowledge, resources, and tools to a novel situ-ation proved instrumental. In the Coast Guard exam-ple (Exhibit 4), the civilian boat operators did not have

expertise in search and rescue, but did have sufficientknowledge to pilot their boats. As emergent responsegroups form, they need to quickly assess a situation,including what actions are required, who can performthem, and what resources are needed. The ability tocontribute to the group’s efforts requires more than justknowledge: the provider must also have the motiva-tion and access to the resources (time, equipment, labor,money, etc.) to make that knowledge available to thegroup. The challenge lies in how to conceptualize exper-tise, and in determining whether expertise should be theprimary basis for assigning tasks to members in emer-gent response groups.

Need for Extension 2: Assessing Credibility in Emer-gent Response Groups. Early TMS studies examineddyads and intact small groups where credibility in mem-ber knowledge was invariably high because of closesocial relationships or past interaction history (Holling-shead 1998a, 2000; Liang et al. 1995). The relative cred-ibility in member knowledge was seen to allow differentindividuals to specialize in different knowledge domains.Credibility in one another’s knowledge was an outcomeof learning either through direct or indirect means (expe-rience, communication, etc.). In such groups, account-ability was also high because parties could assume futureinteraction in which punishments and rewards couldenforce and reinforce expected behaviors.Recent TMS literature on organizational groups as-

sumes credibility in member knowledge as a core ele-ment in successful coordination (Lewis 2003, Lewiset al. 2005, Moreland and Myaskovsky 2000). There is,however, little discussion in current TMS literature abouthow groups develop credibility in member knowledge inhigh-risk situations of fleeting membership, fluid groupboundaries, and nonroutine tasks. The emergent groupthat formed between EPAY and the Red Cross (Exhibit2) was neither based on prior interaction history noron future plans of mutual cooperation. Moreover, thetwo organizations, one for profit and the other nonprofit,had fundamentally different interests. The challenge ishow emergent response groups function effectively whenexpertise cannot be validated.

Need for Extension 3: Expertise Coordination in Em-ergent Response Groups. The TMS of a group is sharedthrough a metastructure describing who knows what thatis understood and agreed on by group members (Bran-don and Hollingshead 2004). These metastructures arerarely made explicit in stable groups because time andmembership stability allow individuals to interactivelycue, evolve, validate, and refine an understanding of oneanother’s expertise (Hollingshead 1998b). The notion ofsharing in the traditional TMS context, then, refers to animplicit cognitive consensus about who knows what.The notion of shared (or team) mental models has

appeared in prior discussions of TMS (Brandon and

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: PerspectiveOrganization Science 18(1), pp. 147–161, © 2007 INFORMS 153

Hollingshead 2004, Levine and Moreland 1999). Sharedmental models are the extent to which individual teammembers’ mental models overlap—that is, the extentto which team members share the same understand-ing of the task, the tools, the team, and the situation(Blickensderfer et al. 1997). It is when group mem-bers have shared mental models that a TMS is mosteffective (Brandon and Hollingshead 2004). Benefits ofshared mental models include more accurate explana-tion and prediction of other group members’ actions andbetter coordination (Mohammed and Dumville 2001).TMS theory suggests that the implicit goal of a groupis to reduce differences in the individual mental mod-els of group members and to develop a shared mentalmodel regarding the group task and who will do what(Brandon and Hollingshead 2004). In emergent responsegroups, shared mental models may be much less likelyto develop.Traditional definitions of groups hold that members

of a group share a common purpose, a shared senseof identity, and a shared perception of interdependencewith others in the group in accomplishing that purpose(Poole and Hollingshead 2005). In an emergent responsegroup, as in the Katrina Wiki example (Exhibit 3), mem-bers pursue different individual purposes simultaneously,rather than pursuing a common group purpose, and theymay not have a sense of group identity. Rather, they mayjust have an ideology about volunteering or democraticcollaboration (Leuf and Cunningham 2001). Interdepen-dencies among members are likely to change quickly.A news correspondents, for example, may help an emer-gent response group to locate victims in need, becominga fleeting member of the group. Just as quickly, though,the news correspondent may move away from the groupto broadcast her story. On the surface, the lack of sta-ble membership suggests that a shared mental modelmay not be viable or even desired in emergent responsegroups. Time may be too precious to seek consensus onevents and actions, and agreements may make the groupless flexible to accommodate to changing inputs. In addi-tion, acute stress negatively affects the ability of groupsto implement their preexisting transactive memory (Ellis2006), let alone try to develop one from scratch. Howdo emergent response groups coordinate their expertisewithout a shared metastructure?We believe that emergent response groups offer the

opportunity for new theory building because such groupschallenge the key assumptions about the operatingenvironment in which TMS develops. As the operat-ing environment is challenged, so are the behaviorsto accomplish the key indicators of TMS. Effectiveresponse in a large-scale disaster requires broadeningthe behavioral repertoire to develop TMS. In §3.1, wedescribe how TMS theory might be extended for eachof the three indicators.

3.1. Extending TMS TheoryWhen membership in a group is fleeting, the notionof the system in a TMS needs expanding beyond thepeople-expertise links that are traditionally the focus ofTMS research (Brandon and Hollingshead 2004). Whenemergent response groups first come together, membersare likely not to ask one another about who knows what;instead, they are likely to ask about what is knownabout the situation and about the actions taken thus far(Dyer and Shafer 2003, Hale et al. 2005). The cogni-tive structure that they develop for the group centersnot around people, but on action-based scenarios thateither have been or might be carried out. These scenariosinclude decisions, actions, knowledge, events, and feed-back (Vera and Crossan 2005). The scenarios are notscripts, because they do not define the roles that peopleplay. Instead, the scenarios are patterns of actions strungtogether to be matched with events (Dyer and Shafer2003). Thus, in the emergent response groups exhib-ited in our three examples (Tables 2–4), we suggest thattheir TMS were likely composed not of people-expertiselinks, but rather of links between the tasks that needed tobe performed and the skills needed to perform the tasks,interwoven into sets of multiple scenarios of actions,unfolding over time and perhaps in parallel.Given this expanded notion of system in a TMS,

we suggest three extensions that facilitate the ability ofmembers of an emergent response group to coordinateduring disasters.

Extension 1: Tailor the Role of Expertise. The rolethat expertise plays in task assignment needs to be mod-ified when applied to emergent response groups. Someindividuals may have considerable expertise relevant tothe goal of the group, but may fail to contribute sim-ply because they lack the means or resources to takeaction on their knowledge (Argote and McGrath 1993).For example, the people who took refuge in the Super-dome in New Orleans had considerable knowledge aboutthe local neighborhoods but were unable to contributethat knowledge to the general good. In fact, much exper-tise that exists in a group may not be actionable for avariety of reasons such as competition, identity threat,power, etc. (Menon et al. 2006). To overcome these bar-riers to contribution, more stable groups may use theirslack resources (e.g., Haas 2006). However, emergentresponse groups lack these slack resources. Willingnessto share expertise, then, may be as important of a fac-tor in determining if a contribution is made as is theexpertise itself (Brandon and Hollingshead 2004). Con-sequently, contributions may not be based on deep taskexpertise. As demonstrated by the civilian operators inthe Coast Guard example (Exhibit 4), the many volun-teers in the Wiki example (Exhibit 3), and the small butagile payment operator EPAY (Exhibit 2), task-relevantexpertise is not a necessary or even a sufficient condition

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: Perspective154 Organization Science 18(1), pp. 147–161, © 2007 INFORMS

for making an important contribution to the emergentresponse group.Individuals in emergent response groups may make

valuable contributions simply by following directivesand providing physical labor without having any requi-site expertise. Individuals may bring value to the groupfrom the networks to which they connect (Jarvenpaa andMajchrzak 2005, Majchrzak et al. 2004). In the debitcard example (Exhibit 2), the representative at the finan-cial institution to whom the Red Cross turned for advicedid not have the scale and scope of resources himself,but he knew where to find them, drawing on his net-work of relationships. Individuals in emergent responsegroups may also bring value from having general, ratherthan narrow specialized, skills. In the debit card exam-ple (Exhibit 2), EPAY did not have the deep exper-tise of LocalTexasBank, which had a long-term workingrelationship with the Red Cross. EPAY employees hadmore general skills, however, particularly an agility toadapt. Those with general skills can provide much valueto the group by being particularly observant, monitor-ing changes in the environment, being creative problemsolvers, or using interpersonal skills for persuading andmotivating people to join the effort. Finally, at the grouplevel, an individual’s expertise may be less relevant thanthe collective expertise of emergent group members, asin the Katrina Wiki example (Exhibit 3).Thus, we suggest that TMS theory may need to recon-

sider how task assignments are made beyond a member’sdeep expertise, to include a group’s collective knowledgeof relationships, tools, and task performance as well asa willingness of group participants to share and to acton that knowledge. That is, capability and motivation,in addition to the domain knowledge typically examinedby TMS theory, are required.Because of the speed of change in a disaster situa-

tion, the TMS notion of expertise as hierarchically orga-nized knowledge in a particular domain may need tobe augmented with a notion of knowledge flexibility,defined as expertise plus the ability to update that exper-tise based on feedback from the environment (Cook andBrown 1999). For example, an individual in an emer-gent response group may be knowledgeable about debitcards, but if that individual is not a good observer andinterpreter of the feedback from the field indicating thatdebit cards are not working for the victims in the wayintended, that person’s knowledge about debit cards is oflimited value, becoming rapidly obsolete as new knowl-edge about how to make debit cards work is gained.Knowledge flexibility may come about by having

higher-order domain principles as suggested by Lewiset al. (2005). In addition, though, knowledge flexibilitymay be derived from individuals’ personality traits thatenable them to cognitively process divergent informationquickly, or through individuals’ experiences with enough

other disasters that they are able to quickly match incom-ing patterns to successful past recovery efforts.In addition to allowing for knowledge flexibility, we

need to augment the notion in TMS theory that mem-ber selection is based on the expertise requirements ofthe collective task. The expertise and experiences thatindividuals bring to the emergent group are likely to beincommensurate, divergent, and potentially irrelevant tothe task at hand, and the group may not have all thenecessary expertise and resources to achieve its objec-tives. In addition, individuals will approach the taskswith institutional and other cultural norms dictated bytheir own experiences, which could make the develop-ment of a shared conceptualization of the collective taskobjective unlikely, if not impossible. Expertise will beconstantly changing and challenged, so the focus of thegroup’s efforts in developing its TMS may not be onagreeing who has expertise in a task-related topic, buton constantly scanning the environment to locate knowl-edge and other resources.

Extension 2: Replacing Credibility in Expertise withTrust Through Action. In emergent response groups, thenotion of credibility in member knowledge as a basis fora TMS found in traditional TMS literature (e.g., Lewis2003, Moreland and Myaskovsky 2000) may need to bereplaced with the notion that TMS develops from a trustthat others will behave in ways that will be helpful to thecommunity—trust that is continuously created and recre-ated through action (Child and Mollering 2003). Trust isdefined as a willingness to hold positive expectations ofanother party’s behaviors, regardless of the inability tomonitor or control that other party (Mayer et al. 1995). Incontrast, traditional TMS literature assumes that efficientcoordination of knowledge is dependent on the group’sability to validate that members have credible knowledgeand are doing the tasks that they are assigned (Brandonand Hollingshead 2004). Such validation occurs throughshared experiences or explicit information about mem-bers’ skills and knowledge (Moreland and Myaskowsky2000). The resource-scarce and time-pressured operatingenvironment of the emergent response group precludessuch validation. In such groups, there is an implicit threatof common fate such that, unless one trusts others andtakes immediate action, significantly more harm (even totheir own security) may result (Meyerson et al. 1996).The threat encourages risk taking, while the generationof action increases the willingness to trust others’ knowl-edge without social proof. In the debit card example(Exhibit 2), only a handshake put the partnership of RedCross and EPAY into full gear; subsequent daily e-mailssuspended doubt and built trust.Trust in emergent response groups takes a form of

swift trust that is developed through task-based action inthe presence of a shared fate or higher-order goal. Swifttrust implies different conceptualizations of trust than are

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: PerspectiveOrganization Science 18(1), pp. 147–161, © 2007 INFORMS 155

typically examined in the TMS literature. Often associ-ated with ad hoc project teams (Harrison et al. 1997,Jarvenpaa and Leidner 1999, Meyerson et al. 1996),swift trust does not assume any shared experience amongthe whole group; it simply includes the positive expec-tation that others will be trustworthy ex ante of anyobserved carryover from past interactions or expectationof future interactions. The concept of swift trust wasoriginally developed to describe the emergence of trustin role-based temporary groups, such as film crews andcockpit crews that come together for a short time andthen disband (Meyerson et al. 1996). Swift trust, how-ever, is not limited to teams that coordinate face to face;it can emerge in dispersed teams that have no opportu-nity for face-to-face encounters (Jarvenpaa et al. 1998,Jarvenpaa and Leidner 1999), as well as in virtual learn-ing communities (Coppola 2004).In emergent response groups, action may initiate swift

trust. The maintenance of trust is likely to be multi-faceted, however, relying on many sources, particularlyas groups face adversity and an inability to observe oreven an awareness of others’ actions. These trust sourcescan also reinforce initial swift trust. If the responsegroup has a member who can be singled out as a coremember (e.g., the Red Cross) or a go between whoassembled the other members, and the other membershave had prior interaction with this core member (e.g.,EPAY and the Red Cross both did business at the finan-cial institution), trustworthiness of this core member isassumed to generalize to trust in the whole group (trustby transference) (Harrison et al. 1997, Meyerson et al.1996). Trust may also be strengthened on the basis ofinstitutional factors, including social and organizationalmembership, rules, and roles, unless a member (e.g.,a person in a police uniform) is observed engaging inactivities that violate expectations (e.g., stealing food)(Kramer 1999, McKnight et al. 1998). Trust may also bebased on dispositional tendencies (Kramer 1999, McK-night et al. 1998). When uncertainty is high, the dispo-sitions of members have a strong impact on determin-ing the level of trust for action, particularly when otherscannot be directly observed acting in a trusting manner(Meyerson et al. 1996). The CEO of EPAY took the highroad by assuming that rescued victims seeking financialrelief were legitimate unless proven otherwise. Finally,where members can visibly see what is happening andperhaps digitally record it, trust can be based on deter-rence (punishment or reputational sanctions) (Lewickiand Bunker 1995).In uncertain conditions, swift trust promotes efficient

response as long as it is neither too low nor too high(Meyerson et al. 1996). Swift trust that is too high isdetrimental, although disappointments can rapidly bringa downward spiral. Also, in emergent response groupscomposed of strangers, swift trust that is too high maypromote opportunistic actions and negative outcomes,

particularly when accountability is low (Langfred 2004).Group members may take advantage of trust placed inthem. Moderate trust is is less amenable to these prob-lems because it promotes alternative paths (e.g., hedges,fail-safe mechanisms) to cover the worst-case scenar-ios that might occur because of the great uncertaintiespresent (Meyerson et al. 1996).Sometimes operating conditions (time urgency, high

emotions, severity of harm) of the emergent responsegroup may be so inhospitable that the group must con-tend with low trust and still engage in transactions anddevelop TMS. In such cases, the group may build redun-dancy in task-expertise scenarios by maintaining parallelactions that help accomplish similar goals. The groupmay also engage in additional activities where specificgoals are set and members’ behaviors are monitoredtoward those goals. The team may set up an informationsystem to track and transmit status information abouteach team member and to make the information avail-able to the whole team. The team may also resort toexternal authorities or plans that instruct the membersfor particular actions even if doing so results in a less-than-optimal response.All in all, the level of trust present in an emergent

response group is likely to affect TMS development, assuggested by existing TMS literature, but the nature ofthat influence is likely to be substantially different thanassumed in the TMS literature. High levels of trust maynot be helpful. Instead, moderate levels may provide suf-ficient predictability with protection from harm. In addi-tion, trust is unlikely to be determined by situationalfactors; instead, trust is likely to be highly dynamic,needing to be constantly created and recreated throughaction (Child and Mollering 2003). The group dynam-ics, not the situation, are likely to affect how trust iscontinuously won (Adler 2001).

Extension 3: Coordinating Knowledge Processes With-out a Shared Metastructure. Instead of the TMS concep-tualization that knowledge coordination occurs througha shared metastructure consisting of a directory of whoknows what and a set of cues for encoding and retriev-ing information from each group member, knowledgecoordination in an emergent response group will needto occur without shared mental models. Based on pastexperiences either as observers of or participants inemergencies, members may have individual mental mod-els about how people should behave in emergency situ-ations and may bring expectations about their own andothers’ roles (Bechky 2006, Faraj and Xiao 2006, More-land and Argote 2003), as well as ostensive routines(Feldman and Pentland 2003) for performing in an emer-gency situation. Emergent response groups, however, arelikely to pose challenges to meeting these expectations.Roles may be in conflict, roles may be left unattended,

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: Perspective156 Organization Science 18(1), pp. 147–161, © 2007 INFORMS

or roles identified as needed may not be matched pre-cisely to the knowledge and motivation that exists at thetime. In the debit card example (Exhibit 2), the role thatrequired a high volume of cash distribution was mis-matched to the capabilities of the initial debit card com-pany. In the Coast Guard example (Exhibit 4), the CoastGuard’s formal role expectation did not include urbanwaterways, or coordinating with civilian boat operatorsin rescue operations.Coordinating an emergent response group in the ab-

sence of a metastructure is likely to create “epistemicdifferences, reputation stakes, and possible blame appor-tionment” (Faraj and Xiao 2006, p. 1155), potentiallymaking the process quite difficult. There may be attribu-tion errors, miscues, communication rework, and signif-icant inefficiency in how knowledge is shared and used(e.g., DeSanctis and Monge 1999). While the group islikely to attempt to create a metastructure (Lewis et al.2005), the instability in roles, task, and membership islikely to rapidly make such a metastructure obsolete. Thegroup, then, may need to evolve not a metastructure, buta set of mechanisms for managing its activities.A group may evolve a variety of mechanisms. Faraj

and Xiao’s (2006) study of medical trauma center prac-titioners as well as studies reviewed by Tierney et al.(2001) of emergent teams responding to large-scale dis-asters demonstrate that such groups evolve the use ofdialogic coordination practices such as joint on-the-spot sense making when metastructures fail. Brown andEisenhardt (1998) suggest that such groups may use sim-ple process structures to coordinate. In the debit cardexample (Exhibit 2), the simple structure of e-mailingan approved list of victim names and daily disburse-ment amounts quickly evolved as the mechanism forcoordination.Emergent response groups may also use a mechanism

of creating a community narrative (Boland and Tenkasi1995), which is a running narrative of the actions takenand not taken, the decisions made, and the theories inuse. Narratives do not represent a single shared under-standing of a domain; rather they represent the mul-tiplicity of events and actions a community is taking,as members are taking them. Narratives may be articu-lated explicitly or understood implicitly. In the KatrinaWiki example (Exhibit 3), the emergent response groupof Wiki contributors created a community narrative oftheir activities by the very act of contributing to theWiki site. Since the site was an organized aggregation ofthe emergent response group’s activities, with a historypage documenting any changes to the page and a dis-cussion page describing discussions that ensued prior tosignificant changes, the Wiki site was simultaneously theproduct of an emergent response group’s work and thenarrative of their actions, decisions, and implied causal

principles. Narratives, when made explicit, may serveas a coordination mechanism by providing an observ-able record of others’ actions that may help membersrecognize a routine, role, or sequence of actions intowhich they might contribute (Bechky 2006, Feldmanand Pentland 2003). Explicitly articulated narratives mayalso make clearer that multiple sequences of actions maybe occurring simultaneously, thus resolving role conflictsby allowing multiple ways to accomplish a task. Forexample, by encouraging civilian boat operators to artic-ulate their plans of action for rescuing victims from thewaterways (e.g., a community narrative), Coast Guardpersonnel were able to recognize the need for differenttypes of support for different plans of action—some boatoperators needed armed guards, others needed experi-enced emergency personnel on board, and others neededthe communication linkage that the Coast Guard couldprovide to local medical personnel.The urgency of time may make it too onerous for

the extra effort of articulating actions as they are beingperformed, yet most emergency response requires somecommunication. Community narratives may then evolvethrough a human intermediary. In Hurricane Katrina’saftermath, ham radio operators, in relaying messages topeople in the field, shared the evolving narrative of whatwas happening and who was doing what about it. InMoynihan’s (2005) study of an emergent group respond-ing to a U.S. poultry disease outbreak, a central dis-patcher became the aggregator of the community nar-rative as members shared the status of avian health onthe farms they visited, the possible reasons they con-sidered that a farm contracted the disease while anotherfarm was spared, and the steps each member planned totake to prevent the next farm from falling victim; thecentral dispatcher continuously shared this informationwith other members as they called in. After a hurri-cane in Florida, trucks with satellite dishes were drivenaround the impacted area so that Florida hurricane dis-aster response teams could periodically synchronize thelatest information they had collected on building condi-tions, using Groove peer-to-peer software; in so doing,they were contributing to an evolving narrative of thegroup’s perspective on the hurricane’s impact and victimneeds.In sum, in the absence of a shared metastructure, em-

ergent response groups need to evolve mechanisms forcoordination. These mechanisms may include simplestructures or more comprehensive narratives. They maycoordinate face to face with dialogue or through interme-diaries, be they human or a website. They may quicklyevolve a routine, may constantly negotiate actions, ormay adopt a set of parallel actions. Given the urgencyof the situation, the mechanism they choose may be lessrelevant than whether they adopt any at all.

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: PerspectiveOrganization Science 18(1), pp. 147–161, © 2007 INFORMS 157

Exhibit 5 Expanding TMS Theory to Emergent Response Group

Characteristics Implications for Implications forof emergent knowledge Current TMS Suggested extensions to TMS research topics inresponse groups∗ coordination theory theory for emergent response groups organization science

TMS Indicator 1: Expertise specialization

• Unstable taskdefinitions andassignments

• Pursuit of multiple,conflicting, andchanging purposesand perspectives

• Fleeting andsometimes unclearmembership

• Geographicallydistributed, diverse,unfamiliar groupmembers

• Volitional memberparticipation, basedon urgent personalneeds

• Action-basedcoordination

• Learning-by-doingmodel

• Minimally acceptabletimely action suffices

• Opportunisticcoordination withemergent leaders,emergentcoordinationprinciples, emergentcoordination channels

Task-relevantexpertiseservesas basisfor taskassignmentandspecialization

• Task-relevant expertise often not present, soany knowledge of relationships, tools,or tasks and ability and willingness to acton that knowledge may serve as basis oftask assignment and specialization

• Knowledge flexibility, sufficiency, andmotivation as additional bases for taskassignment and specialization

• Evolving nature ofexpertise in a group

• Converting focus ondomain knowledge toactionable knowledge

• Keeping volunteersengaged when theirinitial needs are met

TMS Indicator 2: Credibility in member expertise

Validation ofexpertiseneeded foreffectivegroupfunctioning

• Replace credibility with trust in action• Moderate levels of trust more conduciveto building a TMS than high levels

• Trust encouraged without observingmember behavior

• Development of swift trust

• Examining nature of trustas it evolves

• How group dynamicsaffect trust

• When trust and expertisecredibility differ

• Conflicts of swift vs.generalized trust

TMS Indicator 3: Knowledge coordination

Shared mentalmodelof whoknows whatnecessary forefficientcoordination

• Knowledge coordination occurs by observingand recognizing action scenarios, identifyingways to contribute to scenarios, and quicklydevising simple coordination mechanisms

• Community-developed narrativesdescribing events & scenarios mayhelp coordination

• Finding ways to overcomecommunication difficulties

• Using IT to facilitate nar-rative evolution and com-munication

• Coordinating with multipleconflicting action scenarios

∗Drabek and McEntire (2003).

3.2. Summary: Emergent Response GroupDynamics Revisited

Our examination of TMS theory is summarized inExhibit 5. Our examination suggests that by expandingthe context in which TMS theory is applied to includeemergent response groups, insights can be gained intotheir internal dynamics. The three indicators of the levelof development of a TMS provide a useful frameworkfor organizing these insights in the exhibit.Traditional TMS research would not be able to ac-

count for emergent response groups being able to func-tion with relatively little expertise and a specializationthat is based on volition and tasks, not expertise. Ourextensions to TMS theory allow a more inclusive con-ceptualization of how tasks are assigned that focuses onactionable knowledge sufficiency of people, tools, tasks,and member volition.Traditional TMS theory would not be able to explain

how members assess the credibility of one another’sexpertise without observing their behaviors. Our exten-sions to TMS theory suggest that credibility in emergentresponse groups is less important for coordination thanmembers’ ability to constantly create and recreate trustwith task-specific action or without observable actionbased on associations.

Traditional TMS theory would not be able to ac-count for emergent response groups coordinating theirknowledge from afar, without knowing each other’s ex-pertise and without agreed-on cues about how to encode,store, and retrieve each individual’s knowledge. Ourextensions of TMS theory suggest that, because in anemergent response group there will be no metastructuresthat everyone agrees to and understands, participants willbring to the situation their theories in use and experi-enced and expected roles for others to play; when thoseroles conflict or are insufficient, the group will evolvecoordination mechanisms that range from simple struc-tures to more comprehensive community narratives.

Practical Implications of Extending TMS Theory. Ourextensions to TMS theory suggest several insights aboutthe internal dynamics of emergent response teams thathave practical implications for disaster preparedness.The current practice of joint training for emergencyresponse is focused on learning what others know tofacilitate coordination with one another when the crisisoccurs. However, in the event of the crisis, these trainedpeople and their expertise are not necessarily present.Our exploration into the internal dynamics of emergentresponse groups suggests that such groups will include

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: Perspective158 Organization Science 18(1), pp. 147–161, © 2007 INFORMS

people who have not been trained together, such as peo-ple who volunteer with relatively little specialized exper-tise in the emergency domain.Therefore, training should focus on learning how to

quickly recognize volunteers’ volition in participating inan emergent group, the tasks they might engage in, andthe support they might need to carry out those tasks.Such training could also help people to recognize thebenefits and dangers of generalized trust. It could alsohelp people to quickly evolve a coordination mecha-nism that does not rely on what people know, but oncompiling and communicating a narrative of the actionsthat volunteers take, so that others are able to assess forthemselves what actions they could take to help. Suchtraining might help emergency personnel recognize theimportant role they might play as intermediaries in thecoordination process, rather than as coordinators them-selves. Finally, such training should not be limited toemergency personnel. Citizens, too, can be empoweredwith such training. These practical suggestions, then,build on the current crisis management literature rec-ommending training for emergence (e.g., Tierney et al.2001) by providing specificity about the content of thattraining.

4. Implications for Organization ScienceTheorizing

Our examination of the internal dynamics of emergentresponse groups through a TMS lens has implicationsfor theorizing about a broader set of groups and orga-nizations. Our three extensions collectively suggest thatthe cognitive models of knowledge sharing and coor-dination developed for intact work groups need mod-ification to apply to rapidly and continually changingenvironments. Even theories of improvisation, with theirfocus on action as preceding thought and retrospec-tive sense making (e.g., Weick 1998), are limited togroups with a well developed understanding of internalresources and materials at hand, proficiency, presenceof associates similarly committed to and competent atimpromptu decision making, and confidence in dealingwith nonroutine events (Weick 1998). Thus, jazz quar-tets, firefighters, and other groups that many scholarswho are interested in processes of improvisation havestudied are quite different from the emergent responsegroups we examine here. Our theorizing extends think-ing about the relations between cognition and action ingroups of highly skilled members to those groups whosemembers may not have individual mental models, letalone shared mental models of the task.Our extensions also have implications for theories of

dynamic organizations. As the marketplace is becomingincreasingly dynamic, many groups, organizations, andcommunities face ever-higher levels of unpredictabil-ity, urgency, and reconfigurability. Process networks in

China form emergent value chain relationships in re-sponse to customer needs for manufacturing a prod-uct (Hagel and Brown 2005). Cross-organizational newproduct–development teams form and disband rapidly,leaving membership unstable and uncertain (Majchrzaket al. 2004). Virtual organizations and communitiesemerge in response to social needs and competitive pres-sures (DeSanctis and Monge 1999). Terrorist and secu-rity threats create the need for emergent response groupscomprising security personnel from ports, private cor-porations, the FBI, and public utilities (Jarvenpaa andMajchrzak 2005). Theories of dynamic organizationsneed to push their definitional limits and theoreticalboundaries to explain human behavior, where humansare struggling to collaborate under extreme time pressureand risk, with inadequate information, with emotionallyladen volition, and with others who have conflicting pur-poses, fleeting involvement, and changing perspectives.As suggested by Peterson and Mannix (2003), such the-ories need to focus on human behavior and interactionand go beyond structural patterns, typologies, and fea-tures of dynamic organizations.A fundamental issue in dynamic organizations is how

to manage the continuous tension between stability andinstability (Brown and Eisenhardt 1998, Peterson andMannix 2003). Previous research on dynamic organi-zations has argued for stability at the organizationallevel and flexibility at the individual and group levels(Dyer and Shafer 2003, Moreland and Argote 2003). Ourexamination of emergent response groups reiterates theneed for theorizing to consider multiple levels of analy-sis; in addition, though, we suggest the possibility thatsuch theorizing may need to reverse the focal point ofstability: individuals, and not the organization, may bethe source for stability. In fact, research studying indi-vidual response to disasters repeatedly finds that peoplemaintain behavioral continuity and remarkable compo-sure when responding (Drabek 1986). Therefore, sincethe organization contributing to and managing a dynamicgroup may need to be as fluid as the group itself, sta-bility may need to be found in the individual: the indi-vidual’s ability to quickly work with others to developcoordination mechanisms, the individual’s willingness tocontribute where applicable and the individual’s abilityto know what she knows herself, how to use her net-work and how to rapidly signal her knowledge to others.The role of the organization, then, may become one ofserving as an intermediary and directory of coordinationmechanisms, resources, action scenarios, and emergentresponse groups that the individual may contribute to,and use, in responding to the changing environmentalconditions.Our extensions also have implications for theories on

trust. The trust literature—even those studies focused ondynamic organizations—frequently theorize about trust

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: PerspectiveOrganization Science 18(1), pp. 147–161, © 2007 INFORMS 159

as a relatively stable social characteristic generalizedfrom patterns of past and current interaction, dispositions,and institutional structures. The literature portrays trustas situationally dependent. As the situational conditions(such as characteristics of the relationships) change, sodoes trust (e.g., Lewicki and Bunker 1995, McKnightet al. 1998). Even the current literature on swift trustassumes predictability and deterrence from the situa-tional conditions of clear roles, intense social interac-tion, social proofs, firm deadlines, and so on (Meyersonet al. 1996, Jarvenpaa and Leidner 1999). Our theoriz-ing, in contrast, suggests that emergent response groupsface conditions that undermine all these situationallybased sources and bases of trust and require trust thatis constantly created through action (Child and Moller-ing 2003). In such situations, swift trust is only a fleet-ing, situationally specific assessment of “what you justdid.” In such high-risk settings, trust comes from pur-posive action that conveys investment and vulnerability.Trust is highly dynamic and is continuously won ratherthan called on (Adler 2001). Our extensions imply ashift in attention from the conditions influencing trust tothe examination of how action in the team creates andrecreates trust, how group dynamics affect swift trust,and how swift trust affects group dynamics. What hap-pens when swift trust conflicts with contextually basedsources and bases of trust? How does a group resolvethese conflicts?There is a plethora of other research questions worthy

of study. Communication difficulties abound in dynamicorganizations, difficulties made even more complex bythe time urgency, rapidly changing conditions that makeinformation quickly obsolete, emotionality of the situ-ation, and an inability to observe the actions of othermembers. Research is needed on how members com-municate their actions and intents in difficult situations,especially when roles imported from previous contextsdo not work. Research is also needed on the role of deepexpertise and how it evolves, if at all, in dynamic orga-nizations when each situation is so different, or whetherthe cognitive rigidity of deep expertise outweighs itsbenefits. Emergent collaboration is emotionally, intel-lectually, and potentially physically draining. What arethe coping strategies that lead to the most effectiveresponse?In our theorizing, we have repeatedly referred to time

pressures and the emotional intensity typical in a dis-aster; the effect these have on knowledge coordina-tion deserves further research. Disaster scale and scopeare likely to have significant impacts, as well. Socialnetworks and institutional norms of the organizationsand communities from which emergent response groupmembers are drawn may affect not only trust, but thespeed of forming groups, role negotiation, and interac-tion patterns. Further research examining these impactsis needed.

5. ConclusionThe bureaucratic model continues to be the mostcommonly used approach for coordinating crossorga-nizational response to major disasters. The NationalPreparedness System perpetuates and continues to insti-tutionalize this approach. This approach is used in spiteof the disaster research evidence indicating that suchan approach is not likely to accommodate the multidi-mensional crisis situation found in large-scale disasters.Organization science is concerned with dynamic orga-nizations, virtual teams, networks of individuals withinand between organizations, interorganizational relation-ships, and coordination within and across organizations.Therefore, it is within the domain of organization sci-ence to contribute to this debate about new alternativemodels for emergency response. However, this potentialcontribution can only be realized if theoretical modelselaborate on the dynamics of the phenomenon, openingup the internal dynamics of the emergent response groupleading to a deeper understanding of how the emergentresponse group functions within the larger institutionalnetwork of organizations, communities, and individuals.With our improved understanding and readiness, whenthe next disaster hits (as it most assuredly will) the criesfor help (as in Exhibit 1) may not go unanswered.

AcknowledgmentsThe authors would like to thank Linda Argote for inspiringthem to write this “Perspective,” and the helpful comments ofthe anonymous reviewers. They would also like to thank mem-bers of emergent response groups and the disaster researchcommunity who so graciously gave their time to help themunderstand the role of organizational science. This researchwas supported in part by a grant to the third author from theNSF ITR 0427089.

ReferencesAdler, P. S. 2001. Market, hierarchy, and trust: The knowledge econ-

omy and the future of capitalism. Organ. Sci. 12 215–234.

Anderson, A. I., D. Compton, T. Mason. 2004. Managing in a danger-ous world: The National Incident Management System. Engrg.Management J. 16(4) 3–9.

Argote, L. 1999. Organizational Learning: Creating, Retaining andTransferring Knowledge. Kluwer, Norwell, MA.

Argote, L., J. McGrath. 1993. Group processes in organizations:Continuity and change. Internat. Rev. Indust. Organ. Psych. 8333–389.

Argote, L., P. Ingram, J. M. Levine, R. L. Moreland. 2000. Knowl-edge transfer in organizations: Learning from the experience ofothers. Organ. Behav. Human Decision Processes 82 1–8.

Bechky, B. A. 2006. Gaffers, gofers, and grips: Role-based coordina-tion in temporary organizations. Organ. Sci. 17(1) 3–21.

Bigley, G. A., K. Roberts. 2001. The incident command system:High reliability organizing for complex and volatile task envi-ronments. Acad. Management J. 44 1281–1300.

Blackler, F. 1995. Knowledge, knowledge work and organizations: Anoverview and interpretation. Organ. Stud. 16 1021–1046.

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: Perspective160 Organization Science 18(1), pp. 147–161, © 2007 INFORMS

Blickensderfer, E., J. A. Cannon-Bowers, E. Salas. 1997. Theoreticalbases for team self-correction: Fostering shared mental models.M. M. Beyerlein, D. A. Johnson, S. T. Beyerlein, eds. Advancesin Interdisciplinary Studies of Work Teams: Team Implementa-tion Issues, Vol. 4. JAI Press, London, UK.

Boland, R. J., R. V. Tenkasi. 1995. Perspective making and perspec-tive taking in communities of knowing. Organ. Sci. 6 350–372.

Brandon, D. P., A. B. Hollingshead. 2004. Transactive memory sys-tems in organizations: Matching tasks, expertise, and people.Organ. Sci. 15 633–644.

Brown, S., K. Eisenhardt. 1998. Competing on the Edge: Strategy asStructured Chaos. Harvard School Publishing, Boston, MA.

Child, J., G. Mollering. 2003. Contextual confidence and active trustdevelopment in the Chinese business development. Organ. Sci.14(1) 69–80.

Clarke, L. B. 1989. Acceptable Risk? Making Decisions in a ToxicEnvironment. University of California Press, Berkeley, CA.

Cook, S. D., J. S. Brown. 1999. Bridging epistemologies: The genera-tive dance between organizational knowledge and organizationalknowing. Organ. Sci. 10 381–400.

Coppola, N. 2004. Building trust in virtual teams. IEEE Trans. Pro-fessional Comm. 47(2) 95–104.

Department of Homeland Security (DHS). 2004. National IncidentManagement System. Government Printing Office, Washing-ton, D.C. http://www.dhs.gov/interweb/assetlibrary/NIMS-90-web.pdf.

DeSanctis, G., P. Monge. 1999. Communication processes for virtualorganizations. Organ. Sci. 10 693–703.

Drabek, T. E. 1986. Human System Responses to Disasters: An Inven-tory of Social Findings. Springer-Verlag, New York.

Drabek, T. E., D. A. McEntire. 2003. Emergent phenomena and thesociology of disaster: Lessons, trends and opportunities from theresearch literature. Disaster Prevention and Management 12(2)97–112.

Drabek, T. E., J. L. Taminga, T. S. Kilijanek, C. R. Adams. 1981.Managing Multiorganizational Emergency Responses: EmergentSearch and Rescue Networks in Natural Disasters and RemoteArea Settings. Bounder Institute of Behavioral Sciences, Univer-sity of Colorado, Boulder, CO.

Dyer, L., R. Shafer. 2003. Dynamic organizations. R. S. Peterson,E. A. Mannix, eds. Managing People in Dynamic Organizations.Lawrence Erlbaum, Mahwah, NJ, 7–40.

Dynes, R. 1983. Problems in emergency planning. Energy 8 633–660.

Dynes, R., K. Tierney. 1994. Disasters, Collective Behavior, andSocial Organization. University of Delaware Press, Newark, DE.

Ellis, A. P. J. 2006. System breakdown: The role of mental modelsand transactive memory in the relationship between acute stressand team performance. Acad. Management J. 49 576–589.

Faraj, S., L. Sproull. 2000. Coordinating expertise in software devel-opment teams. Management Sci. 46 1554–1568.

Faraj, S., Y. Xiao. 2006. Coordination in fast-response organizations.Management Sci. 52 1155–1189.

Feldman, M. S., B. T. Pentland. 2003. Reconceptualizing organiza-tional routines as a source of flexibility and change. Admin. Sci.Quart. 48 94–118.

Foreman, T., M. M. Ahlers, A. Cooper. 2005. People making deci-sions hesitated. http://www.cnn.com.

GAO (U.S. Government Accountability Office). 2005. Hurricane Kat-rina: Providing oversight of the nation’s preparedness, response,and recovery activities. GAO-05-1053T (Sept. 28).

Haas, M. R. 2006. Knowledge gathering, team capabilities, andproject performance in challenging work environments. Man-agement Sci. 52 1170–1184.

Hagel, J., J. S. Brown. 2005. The Only Sustainable Edge. HarvardBusiness School Publishing, Boston, MA.

Hale, J. E., R. E. Dulek, D. P. Hale. 2005. Crisis response communi-cation challenges. J. Bus. Comm. 42 112–134.

Harrison, R. T., M. R. Dibben, C. M. Mason. 1997. The role of trust inthe informal investor’s investment decision: An exploratory anal-ysis. Entrepreneurship Theory and Practice (Summer) 63–81.

Hollingshead, A. B. 1998a. Communication, learning, and retrievalin transactive memory systems. J. Experiment. Soc. Psych. 34423–442.

Hollingshead, A. B. 1998b. Retrieval processes in transactive memorysystems. J. Personality Soc. Psych. 74 659–671.

Hollingshead, A. B. 2000. Perceptions of expertise in and transactivememory in work relationships. Group Processes and IntergroupRelations 3 257–267.

Hollingshead, A. B. 2001. Cognitive interdependence and convergentexpectations in transactive memory. J. Personality Soc. Psych.81 1080–1089.

Jarvenpaa, S. L., D. E. Leidner. 1999. Communication and trust inglobal virtual teams. Organ. Sci. 10(6) 791–815.

Jarvenpaa, S. L., A. Majchrzak. 2005. Developing individuals’ trans-active memories of their ego-centric networks to mitigate risksof knowledge sharing: The case of professionals protectingcybersecurity. 26th Internat. Conf. Inform. Systems Proc.

Jarvenpaa, S. L., K. Knoll, D. Leidner. 1998. Is anybody out there?The antecedents of trust in global virtual teams. J. ManagementInform. Systems 14(4) 29–64.

Kramer, R. M. 1999. Trust and distrust in organizations: Emerg-ing perspectives, enduring questions. Annual Rev. Psych. 50569–598.

Kreps, G. 1984. Sociological inquiry and disaster research. AnnualRev. Sociol. 10 309–330.

Langfred, C. W. 2004. Too much of a good thing? Negative effectsof high trust and individual autonomy in self-managing teams.Acad. Management J. 47(3) 385–399.

Leuf, B., W. Cunningham. 2001. The Wiki Way: Quick Collaborationon the Web. Addison-Wesley, Boston, MA.

Levine, J. M., R. L. Moreland. 1999. Knowledge transmission inwork groups: Helping newcomers to succeed. L. L. Thompson,J. M. Levine, D. M. Messnick, eds. Shared Cognition in Orga-nizations: The Management of Knowledge. Lawrence Erlbaum,London, UK.

Lewicki, R. J., B. B. Bunker. 1995. Trust in relationships: A model oftrust development and decline. B. B. Bunker, J. Z. Rubin, eds.Conflict, Cooperation, and Justice. Jossey-Bass, San Francisco,CA.

Lewis, K. 2003. Measuring transactive memory systems in the field:Scale development and validation. J. Appl. Psych. 88 587–604.

Lewis, K., D. Lange, L. Gillis. 2005. Transactive memory systems,learning, and learning transfer. Organ. Sci. 16 581–598.

Liang, D., R. Moreland, L. Argote. 1995. Group versus individualtraining and group performance: The mediating role of transac-tive memory. Personality Soc. Psych. Bull. 21 384–393.

Majchrzak, A., L. Cooper, O. Neece. 2004. Knowledge reuse for inno-vation. Management Sci. 50 174–188.

Majchrzak, A., A.Malhotra, J. Lipnack, J. J. Stamps. 2004. Can absencemake a team grow stronger? Harvard Bus. Rev. 82 131–137.

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.

Majchrzak, Jarvenpaa, and Hollingshead: PerspectiveOrganization Science 18(1), pp. 147–161, © 2007 INFORMS 161

Mayer, R. C., J. H. Davis, F. D. Schoorman. 1995. An integra-tive model of organizational trust. Acad. Management Rev. 20709–734.

McKnight, D. H., L. L. Cummings, N. L. Chervany. 1998. Initial trustformation in new organizational relationships. Acad. Manage-ment Rev. 23 473–490.

Menon, T., L. Thompson, H.-S. Cho. 2006. Tainted knowledge vs.tempting knowledge: People avoid knowledge from internalrivals and seek knowledge from external rivals.Management Sci.52 1129–1144.

Meyerson, D., K. E. Weick, R. M. Kramer. 1996. Swift trust andtemporary groups. Trust in Organizations: Frontiers of Studyand Research. Sage Publication, Thousand Oaks, CA.

Mohammed, S., B. C. Dumville. 2001. Team mental models ina team knowledge framework: Expanding theory and mea-surement across disciplinary boundaries. J. Organ. Behav. 2289–106.

Moreland, R. L., L. Argote. 2003. Transactive memory in dynamicorganizations. R. S. Peterson, E. A. Mannix, eds. Leadingand Managing People in the Dynamic Organization. LawrenceErlbaum, Mahwah, NJ.

Moreland, R. L., L. Myaskowsky. 2000. Exploring the performancebenefits of group training: Transactive memory or improvedcommunication? Organ. Behav. Human Decision Processes 82117–133.

Moynihan, D. P. 2005. Leveraging collaborative networks in infre-quent emergency situations. Collaboration series: Researchreport. IBM Center for the Business of Government (June).

Nagle, K. J. 2005. Katrina’s federal heroes. Traffic World (Sept. 19).

Neal, D., B. Phillips. 1995. Effective emergency management: Recon-sidering the bureaucratic approach. Disasters 19 327–337.

New York Times. 2005. Government prepares for next big disaster.(Dec. 31).

New York Times. 2006. Red Cross sifting internal charges over Katrinaaid. (March 24).

Peterson, R. S., E. A. Mannix. 2003. Leading and Managing Peoplein the Dynamic Organization. Erlbaum, NJ.

Poole, M. S., A. B. Hollingshead. 2005. Theories of Small Groups:Interdisciplinary Perspectives. Sage, Thousand Oaks, CA.

Quarantelli, E. L. 1998. What Is a Disaster? Routledge, New York.

Quinn, J. B., P. Anderson, S. Finkelstein. 1996. Making the most ofthe best. Harvard Bus. Rev. 74(2) 71–80.

Tierney, K. J., J. E. Trainor. 2004. Networks and Resilience in theWorld Trade Center Disaster. Center for Earthquake Engineer-ing Research, University of Buffalo, Buffalo, NY.

Tierney, K. J., M. K. Lindell, R. W. Perry. 2001. Facing the Unex-pected: Disaster Preparedness and Response in the UnitedStates. Joseph Henry Press.

Time. 2005. Four places where the system broke down. 166(Sept. 19)34–41.

Trainor, J. E. 2004. Searching for a System: Multi-OrganizationalCoordination in the September 11th World Trade Center Searchand Rescue Response. Doctoral dissertation, Department ofSociology and Criminal Justice and Disaster Research Center,University of Delaware, Newark, DE.

U.S. Congress. 2005. House Ways and Means Committee, Subcom-mittee on Oversight. Hearing to Review the Response by Char-ities to Hurricane Katrina. (Dec. 13).

Vera, D., M. Crossan. 2005. Improvisation and innovative perfor-mance in teams. Organ. Sci. 16 203–224.

Wachtendorf, T. 2004. Improvising 9/11: Organizational Improvisa-tion Following the World Trade Center Disaster. Doctoral dis-sertation, Department of Sociology and Criminal Justice andDisaster Research Center, University of Delaware, Newark, DE.

Wachtendorf, T., J. M. Kendra. 2005. Improvising disaster in thecity of jazz: Organizational response to Hurricane Katrina,http://understandingkatrina.ssrc.org/Wachtendorf_Kendra/.

Wegner, D. M. 1987. Transactive memory: A contemporary analysisof the group mind. B. Mullen, G. Goethals, eds. Theories ofGroup Behavior. Springer-Verlag, New York.

Wegner, D. M. 1995. A computer network model of human transactivememory. Soc. Cognition 13 319–339.

Weick, K. E. 1998. Introductory essay: Improvisation as a mindsetfor organizational analysis. Organ. Sci. 9 543–555.

Dow

nloa

ded

from

info

rms.

org

by [

131.

151.

244.

7] o

n 06

Nov

embe

r 20

14, a

t 22:

27 .

For

pers

onal

use

onl

y, a

ll ri

ghts

res

erve

d.