Review: A Cognitive-Affective Model of Organizational ... · Te™eni/Organizational Communication and IT 252 MIS Quarterly Vol. 25 No. 2/June 2001 edge of the mechanisms that guide

Te�eni/Organizational Communication and IT

MIS Quarterly Vol. 25 No. 2, pp. 251-312/June 2001 251

MISQ REVIEW

REVIEW: A COGNITIVE-AFFECTIVE MODEL OFORGANIZATIONAL COMMUNICATION FORDESIGNING IT1, 2

By: Dov Te�eniCenter of Global Knowledge ManagementBar-Ilan UniversityRamat-GanISRAEL [email protected]

Abstract

There are several theories available to describehow managers choose a medium for communi-cation. However, current technology can affect notonly how we communicate but also what wecommunicate. As a result, the issue for designersof communication support systems has becomebroader: how should technology be designed tomake communication more effective by changingthe medium and the attributes of the messageitself? The answer to this question requires a shiftfrom current preoccupations with the medium of

1Richard Watson was the accepting senior editor for thispaper.

2MISQ Review articles survey, conceptualize, andsynthesize prior MIS research and set directions forfuture research. For more details see

http://www.misq.org/misreview/announce.htmlThe associated web site for this paper is located at

http://misq.org/misreview/teeni.shtml

communication to a view that assesses thebalance between medium and message form.There is also a need to look more closely at theprocess of communication in order to identifymore precisely any potential areas of computersupport.

This paper provides the spadework for a newmodel of organizational communication, and usesit to review existing research, as well as to sug-gest directions for future research and develop-ment. Beginning with the crucial aspects ofaction, relationship, and choice, an integratedmodel of how people communicate is developed.This model incorporates three basic factors:(1) inputs to the communication process (task,sender-receiver distance, and values and normsof communication with a particular emphasis oninter-cultural communication); (2) a cognitive-affective process of communication; and (3) thecommunication impact on action and relationship.The glue that bonds these factors together is a setof communication strategies aimed at reducing thecomplexity of communication.

The model provides a balance between relation-ship and action, between cognition and affect, andbetween message and medium. Such a balancehas been lacking in previous work, and we believeit reflects a more realistic picture of communica-tion behavior in organizations. A set of proposi-tions generated from the model sets an agendafor studying the communication process as well asits inputs and outputs. Furthermore, this knowl-


252 MIS Quarterly Vol. 25 No. 2/June 2001

edge of the mechanisms that guide behavior isused to demonstrate the potential for developingdesign principles for future communication supportsystems.

Keywords: Organizational communication,communication complexity, cognition, affect,organizational memory, design

ISRL Categories: HA08, AC0401, HA10, AD0518

Introduction

Motivation, Scope and Contribution

Nowadays, managers have at their disposal awide variety of communication technologies fromwhich to choose. A number of recent studieshave reviewed and extended theories of howmanagers choose a medium for a specific situa-tion.3 Nevertheless, current technology can alsoaffect what we communicate, as well as how wecommunicate it. Thus, the question for designershas become broader: how should technology bedesigned to make communication more effectiveby changing not only the medium, but alsoattributes of the message itself?

A short example of organizational communicationcan illustrate how we choose the medium andmessage form. Table 1 shows 10 messagesrecorded on three consecutive days. Threemessages are taken from a diary in the productionroom. The other seven use a variety of othermedia: e-mail, face-to-face private meeting, typedmemo, phone call, and voice mail. The messages

of the first two days center on the production oftwo paper products, of which one is red and theother blue (product #8123). A problem hasoccurred and it requires action and communi-cation to solve it. We shall refer to this examplethroughout the article, but for now, it will suffice tonote that people have different communicationgoals: they may request the next shift to takeaction on product #8123, coordinate the team-work, build a relationship with another employee,and motivate workers. People also choose to usedifferent media for different goals. Moreover,people choose different forms of a message (e.g.,the level of formality) but also make clever adapta-tions to given situations and media. For example,using a diary to convey a happy greeting with asmiling face effectively conveys an emotionthrough a medium that is usually expected tocommunicate short task-oriented messages. (Thetypo�product 1823�is intentional to demonstratelater how technology can help reduce errors incommunication.)

The model developed below attempts to explainhow people choose the message form and themedium according to goals and situations.Following on from this model, if correct, it mightthen be possible to design a computer-basedcommunication system. For example, the diary,as part of a sophisticated organizational memory,could recognize that Jack is a new worker andsupplement the message of the April 3 morningshift with additional context information such asproduct name (rather than just #8123) and detailsabout the blue dying color. More generally, amodel of effective communication could beincorporated into communication technology so asto adapt messages. This could be achieved byautomatically recommending to the sender theoptimal amount of context information in themessage.

A model of user behavior for guiding design needsto satisfy several conditions. It should describe notonly the product, but also the process of com-munication, in order to identify opportunities forcomputer support. For example, the developers ofColab (one of the more ambitious collaborativesupport systems) provide an insightful analysis of

3Webster (1998) provides a comprehensive overview oftheories that describe how managers choose a mediumfor a specific situation. Among these theories are thoseconcerning media richness (Daft and Lengel 1984, 1986)and social presence (Short et al. 1976), theories relatedto media richness (Rice et al. 1989; Straub 1994;Trevino et al. 1987), further theories such as thoserelating to physical accessibility of the medium oravailability in space and time (Markus 1994a; Reinschand Beswick 1990; Rice and Aydin 1991), and indeed,more recent suggestions such as a task closure model(Straub and Karahanna 1998) and a combined view ofutility and norms (Kraut et al. 1998).


MIS Quarterly Vol. 25 No. 2/June 2001 253

Table 1. Messages Illustrating Communication Goals, Media, andForms of Message�

Diary Entries Ad Hoc Communication

(1) April, 2, evening shift �Smith: Joey spilt tea ontothe pulp for product #8123. We had to leave tray #5open to dry (went by themanual p. 501). We did goon with the blue order andfinished it, but take care ofproduct 1823 first thing (youwill need Joey or someother painter).

(2) E-mail from Smith to contract manager: Delay in product #8123to Macy�s. Expected to be completed tomorrow a.m.(3) E-mail from contract manager to logistics (threaded to previouse-mail from Smith): postpone delivery for Macy�s to Monday.(4) Face-to-Face Smith and contract manager: �...I know what you�rethinking about Joey. I want to ask you not to make a big deal out ofall this. The guy felt bad and has already volunteered to workovertime. BTW, do you know the dining room is closed; there reallyis no place to get a snack at night.�(5) E-mail from contract manager to Smith: I�ve never met Joey. Askhim to stop by so I can get to know him.

(6) April, 3, morning shift �Mike: back on schedule. We finished product #8123left over from yesterday inaddition to planned produc-tion, but we are lower thanexpected on blue!

(7) Typed memo from contract manger to all employees:Effective immediately 4/3/2000Please refrain from bringing in drinks or food intoThe production rooms. The dining room will beOpen 24 hours a day.

(8) Phone-call Mike to Joey: �Hi Joey, it�s Mike. I just heard therewas a delay with the red order. I know you must be very tired, butcould you possibly come in for a few hours. I need to set up a teambut can�t find a painter....Great, thanks Joey. So what time can youcome in?�

(9) April, 4

Happy Birthday, Jack!

(

(10) Voice mail from CEO to all employees: Good morning. I want toremind you about the European visit tomorrow. We all need to be atour best. You must believe me when I say there will be no layoffsas a result of this merger. I have negotiated this issue to the verylast detail telling the newcomers that we have always been familyand that this is the way we stay!

�This example follows scenarios of communication found in Saunders and Jones (1990) and Robinson et al. (2000).

users� communication failures, which they attributeto the designers� need for control over the com-munication process (Tatar et al. 1991). Second,rather than building on either cognitive or affectiveaspects of communication, the model shouldcapture both aspects, so as to build a moreaccurate representation of actual behavior. Paststudies have tended to concentrate on oneaspect, rather than the other. This is nowchanging as researchers attempt to bring togetherrelational communication in organizations, actionrelated communication, and communication tech-nology for collaborative work (see, for example,Fulk 1993; Kraut et al. 1998; Poole and Jackson

1993; Sitkin et al. 1992; Webster and Trevino1995). This paper attempts to go one step furtherby offering a model of organizational communi-cation that integrates action and relationshipsthrough a set of cognitive and affective strategies.A third requirement of a model leading to designimplications is that it should explain how a singlemessage is produced if we are to support mes-sage production (Rice 1992; Webster and Trevino1995). Ideally, such support would be an optimalconfiguration of medium and message attributes.

Our first contribution is a review of the literatureleading to the development of the model. In his



seminal book Organizations in Action, Thompsonsets out to identify "a framework which might linkat important points several of the now indepen-dent approaches to the understanding of complexorganizations" (1967, p. viii). He considers hisframework to be a conceptual inventory capable ofgenerating propositions and demonstrated theplausibility of these propositions by using illus-trative studies without any systematic evidence.This paper is a step toward a conceptual inventoryof computer mediated communication in organiza-tions. It is an attempt both to assemble pastresearch on communication from diverse sourcesand to propose a new understanding of organi-zational communication. Not only does it bringtogether different perspectives from differentsources so as to provide a wider reference forfuture research, but it reveals elements of thecommunication process which can refine previousfindings and help stimulate new work. Forexample, a statement such as �face-to-face and e-mail rather than a typed letter is used to conveyinformality� (e.g., Trevino et al. 1987) may not betrue in an organization that imposes formal e-mailas its standard means of communication. On thecontrary, face-to-face communication may sym-bolize a very formal event.

The second main contribution offered by thispaper is in providing a prolegomenon of a newmodel that (1) is more balanced in its treatment ofaction and relationship, of cognition and affect,and of message and medium, and (2) digs deeperto reveal the mechanisms by which people chooseto behave. The result is a more complex descrip-tion of communication behavior that has multiplepurposes (e.g., to accomplish a task and maintaina good relationship) and uses a range of cognitiveand affective strategies. If indeed it presents amore realistic view of communication behavior, itshould be more capable of informing design. Theeconomic value of improved design of communi-cation technology should be clear. Thompson�swidely accepted theory assumes that the cost ofcommunication should be minimized, and to thisend, organizations are designed to reduce com-munication when coordination can be achievedwithout it. The balanced view of action and rela-tionship presented here reinforces the need tominimize the cost of communication, but alsoincludes in it the cost of low commitment to action

due to poor relationships between communicators.Hence, reducing human communication or auto-mating it may damage the organization when ithampers communication intended to build a rela-tionship. Nevertheless, the benefits and costs oforganizational communication are evidently veryhigh. The finding that managers spend around75% of their time communicating has not changedover the past 30 years, from Mintzberg (1973) toRice and Shook (1990). Improving organizationalcommunication should, therefore, be extremelyvaluable (see also Carlson and Davis 1998).

The Proposed Model

Organizational communication is seen from athree-fold perspective: action, relationships, andchoice. Organizations must focus on action, andcommunication plays a pivotal role in organi-zations, and may even be seen as the foundationfor most organizational action (O�Reilly and Pondy1979; Weick 1979). Hence, it must be assumedthat organizational communication eventuallyleads to action, although not all communicationcan, nor should it be, associated directly with aspecific action. In other words, communication isseen as taking action and organizations are seenas collections of communicative acts (Winogradand Flores 1986). This perspective helps to iden-tify the goals of communication insofar as theyrelate to different types of action while it also helpsto define effective versus poor communication.

Second, organizations may be described asentities engaged in social, as well as economic,exchange (Blau 1964). Since they cannot existwithout social communication, action-orientedgoals are complemented by the relationship-oriented goals of communication.

Third, a communicator will generally choose howto communicate. We use a combination of socialand utilitarian values to describe how peoplechoose their communication behavior, includingtheir choice of communication media.4

4Indeed, rational-choice models for using communicationtechnology, such as media richness theory (Daft andLengel 1986), have been influential. However, recent



Within the perspective of choice, action, andrelationship, we develop a model that has threemain factors, each of which includes severalelements (shown in Figure 1):

� Inputs to the communication process: (1) taskattributes, (2) distance between sender andreceiver, and (3) values and norms of communi-cation;

� A communication cognitive-affective processthat describes the choice of (1) one or morecommunication strategies, (2) the form of themessage, and (3) the medium through which itis transmitted; and

� The communication impact: (1) the mutualunderstanding and (2) relationship between thesender and receiver.

Looking back at Table 1, the example demon-strates several communication goals, forms ofmessage, and media. Communication strategies,however, are less obvious. For example, in tryingto influence the employees, the CEO takes theirperspective in the voice mail about the Europeantakeover. Below we enumerate several othercommunication strategies and show how theyaffect the choice of medium and message.

We use extensively the notion of communicationcomplexity to explain the choices of strategies,messages, and media. Communication com-plexity results from the use of limited resources toensure successful communication under problem-atic and uncertain conditions. It grows as thedemands of the communication process onmental resources approach their capacity (e.g.,

Rasmussen 1986). The sources of communica-tion complexity can be categorized as cognitivecomplexity, dynamic complexity, and affectivecomplexity.5

Cognitive complexity is a function of

(1) the intensity of information exchanged (inter-dependency) between communicators, whichincreases the probability of misunderstanding(Straus and McGrath 1994),

(2) the multiplicity of views held by the com-municators, which increases the plausibility ofunderstanding the message in a differentcontext than intended (Boland et al. 1994),and

(3) the incompatibility between representationand use of information, which requires theinformation communicated to be translatedbefore it can be used, and increases thedemands on resources and the probability oferror (Barber 1988; Norman 1990).

Dynamic complexity refers to how far the com-munication process depends on time constraints,unclear, or deficient feedback and changes duringthe process. Dynamic complexity increases thelikelihood of misunderstanding the required action(Diehl and Sterman 1995). For example, when thereceiver�s behavior is highly unpredictable (e.g.,lapses of attention), the communicator needs toadapt the communication process to fit in with thenew behavior.

Affective complexity, meanwhile, refers to how farcommunication is sensitive to attitudes or changesin disposition toward the communication partner orthe subject matter. It is typified by relationaloriented obstacles such as mistrust and affectivedisruptions (Salazar 1995). research has shown that these models alone cannot fully

explain empirical findings about the use of communi-cation technology (El-Shinnawy and Markus 1997; Fulkand Boyd 1991; Fulk et al. 1990; Huang et al. 1998; Lee1994; Ngwenyama and Lee 1997; for comprehensiveaccounts of evidence on media choice see also Markus1994a; Straub and Karahanna 1998; Webster 1998;Zack 1993). In this paper, social-influence accounts ofmedia selection are advanced, not so much as analternative approach to rational-choice models (e.g.,Donabedian et al. 1998), but rather as a complement ofrelationship goals to task goals (e.g., Webster andTrevino 1995).

5Cognitive and dynamic complexity are action oriented,corresponding to coordinative, component, and dynamiccomplexity in defining task complexity (Wood 1986).



Goal Strategies

Medium

MessageForm

Communication process

56

43

2Task

Values and Norms

Mutualunderstanding

Relationship

Communicationinputs

8, 9,10 1

Communicationimpact

7Distance Goal Strategies

Medium

MessageForm

Communication process

56

43

2Task

Values and Norms

Mutualunderstanding

Relationship

Communicationinputs

8, 9,10 1

Communicationimpact

7Distance

Figure 1. A Cognitive-Affective Model of Organizational Communication

A simple task can demonstrate these sources ofcommunication complexity. Joe asks Rita to helphim bring his new boat to harbor through a narrowand long water passage. They decide to row theboat and adjust the boat�s course, wherenecessary, by rowing faster on one side andholding the oar steady on the side to which theywish to turn. They each take a side of the boatand row in parallel, both maintaining eye contactand each deciding on the rowing pace byestimating the leeway between the boat and thebank. Now suppose that Joe cannot see what ishappening on Rita�s side and vice versa. Theyneed to communicate continuously to inform oneanother of the leeway on each side in order towork harmoniously. The intensity of the com-munication is higher because Joe must rely oninformation from Rita before he can act. Cognitivecomplexity is, therefore, higher than in theprevious scenario. Now let us suppose, further,that it takes a few seconds to see the effect ofadjusting the speed of rowing on the change ofcourse. Joe says �Right� but sees no immediatereaction (feedback) and is left uncertain as towhether Rita heard or understood his message.Dynamic complexity is therefore high and Joe mayconsequently ask Rita to confirm by saying �OK�whenever she gets a message. Finally, let us saythat Joe is not sure about how Rita feels towardhim today and suspects that she may not mindterribly if the boat gets scratched. So when he

screams �Right� and she answers slowly �OK,� heis not sure that he can rely on her intent to followwith appropriate action. This is a state of highaffective complexity.

The three factors and their elements are mappedin Figure 1 and described in Table 2. Later, theattributes or classes of the model are definedmore precisely. The paper proceeds as a journeyfrom a more abstract discussion of communicationgoals rooted in philosophy, through to an analysisof cognitive and affective strategies built onbehavioral sciences, to the more concrete designimplications with regard to information and com-munication technologies. Figure 1 serves as amap to keep on track. The criterion for choosingthe landmarks for the central path concerns howto best uncover the process of communication sothat others can forge new paths along similar linesfor new conditions. For example, organizationalpolitics, which are left out of the analysis, undoub-tedly play a role in shaping communication, butthey too could be investigated in the future, usingthe same rationale developed to analyze the effectof culture. So while the review of the literatureattempts to be comprehensive within the boundaries sketched out in Figure 1, the developmentof the model concentrates on representativeelements. The choices of these elements aredescribed in more detail later.



Table 2. Glossary of Elements in the ModelFactor Element Description: Classes and Attributes of Elements

Impact Mutualunderstanding

The communicative act is judged to be comprehensible andtrue.

Relationship The communicative act is judged to be trustworthy andappropriate.

Process Communicationgoals

The sender's intended impact of communication on thereceiver: instruct action, manage interdependent action,manage relationships, and influence.

Communicationstrategies

Methods of coping with communication complexity to achievecommunication goals: contextualization, affectivity, control,perspective taking, and attention focusing.

Message form Characteristics of the form of the information communicated:size, distribution, organization, and formality of the message.

Medium Characteristics of the physical medium on which the mes-sage is transmitted: channel capacity, interactivity, andadaptiveness.

Inputs Task Characteristics of the task situation: analyzability, variety, andtemporal demands.

Sender/receiverdistance

The relative situations of sender and receiver: cognitive andaffective.

Values andnorms

Cultural values are stocks of knowledge that guide behaviorof communicators belonging to that culture: independence-interdependence.

Communicationcomplexity

Cognitivecomplexity

The complexity due to the intensity of information exchange,the multiplicity of views and the incompatibility betweenrepresentation and use of information.

Dynamiccomplexity

The complexity due to time constraints, deficient feedback,and changes during the process.

Affectivecomplexity

The complexity due to sensitivity to attitudes and changes indispositions.

Each of the numbered arrows in Figure 1 repre-sents a general proposition developed below. Theexposition of the model follows Figure 1 from rightto left, beginning with communication impact(proposition 1) in the next section of the paper.The heart of the discussion is presented in thesection that explains the elements of the com-munication process, summarizes the principles ofbehavior assumed in this model, and proposeseffective combinations of goal, strategy, medium,and message attributes (propositions 2 through 7).We then examine how inputs into the communi-

cation process can affect it by determining aperson�s goal priorities and setting limitations onbehavior (propositions 8 through 11). Specialattention is devoted in this section to the inter-cultural effects on the communication process dueto the growing importance and recent interest inglobal communication. The final section looks atthe model's implications for research and design.It takes the model's three factors (Figure 1) as aspecification of required functionality, adds thenotion of organizational memory as a necessaryresource, and postulates some general design



guidelines for future communication supportsystems.

The last introductory note specifies the level of thetheory (Klein et al. 1994). In Table 1, messages1, 2, 3, and 8 are all associated with the task ofproducing and delivering product #8123 withminimal delay. These four messages may be seenas a higher-level group of individual messagesthat are all influenced by certain attributes of thecommon task. The model described below is amixed-determinants model in the sense that theelements of a message (e.g., its medium) aredetermined by other elements of the message (itsgoal) but also by elements of the common task(e.g., the urgency to complete the task withminimal delay). The section on the communicationprocess develops the relationships between ele-ments of a message, assuming that the messagesare independent of group effects. The section oninputs to the communication process introducesthe effects of inputs such as task attributes onmessages, assuming that messages are relativelyhomogeneous with respect to the inputs.6 Forexample, all messages associated with an urgenttask will usually be communicated by phone or e-mail but not by typed memos. The effects of bothlevels may interact. For example, messages thatare not only associated with an urgent task (agroup level) but also involve persuasion (amessage level) will be communicated by phonerather than by e-mail more frequently than urgentmessages not involving persuasion.

Elements of CommunicationImpact

The Theory of Communicative Action

A definition of communication impact necessarilybegins with an explanation of what is actually

meant by communication. Of the many definitionsof communication, we have sought one whichemphasizes goal driven behavior, which is latershown to be the basis for choosing strategies soas to impact action and relationship. Gerald R.Miller claims that, �In the main, communicationhas its central interest in those behavioralsituations in which a source transmits a messageto a receiver(s) with conscious intent to affect thelatter�s behavior� (1976, p.92). By analyzing themessage, its communicative impact can beassessed to the extent that the sender�s intent(goal) has been understood and accepted oncethe message has been received. Furthermore, itis only the immediate impact of the communicativeact on the receiver�s desire to react that isinvestigated.7

The model of organizational communicationproposed here is a pragmatic theory of a rationalcommunicator who uses resources to implementcommunication goals. We build on the theory ofcommunicative action (Habermas 1984), whichsimilarly situates social communication within atop-down hierarchy of goals and resources in thecontext of social norms and cultural values.However, to be able to move from a theory ofcommunication to the design of systems thatsupport communication, the top-down view mustbe complemented with a bottom-up view of howlimited resources are used to achieve goals.

6We use the terminology offered by Klein et al. (1994).In mixed-determinants models, determinants from avariety of levels may influence the dependent variables.Homogeneity with respect to a construct implies that thegroup members� values on a given construct aresufficiently similar so that it is meaningful to talk aboutthe group construct.

7An in-depth examination of communication makes itnecessary to concentrate on the direct impact of com-munication to keep the discussion manageable. Forexample, Computer-Supported Cooperative Work litera-ture usually treats the impact of using communicationsystems on action and relationship. Hollingshead andMcGrath (1995) provide several instances of suchimpacts. Regarding action, communication affects(1) the task product, e.g., time, number and quality ofdecision solutions, and (2) the users� reactions, e.g.,satisfaction and rated effectiveness. Regarding rela-tionship, communication affects (1) the relationsbetween actors, e.g., attraction and feeling of alienation,and (2) the pattern of interaction, e.g., total or non-taskamount of participation. This paper examines the impactof the communicative act on the receiver�s under-standing of the message and on the receiver�s relation-ship with the sender, but does not examine any aspectsof organizational impact, such as decision quality.Ideally, the individual and organizational levels ofanalysis can be integrated (O�Reilly et al. 1987).However, for practical reasons, this paper is limited inscope and assumes that better communication will leadto better decision quality.



These two views explain, respectively, what andhow people communicate, and, when examined inthe context of a particular situation, they providethe necessary knowledge for design. For example,if we assume that senders exercise more controlover interpersonal messages when they feel thereceiver's reaction is less predictable, a designerwho knows this may build into the system bothhigh and low levels of control (e.g., instant versusdelayed feedback). Moreover, the system couldbe designed to set the default level of controlaccording to the level of uncertainty about thereceiver's reaction (e.g., according to how closelythe sender and receiver have worked together).But such design decisions rely on knowledge ofhow people communicate and, therefore, extendsthe scope of the theory of communicative action.Below we simplify the framework used by Haber-mas and argue why this framework is appropriatefor a model of organizational communication,while we also indicate in general terms what mustbe adapted.

Habermas (1998) developed his theory of com-municative action as an element of a moregeneral theory of society.

[The] concept of communicative actionrefers to the interaction of at least twosubjects capable of speech and actionwho establish interpersonal relationships.The actors seek to reach an under-standing about the action situation andtheir plans of action in order to coordi-nate their action by way of agreement(Habermas 1984, p. 86).

Communicative action (a behavior) takes place inrelation to three additional factors�culture,society, and person�that together constitute the�lifeworld,� which serves as the context forcommunication. These four factors are organizedin Figure 2, alongside the corresponding productsby which action is oriented: resources, values,norms, and goals. In order for goals to beachieved, coordination between communicators isnecessary, as is the commitment of communica-tors to behave in certain ways, which is assumedto be part of establishing relationships.

Habermas further talks about two characteristicsunderlying communicative action: rationality andcomplexity. Communicative action is based on the

senders� obligation to provide the reasoning forthe validity of their claims, and furthermore, theseclaims may be rooted in or regulated by the life-world shared by the communicators. The com-plexity of implementing communicative actiongrows with the need for coordination and type ofaction orientation (goals, norms, and values).These two characteristics are shown, respectively,as a top-down and bottom-up arrow in Figure 2.The distinction between goals and resources iscommon in social (and organizational) analysis,but the difference between norms and valuesrequires explanation. Norms are the orders ofinterpersonal relationships accepted by society forthe purposes of regulating practices and habitualbehavior. Norms are expected to be validated notonly against standards of rationality but alsoagainst standards of relationships. In contrast,engrained in the culture are values of objects ofutility that are not usually put to tests of validity.

To adopt communicative action for organizationalcommunication, we evaluate whether the systemdepicted in Figure 2 is appropriate in the organiza-tional setting. We regard an organization as asocial system that can be characterized byresources, goals, norms, and culture. Moreover,we assume that organizations rely on coordinatedaction and, indeed, act under norms of rationalityin the sense that actors of the organization areguided by their goals, norms and culture on howto act in order to produce desired outcomes(Thompson 1967). Communication complexityhas already been discussed. In other words,looking at Figure 2, one can substitute �person�with an actor in an organization, replace �society�with organization, and consider �culture'�as bothorganizational and national culture. Finally, in thisorganizational setting, we take communicativeaction to be an ideal form of organizational com-munication. Indeed, several studies of informationsystems have recently used the theory of com-municative action to understand organizationalcommunication (e.g., Ngwenyama and Lee 1997;Ngwenyama and Lyytinen 1997).8

8Recent studies have used Searle�s (1969) theory ofspeech acts (e.g., Moore 1998; but see criticism byJanson and Woo 1996), Habermas�s (1984, 1987)theory of communicative action (e.g., Mingers 1995;Ngwenyama and Lee 1997; but see criticism bySharrock and Button 1997), and a combination of bothSearle and Habermas (e.g., Schoop 1997).



PersonGoals

BehaviorResources

SocietyNorms

CultureValues

Rational it y

Com

plex

ity

PersonGoals

BehaviorResources

SocietyNorms

CultureValues

Rational it y

Com

plex

ity

Figure 2. A Simplified View of the Lifeworld in Communicative Action by Habermas

The framework in Figure 2 must, however, beadapted to form the basis for design. First, whilethe theory of communicative action has littleconnection with the physical aspects of com-munication, a model leading to design mustaddress the interaction between the communi-cator and the media. The levels of context (inFigure 2) are necessary but not sufficient fordesigning information technology; they must becomplemented with lower levels of abstraction thatdescribe the physical functions and form ofcommunication (Rasmussen 1986). Therefore, inour model, behavior will be described in greaterdetail as communication strategies, medium, andmessage. Second, Habermas' categories of goalsand actions, which he sees as universal, may beinadequate to capture the idiosyncrasies of aspecific setting such as an organization (Sharrockand Button 1997). Therefore, we use the prin-ciples of communicative action, but do not adoptthe detailed classifications. Third, in the theory ofcommunicative action, a situation represents atemporally and spatially defined segment of the

lifeworld that is delimited in relation to goals, butthere is little concern with how the situation affectsbehavior. Therefore, in our model, situation will becharacterized more specifically as task andsender-receiver distance to demonstrate how theyaffect behavior.

Communication Impact Definedas Mutual Understanding andRelationship

The first implication of the theory of communi-cative action is a definition of communicationimpact. Habermas (1984, 1987) claims that fourconditions are necessary for a communicative actto take place:

(1) the act must be comprehensible, so that thereceiver can understand the sender;

(2) the act must be true, so that the receiver canshare the sender�s knowledge;



(3) intentions must be expressed truthfully, sothat the receiver can trust the sender; and

(4) the act must be appropriate within somenormative context so that the receiver canagree with the sender within this valuesystem.

Communication is at once an act of building amutual understanding between sender andreceiver and building a relationship between them.The relationship is necessary for gaining a com-mitment from the receiver to the sender or thelarger group to which they belong. In fact, it is theact of communication more than the informationalcontent that produces this commitment (Huff et al.1989). Although some may regard the third con-dition to be somewhat naïve, truthful expressionsof intentions are the basis for the commitmentnecessary in social exchange (Blau 1964).

These four conditions of validity allow us todevelop two mirror perspectives: (1) definingmutual understanding and relationship as the im-pact of successful communication and (2) charac-terizing impediments to action and to relationshipsas poor communication. While Habermas com-bines understanding and relationship, we separatethe two, acknowledging that they are interwovenyet assuming that people can distinguish betweenthe two. Mutual understanding refers to the firstand second conditions (a comprehensible andtrue message) and is associated with the action-oriented aspect of communication. Mutual under-standing includes not only the receiver�s under-standing of the message, but also the sender�sawareness that the recipient of the message hasunderstood it. Relationship refers to the third andfourth conditions (trustworthiness and appro-priateness) and is associated with the dimensionof relationships within the communicative act(Habermas 1984). Successful communicationnecessitates both aspects, although the marginalimpact of a single communicative act on rela-tionship may be smaller than that on mutualunderstanding and may depend on the precisecommunication goal. For example, if the sender�sgoal is to convey the price of a product, thenmutual understanding is the desired impact. Onthe other hand, if the goal is to influence the

receiver to purchase the product, then thecommunication must not only be comprehensible,but also appropriate.

Relationship and mutual understanding are, ofcourse, closely interrelated. Indeed, work rela-tionships, and more specifically, mutual trust,facilitate a more productive flow of information(Hart and Saunders 1997; Nelson and Cooprider1996). Moreover, relationships build trust, whichcan be described as the confidence that thereceiver will fulfill obligations and behave in a fairand predictable manner (Anderson and Weitz1989). Communication is, therefore, more effec-tive when trust and commitment are high (e.g.,Dore 1983; Williamson 1975). Explanations per-ceived to be timely and sufficiently detailed toallow for adequate understanding of the messageare more likely to lead to trust (Whitener et al.1998). In fact, any communication betweenmanagers and subordinates that appears to beopen builds trust (Gabarro 1978). On the otherhand, faulty communication and unsuccessfulinteraction make it impossible to reduce psycho-logical distances between people (Schein 1996).Thus, mutual understanding and relationship areintimately related.

The four conditions of communication invalidityalso allow us to investigate poor communicationas an impediment to action and relationship. Sucha perspective is useful for two reasons: in prac-tical terms, failures in communication are ofteneasier to measure than successful communi-cation, while conceptually, a focus on communi-cation invalidity can serve to explain behavior thatattempts to overcome impediments to action. Inline with the action perspective adopted here, poorcommunication can be defined as an impedimentto action, that is to say, any exchange of infor-mation that leaves the receiver unable, unwilling,or unsure of how to proceed with the sender�sintended impact. Taking a similar approach, DeBono (1976) considers thinking from the point ofview of action and defines an effective explanationas one that allows an individual to decide on whatto do next.

High levels of communication complexity can leadto communication failures. Cognitive complexity



and dynamic complexity impede mutual under-standing by making it difficult to understand andshare knowledge, which are the first two condi-tions of successful communication noted byHabermas (1984). Affective complexity primarilyimpedes the processes of building trust andsetting an appropriate normative context of thecommunication that is acceptable to both receiverand sender. These are the third and fourth condi-tions of successful communication. Overcominghigh complexity is thus central to successful com-munication. In theories of human information pro-cessing, it is generally agreed that higher com-plexity results in poorer performance, althoughextremely low levels of complexity may reduceperformance by failing to arouse sufficientattention (Rasmussen 1986; Schroder et al.1967). We adapt this general observation to thespecific domain of communication in proposition 1and treat it here as an axiom from which otherpropositions are derived.9

Proposition 1: Beyond some minimal valueof communication complexity, a reduction init will result in higher levels of mutualunderstanding and relationship.

In sum, the impact of a communication is linked tothe receiver�s intent to react. The theory of com-municative action is used (1) to define communi-cation impact as mutual understanding and rela-tionship and (2) to define impediments to actionand relationship, and characterize their causes ascognitive complexity, affective complexity, anddynamic complexity. Reductions in communicationcomplexity are expected to improve communi-cation.

The Communication Process

The discussion of the communication processbegins with communication goals (see Figure 1)and proceeds by incorporating three inter-wovenelements, as follows:

� the communication strategies for a given com-munication goal;

� the medium on which the message is trans-mitted;

� the form in which the information is packed intoa message.

Although not necessarily sequential, they will beexplained one by one in the sections that follow(Table 3 provides a glossary of all the processelements). Once the first two elements (goals andstrategies) have been discussed, we will elaborateon the assumptions behind behavior and choice,and will use these to explain how goals affect thechoice of strategies. The same rationale forchoice is later used to explain how strategiesaffect the choice of message and medium, andhow inputs affect the process (discussed in thesection on inputs to the communication process).

Goals

Habermas (1987) discusses four broad socialprocesses that require communication: reachingunderstanding, coordinating action, building rela-tionships (socialization), and strategically influen-cing others. Several empirically derived classifi-cations of organizational communication goalsexist (see recent classifications in Carlson andDavis 1998; Kettinger and Grover 1997; Mackay1988; Orlikowksi and Yates 1994; Poole andHirokawa 1996; Te'eni and Schwartz 2000). Theyare diverse and more elaborate than Habermas�four processes. This is presumably because theyattempt to map the communication goals to amultitude of observable activities within theorganization, e.g., processing of information vs.choice making. In the interest of parsimony intheory development, we choose here to build onthe theoretical foundations in communicativeaction, adapting them to the organizational

9The discussion is framed as a set of propositions, thefirst of which is taken here to be axiomatic and theremaining propositions are part of an expandable set oftestable theorems (Blalock 1969). In discussing how toconstruct verbal theories, Blalock argues that �Axiomsare propositions that are assumed to be true. Theorems,on the other hand, are derived by reasoning, or deduced,from the axioms� (p. 10). He further suggests two rulesfor stating theories in verbal form: (1) axioms should bestatements that imply direct causal links and(2) theorems should be stated in a testable form.



Table 3. Glossary of Attributes in the Communication Process

Element Attributes Definition

Goals

Instructing action Getting the receiver to act according to the sender�s wishes.

Managing inter-dependent action

Coordinating interdependent actors.

Managing relationships Fostering relationships between people at work.

Influencing Attempting to influence behavior and attitude to conform to the sender�swishes but realizing the receiver can behave differently.

Strategies

Contextualization Provision of explicit context in messages.

Affectivity Provision of affective components (emotions, moods) in messages.

Control�testing andadjusting

Testing and adjusting communication according to feedback during theprocess.

Control�planning Planning the pattern of communication and contingencies ahead of theprocess.

Perspective taking Considering the receiver�s view and attitude.

Attention focusing Directing or manipulating the receiver�s information processing.

Media

Interactivity The medium�s potential for immediate feedback from the receiver.

Channel capacity The medium�s potential to transmit a high variety of cues and languages.

Adaptiveness The medium�s potential to adapt a message to a particular receiver.

Message

Size Number of semantic units.

Distribution Number of destinations to which the message is sent.

Organization The extent to which the message is ordered to support mutualunderstanding.

Formality The abstraction of a description toward closure of action according tothe rules of communication in the particular organizational setting.



setting.10 Furthermore, as we define each of thefour communication goals, we also relate it tocommunication complexity.

The simplest instance of reaching understandingis in instructing action, which has the goal ofgetting the receiver to act according to thesender�s wishes (Habermas calls this �instrumen-tal action�). One such action is a request forinformation. Although most of the empirical classi-fications of organizational communication men-tioned above appear to classify informationseeking as a distinct class of action, we remainconsistent with the theory of communicative actionand include a request to provide information as aspecial case of instructing action. Furthermore,there is no inherent source of complexity in thiscommunication. Indeed, other goals will be seento be more complex by comparison.

The aim of managing interdependent action is tocoordinate interdependent actors. In comparisonto instructing action, managing interdependentaction is higher in terms of cognitive complexity.According to Thompson (1967), interdependentaction generates cognitive complexity due tointerdependence between actors that can beeither serial (one�s action depends on the other�saction) or reciprocal (one�s action both dependson and affects another�s action). Furthermore,when the task requires simultaneous interactionsbetween actors, dynamic complexity increases too(Van de Ven et al 1976).

The purpose of managing relationships is to fosterrelationships between people at work (we use theterm managing relationships rather than buildingor maintaining relationships to connote the wholerange of actions beyond initiating creating arelationship, e.g., preserving, strengthening, and,if necessary, severing relationships). The com-munication needed to manage relationships is notaction-oriented (Scollon and Scollon 1995; Streetand Cappella 1985; Weedman 1991). This goal is,therefore, closely related to affective complexitydue to the possible dispositions between senderand receiver. Moreover, the absence of any focuson action often widens the possible meaning ofthe message, making it more subjective. Indeed,managing relationships is most frequently invokedin situations of changing or deteriorating relation-ships (Lee and Jablin 1995). Where such com-munication is irregular, and the interpersonalcontext of the sender-receiver communication isless established, the uncertainty of the receiver'sreactions to the communication is high. Thisuncertainty also increases dynamic complexity.

Influencing can be either action oriented or rela-tionship oriented. Influencing is about attemptingto influence behavior and attitude in order toconform to the sender�s wishes but realizing thereceiver can behave differently. Influencing isoften concerned with resolving conflicts and, thus,it reflects high interdependence between com-municators, more so than thinking collectively(Straus and McGrath 1994). Moreover, the needfor influencing assumes a multiplicity of views orpreferences held by the communicators, whichneed to be connected. Influencing is, therefore, ofhigh cognitive complexity. Furthermore,influencing assumes that the receiver�s behaviorcannot be determined in advance and, therefore,that communication should be sensitive tochanges in the behavior. Influencing includesattempts to change behavior by bringing about analteration in attitude. Hence, influencing maydepend on the receiver�s dispositions toward thesender or the subject matter and is, therefore,usually high in affective complexity.

Table 4 lists the four communication goals and thecorresponding sources of communication com-plexity to which they are most susceptible.

10While most recent classifications of organizationalcommunication differentiate between communicationaround current action and communication for somefuture action, which includes knowledge acquisition(Carlson and Davis 1998; Kettinger and Grover 1997;Poole and Hirokawa 1996; Te'eni and Schwartz 2000),we decided to keep to the theoretical framework ofcommunicative action, in which the situation of socialaction is defined by temporal and spatial conditions. Thedistinction between communication about current andfuture action (e.g., formal procedures for actions) will,therefore, be characterized in the present study by anattribute of the task element called temporality (dis-cussed in the section on inputs to the communicationprocess).



Table 4. Communication Goals and Corresponding Sources of Copmlexity

Goal Inherent Source of Complexity

Instructing action (includesrequesting information)

Managing interdependentaction

High cognitive complexity due to interdependency and multiplicity ofviews. High dynamic complexity when interdependent action is donein parallel.

Managing relationships High affective complexity due to dependency on interpersonaldispositions. High dynamic complexity when reactions are uncertain.

Influencing High cognitive complexity due to multiple views and high affectivecomplexity due to dependency on interpersonal dispositions.

Communication Strategies

The four classes of communication goals(Table 4) and their inherent sources of communi-cation complexity are used below to explain thechoice of communication strategies. Communi-cation strategies may be defined as the means bywhich communication goals can be fulfilled.11

Such strategies are needed to overcome thecomplexities of the communication process. Thesix communication strategies discussed below aresummarized in Table 3. They have evolved fromprevious research in communication, human infor-mation processing, and organizational behavior,as well as from our own observations. The termscommunication strategies, patterns of communi-cation behavior, communication preferences, andcommunication styles were used interchangeablyto search the literature. Most of the strategiesoutlined have been adapted to fit in with the levelof analysis of a message and the cognitive-affective orientation adopted in this paper.

The list of strategies selected here is likely to beincomplete. However, it is at least representativeof the three different stages in the communicationprocess: the creation of a message by thesender, the transmission of the message between

sender and receiver, and the receipt of themessage by the receiver. Communication isachieved by being more or less active in thefollowing activities: (1) creation�building cognitiveand affective components in the message;(2) transmission�controlling the informationtransmission through planning and throughtesting; and (3) receipt�considering the receiver�sperspective on the issue and affecting thereceiver�s information processing. These stra-tegies can be seen as the centerpiece of thecommunication process.

ContextualizationContextualization may be described as the pro-vision of explicit context in the message. Itrequires the sender to build an explicit inter-pretation of the issue as opposed to noting onlythe desired reaction or core message. Thus,contextualization is central to theories of compre-hension and is necessary for improved problemsolving performance (Kintsch 1988; Mayer 1985;van Dijk and Kintsch 1983). Context is usuallyconstructed through layers around the coremessage that explains, among other issues, thefollowing: how an action can be performed; howit can be broken down into sub-actions; how theaction answers its motivation; what informationmay be related to the message; what alternativeinterpretations are possible. Piaget, for example,discusses understanding as "the 'how' and 'why' ofthe connections observed and applied in action"(1978, pp. 218). In information systems, the con-

11 The term �communication strategies� is borrowed fromthe literature on inter-language communication to denoteways in which people tackle and overcome problems inpursuing a communicative goal (Kasper and Kellerman1997).



notation of contextualization is narrower: it seeksto elucidate the situation in which the messagewas created, detailing such issues as who iscommunicating with whom, when, and under whatconditions (Schwartz and Te�eni 2000). Message#3 in Table 1 is an instruction to postpone delivery(the core of the message) and message #2 (whichis threaded) gives the reason for the instruction(the context). Additional context information aboutthe message creation is the sender (contractmanager), receiver (logistics) and date (April 2).

AffectivityAffectivity may be seen to be the inclusion ofaffective components in the message that des-cribe emotions and moods, not necessarilypleasant ones (Schwarz 1990). Emotions aremore intense, relatively short-lived, and usuallyprompted by a clear trigger, such as excitementabout the prospects of success, an apology, andthe pleasure of meeting someone. Moods, suchas the state of feeling good, are rather longerterm affective states, usually with no salientcause. Schlosberg (1952) has mapped affectaccording to two dimensions: attention-rejectionand pleasantness-unpleasantness. This was laterreconfirmed for communication that is both non-verbal (Green and Cliff 1975) and verbal (Osgood1969). Affectivity can be used to motivate, e.g., tosustain favorable attitudes and dispositions, andto inform, e.g., provide information about thesubjective evaluation of a product. Indeed, such aquality is needed to cope with potential com-munication problems due to affective complexity.

Control by Testing andControl by PlanningControl is largely a matter of overseeing and, ifnecessary, adjusting the communication processto assure effective communication. For example,Street and Cappella (1985) note the need formaintaining coherence in discourse and managingdominance and control, and Clark and Brennan(1991) emphasize the continuous need to coor-dinate content and process. Moreover, differentlevels of control are needed for different types ofgoals (Jordan 1998). Control can take two basicforms: (1) planning the pattern of communication,and if necessary contingency patterns, ahead of

the process, and (2) testing and adjusting on thebasis of feedback during the process (online).Indeed, people are capable of recognizing andadopting spontaneous versus planned communi-cation, depending on whether the interdepen-dence between them is parallel or sequential (Lea1991).

For planned control, one needs to considerwhether the communication process is pre-determined, leaving the locus of control with thesender, or whether it is flexible, leaving open theprogression of communication and letting controlshift from one partner to another. A particularcharacteristic of control through planning is theclear designation of who does what in the com-munication process and a distinction between planand implementation. In planned control, messageredundancy, especially repetition of key ideas, isused to ensure successful communication (Mayer1985). For example, senders sometimes copymessages to other people with the sole intentionof pressuring the receiver to take action (Philipsand Eisenberg 1993). Additionally, the sendermay send the same message several times,rephrase messages, and summarize previouscommunication. Important characteristics of thiscontrol behavior are, therefore, redundancy andrepeated communication (�I�m sending thismessage again to your other address,� �Attachedis a summary of our phone conversation�) andmultiple recipients (including multiple copies andblind copies).

In seeking to achieve control through testing andadjusting, the sender plays an active part inensuring that the process works well. Timelyfeedback is, of course, essential for effectivecontrol (Te�eni 1992). For example, the senderrepeatedly asks the receiver if the communicationwas successful and adjusts the messageaccording to the receiver�s reaction. Characteristicof this control behavior are online tests of thecommunication such as �Did you get the mes-sage?� and �Do you understand the message?�

Perspective TakingPerspective taking is concerned with whether thereceiver�s view and attitude are a target of thecommunication or whether they are left outside



the scope of communication. This strategy in-cludes both cognitive and affective aspects of thereceiver�s perspective. Krauss and Fussell(1991b) argue that perspective taking, in whichthe sender actively considers (imagines) thereceiver�s point of view, is necessary for thecommunication to be comprehensible. At a mini-mum, it requires you to consider what your partnersees and hears of your message (Schober 1993).In a similar but broader sense, Scollon and Scol-lon (1995) use the term involvement to describe acommunication strategy in relational communi-cation. They too stress the sender�s involvementin the receiver�s world, but include in it the waythat world is seen publicly. Perspective taking canbest be demonstrated by taking interest in thereceivers� viewpoints, inquiring about their affairsand attitudes and supporting them, sharing com-mon beliefs and talking in a personal style (e.g.,message #10 in Table 1). It usually includes thesender�s expression of attitude that can be char-acterized by the use of magnifying adverbs andattitudinally loaded words (Eggins and Martin1997).

Attention FocusingIn attention focusing (also known as �flagging� or�contextualization cues�), the sender attempts todirect or even manipulate the receiver�s pro-cessing of the message. In organizational theory,attention focusing is a well-known strategy usedby managers who wish to direct knowledgeacquisition (Simons 1991). At the level of com-municating a message, attention focusing mayinvolve highlighting parts of the message,switching from small to large letters, shouting aftertalking softly, etc. (Gumperz 1982). This processinvolves the use of various techniques to affectinformation processing, such as switching format(size, uppercase, underline, etc., such as inmessage #7 in Table 1), switching styles, andcreating an unusual composition (e.g., sequenceof sentences, repeats, headings). Using some ofthese techniques in moderation is common, butdeliberate, sophisticated and pervasive use ofmultiple techniques cannot be taken for granted.Moreover, in itself, attention focusing introducescomplexity and should be applied with care. Forinstance, some people take offense in a switchfrom lower to uppercase to denote urgency.

Principles of Behavior Assumedin the Model

Having defined communication strategies (sum-marized in Table 3), it is now possible to discusshow the sender chooses to use them. Thesender�s choice of how to communicate is centralto this model: it is used first to explain strategyselection, then to explain medium and messageselection, and, finally, it is expanded to explainhow communication inputs affect the communi-cation process. Following Beach and Mitchell(1998), the choice of strategies involves two typesof tests: (1) a filtering of admissible strategies(e.g., affectivity may be banned, de facto, informal communication) and (2) a cost-benefit(profitability) analysis in which benefits (e.g.,accuracy in message comprehension) areweighed against costs (e.g., time spent). Forexample, perspective taking requires time and willbe selected only if the probability and cost ofcommunication error to the sender for a particularmessage is high enough to justify the effort. 12

Moreover, the cost-benefit approach is also appro-priate to the relationship impact of communication(Blau 1964). Communication complexity, there-fore, plays a major role in the choice of strategiesbecause it reflects the plausibility of errors ordifficulties in communication, while strategies arethe means by which complexity is reduced. Lateron, when we talk about the medium, the choice ofstrategy will be tied to the choice of medium,changing the costs and benefits, or even thefeasibility of strategies (e.g., it may be infeasible tocontrol by testing and adjusting an airmail letter).Similarly, certain inputs to the communicationprocess may make strategies infeasible (e.g., acomplete lack of knowledge about the receivermay make perspective taking infeasible).

12Such two-stage models are common in decisionmaking (Beach and Mitchell 1987, 1998). The mech-anism for choosing a strategy includes two aspects, thefirst of which involves a filtering of acceptable strategiesand the second of which implies a cost-effectivenessselection of preferred alternatives. The first stage, byimplication, suggests a yes-no definition of feasibility,while the second is a matter of degree, and of weighingcosts and benefits. The costs are usually personal costs,such as mental effort, and the benefits are accuracy andsalience of decisions (Payne et al. 1993).



Underlying the application of this framework areseveral principles of behavior that apply generallyto information processing but have also, onoccasions, been applied to communication. First,people avoid or may attempt to reduce effort andcomplexity so that they will select the least costlystrategy that achieves their goal. Describing therationale for choosing how to communicate,Kasher (1977, p. 231) describes it as �given abasic desired purpose, the ideal speaker choosesthat linguistic action, which he believes, mosteffectively and at least cost attains that purpose.�This principle explains, for example, why peopleemploy communication strategies in the firstplace.

The second principle refers to shifts betweenstrategies. People represent action at multiplelevels of abstraction, and at any one moment, oneof these levels is their focal level (Rasmussen1986; Vallacher and Wegner 1987). People tendto remain on higher rather than lower levels ofabstraction, but shift their attention to a lower levelof abstraction when complexity increases andbreakdowns occur. Berger (1998) uses thisprinciple to explain patterns of communicationbehavior, and we use it here to explain shiftsbetween strategies (control and contextualization),describe the choice of message form, and derivedesign implications. Moreover, from this principleit follows that whenever people are able to do so,they strive to attain closure on actions at lowerlevels so that they can proceed with the higherlevels. This explains, for example, why peoplechoose less effective but more readily availablemedia (Straub and Karahanna 1998).

Third, people have expectations, often culturalones, about representations and the way theycorrespond to goals (in other words the affordancerendered by a representation or technology, e.g.,Norman 1990). For example, some people expectto read a message from left to right, and othersfrom right to left. Representations that are incom-patible with these expectations require additionaleffort or increase the probability of error, and are,therefore, rejected in favor of compatible repre-sentations (Barber 1988). Fourth, people adapt to

incompatible representations by transforming orrestructuring the representation or by employingcompensatory strategies, e.g., re-reading themessage. This last category may take longer toappear than the first three reactions and will bereferred to as secondary behavior which is able todescribe adaptations to incompatible medium andmessage (propositions 4 and 6). These principlesof behavior are treated here as assumptions,providing the rationale for developing thepropositions below. Therefore, the propositionsabout the choice of behavior describe how peopleprioritize goals and choose strategies according tothese assumptions. However, the propositionsabout effective configurations of messages andmedia are normative in the sense that they statewhich configurations are more effective inachieving higher impact for certain situations(these are propositions 3 and 5). Accordingly, wewill use different formats for the descriptive andnormative propositions.

Proposition 2: Certain StrategiesAre Selected for Certain Goals

At this stage in the paper, we develop the generalproposition that certain strategies are selected forcertain goals. This general proposition is com-posed of five specific propositions, each of whichdetermines what strategy can improve communi-cation by looking at the sources of complexityinherent in the communication goals (the specificpropositions are depicted as arrows labeled Athrough E in Figure 3). Thus, we examine thepotential benefits of each strategy in coping withcognitive, affective, and dynamic complexity. Thisline of thought is depicted in Figure 3 (whichexplodes the link between goals and strategiesshown in Figure 1) and applied below to each ofthe strategies with the exception of attentionfocusing. The latter is needed whenever com-plexity is high in relation to the receiver�s attentionspan and is primarily a function of environmentalfactors such as information overload (discussed inthe section on inputs to the communicationprocess). A comprehensive view of the strategiesis then taken and a possible tradeoff betweenstrategies is demonstrated.



InterdependencyMultiplicity

Dynamic complexity

Affective complexity

Instruct action

Manageinterdependent action

Manage relationship

Influence

Control--testing and adjusting

Control--planning

Contextualization

Perspective taking

Affectivity

Attention focusing

GoalsGoals StrategiesStrategiesC

D

A

E

B

Incompatibility

Cognitive complexityInterdependency

MultiplicityMultiplicity

Dynamic complexity


Instruct action

Manageinterdependent action

Manage relationship

Influence

Control--testing and adjusting

Control--planning

Contextualization

Perspective taking

Affectivity

Attention focusing

GoalsGoals StrategiesStrategiesC

D

A

E

B

Incompatibility

Cognitive complexity

Figure 3. The Complexity of Communication Goals Affects Strategy Selection

Specific Propositions onChoice of StrategiesContextualization is necessary when a messageis liable to be misunderstood (Gumpertz 1982).Such a misunderstanding occurs most frequentlywhen cognitive complexity is high, for examplewithin non-routine situations involving a morecomplex exchange of views (Daft and Lengel1984). While reducing cognitive complexity bysimplifying the situation may be seen to be highlydangerous, providing information rich with contextis more effective (Janis 1989).

Proposition 2A: Contextualization is selectedfor communication goals characterized byhigh cognitive complexity.

For example, managing interdependent action,which can be ranked highly in terms of cognitivecomplexity, is particularly susceptible to misunder-standings. The strategy of building context intothe message decreases the probability of mis-understanding and thereby increases the proba-bility of accomplishing the goal of thinking collec-

tively. Indeed, Tyre and von Hippel (1997) showedhow the explicit presentation of multiple contextstriggers different, more effective thinking.Influencing, too, leads to high cognitive com-plexity, as it assumes a difference of opinions orpreferences, but also high affective complexity(Petty and Cacioppo 1986).

Affectivity is needed when affective complexity ishigh. Such a strategy is often associated withbuilding relationships, a process that depends onattitudes and trust (Scollon and Scollon 1995).

Proposition 2B: Affectivity is selected forcommunication goals characterized by highaffective complexity.

Control (either by planning or by testing andadjusting) is required when a given situation isperceived to be complex and when the probabilityof communication error is high (Srinivasan andTe�eni 1995). Of the two types of control, testingand adjusting is needed when dynamic complexityis high. Predetermining the process is counter-



productive when it is difficult to predict the patternof communication, such as in first acquaintances.In particular, goals that require a high degree ofparallel interdependence, such as influencing andcollective thinking, can be less effective when theprocess is predetermined (Conger 1998). Simi-larly, communication for monitoring generallyrelates to unplanned events and, therefore,requires immediate adjusting if the communicationis faulty.

Proposition 2C: Control by testing andadjusting is selected for communicationgoals characterized by high cognitive com-plexity coupled with high dynamic com-plexity.

In contrast with monitoring and influencing, settingprocedures usually assumes a relatively smallamount of dialogue. The mode of communicationis often one to many, and the receivers will usuallyreceive the information at a later time, perhapsmonths after it was issued. Dynamic complexity isusually low.

Proposition 2D: Control by planning isselected for communication goals charac-terized by high cognitive complexitycoupled with low dynamic complexity.

Perspective taking is needed when the receiver�sview may distort or reject the intended meaning ofthe message if its contents are not adapted to fitthe receiver's view (Krauss and Fussell 1991b).Perspective taking is applicable in situationswhere there are inconsistent views and is mostprobable when a multiplicity of views (high cogni-tive complexity) is coupled with high affective com-plexity. This communication strategy is. therefore,of paramount importance in influencing, where allpartners feel they are entitled to their views anddo not feel an obligation to conform. In suchcases, dismissing one�s partner�s views iscounterproductive (Conger 1998).

Proposition 2E: Perspective taking isselected for communication goals charac-terized by high cognitive complexitycoupled with high affective complexity.

A Comprehensive View of StrategiesThe general proposition that certain strategies areappropriate for certain goals has been articulatedby five specific propositions (2A through 2E)shown in Figure 3. The same rationale can beused to develop additional propositions as welearn more about the communication strategiesand goals. Another important direction to examineis the relationship between strategies as com-municators will usually employ several strategies.One example is the tension between control andcontextualization. Control is needed in order toregulate actions and correct them if necessary,and it requires a high level of abstraction. Incontrast, contextualization supports comprehen-sion, and it requires a lower level of abstraction.According to the second principle of behavior (seethe earlier section on principles of behaviorassumed in the model), people will shift fromcontrol to contextualization only when a miscom-munication occurs, and will shift back oncecomprehension is achieved.

Corollary: Contextualization is selected overcontrol when mutual understandingdecreases.

Media: Certain Medium Attributes AreMore Effective for Certain Strategies

Given a particular set of communication stra-tegies, a particular medium and message formneed to be chosen for it to implemented. The bi-directional links between strategies and medium(propositions 3 and 4), between strategies andmessage (propositions 5 and 6), and betweenmedium and message (proposition 7), as shownin Figure 4, are discussed below. We begin with areview of past research on attributes of themedium.

Attributes of Medium and Reviewof Recent StudiesNowadays, a variety of communication tech-nologies are available, including letters, memos,faxes, telephone, e-mail, voice mail, and, veryshortly, video conferencing and the mobile video



Dynamic complexity



Channel capacity

Interactivity

Adaptiveness

Size

DistributionOrganization

Formality


Control--planning

Contextualization

Perspective taking

AffectivityAttention focusing

Strategies

Medium

Message form

7

4

6

5

3

Dynamic complexity



Channel capacity

Interactivity

Adaptiveness

Size

DistributionOrganization

Formality


Control--planning

Contextualization

Perspective taking

AffectivityAttention focusing

Strategies

Medium

Message form

7

4

6

5

3

Figure 4. Strategies Affect and Are Affected by Medium and Message

phone. The most common classifications ofmedia build on the following three dimensions ofmedia richness:13

� Interactivity�the potential for immediate feed-back from the receiver. It is manifested bysimultaneous, synchronous, and continuousexchange of information (Zack 1993).

� Channel capacity�the potential to transmit ahigh variety of cues and languages (Daft andLengel 1984).

� Adaptiveness�the potential to adapt (per-sonalize) a message to a particular receiver(Daft and Lengel 1984).

Consider the voice mail of the CEO in theintroductory example (Table 1). The CEO, who istrying to reassure the employees chooses amedium that has higher channel capacity than,say, a written memo (voice carries with it subtlesignals that are difficult to emulate in writing). Arecorded video (audio-visual) message wouldhave an even higher channel capacity. Techno-logies that allow the sender to insert the name ofthe employee (from a file of names recorded in theCEO's voice) would increase adaptiveness, butonly superficially (much like personalized junk

13Most classifications or ordering of media build onmedia richness characteristics, which integrate the levelof interactivity, number of channels supported, capacityto transmit a high variety of languages, and ability topersonalize messages (Daft and Lengel 1984, 1986;Daft et al. 1987). See Sillince (1997) for an expandedlist. In order to develop a more suitable basis fordesigning these attributes, we first examine eachattribute separately, and then the interaction betweenattributes, concentrating on the three attributes mostwidely addressed (the capacity to transmit languagevariety can be subsumed under channel capacity, seeDennis and Kinney 1998).



mail). Furthermore, the recorded message is oflow interactivity in comparison with, say, a videoconference, where the CEO could present theissue and then take questions about it.

Table 5 shows the results of recent studiesrelating to the quality of communication withmedium attributes. Where possible, the results areorganized according to the three attributes. Thosestudies that did not distinguish between theseattributes are grouped under combined attributesas face-to-face (FtF) communication versuscomputer-mediated communication (CMC) suchas e-mail. In the table, impact is divided in two:action-oriented and relationship-oriented, Further-more, this and other reviews in the paper areintentionally limited to research published in thelast 15 years in view of the dramatic changes incommunication technology. Exceptions are citedonly to emphasize differences or similarities withthe communication world of the 1960s and 1970s.

The review demonstrates that the impact of mediaon both action and relationship is inconclusive. Ina period of rapid advances in information andcommunication technologies, this may be seen toreinforce earlier conclusions regarding thecontradictory results of media richness a decadeago (Markus 1994a, Rice 1992). One way toresolve the contradictions with regard to effects ofmedia is to examine further contingencies on goal-related factors, e.g., type of task (D�Ambra et al.1998) and level of interdependency in task (Strausand McGrath 1994), or upon input factors, e.g.,sender-receiver distance (Zmud et al. 1990).Another way is to break away from media richnessas an integrated perception of the potential reduc-tion in task equivocality (Daft and Lengel 1984),and to investigate separately what each attributeof the medium affords. In this study we take bothdirections by bypassing media richness (as aconstruct, not the principles of the theory) toconcentrate on the three attributes of media, andby looking closely at the contingent effects ofmedia on communication strategies and inputs.

Contradictory results remain, however, even whenexamining each attribute separately. In terms ofinteractivity, interactive brainstorming (versus non-interactive) is both more productive (Valacich et

al. 1994) and less productive (Pinsonneault et al.1999). Channel capacity too shows an incon-sistent impact, not only on mutual understanding,but also on relationship. The review also points atthe paucity of behavioral research on the impactof adaptiveness, which contrasts with the techno-logical efforts placed on personalization of com-munication technologies (discussed in the sectionon implications and conclusions). In the case ofrelationship-oriented impact, there seems to be adifferential effect between higher and lowerchannel capacity over time, but even this effect isinconsistent with regard to video versus FtF. Oneexplanation may be that there are primary effectsthat are immediate and secondary effects (adapta-tions to initial states) that take longer. In ourmodel, these are propositions 3 and 4, respec-tively. In general, the observations of maturebehavior will include adaptations, while observa-tions of initial communication patterns will not. Forexample, with time, people find ways of trans-mitting feelings in e-mail, even though it would notbe the medium that would be considered com-patible with the strategy of affectivity.

Proposition 3: The Choiceof Medium AttributesThe choice of medium according to strategies(proposition 3) is shown in Figure 4 as the arrowpointing to medium. It reflects the earlier dis-cussion of the two types of tests: admissible andprofitable. Is the particular medium admissible forthe strategy (e.g., control by testing and adjustingis infeasible on messages sent by �snail mail�)and is this medium best suited for the strategy(e.g., conveying emotions can be done through e-mail but face-to-face enables you do so moreeasily and more effectively)?

Interactivity facilitates control through testing andadjustment because of its ability to provide instantfeedback (Smith and Vanecek 1990). Moreover,high interactivity implies that the sender controlsthe pace of the communication, while low inter-activity leaves the decision with the receiver,reducing the sender�s control. Such dynamic con-trol is a necessary dimension in a person�s abilityto cope with dynamic complexity (McLaughlin1984). Interestingly, interactivity is itself a source



Table 5. Recent Studies on Medium EffectsMediumAttribute Action-Oriented Impact Relationship-Oriented ImpactInteractivity Immediate feedback improves under-

standing (Clark 1992; Dennis and Kinney1998). Immediate feedback speeds com-munication, which in turn, improves under-standing (Clark and Brennan 1991; Walther1992). CMC with blocked concurrent inputproduces less non-redundant communi-cation (Valacich et al. 1993).

Interactivity is important for affect inCMC (Kiesler et al. 1985).

Channelcapacity

Multiple cues can improve but also hinderunderstanding (Dennis and Kinney 1998).Higher channel capacity can speed but alsoslow down communication (Chapanis 1988;Sproull and Kiesler 1992). Higher capacityreduces explicit control (Kraut et al. 1998).Video conferencing produces more aware-ness and conversational fluency than voicealone (Tang and Isaacs 1992), particularlyin larger groups (Daly-Jones et al. 1998). Incomparison to audio-only, video has noeffect on mutual understanding (Gale 1990)or some improvement but less than FtF,when it is high quality (Doherty-Sneddon etal. 1997).

Mixed results on whether multiplecues seem less or more friendly(Fulk and Collins-Jarvis in press;Walther 1992, 1995). Low capacitychannels reduce social cues (Sproulland Kiesler 1992) but not if com-municators sense a social identitywith the communicating parties (Leaand Spears 1991; Spears et al.1990). Video conferencing iseffective in promoting social activity(Fish et al. 1993). Video vs. FtF shows no effect oninitial trust (Muhlfelder et al. 1999).

Adaptiveness None found. Voice-mail seems more personalthan e-mail (Adams et al. 1993).

Combined(FtF versusCMC)

FtF produces better mutual understandingthan CMC (Straus and McGrath 1994) andonly so for preference tasks (Tan et al.1999). Mixed results: FtF produces morevalid and novel arguments than CMC(Kiesler et al. 1985; McGuire et al. 1987)and equally valid arguments (El-Shinnawyand Vinze 1998). CMC causes informationsuppression (Hollingshead 1996). CMCgenerates less communication than FtF inhierarchical teams (Hedlund et al. 1998;Hightower and Sayeed 1996). CMC gene-rates more productive brainstorming byreducing production blocking (Gallupe et al.1994; Valacich et al. 1994) but also lessproductive brainstorming by increasing dis-traction and complexity (Pinsonneault et.al.1999). CMC generates more communica-tion in organizations (Schultze and Vanden-bosch 1998). CMC generates more com-munication but processed inaccurately(Dennis 1996). CMC generates more biaseddiscussions, especially when informationload is high (Hightower and Sayeed 1995). Technology-performance relationshipdepends more on experience than type oftask (Hollingshead et al. 1993).

Compared with CMC, FtF is ratedless relationship oriented and lessexpressive of affect (Hollingshead etal. 1993; Lea and Spears 1991;Walther 1995). Compared with CMC,FtF is rated less relationship orientedbut produces more total communi-cation (Hiltz et al. 1986). However,Siegel et al. (1986) found them to beequally task oriented. Use of e-maildecreases communication andamount of greetings (Sarbaugh-Thompson and Feldman 1998).CMC produces less trust than FtF(Rocco 1998) and poorer social life(Markus 1994b) and closer ties with-in coalitions but more social unrestoverall (Romm and Pliskin 1998).CMC over time increased socialorientation, trust, and informality(Walther and Burgoon 1992) butslower than FtF (Chidambaram1996).A synchronous web-base conferenceproduces less relational links thanFtF (Warkentin et al. 1997).



of dynamic complexity because of its potentiallyspontaneous, unpredictable progression of com-munication and the possibility of interruptions(e.g., �chat� is more interactive and moreunpredictable than asynchronous e-mail). Hence,unless used only when needed, interactivity mayprove to be a liability.

Proposition 3A: For control by testing andadjusting, high, rather than low, interactivityis more effective.

Although it may be hard to show when highchannel capacity is counterproductive (Rice 1992;Tan et al. 1999b), it is possible to determine forwhich strategies it is especially useful. Contex-tualization deals with high cognitive complexitythrough the provision of multiple layers of context,multiple views, and, in general, more task-relatedinformation than communication without contex-tualization. It necessarily follows that such com-munication requires greater channel capacity.Indeed, in this regard, media richness theorypredicts that communication aimed at resolvingambiguity and explaining interpretations willrequire interactive media and high channel capa-city media (Daft and Lengel 1984). Furthermore,in a survey of a large petrochemical company,Russ et al. (1990) found that managers selecthigh channel capacity media for equivocal mes-sages and low channel capacity media for lessequivocal messages, as did Daft and Lengel(1986; Daft et al. 1987). Taking a different pers-pective of video versus audio, Whittaker (1995)sees video as providing data about the objects ofdiscussion rather than adding non-verbal cuesabout the communicators, but neverthelessregards this additional channel capacity as a toolfor building a shared context.

Proposition 3B: For contextualization, high,rather than low, channel capacity is moreeffective.

Similarly, affectivity copes with high affectivecomplexity by adding emotions and dispositionsinto the message. In comparison to cognitivestrategies, though, affectivity is relatively sensitiveto how the affect is transmitted and received(Wallbott and Scherer 1986). Such communi-

cation often requires a wide variety of signs thatcan be transmitted only on channels of highcapacity (e.g., Carnevale and Isen 1986). Themedia richness theory asserts that high channelcapacity, e.g., face-to-face oral communication, isnecessary to enable social cues such as facialexpressions, body language, and tone of voicethat are absent in written communication or CMCsuch as e-mail. However, high channel capacitymedia have only the potential for enriching com-munication. They cannot ensure a richer com-municative act in reality (Ngwenyama and Lee1997). There is, nevertheless, some evidence thathigh channel capacity is perceived to be moreeffective and more appropriate for affective com-munication, primarily because of the complexity offeelings and importance of non-verbal messages(Sproull and Kiesler 1992; Westmyer et al. 1998).It is interesting to note that the use of audio-visual,in comparison to audio only, has been shown tobe effective in terms of informal communication,which often includes social information (Bly et al.1993; Fish et al. 1993).

Proposition 3C: For affectivity, high, ratherthan low, channel capacity is more effec-tive.

Adaptiveness is necessary in strategies thatattempt to tailor the message to a personalperspective. For example, when compared with e-mail (higher adaptiveness) and face-to-face con-versation (highest adaptiveness), a bulletin boardis unable to support effectively any form of adap-tiveness. In practical terms, it is sometimes hardto separate adaptiveness from channel capacity.In a simple case, when the sender refers in themessage to the receiver�s view too, then thechannel should have the capacity to transmitmultiple views. In a more realistic case, perspec-tive taking includes affective as well as cognitivereferences to the receiver�s world. In such cases,the channel must also have the capacity to sup-port the personal touch by tone, pronunciation,and other non-verbal gestures. A useful exampleis voice-mail, which eliminates the synchronicity ofa telephone conversation but includes its channelcapacity and adaptiveness. E-mail, in compari-son, can be adapted to the receiver but has lesschannel capacity. Adams et al. (1993) have



compared the two media and found voice-mail tobe more personal than e-mail according to mostpeople. Quite possibly, this may be because mostpeople cannot separate adaptiveness from thecapacity to transmit a variety of social cues.

Perspective taking usually requires the sender tounderstand new viewpoints and adapt themessage accordingly (Goldberg 1990). It alsorequires the sender to adapt the message to makeit more personal. In contrast, managers conveyingto a group the formal structure of authority andcode of behavior (e.g., setting procedures androles) will usually select a written rather than face-to-face mode of communication.

Proposition 3D: For perspective taking,high, rather than low, adaptiveness is moreeffective.

Proposition 4: Adaptations toNon-compatible MediaAdaptations are necessary when the medium isnot compatible with the strategy, i.e., inadmissible,high on cost, or low on benefit (see the earliersection on assumed principles of behavior). Thissection examines the secondary effect of mediumattributes on communication strategies or, in otherwords, how users adapt to the media (DeSanctisand Poole 1994). This is shown in Figure 4, whereincompatible media increase complexity, although,compared with the effect of goals on strategies inFigure 3, the main source of complexity is now themedia attributes rather than the communicationgoals. The first implication of reduced or insuffi-cient channel capacity, as perceived by thesender, is that of higher dynamic complexity as aresult of the lack of feedback (O�Connaill et al.1993). The effect of higher dynamic complexity isa perceived need for higher control to overcomepossible disruptions of the communication flow(Rutter and Stephenson 1977). This reaction willbe particularly strong if the cost of control remainsrelatively low, for example, in a system thatenables an automatic reminder if there is no replywithin a pre-specified time. The choice of theparticular type of control strategy to use dependson the interactivity of the medium available. Wheninteractivity is high, control through testing andadjusting will be used to cope with reduced

channel capacity (Doherty-Sneddon et al. 1997).For example, field workers used more explicitcontrol when channel capacity was reduced fromvideo to audio (Kraut et al. 1998). When inter-activity is low, however, senders may choose tocope with reduced channel capacity by increasingcontrol through planning, e.g., by explainingprocedures in different ways to increase theprobability that one of them will be understood.

Proposition 4A: Senders will adapt to lowchannel capacity coupled with highinteractivity by increasing control throughtesting and adjusting.

Proposition 4B: Senders will adapt to lowchannel capacity coupled with low inter-activity by increasing control throughplanning.

A second result of reduced channel capacity ishigher affective complexity. Limiting the means ofconveying emotions becomes threatening to thesender when there is uncertainty about thereceiver�s response and, at the same time, littleability to control the communication. This scenariomay explain Lea and Spears� (1991) notion ofincreased social cues on e-mail only when thereis an established social identity. Moreover, even ifsocial identity has been established, only highinteractivity will allow the sender to securelyincrease affectivity that uses limited cues (e.g.,only words) in order to compensate for the lack ofcustomary cues of affectivity (Walther andBurgoon 1992). The findings on this matter are,however, mixed (Sarbaugh-Thompson andFeldman 1998).

Proposition 4C: Senders will adapt to lowchannel capacity coupled with low inter-activity by decreasing affectivity.

Proposition 4D: Senders will adapt to lowchannel capacity coupled with high inter-activity by increasing affectivity.

Comprehensive View of MediumTable 6 summarizes the specific propositionsabout how strategy dictates medium attributes and



Table 6. Summary: Certain Strategies Require Certain Media Attributes(Proposition 3) and Certain Media Induce Certain Strategies (Proposition 4)

Specific Proposition Medium Strategy

Strategy÷Medium 3APrimary

High interactivity Control by testing and adjusting

3B High channel capacity Contextualization

3C High channel capacity Affectivity

3D High adaptiveness Perspective taking

Medium÷Strategy 4ASecondary

Low channel capacity + highinteractivity

Increase control by testing andadjusting

4B Low channel capacity + lowinteractivity

Increase control by planning

4C Low channel capacity + lowinteractivity

Decrease affectivity

4D Low channel capacity + highinteractivity

Increase affectivity

how media trigger adaptations through the stra-tegies. More complex propositions about combina-tions of attributes can be developed in the future.The combination of high interactivity and highchannel capacity is particularly potent when highdynamic complexity is coupled with high affectiveor cognitive complexity. For example, contex-tualization and control through testing andadjusting are needed for understanding complexproblems. Interactive hypermedia is a primeexample of high interactivity and high channelcapacity. (Hypertext is a device that not onlyorganizes but also communicates multiple andrelated views.) Indeed, in an experiment thatrequired contextualization, interactive hypertextwas shown to be more effective than a linearinteractive support system (Mao et al. 1996).Another example of high dynamic and cognitivecomplexity is parallel work such as collectivethinking. The channel capacity of video con-ferencing provides the necessary cues to enableeffective feedback and the interactivity makes itpossible to react in time (Daly-Jones et al. 1998).

If communication technologies could be seen asbeing different combinations of interactivity,capacity, and adaptiveness, then some of these

technologies would be incompatible with ourexpectations; expectations formed by media towhich we are accustomed (face-to-face, of course,is the one we literally grew up with). This createscertain problems that have to do with unbalancedadaptation. For example, as has been notedabove, people express their emotions through e-mail. Like a letter, e-mail has low channelcapacity. However, it has little control throughplanning. Like face-to-face, it enables instantresponses but without non-verbal feedback usedfor control through testing and adjusting. E-mail isleft uncontrolled, which may explain the contra-dictory results that even when CMC does notreduce social communication, it disturbs rela-tionships (e.g., Markus 1994b; Romm and Pliskin1998).

Message: Certain MessageAttributes Are More Effectivefor Certain Strategies

Attributes of Message FormFour attributes of message are defined here andthen linked to strategies. Message form is a term



that characterizes the configuration and style ofthe information communicated. In contrast to thechoice of media, the choice of the message formhas received little attention in information systemsresearch. Research in communication theory hasalso neglected it as evident by the call to �devotemore energy than they have in the past to a closestudy of messages themselves� (Stohl andRedding 1987, p. 494). Past characterizations ofmessage and the more recent work in CMCsuggest four attributes of a message: size, dis-tribution, degree of organization (structure), anddegree of formality. Here, these attributes areredefined when necessary to ensure they do notoverlap with attributes of the medium.14

Message size is a function of the number ofsemantic units such as words or sentences (Daftand Lengel 1986). Distribution is the number ofdestinations to which the message is sent. Sizeand distribution are the most popular measures ofcommunication in CMC (Rudy 1996), perhapsbecause they are relatively easy to retrieve fromsystem logs with little need to code the materialmanually.

The degree of message organization may bedefined as the extent to which the message issystematically ordered to support mutual under-standing, by explaining how message can beunderstood (the word �structure� is avoidedbecause it is usually associated with the task orprocess). For example, the next paragraph(starting with �A highly organized message�) isdivided into four ordered dimensions, creating aclear two-level hierarchy in the message thatguides the reader from upper to lower level, andproceeds from the first to the fourth itemsequentially.

A highly organized message may, therefore, becharacterized in terms of several, but notnecessarily all, of the following dimensions:

(1) An obvious set of ordered and clearlydistinguished elements that can easily bedifferentiated and discriminated (Schroder etal. 1967) (e.g., paragraphs with an openingthat indicates the theme or sections withsubtitles or numbering);

(2) A clear allocation of tasks between senderand receiver so that the latter can imme-diately understand the action required (e.g.,the sender provides information and thereceiver is expected to take action).

(3) A clear structure of and access to differentlevels of context to easily grasp and navigatethe macrostructure (van Dijk and Kintsch1983) (e.g., explanations as footnotes,references to documents that provide moredetails or a more complete rationale, hyper-text style access to more details);

(4) A familiar or standard format for immediaterecognition to avoid searching or learning(Berlyne 1960) (e.g., each of the diary entriesin Table 1 begins with date, shift, and author).

The fourth attribute of message is formality, whichdenotes interactive closure toward some organiza-tionally accepted representation of action (Iedema1999).15 One scenario of such closure is the

14Formality is an essential characteristic of com-munication, but has received only casual attention inempirical research (Stohl and Redding 1987). Influentialsources on formality are Downs (1967) and Melcher andBeller (1967). Both sources characterize formality as afunction of the organizational setting, namely thecapacity in which the sender acts. Stohl and Reddingcompiled a list of message classification schemes, butagain, most of these schemes incorporate other con-structs of our model (goals and input factors). However,characterizations of messages that do not clearlydistinguish between message and medium would beinappropriate for a model (such as ours) that does. Forexample, information richness has been defined as �theability of information to change understanding within atime interval. Communication transactions that canovercome different frames of reference or clarifyambiguous issues to change understanding in a timelymanner are considered rich� (Daft and Lengel 1986, p.560). In contrast, we seek constructs that are evident inthe message itself, rather than which remain a judge-ment of its ability to achieve a goal and, consequently,blur the distinction between medium. Clearly, this issuedeserves more attention. For example, from the socialinfluence perspective (Fulk et al. 1990) and symbolicmeaning perspectives (Sitkin et al. 1992; Trevino et al.1987), formality can be conveyed by the choice ofmedia. It may be that a separate construct of formalitycould be defined in the context of the medium too.

15This definition does not preculde formalities dictatedby more general norms outside the organization (e.g.,language), but for simplicity they are subsumed under�organizationally accepted representation of action.�



progression from stories, through investigationand experimentation, to (probable) facts. Thesefacts are represented as formal, abstract, andaccepted (Latour 1990). Formality, therefore,implies an abstraction of a more concrete descrip-tion. A given issue embedded in a particularcontext can be more formally represented. Thisallows it to describe a principle relevant to a moregeneral situation in a way that conforms to therules of communication in the organizationalsetting. In the example of Table 1, Smithinformally tells the contract manager about howJoey spilt tea in the production room, suggestingthat this would not happen had the dining roombeen open. After some processing (formalityusually involves careful effort), the contractmanager issues a formal memo that is moreabstract and which focusses on the desiredaction. This is written in a style that signalsconformity with the organization. In practicalterms, it is easier to measure conformity withformal rules of communication than closure towardconformed action (c.f., Irvine 1979).

Proposition 5: The Choiceof Message AttributesHaving described the four attributes of messageform, we now turn to the links between messageand strategies. Affectivity relies on a sense of sin-cerity, which may be lost if the receiver sees thesame emotions shared with others. High degreesof affectivity may also rely on intimacy, which isclearly incompatible with wide distribution.

Proposition 5A: For affectivity, a small,rather than wide, distribution is moreeffective.

Message organization becomes profitable whenthe effectiveness gained in terms of reducedcomplexity outweighs the usually substantial effortinvolved. High message organization is, therefore,found only when communication complexity ishigh and time and effort are affordable. Forexample, contextualization, which is associatedwith higher complexity, requires a well-organizedmessage to improve comprehension (Mayer1985).

In achieving control through planning, the effortinvested in planning frequently implies a corres-ponding effort in organizing the message.Planning will usually result in a highly organizedmessage that enables tasks to be clearly allo-cated between sender and receiver, a necessarypart of the plan. For example, this paper is anunusually complex message but it can demon-strate how message organization is linked tostrategies. For example, each proposition isalways preceded by an explanation. A set ofinstructions to the reader could support controlthrough planning by anticipating that some(hurried) readers will read only the propositions,and look for more details or examples only whenthey are not sure they understand the message.The instructions would inform the readers thatwhen they feel uncertain about a proposition, theyshould look for the explanation in the paragraphspreceding the proposition.

Proposition 5B: For contextualization, high,rather than low, message organization ismore effective.

Proposition 5C: For control by planning,high, rather than low, message organizationis more effective.

High affectivity is more likely to benefit from areduction in formality (Morand 1995). Indeed,formality provides a direct contradiction tospontaneity and personal attention, which are bothexpected in sincere affectivity and involvement.

Proposition 5D: For affectivity, low, ratherthan high, formality is more effective.

In general, contextualization requires more speci-fic, concrete information than does control (Te�eni1992). Contextualization provides details andexplains how to take action. Effective contextuali-zation often relies on examples, step-by-stepprocedures, and even rich scenarios, presentedas narratives. These messages are usually repre-sented with low formality. On the other hand,control operates on a higher level of abstraction(discussed in the earlier section on the com-prehensive view of strategies) and often relies onsimple rules or more general principles of actionwhich are represented with higher formality.



Proposition 5E: For control, high, ratherthan low, formality is usually more effective.

Proposition 5F: For contextualization, low,rather than high, formality is usually moreeffective.

Proposition 6: Adaptations toNon-compatible Message FormsHaving looked at how strategies dictate messageattributes, we now consider how messageattributes affect strategy selection. One directresult of larger messages is higher cognitivecomplexity due to information overload, which canbe reduced by attention focusing.

Proposition 6A: Senders will adapt to a longmessage by increasing attention focusing.

Poor message organization also increasescognitive complexity. Control through testing andadjusting is a common reaction to a disorganizedmessage, but can be accomplished only whenusing interactive media. For example, in situationsof stress, messages are often disorganized andcontinuous control with feedback is usually theonly way to cope with the complexity.

Proposition 6B: Senders will adapt to lowmessage organization by increasing controlthrough testing and adjusting, providedmedia interactivity is high.

Comprehensive View of MessageTable 7 summarizes links between strategies andindividual attributes of a message. We alsoexamined the interactions among message attri-butes, but found little evidence of such inter-actions. It may be reasonable to expect size andformality, which are both higher on cognitivecomplexity, to increase message organization.Moreover, propositions 5 and 6 refer to individualstrategies, but communication involves combina-tions of strategies too. Indeed, the corollary toproposition 2 predicts that, in shifts betweencontrol and contextualization, the latter will replacethe former when communication breaks down.Propositions 5E and 5F, therefore, imply a corres-ponding change from high to low formality. Recallthe instructions to readers on how to go from theproposition to its preceding paragraph when theydo not understand the proposition. The propo-

sitions, which are more abstract and phrased inthe accepted terminology of the model, are ofgreater formality than their preceding paragraphs,which should be more concrete and possiblyinclude specific examples. One of the implicationsfor design (discussed below) is that these shiftsfrom one level to another should be supported byhigher message organization but also withcorresponding changes of formality.

Links Between Media andMessage Form

Choice mechanisms involving admissibility andprofitability can also explain the interactionsbetween medium and message form. Forexample, e-mail facilitates an increase in messagedistribution and a commensurate decrease in dis-tribution costs, and this makes distribution moreattractive to the user. Indeed, past research hasshown that CMC increases the distribution ofmessages (Palme 1985; Phillips and Eisenberg1993; Sproull and Kiesler 1992). Similarly,interactivity increases the time-related cost of longmessages because of the online nature of thedialog but not necessarily the velocity of mes-sages (Jones et al. 1993). In fact, there is someevidence that this is what happens with the use ofCMC (Trevino et al. 1987). Social norms of usingcertain interactive media, such as small talk at thebeginning of a face-to-face meeting or phoneconversation, may, however, moderate this effect.

Proposition 7A: When interactivity is high,senders will exchange shorter, rather thanlonger, messages.

Low channel capacity generates a sense of limitedfeedback and higher risk of failure in communi-cation that calls for more control (propositions 4Aand 4B). Moreover, according to Proposition 5E,higher formality is associated with increasedcontrol while low formality introduces a higherchance of failure (e.g., getting angry, loosingcontrol, and offending your partner). As a result,low channel capacity may induce higher formality,particularly through more stringent rules of com-munication (see also O�Connaill et al. 1993).

Proposition 7B: When channel capacity islow, senders will exchange messages ofhigher, rather than lower, formality.



Table 7. Summary: Certain Strategies Require Certain Message Attributes(Proposition 5) and Certain Message Attributes Induce Certain Strategies(Proposition 6)

Specific proposition Message Strategy

Strategy÷Message 5APrimary

Small distribution Affectivity

5B Highly organized message Contextualization

5C Highly organized message Control by planning

5D Low formality Affectivity

5E High formality Control

5F Low formality Contextualization

Message÷Strategy 6ASecondary

Long messages Increase attention focusing

6B Low message organization +high interactivity

Increase control by testing andadjusting

Inputs to the CommunicationProcess

Attributes of Task, Sender-ReceiverDistance, and Values and Norms

Research into communication shows that thecommunication process is affected by a host ofinputs (e.g., Carlson and Davis 1998; Ehrlich1987; Fulk et al. 1991; Markus et al. 1992),although exactly how it is affected is not alwaysclear. This section attempts to examine theseeffects on the communication process elementsidentified above, namely goals, strategies, media,and messages. As noted earlier, the formulation ofthe communication context developed in thetheory of communicative action (see the lifeworldin Figure 2) must be adapted to the organizationalsetting so that its effects on behavior can beanalyzed. To do so, we have adopted threecommon classes of inputs to communication:(1) task, (2) sender and receiver characteristics,and (3) values and norms (c.f., McGrath 1984). Itis obvious that all three inputs can affect com-munication. Communication goals are shaped bytasks, as well as by the relative situation betweenthe communicators, be it physical (in space and

time) or psychological and social. In addition, thevalues and norms of communicators dictate cer-tain communication patterns. In the introductoryexample (Table 1), there is a physical-timedifference between the shift workers communica-ting through the diary, which makes the messagenecessary in the first place. The happy birthdaygreeting is welcomed under an organizationalnorm of congratulating colleagues on theirbirthday, but in a more conservative organization,it might be considered as a breach of privacy.

In this section, we first select the attributes ofthese three inputs (shown in Table 8) and thenuse them to review recent literature. The reviewis necessarily restricted to these elements.Furthermore, after the review of each input,several propositions are constructed that link theinput to the process elements. Of the manypossible propositions, we concentrated on linksthat could be developed using the admissibilityand profitability choice mechanisms and thenotion of communication complexity. The propo-sitions thus refer to a yet smaller subset of inputsand process elements (e.g., values but notnorms), albeit one that is sufficient to demonstratethe mechanism by which the inputs affect theprocess.



Table 8. Glossary of Input Attributes

Element Attribute Definition

Task

Analyzability The ability to define procedures needed to complete the task.

Variety The variation among different instances of the task.

Temporality The time related demands to complete the task.

Sender-receiver distance

Cognitive distance The gap between the sender's and receiver's interpretations beforetransmitting the message.

Affective distance The emotional gap between the sender and the receiver beforetransmitting the message.

Values

Interdependence A tendency to think and act as a person independent of others or asa person inter-dependent on others.

TaskIn the information processing view of organiza-tions, tasks are usually classified according to twofundamental dimensions: task analyzability andtask variety (Daft and Lengel [1986], but seeDennis and Kinney [1998], who consider only theformer). Task analyzability characterizes theability to define the procedure (algorithm) neededto complete the task. Task variety describes thevariation among different instances of the task. Athird dimension that has become prominent is tasktemporality, which includes the time during whichthe task must be completed as well as howtemporal patterns affect the task demands.

The operational definitions of analyzability andvariety can be adopted from the media richnesstheory. Task variety is the frequency of unex-pected or novel events encountered in a taskinstance. Meanwhile, task analyzability belongs toa person�s knowledge of the exact procedures foraccomplishing the task. Both elements have beenmeasured by questionnaires (Daft and Macintosh1981), expert judgements (Daft et al. 1987), andcoded interviews (Donabedian et al. 1998). Timeto complete can be measured by the number oftemporal units (e.g., minutes or weeks), andtemporal patterns have been linked to stages of adecision task (Saunders and Jones 1990).

Sender-Receiver DistanceCommunication is shaped by senders andreceivers and is, therefore, likely to be affected bytheir individual styles of method or informationprocessing. For example, individuals who areable to cope with high cognitive complexity planmore complex communication (Waldron andApplegate 1994) and those who are more inde-pendent seek and provide more information(Oetzel 1998). It may be even more important,however, to discern the relative characteristicsthat create a distance between sender andreceiver. While individual characteristics intro-duce variations in the process that may some-times be ignored when analyzing of communi-cation, distance must be overcome, one way oranother, to enable communication to take place.The term distance usually connotes physicaldistance and, indeed, numerous studies havelooked at the effect of physical distance oncommunication in general and on computer-mediated communication in particular. Equallyimportant, though, is psychological distance,which may exist regardless of the physicaldistance.

In order to capitalize on the model developedabove, the sources of psychological distance areabstracted from the organizational context and



framed as primarily cognitive or affective. Asender-receiver cognitive distance may be definedas the initial gap between the sender's andreceiver's interpretations before transmitting themessage (Hutchins 1991). This may result fromdifferences in current information or from differentways of thinking and communicating. Such dif-ferences arise from different work functions,different nationalities, different organizationalcultures, differences in languages, etc.

A sender-receiver affective distance, meanwhile,is the initial emotional gap (negative relationship)between sender and receiver before transmittingthe message. It involves feelings and attitudesbetween two parties. Strangers to each other forman affective distance on instinct or stereotypicalimpressions, for instance. Communicators with ahistory of interaction form an affective distance onthe basis of the relationship that results fromprevious communications.

Values and NormsValues and norms define the outer (or highest)layer of the communication context (Figure 2) andare probably the hardest to relate systematically tothe communication process. A complete discus-sion of the effects of values and norms on com-munication lies beyond the scope of this paper,but we need to demonstrate their importance andto consider how they affect the process in order toexplain the underlying mechanisms of the model.We, therefore, make some simplifications andcompromises. A striking example of the effect ofnorms on communication is described in theColruyt case study (Janson et al. 1999). Colruyt(a chain of stores) has developed an organiza-tional culture that emphasizes maximum workerparticipation, so that individual limitations might beovercome. Thinking collectively had been pre-ordained by the CEO, and so workers had toadopt a strategy of controlling communication toensure that workers understood each other.Perhaps the most vivid example reported con-cerns a new employee who became embroiled inan emotional exchange of e-mails with hercoworkers. A senior manager who witnessed thisresponded on the e-mail system to all involved

(with copies to other managers) that �one shouldrefrain from flaming��getting emotional on thesystem.� He further specified that the correct wayof handling issues is to make a personal appoint-ment to discuss the matter� (Janson et al. 1999, p.191). Thus norms clearly affect goal priorities anddictate appropriate communication strategies andmedia. However, this example also points at thedifficulty of abstracting such effects to linksbetween process elements and general attributes(leading to operational definitions) of values andnorms.

The first simplification we make (following theframework adapted from Habermas in Figure 2) isthat values are associated with culture, whilenorms are associated with organizational rules ofconduct (formal and informal). Organizationalnorms of communication and culture have beenseen in past research to dictate decisions onmessage and medium that transcend taskattributes (Fulk and Boyd 1991; Fulk et al. 1990),and we extend these decisions to include thechange of goal priorities and selection of stra-tegies. We demonstrate these effects with onedimension of culture, independence-interdepen-dence, which is closely related to individualism-collectivism.

Individualism-collectivism is a major dimension forexplaining similarities and differences in communi-cation behavior (Gudykunst 1998; Gudykunst andMatsumoto 1996). Furthermore, it is probably theone most used in research on CMC (Rice et al.1998; Tan et al. 1998a). In broader terms, nationalculture is a set of unique values that guides thebehavior of people belonging to that culture(Triandis 1995). Dimensions of national culturesuch as individualism-collectivism are likely to bewidely applicable, due to the growing interest ininternational communication and the growingimportance of intercultural communication in multi-national organizations. We hasten to add, how-ever, that this and other dimensions have beencriticized for their weak theoretical basis. Inbuilding the theoretical development, therefore,we follow others in combining individualism- col-lectivism with theories of cultural psychology (e.g.,



Gelfand and Christakopoulou 1999). This explainsthe use of independence-interdependence.

A fundamental difference between Western(labeled individualistic or independent) andEastern (labeled collectivistic or interdependent)cultures is the different values these culturesplace on being separated from or connected toothers. These different values produce two distinctways of seeing oneself: one independent (charac-teristic of Western cultures) and the other inter-dependent (characteristic of Eastern cultures).These ways of seeing oneself reflect the focus ofindividuality in the respective cultures. Moreover,they produce different goal priorities and differentcognitive and affective behaviors (Markus andKitayama 1991; Trompenaars 1998). We build onthis basic difference in individuality in order toanalyze how individuals belonging to an indivi-dualistic versus a collectivistic culture differ in theircommunication processes.

In individualistic (independent) cultures, indivi-duals direct their thinking toward their inner worldand what makes it different to others. As a result,these individuals are more likely to seek infor-mation about themselves than about the groupthey belong to (Markus and Kitayama 1991). Bycontrast, in collectivistic (interdependent) cultures,individuals see themselves as part of a socialrelationship in which one�s own thinking andfeeling are interdependent with regard to those ofothers in the group. As a result, these individualsseek information about others in the group as wellas themselves (Markus and Kitayama 1991).

Independence describes cultures in which indivi-dual goals are dominant and ties between indivi-duals are weak (except, of course, the immediatefamily unit). Interdependence describes thosecultures in which people tend to function as stronggroups and maintain such ties for very longperiods (Hofstede 1991). Interdependence isclosely associated with low-context communica-tion and collectivism with high-context communi-cation (Gudykunst and Ting-Toomey 1988; Hall1976). In high-context communication, less of theinformation communicated is in "the coded,

explicit, transmitted part of the message" (Hall1976, p. 79). Low-context communication, incontrast, is explicit and usually direct, precise, andconsistent with one's feelings.

A General Explanation of HowInputs Affect Process

This section provides a general rationale forpredicting the effects of inputs on the process, asoutlined in Figure 5. Two routes are shown byarrows going from the input box to the processbox through two types of explanations: goalpriority and complexity (shown in ellipses). Thelower route of complexity involves the admissibilityand profitability mechanism for choosing com-munication strategies shown in Figure 3. Likeother applications of the cost-effectivenessapproach to human-information processing, thecost side of the equation has dominated theexplanations of the choice of media (Reinsch andBeswick 1990). For example, workers whoexperienced difficulties using e-mail for communi-cation used it less than those who communicatedflawlessly (Lantz 1998). Nevertheless, inputs mayaffect both the relative costs (e.g., time and effort)and the benefits (e.g., the probability ofcommunication failures), thereby affecting thecommunication process. In the upper route of goalpriority, inputs affect the process by setting orchanging the sender�s goal priorities, e.g., makingit more important for the person to engage incommunication for managing relationships. Whilethe discussion of the communication process(presented earlier) assumes the goals to be given,the discussion of inputs should consider thepossibility of changes in the sender�s goalpriorities. Furthermore, such changes may bereflected in corresponding changes in the relativefrequencies in which the goals are adopted.

In reality, the communication process may beaffected simultaneously by several inputs. Indeed,a recent study by Kraut et al. (1998) demonstratesthe multiple effects of communication inputs in asingle case. Their study revealed that the adoptionof a new communication medium was influencedby a change in the benefits of media as more col-



Cognitive Complexity

Dynamic complexity


Goal priorityTaskAnalyzabilityVarietyTemporality

Distance

Cognitive distanceAffective distance

Goals

Strategies

Priority and complexityProcessInputs

ValuesIndependence -

Interdependence

8

99

10

Cognitive Complexity

Dynamic complexity


Goal priorityTaskAnalyzabilityVarietyTemporality

Distance

Cognitive distanceAffective distance

Goals

Strategies

Priority and complexityProcessInputs

ValuesIndependence -

Interdependence

8

99

10

Figure 5. Inputs Affect Communication Process

leagues began to use CMC but also directly bymaking CMC an acceptable norm of com-munication. Furthermore, in addition to the tworoutes in which inputs affect the process (shownin Figure 5), inputs may also determine thefeasibility of certain media, e.g., a physicaldistance precludes an immediate face-to-facemeeting.

Task Attributes Affect theCommunication Process

ReviewTable 9 provides an outline of recent studies thatlink task attributes to communication. Task hasbeen a primary interest of group work in general(McGrath 1984) and of computer supported groupwork in particular (DeSanctis and Gallupe 1987;Zigurs and Buckland 1998). Most of the studiesreviewed measure performance according tovarious combinations of task and communicationmedium, where performance includes quality of

decision output (or group products) and quality ofcommunication such as accuracy and conver-gence (Smith and Vanecek 1990). Table 9 islimited to recent studies that clearly link the task tothe quality of communication. More generalreviews of communication media, task, and groupperformance can be found in Hwang (1998) andStraus and McGrath (1994).

The review suggests the importance of con-sidering task requirements in terms of interdepen-dencies between workers. The higher the inter-dependency, the higher the cognitive complexity,and the more intensive the need for managingcollective action becomes. Higher interdepen-dency is assumed in judgmental tasks, whichusually imply more intensive influencing.

The review also shows an understandable con-cern with regard to the effect of task on action-related impact but no concern with relationship.This has the danger of misrepresenting the fulleffect of task attributes, for example, by dis-counting the effects of analyzability on communi-



Table 9. Recent Studies of Task and Communication QualitySituationalAttributes Use and Action-Oriented Impact

Relationship-Oriented Impact

Taskanalyzability

Immediate feedback and multiple cues improve under-standing particularly for equivocal tasks (Dennis and Kinney1998; Straus and McGrath 1994) but also asynchronouscommunication improves idea exchange in less equivocaltasks (Shirani et al. 1999). CMC improves communicationonly for low interdependence tasks (meta-analysis by Hwang1998; Daly 1993; El-Shinnawy and Vinze 1998). Groupsupport systems with FtF outperformed those without, onlyfor highly equivocal tasks (Tan et al. 1999a) and reduceddifferentials (Tan et al. 1999b). Richer media are preferredoverall but not as a function of analyzability (D�Ambra et al.1998). Video conference chosen for routine tasks (Webster1998).

None found.

Task variety Variety requires more information (Daft and Macintosh 1981).Use of inter-organizational e-mail grows with uncertainty(Kettinger and Grover 1997).

None found.

Tasktemporality

Communication under time pressure is faulty: none found.Use of phone relative to e-mail increase with urgency(Wijayanayake and Higa 1999).Communication patterns change over time vis-a-vis the task(Jones et al. 1994), and these temporal effects interact withthe type of medium (Saunders and Miranda 1998) and withcommunication genre (Orlikowski and Yates 1994). Theperceived appropriateness of media grows with experience(King and Xia 1997).

Relational communi-cation is unsuccess-ful in short time(Walther 1995).

cation problems, which in turn may damagerelationship (Berger 1998). The uncertainty reduc-tion theory (Berger and Calabrese 1975), whichwas developed independently of task variety andanalyzability, may provide the link. This theorypredicts that the higher uncertainty about actionsand attitudes of communicators increases infor-mation seeking, which will eventually decreaseuncertainty and, consequentially, strengthen rela-tionships (this is revisited below in the discussionof sender-receiver distance). However, to useuncertainty reduction theory for establishing thelink between task variety and relationship, thetheory will have to incorporate task variety (Bergerand Gudykunst 1991).

PropositionsWe analyze the effects of task on communicationby looking at the effects on goal priority andcomplexity (Figure 5). The task perspective oforganizational communication regards the infor-mation communicated as a response to uncer-tainty. Uncertainty is defined as �the differencebetween the amount of information required toperform the task and the amount of informationalready possessed by the organization� (Galbraith1977, p. 36). Thus the task-related need forinformation is a function of both the task (varietyand analyzability) and the goal (performanceaspirations).

The straightforward effect of variety on com-munication is to increase the priority of requesting



information (which was defined as a special caseof �instructing action� goal) because moreinformation is needed to describe higher variationamong instances. This is consistent with mediarichness theory, which claims that high taskvariety requires larger amounts of information(Daft and Lengel 1986).

Proposition 8A: Higher task varietyincreases the frequency of requestinginformation.

The effects on strategies is explained through theprofitability test in which complexity is the primarycost and accurate goal achievement is theeffectiveness (Beach and Mitchell 1987). Low taskanalyzability increases the probability of misunder-standing how to proceed with action (Daft andLengel 1984). This increases the cognitive com-plexity of the communication because of theambiguity and multiplicity of meanings, which inturn increases the benefits of contextualization(Boland et al. 1994; Gumperz 1982). This too is inline with media richness theory, which claims thatlower task analyzability requires richer information(Daft and Lengel 1986).

Proposition 8B: Lower task analyzabilityincreases the use of contextualization.

Of all temporal attributes, we concentrate on theone most directly associated with the model,namely, the time available to complete the task.The relationship between time-to-complete andcommunication complexity is curve-linear: com-munication in either very short or very long timespans is more difficult than in intermediate timespans. Under time pressure, communication willbe stressful as cognitive demands exceed cogni-tive resources and feedback (which furtherconsumes cognitive resources) may be infeasibleor unclear due to time constraints and, therefore,dynamic complexity will be high.

On the other hand, when tasks stretch over longperiods of time, communication is often out ofcontext and out of control. To the receiver, a mes-sage relating to historical events will usually beout of context, making it difficult to comprehendunless effort is made to get back into the problem.

In fact, people are reluctant to invest this setupcost, often avoiding communication or misunder-standing. Further, communication over longperiods of time may spiral out of control unlesseffort is made to manage the communication overtime, and individuals remember when to initiatecommunication, remember to respond, and detectproblems in communication where feedback isoften irregular and delayed. Saunders and Jones(1990) propose a model of information acquisitionin decision tasks in which communication patternschange during the life cycle of the task. Forexample, communication about the task may beintense at some initial stage, stop for a while, andthen resume sporadically, and not be sufficientlysalient to regain the communicator's attention.These changes of patterns are difficult to control.Thus, in extremely long time spans, both cognitivecomplexity and dynamic complexity are high.Typical examples are organizational procedures,which require a careful planning of the messageand what mistakes may arise as a result of timechanges (similar to temporal distances betweencommunicators discussed below). Furthermore,as control by planning is associated with highermessage organization, this would also explain thetypically high organization of procedures.

Proposition 8C: A short time-to-completethe task increases the use of control bytesting and adjusting.

Proposition 8D: A very long time-to-complete the task increases the use ofcontrol by planning.

Future work can develop more complex proposi-tions about the interaction between task analyz-ability and communication goals. When the goalof communication is to influence or to manageinterdependent action for tasks of low analyz-ability, the result is high cognitive and dynamiccomplexity (see the earlier section on assumedprinciples of behavior). Thus the interaction of thetwo sources of complexity produces higher com-plexity and, as a result, a higher chance of mis-understanding. This will result in a more intenseuse of communication strategies to cope with thecomplexity and thereby place additional demandson medium and message form. For example,



tasks that involve value judgements in collectiveaction will require negotiations and influencing,which in turn will require high channel capacity,interactivity, and often adaptiveness (Straus andMcGrath 1994). Higher interdependence willusually lead to a higher need for managing rela-tionships and a higher use of perspective takingand affectivity strategies. These too require highchannel capacity. The accumulated demands onthe use of strategies will require very rich com-munication media such as face-to-face. Thisexplains the role of task interdependence in theresults of Table 9.

Sender-Receiver Distance Affects theCommunication Process

ReviewTable 10 reviews recent studies on the effects ofsender-receiver distance, classifying the studiesinto physical and psychological distance. In theorganizational setting, psychological distance isalso divided into organizational and culturaldistance. From the table, it can be seen that,overall, sender-receiver distance is a determiningfactor in the communication process, but most ofthe studies (especially those looking at psycho-logical distance originating from intercultural dif-ferences) have concentrated on task orientedimpact.

One of the more powerful methods of studyingsender-receiver effects is to analyze patterns ofcommunication in organizations. Surprisingly, thistype of research has declined dramatically sincethe 1970s (see O�Reilly et al. 1987) and has notrevived despite significant changes in powerstructures within modern organizations.

PropositionsThe distance between sender and receiverchanges both goal priorities and considerations ofcompatibility and profitability (Figure 5). Theeffects of physical distance are, on the one hand,extremely dependent on the advances in tech-nology, and on the other hand, largely dependenton human physiology, which is beyond our model.We, therefore, concentrate on the cognitive andaffective distances. From a cognitive perspective,

greater differences or distances between sender-receiver worldviews, values, languages, and othercommon factors pertinent to informationprocessing will increase cognitive complexity ofthe communication and lower the plausibility ofmutual understanding. A greater cognitive dis-tance may also be associated with higheruncertainty about what the receiver knows and,therefore, higher cognitive complexity (Kraut andHiggins 1984). Indeed, uncertainty reductiontheory (Berger and Calabrese 1975; Berger andGudykunst 1991) predicts that uncertainty aboutthe receiver will induce more information seeking.When communication diverges from mutualunderstanding, a shared context needs to becreated (Goffman 1981; Krauss and Fusell1991a). In contrast, communication betweenestablished work groups can be less explicitwithout hindering mutual understanding (Bernstein1964). For example, distances generated by inter-cultural differences have been shown to dependon the level of intersection between the pheno-menal fields of the sender and receiver (Haworthand Savage 1989). Furthermore, different back-grounds with different worldviews and experienceswill stimulate a more intense exchange of ideasand perspectives (Cox and Blake 1991; Markus1990). Hence, a higher priority on goals of seekinginformation.

Proposition 9A: Greater cognitive distanceincreases the use of contextualization.

Proposition 9B: Greater cognitive distanceincreases the frequency of requesting infor-mation.

Parallel to the effect of cognitive distance oncognitive complexity, a greater affective distancemay increase affective complexity by increasingthe sender's anxiety over the receiver's reactions(Stephan and Stephan 1985). Such anxiety willusually result in cognitive biases such as selectiveinformation processing. Affective distance mayalso imply a lack of trust between communicators(recall that subsequent trust following the com-munication is part of impact). Low trust reducesthe likelihood of information exchange (Williamson1975). This finding demonstrates how commonpractices may be ineffective and create a �Catch22" situation: cognitive distance coupled with lowtrust will reduce the likelihood of informationexchange where it is most needed in order to build



Table 10. Recent Studies on the Effects of Sender-Receiver Distance

Sender-ReceiverDistance Use and Task-Oriented Impact Relationship-Oriented Impact

Sender-receiverphysical distance:Space

Time

Distance reduces amount of communica-tion but moderated by CMC availability(Sarbaugh-Thompson and Feldman 1998;Sproull and Kiesler 1991). However,distance has no effect (Valacich et al.1993). Distance is one of three majordeterminants of media choice (Caldwell etal. 1995).Distance affects media choice (Reinschand Beswick 1990; Webster and Trevino1995).CMC is particularly useful to communicatebetween shift workers (Huff et al. 1989).Temporal unavailability leads to less inter-active media to promote task closure(Straub and Karahanna 1998).

Distance reduces non-taskrelated communication(Sarbaugh-Thompson andFeldman 1998).CMC reduces impact ofdistance on building rela-tionships (reviewed inMcKenna and Bargh 2000).

Sender-receiverpsychologicaldistance:organizational

Direction of communication affects mediachoice (Zmud et al. 1990). Richest exchange of information isbetween supervisor and subordinates(Allen and Griffeth 1997). Rich media wasbelieved to improve comprehension whenseeking information from outside theorganization (Lee and Heath 1999). Conflicting results on preference for e-mailvs. phone as a function of distance(Wijayanayake and Higa 1999). E-mail ispreferred when effective distance is high(Markus 1994b). Less shared informationleads to higher rate of relevant information(Hightower and Sayeed 1995; Stasser andTitus 1987).�

An awareness of sender-receiver relations is essentialfor successful relationships(Gabarro 1990). Sharedknowledge mediates effect ofmutual trust on performance(Nelson and Cooprider 1996).

Sender-receiverpsychologicaldistance:intercultural

Incompatible cultural patterns of sharinginformation lead to less effective communi-cation (Brett and Okumura 1998; Ohbuchiand Takahashi 1994). No support for effectof intercultural distance on frequency ofinformation seeking (reviewed in Bergerand Gudykunst 1991).

None found.

�Older but unchallenged observation: direction of communication affects amount of information transmitted (O�Reilly andRoberts 1974).



trust. Indeed, mis-communication will be higherwhen inter-cultural distance is greater, because ofdifferent languages, different patterns of usinglanguage, different values and beliefs, and dif-ferent attitudes to communication (Larkey 1996).

Cross-cultural communication, which is charac-terized by a combination of cognitive and affectivedistances, is often problematic. Moreover, dif-ferent reactions to misunderstandings may inten-sify the impact of those misunderstanding,creating not only dissatisfaction but outright hos-tility, which impedes communication even further(Pettigrew and Martin 1989). However, there is aneed to test empirically the effect of affectivedistance on the frequency of information seeking(Kellermann and Reynolds 1990).

Proposition 9C: Greater affective distancereduces the frequency of requesting infor-mation.

Communication Values and NormsAffect the Communication Process

ReviewThe review of recent studies on norms and valuesis organized by organizational and national(usually cross-cultural) studies (Table 11). Thestudies of organizational effects on CMC havefocused on differences in the adoption of themedia. We could not find a similar concern withgoal priorities and the message form and content,but this may reflect an inadequate search. Bycomparison, the studies of national characteri-zations are more varied. On the one hand, thereis evidence of diversity in the way different cul-tures prioritize communication goals and choosemedia and message. On the other hand, it wouldseem that CMC can moderate the cultural dif-ferences (Suzuki 1997; Tan et al 1998b; Watsonet al. 1994; but see Rice et al. 1998). Further-more, in the cross-cultural studies, researchershave used more structured measures (ques-tionnaires), which may make it easier togeneralize across studies.

PropositionsAs noted above, to keep the scope manageable,the propositions concern only one dimension ofvalues: interdependence. Recall that comparedwith independent cultures, individuals in inter-dependent cultures are more likely to think andfeel interdependently with regard to others in agroup. Interdependence dictates, therefore, adifference between in-group behavior and out-group behavior that is less pronounced inindividualistic cultures (Espinoza and Garaza1985). Indeed, this difference would qualify certainpatterns of communication found with collectiviststo in-group behavior where the psychologicaldistance (primarily affective distance) does notexceed some threshold. Interdependence furtherimplies a higher frequency of relationship goals incommunication and a higher rate of affectivity(Gudykunst and Ting-Toomey 1988; Trompenaars1998). Finally, there is also some indication oflower formality in interdependence (Trompenaars1998).

Proposition 10A: Individuals in interdepen-dent cultures tend to engage in more fre-quent communication for managing rela-tionships, provided the affective andcognitive distance does not exceed somethreshold.

Moreover, individuals will usually exercise moreperspective taking and exchange more informaland personal information between group members(Gudykunst et al. 1987). Interestingly, they tendto not disclose information outside the group butonly within the group (Triandis 1989). Collectivistsnot only seek greater involvement, but also aremore capable of doing so (Markus and Kitayama1991). For example, collectivists (Greeks) provedto be better than individualists (Americans) atconsidering their counterparts' interests in anegotiation task (Gelfand and Christakopoulou1999). However, the evidence is still inconclusive(Singelis and Brown 1995), and it is not clearwhether this prediction should also be qualified bythe psychological distance between partners.

Proposition 10B: Individuals in interdepen-dent cultures tend to use more perspectivetaking.



Table 11. Recent Studies on the Effects of Communication Norms and Values

Use and Task-Oriented Impact Relational-Oriented Impact

Organizationalnorms

Social factors affect attitudes and useof CMC (Fulk 1993; Fulk and Boyd1991; Fulk et al. 1990; Kraut et. al1998).Organizational norms have a smallereffect on e-mail than other media(Kiesler 1986).

Coworkers with similar attitudes useCMC to create �ego networks� orrelational communication networks(Fulk and Ryu 1990; Rice and Aydin1991). Hacker�s culture leads to a�flight� from relationship with people(Turkle 1984).

National culture Japanese (collectivists ) use explana-tions and information seeking morefrequently than Americans (indivi-dualists) (Neuliep and Hazleton 1985).Culture moderates the impact of CMC(Watson et al. 1994). Individualism-collectivism did not moderate onpreference and richness assessment ofmedia (Rice et al. 1998). Japanese ratee-mail lower than FTF and fax,although fax was not considered moreuseful (Straub 1994). CMC moderatedthe impact of culture on rounds ofinformation in a group decision (Tan etal. 1998b).

Different cultures build relationshipsand trust in different ways (Doney etal. 1998). Different cultures agree onsome bodily expressions of emotionsbut often disagree on expressionintensity and differ in the use of emo-tional communication strategies suchas control (review by Matsumoto etal. 1989).Individualism-collectivism (American-Japanese) moderates the effect ofsocial identification on relationalcommunication (Suzuki 1997).Relation-oriented cultures prefer e-mail over fax (Rowe and Struck1999).

Table 12 summarizes propositions 8, 9, and 10,which refer to the effect of individual inputs. Thecombined effects of inputs are beyond our scope,but, clearly, they occur and must be researched.Moreover, combinations such as high timepressure and great psychological distance mayproduce critical communication complexity that willfail without appropriate support.

Implications and Conclusion

Thus far, we have proposed a model of organi-zational communication, reviewed recent publica-tions from several distinct fields of research, anddemonstrated how the model can be used togenerate propositions (see Figure 1). Our reviewsuggests that information systems research has

concentrated primarily on communication inputsand impact. The proposed model, on the otherhand, opens up the black box of the communi-cation process by seeking to define those choicemechanisms relating to cognitive and affectivegoal-based strategies, media, and messages. Theview presented is one in which action-orientedimpact is complemented with relationship-orientedimpact. Hence, the paper�s main contribution is inorganizing diverse research into a coherentframework. This enables us to generate a novelunderstanding of goals and strategies, messageforms and media, and multi-purpose communi-cation. The argument here is that this frameworkis a more realistic and more informative view ofcommunication. In consequence, this sectiontakes this view one step further by exploring theimplications of the model on future research andpractice.



Table 12. Summary: Propositions 8, 9, and 10�Task, Sender-Receiver Distance, andCommunication Norms Affect the Goal Preference and Strategy Selection

SpecificProposition Input Component of Process Affected

Task 8A Higher task variety More frequent requests of information

8B Lower task analyzability Higher contextualization

8C Short time to complete task Higher control by testing and adjusting

8D Very long time to complete task Higher control by planning

Distance 9A Greater cognitive distance Higher contextualization

9B Greater cognitive distance More frequent requests of information

9C Greater affective distance Less frequent requests of information

Values 10A Greater interdependence More frequent managing relationships

10B Greater interdependence More perspective taking

Implications for Research

The proposed model brings one closer to theory-based empirical research, which is urgentlyneeded (Steinfield and Fulk 1990). The primaryfocus on the communication process, whichgoverns this paper, means that it has beennecessary to limit the scope of the model, and thepropositions are, therefore, restricted to a subsetof potential elements and relationships. The mostobvious implication for future research is to enablethe generation of hypotheses from the proposi-tions developed above. The operational definitionsproposed here, or dealt with in the cited research,suggest that such empirical work is feasible.Indeed, preliminary field work suggests that theelements of the communication process (e.g.,attributes of the message) can be measuredsatisfactorily (Te�eni et al. forthcoming). Two otherimplications are developed around (1) refining themodel and (2) exploring new perspectives.

Refining the Model

The Dynamic Impact of Communicationon the Communication InputsThe broken arrow in Figure 1 signifies a dynamicprocess, in which the impact of communication

feeds back over time into the situational context toaffect the sender-receiver distance. It may alsoaffect the task through growing experience and,over even longer periods of time, the norms andvalues. The feedback relationship reveals the fullcomplexity of communication when it is treateddynamically, but it should also be noted that itnecessarily implies an elevation of our discussionat the level of a message to the level of an indivi-dual (the sender). We accept here that com-munication creates a shared meaning by buildinga social context (Sproull and Kiesler 1992) and acognitive context (Kintsch 1988). Indeed, as usersgain experience, they will place increasing valueon the impact of communication (King and Xia1997). We can further assume that the knowledgelearned from the integration of multiple messagesresides in an individual's memory, so that we canbriefly sketch future directions of research onthese impacts of communication and, in the nextsection, discuss some practical implications.

Several research directions can be pursued toexplore this feedback loop. One is the inter-mediate effect of mutual understanding and rela-tionship with regard to cognitive complexity (Krautand Higgins 1984) and affective complexity(Gudykunst and Shapiro 1996; Stephan andStephan 1985). Perceptions of mutual trust and



mutual understanding grow with communication,provided that the parties of the communicationprocess perceive it to be successful, but this takestime (Walther 1995). A second direction is theinteraction of such an impact with the medium.The Internet seems to have reduced perceptionsof communication complexity and, consequently,the sender-receiver distance (McKenna and Bargh2000). Although computer-mediated communi-cation has been shown to increase trust, it issometimes short lived (Jarvenpaa and Leidner1999). As we discuss in the next section, infor-mation technology can go further to establish andmaintain a shared understanding, therebyreducing the cognitive distance. A third directionuses structuration theory to analyze the impact ofcommunication technology on norms (Orlikowski1992) and the effects over time on genres ofcommunication (Orlikowski and Yates 1994; Yateset al. 1999; but see critic in Banks and Riley1993).

Refining Elements and Attributes:Goals, Strategies, MessageOur review suggests important areas for furthertheoretical development. Flores (1998) hasalready stated that we know more about taskrelated aspects of communication (action) thanabout commitment in communication (rela-tionship). We should, therefore, begin by refiningthe goal of managing relationships so as to distin-guish, for example, between building, maintaining,severing, and controlling relationships (Rogersand Farace 1975). Similarly, new affectivestrategies may be articulated such as a strategy ofaffective control (as opposed to the predominantlycognitive control we have discussed above).Research has shown, for example, that controlover social interactions is relaxed in CMC (Siegalet al. 1986).

An important area of research is the interactionbetween task-related goals and relational goalsand between affective and cognitive strategies.For example, Goffman (1981) describes theinvolvement-independence strategy, in which thegoal of maintaining �face� incorporates a tendencyto consider one�s partner's thoughts and feelingsto a greater extent (involvement) or lesser extent(independence). It has been defined as �the

negotiated public image, mutually granted eachother� (Scollon and Scollon 1995, p. 35). As such,it may have special relevance to new forms ofcommunication in the World Wide Web (Flores1998). We should note, however, that this goesbeyond our model to include what Habermas hasreferred to as the dramaturgical model of action,which may have different validity criteria. Anotherexample of an affective-cognitive strategy is thatof ambiguity in messages, which serves to pro-mote the goal of influencing (Eisenberg 1984) butmay also be effective in maintaining face.

The summary of recent work on the effects ofmedia shows that it is necessary to study theimpact of interactivity on relationship and theimpact of adaptiveness on both relationship andunderstanding (see Table 5). Moreover, there isclear need for more empirical research on mes-sage form. Meanwhile, in the review of taskattributes, there are no studies of task effects onrelationship (see Table 9). This is an area ofpotential research that is of particular significanceto virtual organizations, which on the one hand willhave to cope with decreasing levels of analyz-ability and, on the other hand, may depend onCMC to form relationships. Such organizations arelikely to experience difficulties in communicationunless steps are taken to cope with communi-cation complexity. The situation of overload withinorganizations, a corollary of task variety, amountof information, and media that promotes distri-bution, is another research area of growingimportance. Rudy (1996) has summarized anddiscussed future directions for research onoverload and, more recently, so have Schultzeand Vandenbosch (1998).

More research is also needed on the effects ofsender-receiver distance. As has been notedabove, the analysis of communication flowsbetween people in organizations has declineddramatically. In light of the potential effects ofsender-receiver distance on communication, suchanalyses would seem to be fundamental to anystudy of interpersonal communication. It may betrue to suggest that the emerging forms oforganizations, in particular virtual and multi-national firms, will trigger a new interest in thisarea of communication (Quinn 1992). Similarly,little research into individual styles in CMC has



been found. This absence may be connected withthe widespread decrease in research intoindividual differences in information systemsduring the 1980s, after a long stream of research(Huber 1983). Nevertheless, cognitive styles thatdictate communication determine, by definition,different preferences and capabilities for differentcommunication strategies. This whole area seemsan untapped avenue for future research that may,after all, lead to individually tailored systems. Thegeneration of such systems may be especiallyrelevant to the new virtual organization in whichthe bulk of the communication cannot rely on face-to-face communication.16

Exploring Other Perspectives:Receiver, Privacy and OthersSeveral other important issues have been omittedfrom the model to keep its complexity manage-able. Perhaps the most immediate need is to addthe receiver's perspective (c.f. Contractor andEisenberg 1990; Rudy 1996). Communicationstrategies have been described from the sender�sperspective, since they have been formulated asmeans for achieving goals determined by thesender. Yet the active receiver becomes a senderherself the moment she responds, and followingthis, the same strategies are employed. Moreover,the same strategies can be applied in a similarfashion to the process of receiving information(this assumes that the sender and receiver sharethe same communication goals). For example,the receiver can also use the attention-focusingand control strategies to improve understanding.Nevertheless, it is necessary to articulate newcommunication strategies for receiving information

designed to achieve goals determined by thereceiver and it is necessary to examine the samestrategy from the receiver�s viewpoint, e.g., howdoes the receiver choose to respond to a requestfor information.

A related perspective is that of privacy, which isconcerned with the right of individuals to deter-mine when, how, and to what extent information istransmitted. Every act of communication dis-closes something of the communicator and oftenthis is regarded as a risky act (Goffman 1981).The research directions recommended above(particularly relational goals and affective control)will demand a better understanding of privacy.Furthermore, CMC intensifies disclosure bymaking recorded information accessible, and notalways in an obvious way. At a message level,privacy links directly to the medium and through itto organizational memory (a topic discussedbelow). The sender�s perception of the communi-cation�s confidentiality depends on media attri-butes (Sillince 1997). It will be important toinvestigate how perceptions of confidentialityaffect communication behavior, and to link themback to attributes of the media. For example,when do people compromise and choose media oflow confidentiality? Can privacy dictate lowchannel capacity? The privacy perspective will belikely to become a crucial aspect in understandingorganizational communication, particularly as theboundaries between the workplace and home(Venkatesh and Vitalari 1992) and between theorganization and its suppliers and customerscontinue to blur. Privacy may be an importantfactor in generalizing our model across theseboundaries.

Once these aspects of receiver and privacy areclarified, it will be possible to analyze moreeffectively other important issues of communi-cation and to integrate them with the elements wehave discussed here. Recent examples includetopics such as suites of communication techno-logies (Ocker et al. 1998), network size (Valacichet al. 1991), message content (Sussman andSproull 1999), gender (Gefen and Straub 1997),and awareness (McDaniel and Brinck 1997).

16The ideas of communication complexity are a naturalentry point for studying the effect of cognitive styles suchas cognitive complexity (see Kelly 1955; Schroder et al.1967). A review of such a broad area lies beyond thescope of this paper, but one example demonstrates thedirection of further research. Mischel's (1973) conceptionof a self-regulatory system defines individual differencesin the degree of control one imposes on informationprocessing. This style is likely to predict the tendency touse the strategy of control. For other directions ofresearch on individual differences in communication, seeGreene and Lindsey (1989), O�Keeffe (1988) and Wilson(1989).



Implications for DesigningCommunication Support Systems

As has been suggested previously, communi-cation in organizations is becoming increasinglycomplex, more intensive, and supported morefrequently by information technology (Huber1990). Systems that support communication will,in the future, have to deal with large communi-cation networks, mobile communication, inter-cultural communication, ubiquitous multi-mediacommunication, and continuous states of a heavyinformation load (or overload). Decomposing thecommunication process into sub-processes bringscloser the possibility of developing more specificdesign guidelines for such systems. The emphasishere is on what functionality needs to be devel-oped, rather than how to develop it, and specifictechnologies are presented only to demonstratefeasible directions. We must, however, stress thetentativeness of implications drawn from anuntested theory, which is obviously a lessdesirable foundation for deriving design impli-cations than one that has been tested. Yet it maybe beneficial to trigger and guide experimentationwith new functionality. Indeed, several influentialarticles have proposed a framework and derivedfrom it prescriptions for designing and imple-menting information technologies for groups(DeSanctis and Gallupe 1987) and organizations(Huber 1990). Moreover, we cannot at this state ofknowledge solve possible tradeoffs betweeneffects that are addressed by individual propo-sitions. This knowledge may come from observingpeople communicate with a variety of advancedtechnologies and consequently revising certainpropositions.

The three factors of the proposed model(Figure 1) frame the discussion about func-tionality. One can conceive of information tech-nology that (1) identifies the inputs (e.g., the initialdistances between communicators), (2) supportsthe formulation of goals and the choice andimplementation of communication strategies,medium, and message, and (3) provides the userwith feedback on impact. Organizational memoryis a key resource in supporting each of thesetypes of functionality. Figure 6 can, therefore, beseen as a general framework for design that takes

off from the systems requirements specified inFigure 1. The discussion of design followsFigure 6.

Organizational MemoryOrganizational communication and organizationalmemory should be tightly interrelated (Anand et al.1998). Recent research on designing communi-cation systems demonstrates this link: Annotate!is a knowledge dissemination system in whichuser appraisals of knowledge items affect thesystem's prioritization of answers to a query(Ginsburg and Kambil 1998), and AIMS is anagent that automatically prioritizes e-mail mes-sages according to personal preferences mappedby organizational categories of messages(Motiwalla 1995). In this section, we wish tocapitalize on a deeper understanding of contextand of the communication model to flesh out new,perhaps speculative, directions on OM design,particularly the part of OM that is built oncommunication and learning.

Organizational memory (OM) is taken here to bea repository of the context of action. Cossette(1998) has distinguished between three types ofcontext: situational, cognitive, and emotional.Situational context includes information about thecommunicators, the place of the interaction, andtime of the interaction. The earlier discussion ofthe message and medium attributes suggests thatthese too are important aspects of the situationalcontext, inasmuch as they affect the meaning ofthe information transmitted. Cognitive contextincludes the communicators� intentions and hopesof the receiver�s ability to understand. Emotionalcontext is the feelings which one communicatorhas toward another and about the issue com-municated, and can be part of the affective infor-mation communicated (Schwarz 1990). Thesethree types of context address, respectively, thedynamic, cognitive, and affective complexity. Theycould possibly be built into OM to cope with thethree sources of complexity by supporting therelevant communication strategies (for a moretechnical examination of the roles of the differenttypes of context, see Fairclough 1992; Hallidayand Hasan 1989).



Identifysituation

Supportcommunication goalsand strategies

Feedbackimpact

Provide medium and message form

Organizationalmemory

Identifysituation

Supportcommunication goalsand strategies

Feedbackimpact

Provide medium and message form

Organizationalmemory

Figure 6. The Functionality of Computer Supported Communication Systems

The diary in Table 1 is not only a medium fordialog but also part of an OM. In other words, OMis not only used to support communication but, infact, can build from it. An important challenge is todistill the different types of context from theoriginal message, store it in OM, and use it to re-contextualize future communication (Schwartz andTe�eni 2000). Furthermore, following the discus-sion of formality, the OM would include informationthat is organized at different levels of formality.Over time, some messages will be graduallyabstracted to become increasingly formal. Asdiscussed above, it may be important to be able totrace back from the formal to the originalmessage, when communication breaks down. TheOM may, therefore, be constructed through stories(such as the one about Joey spilling tea), facts(such as product #8123 is incomplete), and moreformal principles (such as no food in production).(For some initial directions on abstracting texts,see Crampes et al. 1998.)

InputsA key role of information technology is to createan awareness of the state of input. Physicaldistances in space and time are usually easier todisplay than psychological distances, which are

often less noticeable. The review of sender-receiver distance and the discussion of proposi-tion 9 demonstrated how these distances are bornand how they affect the communication process.Technology�s role in creating awareness has, inthe past, concentrated on the perceptual issues ofcommunication. Clark's contribution theory ofdiscourse (Clark and Brennan 1991) suggests thatit is crucial to monitor the receiver's attention andunderstanding for effective communication,resulting in several attempts to design appropriatesupport such as Portholes (Dourish and Bly 1992),ClearBoard (Ishii et al. 1992), and Peepholes(Greenberg 1996). Similarly, it should be possibleto display semantic distances between sender andreceiver, for example those concerning differencein terminology. For instance, kMail (Schwartz andTe�eni 2000) uses previous knowledge to showwhether or not the same terms are shared bysender and receiver, assuming that a visualdisplay of differences enhances the awareness oftheir existence.

Organizational memory can play a major role incharacterizing the workers involved in the com-munication and then characterizing the psycho-logical distance between them (Anand et al.1998). Simple examples are identifying communi-



Table 13. Computerized Support of Communication Strategies

Contextualization Organizing and retrieving the context information, structuring thecontext information presented

Control Provide feedback on the communication process by manipulating themedia, displaying receiver�s reactions, recording communication,monitoring progress

Attention Focusing Formatting and structuring the information presented, remote control ofinformation presented

Affectivity Templates of appropriate affectivity and feedback on current message(e.g., language check)

Perspective taking Presentation of receiver�s views (e.g., cognitive maps or physicalobjects) to sender at the time of message preparation

cation between different national cultures,between different organizational cultures, orbetween users with different profiles. In theseexamples, an organizational memory is neededwhich is able to characterize the users and matchthem (e.g., DIRE, in French 1994). More complexexamples of such a process may involvedifferences between people who have, in the past,ascribed different meanings to current terms.Furthermore, some errors in communication canbe detected (or suspected) on the basis of theimmediate or related context stored in the OM. Atrivial example is that the mistake in Table 1 aboutproduct 1823 (should be 8123) could be correctedautomatically from its context.

Communication Strategies and GoalsOur model identifies a list of communication stra-tegies, all of which are candidates for computersupport (see Table 13), but also identifies theconditions in which they should be activated inresponse to communication complexity inherent inthe process (proposition 2) and induced by inputs(propositions 8, 9, and 10). Contextualizationdepends on information retrieval and, as a result,is likely to be the most promising direction forcomputer support. Nevertheless, this processposes two main problems: providing complete andrelevant information and providing it only whenneeded. The first problem implies knowledge

structures that on the one hand enable effectiveretrieval and, on the other hand, enable effectivepresentation to the user as an accessible, multi-layered knowledge structure of context such ashypertext.

The second problem (timing) calls for mechanismsthat detect the conditions under which contex-tualization is needed to avoid information over-load. Spider (Boland et al. 1994) is designed topresent context in a variety of forms so that it canlead more efficiently to better, richer communi-cation. The system displays the differentrationales behind an issue in the form of cognitivemaps that highlight the similarities and differencesin the communicators� perspectives. Thus, mes-sages are richer in context but (more importantly)are displayed in a fashion that is manageable.Mao et al. (1996) show how contextualized accessimproves problem solving with Hyper-Finalyzer.The very same idea can be used in person toperson communication. Indeed, Kock (1998)looked at how expert systems technologiessupport contextualization in group-activities.

Control of communication can be enhanced byseveral techniques. One is to emulate the capacitybuilt into face-to-face communication (which wouldbe more effective than the secondary reactionpredicted by propositions 4A and 4B). Effectiveeye contact, which helps maintain control by



informing the sender what message the receiversare sensing, can be emulated by advanceddisplays that create and update a picture of thespatial relationships between communicatorsduring the session (Fussel and Benimoff 1995;Muhlbach et al. 1995). However, the quality ofmedia, with respect to channel capacity, inter-activity, and adaptiveness, has to be very high(Doherty-Sneddon et al. 1997; see a review oftechnical requirements by Patrick 1999). Anotherdirection for supporting control (proposition 4A) isto organize simultaneous feedback duringmessage production and manage it throughout thesession (Herring 1999). For example, messagesthat include dynamically created hypertext links tothe organizational memory should be shown to thesender before transmission (Schwartz and Te�eni2000). The management of feedback throughoutthe session is particularly important when thereare several participants and each one may havehis or her own space on the screen and whenthere are several streams of conversation andeach stream must be associated with its particularfeedback. Herring suggests that two differentviews should be possible, one linear (reflecting thetime dimension), and the other non-linear(reflecting topical progression).

A third direction is to support planned control(proposition 4B). CMC generally provides morecontrol than other media with similar channelcapacity because it can store information to allownon-simultaneity (Hesse et al. 1988). Controlthrough planning cannot guarantee perfect imple-mentation of the plan (e.g., directing the reader tolook for the explanations in the paragraphs pre-ceding the proposition, but the reader may notsearch and find the explanations). Nevertheless,interactive technology, unlike a printed message,can be designed to guide the reader when thesituation arises.

Information technology supports attention focusingthrough formatting effects, multi-modal messagesthat include synchronized voice and motion, andpointing by remote control (e.g., pcANYWHERE).Video conferencing, too, has proven an effectivetechnique to focus attention or at least to createan awareness of low attention (Daly-Jones et al.1998). However, these techniques may notsuffice to draw and retain attention as communi-

cation of higher complexity, from more sources,and on multiple media compete over the user�slimited resources. One possible direction that maybe taken is to develop technologies that providemore integrative solutions, using combinations ofmessage and medium, to grab more attentionfrom the user and allocate it efficiently.

Computer support for perspective taking can beachieved indirectly by presenting the receiver�sperspective, thereby triggering the sender to usethe strategy of perspective taking. Presentingyour communication partner�s cognitive map isone possibility (Boland et al. 1994). Designersand scholars often use whiteboards to createdrafts of message and invite others to add theirperspective by moving, erasing, and adding ideasaround the board. Flatland is a computer- basedwhiteboard that groups ideas and presents themin different perspectives (Mynatt et al. 1999).

Computer support for affectivity is likely to be themore difficult to achieve, because much of ouraffective communication in organizational life istraditionally non-verbal, instinctive, and, often,intentionally non-documented. Using informationtechnology to add social and emotional material tomessages is directly tied to how well we manageto incorporate it in organizational memory asemotional context.

Proposition 2 (about the appropriateness ofcommunication strategies shown in Figure 2) isyet to be tested, but if shown to hold true, theproposition can be used in designing systems thatautomatically recommend the use of communi-cation strategies. If organizational memory isattributed with communicative characteristics thatmatch those described in the model (e.g.,AnswerGarden in Ackerman 1998), the systemcan use either stereotypic knowledge or individualknowledge to assess when strategies are more orless effective and recommend them accordingly.Moreover, when communication goals are deter-mined, it may be possible to match strategies togoals. Some early work on Coordinator (Winogradand Flores 1986) shows it is possible to assign toeach message its purpose but it burdens thesender too much. Advances in technologies thatdetermine the user�s communication goals couldbe used to match appropriate strategies, and



provide feedback to support control (see, forexample, Ardissono and Sestero 1996). Indeed,Collagen models collaborative discourse on thebasis of goals that are either voiced by the user ordetected automatically by an intelligent agent(Rich and Sidner 1998).

Message and MediumPropositions 3, 5, and 7 stipulate effective combi-nations of strategy, message, and medium,namely, compatible designs of medium and mes-sage. If shown to be correct, these combinationscould also be incorporated in support systems. Inthe future, unified interfaces, providing a gatewayto alternative media (e.g., computer, fax, phone,videoconference) and alternative forms of mes-sages (templates, hypermedia structures) willprobably be commonplace. The most direct impli-cation of the corollary to proposition 2 (about shiftsfrom control to contextualization) and of thediscussion on the corresponding shifts in formalityis the need to support an easy transition betweenlevels of formality. We have already noted that theOM should preserve the progression of infor-mation items from low to high formality (e.g.,stories, facts, and abstract principles). In addition,however, the support system should be designedto supply the right level at the right time. Further-more, such systems should be media sensitive(Trevino et al. 1990). Intelligent systems canrecommend effective choices, present defaultdesigns or point at poor choices by using theproposed model to identify effective combinationsof medium and message for given goals andstrategies. For example, Kennedy et al. (1998)use a model-based approach to generate compa-tible message formats.

Feedback on ImpactUltimately, feedback on the impact of com-munication must come from the user�s own reac-tion, but future systems may serve as effectiveproviders of this feedback to the sender. The OMcan be designed to include results of successesand failures of communication that are provided tothe sender at the appropriate time. Little researchhas been carried out in this area, but as communi-cation support systems become more common,the importance of informing senders of impact willgrow. Some form of feed-forward may be possible,

e.g., a simulation of probable errors due to highcognitive distance. Clearly, there is still much todo in terms of developing ways of identifying andreporting on communication failures (see thediscussion in the elements of communicationimpact section).

Conclusion

Globalization, competition, technological sophis-tication, and speed have increased the complexityof organizations. If, indeed, organizations thriveon communication, then clearly, communicationshould enable them to cope with such complexity.The information systems field can, and indeedmust, play a role in enabling effective communi-cation, but for this, such a field needs to informthe design of information technology on the basisof a realistic model. The proposed model hassought to draw a balance between relationshipand action, cognition and affect, message andmedium. As noted in the introduction, we believethat such a balanced view provides a morerealistic view of organizational communication andavoids possible pitfalls in prescriptions, e.g., mini-mize communication by disregarding the need forbuilding relationship. Underlying this approach isthe realization that we are facing new forms ofcommunication and new forms of organizations(Fulk and DeSanctis 1995).

It is important to invent new designs that are ableto support new forms of communication, but it isessential that this should be done only on thebasis of a better understanding of what needs tobe accomplished. In this paper, an attempt hasbeen made to achieve such an understanding byabstracting patterns of current behavior, ratherthan simply by speculating on what can be donewith emerging technologies, such as virtual reality.Computer-mediated communication has beendescribed as a different state of communication,which �may change the psychology and sociologyof the communication process itself �[creating] anew linguistic entity with its own vocabulary,syntax and pragmatics� (Rice and Love 1987, p.86). In order to direct the new communicationprocess, however, it is necessary to understandthe way people choose to behave. Only then willit be possible to design support that is more



relevant to actual communication behavior.Furthermore, it is time to reconsider the metaphorof lean-rich media, which climaxes with face-to-face communication. Intelligent communicationsupport systems may be better thought of asproviders of optimal levels of interactivity, channelcapacity, and adaptiveness in conjunction withrecommendations of optimal message form. Atthe same time, however, it is important to note twocaveats. One is the simplified view of organi-zational communication adopted here in whichorganizational politics are ignored. Organizationalgames in which communication is a medium ofpower may certainly distort the behavior describedabove (Frost 1987). Secondly, technology not onlyfacilitates communication but also creates newrealities that may trigger unproductive communi-cation behavior (Spears et al. 1990; Sproull andKiesler 1992). We may, for example, learn to relyon machines for relating to other people instead oflearning how to relate (McLeod 1999). Thus, therather optimistic view of technology should betaken with a grain of salt. In fact, under certainconditions it may be more effective to train peoplehow to communicate rather than delegate com-munication to machines.

Enterprises of the future are likely to rely evenmore heavily on virtual organization. Trust will becrucial. At the same time, however, they may findit more difficult to develop trust between peoplewho hardly ever meet (Handy 1995). Thus, com-munication is expected to play a growing role inpromoting not only task-oriented goals, but alsorelationship-oriented ones. In this respect, webelieve that, increasingly, organizations will needto design communication support systems basedon cognitive and affective models in order tofacilitate better operations and working rela-tionships within such virtual organizations.

Acknowledgments

Many good colleagues have constructivelycriticized this manuscript in the last two years, andI am very grateful to each and every one. Inparticular, I thank the attendees of severalresearch seminars at Oxford (England), Universityof Warwick (England), Case Western ReserveUniversity (USA), New York University (USA), and

University of Waterloo (Canada). I also thankPhillip Ein-Dor and my colleagues at the Center ofGlobal Knowledge Management at Bar-IlanUniversity (Israel) for their valuable comments. I�mindebted to Rick Watson for his insightful guid-ance and continued encouragement. I would alsolike to thank Gerard Sharpling, Sharon Te�eni, andTali for editing the many pages, illustrations andreferences. The editing of this manuscript hasbeen supported by the Schnitzer Foundation forResearch at Bar-Ilan.

References

Ackerman, M. S. �Augmenting OrganizationalMemory: A Field Study of Answer Garden,�ACM Transactions on Information Systems(16:3), 1998, pp. 203-224.

Adams, D. A., Nelson, R. R., and Todd, P. A. �AComparative Evaluation of the Impact of Elec-tronic and Voice Mail on Organizational Com-munication,� Information & Management (24:1),1993, pp. 9-21.

Allen, D. G., and Griffeth, R. W. �Vertical andLateral Information Processing: The Effects ofGender, Employee Classification Level, andMedia Richness on Communication and WorkOutcomes,� Human Relations (50:10), 1997,pp. 1239-1260.

Anand, V., Manz, C. C., and Glick, W. H. �AnOrganizational Memory Approach to Informa-tion Management,� Academy of ManagementReview (23:4), 1998, pp. 796-809.

Anderson, E., and Weitz, B. �Determinants ofContinuity in Conventional Industrial ChannelDyads,� Marketing Science (8:4), 1989, pp.310-323.

Ardissono, L., and Sestero, D. �Using DynamicUser Models in the Recognition of the Plans ofthe User,� User Modeling and User AdaptedInteraction (5), 1996, pp. 157-190.

Banks, S. P., and Riley, P. �Structuration Theoryas an Ontology for Communication Research,�Communication Yearbook (16), 1993, pp. 167-196.

Barber, P. J. Applied Cognitive Psychology: AnInformation-Processing Framework, Methuen,London, 1988.

Beach, L. R., and Mitchell, T. R. �Image Theory:Principles, Goals, and Plans in Decision



Making,� Acta Psychologica (66:3), 1987, pp.201-220.

Beach, L. R., and Mitchell, T. R. �The Basics ofImage Theory� in Image Theory: Theoreticaland Empirical Foundations, L. R. Beach (ed.),Lawrence Erlbaum, Mahawa, NJ, 1998, pp. 3-18.

Beer, M. Organizational Change and Develop-ment, Goodyear, Santa Monica, CA, 1980.

Berger C. R. "Message Plans, CommunicationFailure and Mutual Adaptation During SocialInteraction,� Chapter 5 in Progress in Com-munication Sciences, Volume XIV, MutualInfluence in Interpersonal Communication:Theory and Research in Cognition, Affect andBehavior, M. T. Palmer and G. A. Barnett(eds.), Ablex Publishing Stamford, CT, 1998,pp. 91-111.

Berger, C. R., and Calabrese R. "Some Explora-tions in Initial Interactions and Beyond: Towarda Developmental Theory of Interpersonal Com-munication,� Human Communication Research(1:1), 1975, pp. 99-112.

Berger, C. R., and Gudykunst, W. B. �Uncertaintyand Communication,� in Progress in Communi-cation Sciences, Volume 10, B. Dervin (ed.),Ablex Publishing, Norwood, NJ: 1991, pp. 21-66.

Berlyne, D. E. Conflict, Arousal and Curiosity,McGraw-Hill, New York, 1960.

Bernstein, B. �Elaborated and Restricted Codes,�American Anthropologist (66:6), 1964, pp. 55-69.

Blalock, H. Theory Construction, Prentice Hall,Englewood Cliffs, NJ, 1969.

Blau, P. M. Exchange and Power in Social Life,John Wiley, New York, 1964.

Bly, S., Harrison, S., and Irwin, S. �Media Spaces:Bringing People Together in a Video, Audio andComputing Environment,� Communications ofthe ACM (36:1), 1993, pp. 28-47.

Boland, R., Tenkasi, R., and Te�eni, D.�Designing Information Technology to SupportDistributed Cognition,� Organization Science(5:3), 1994, pp. 456-475.

Brett, J. M., and Okumura, T. �Inter- and intra-Cultural Negotiation: U.S. and Japanese Nego-tiators,� Academy of Management Journal(41:5), 1998, pp. 495-510.

Caldwell, B. S., Uang, S. T., and Taha, L. H.�Appropriateness of Communication Media Use

in Organizations: Situation Requirements andMedia Characteristics,� Behaviour and Infor-mation Technology (14:4), 1995, pp. 199-207.

Carlson, P. J., and Davis, G. B �An Investigationof Media Selection Among Directors andManagers: From �Self� to �Other� Orientation,�MIS Quarterly (22:3), 1998, pp. 335-362.

Carnevale, P. J. D., and Isen, A. M. �The influ-ence of Positive Affect and Visual Access onthe Discovery of Integrative Solutions inBilateral Negotiation,� Organizational Behaviorand Human Decision Processes (37:1), 1986,pp. 1-13.

Chapanis, A. �Interactive Human Communica-tion,� in Computer Supported CooperativeWork: A Book of Readings, I. Grief (ed.),Morgan Kaufmann Publishers, San Mateo, CA,1988, pp. 127-142.

Chidambaram, L. "Relational Development inComputer-Supported Groups,� MIS Quarterly(20:2), 1996, pp. 143-163.

Clark, H. H. Arenas of Language Use, The Uni-versity of Chicago Press, Chicago, IL, 1992.

Clark, H. H., and Brennan, S. E. �Grounding inCommunication,� in Perspectives on SociallyShared Cognition, L. B. Resnick, J. Lewine,and S. D. Teasley (eds.), American Psycholo-gical Association Press, Washington, DC,1991, pp. 127-149.

Conger, J. A. �The Necessary Art of Persuasion,�Harvard Business Review (76:3), 1998, pp. 84-95.

Contractor, N. S., and Eisenberg, E. M. �Com-munication Networks and New Media in Organi-zations,� in Organizations and CommunicationTechnology, J. Fulk and C. Steinfield (eds.),Sage Publications, London, 1990, pp. 143-172.

Cossette, P. �The Study of Language in Organi-zations: A Symbolic Interactionist Stance,�Human Relations (15:11), 1998, pp. 1355-1377.

Cox, T. H., and Blake, S. �Managing CulturalDiversity: Implications for Organizational Com-petitiveness,� Academy of Management Execu-tive (5), 1991, pp. 45-56.

Crampes, M., Veuillez, J. P., and Ranwez, S.�Adaptive Narrative Abstraction,� in Pro-ceedings of HyperText �98, ACM Press, NewYork, 1998, pp. 97-105.

Daft, R. L., and Lengel, R. H. �Information Rich-ness: A New Approach to Managerial Behavior



and Organizational Design� Research inOrganizational Behavior, 1984, pp. 191-233.

Daft, R. L., and Lengel, R. H. �OrganizationalInformation Requirements, Media Richness andStructural Design,� Management Science(32:5), 1986, pp. 554-571.

Daft, R. L., Lengel, R. H., and Trevino, L. K.�Message Equivocality, Media Selection andManager Performance: Implications for Infor-mation Systems,� MIS Quarterly (11:3), 1987,pp. 355-366.

Daft, R. L., and Macintosh N. B. �A TentativeExploration into the Amount and Equivocality ofInformation Processing in Organizational WorkUnits,� Administrative Science Quarterly (26:2),1981, pp. 207-224.

Daly, B. L. �The Influence of Face-to-Face versusComputer-Mediated Communication Channelson Collective Induction,� Accounting, Manage-ment and Information Technology (3:1), 1993,pp. 1-22.

Daly-Jones, O., Monk, A., and Watts, L. �SomeAdvantages of Video Conferencing Over High-Quality Audio Conferencing: Fluency andAwareness of Attentional Focus,� InternationalJournal of Human-Computer Studies (49:1),1998, pp. 21-58.

D�Ambra, J., Rice R. E., and O�Connor M."Computer-Mediated Communication andMedia Preference: An Investigation of theDimensionality of Perceived Task Equivocalityand Media Richness,� Behaviour and Infor-mation Technology (17:3), 1998, pp. 164-174.

De Bono, E. Practical Thinking, Pelican Books,Middlesex, England, 1976.

Dennis, A. R. �Information Exchange and Use inGroup Decision Making: You Can Lead aGroup to Information, But You Can�t Make ItThink,� MIS Quarterly (20:4), 1996, pp. 433-457.

Dennis, A. R., and Kinney, S. T. �Testing MediaRichness Theory in the New Media: The Effectsof Cues, Feedback and Task Equivocality,�Information Systems Research (9:3), 1998, pp.256-274.

DeSanctis, G., and Gallupe, R. B. �A Foundationfor the Study of Group Decision Support Sys-tems,� Management Science (33:5), 1987, pp.589-609.

DeSanctis, G., and Poole, M. S. �Capturing theComplexity in Advanced Technology Use:

Adaptive Structuration Theory,� OrganizationScience (5:2), 1994, pp. 121-147.

Diehl, E., and Sterman, J. D. �Effects of Feed-back Complexity on Dynamic Decision Making,�Organizational Behavior and Human DecisionProcesses (62:2), 1995, pp. 198-215.

Doherty-Sneddon, G., Anderson, A, O�Malley, C.,Langton, S., Garrod, S., and Bruce, V. �Face-to-Face and Video-Mediated Communication:A Comparison of Dialogue Structure and TaskPerformance,� Journal of Experimental Psycho-logy: Applied (3:2), 1997, pp. 105-125.

Donabedian, B., McKinnon, S. M., and Bruns, W.J. �Task Characteristics, Managerial Socializa-tion and Media Selection,� Management Com-munication Quarterly (11:3), 1998, pp. 372-400.

Doney, P. M., Cannon, J. P., and Mullen, M. R.�Understanding the Influence of NationalCulture on the Development of Trust,� Academyof Management Review (23:3), 1998, pp. 601-620.

Dore, R. �Goodwill and the Spirit of MarketCapitalism,� British Journal of Sociology (34:4),1983, pp. 459-482.

Dourish, P., and Bly, S. �Portholes: SupportingAwareness in a Distributed Work Group,� inProceedings of ACM CHI�92 Conference onHuman Factors in Computing Systems, P.Bauersfeld, J. Bennett, and G. Lynch (eds.),ACM Press, New York, 1992, pp. 541-547.

Downs, A. Inside Bureaucracy, Little Brown,Boston, 1967.

Eggins, S., and Martin, J. R. �Genres and Regis-ters of Discourse� in Discourse as Structureand Process, T. A. van Dijk (ed.), Sage Publi-cations, London, England, 1997, pp. 231-256.

Ehrlich, S. F. �Strategies for Encouraging Suc-cessful Adoption of Office Communication Sys-tems,� ACM Transactions on Office InformationSystems (5:4), 1987, pp. 340-357.

Eisenberg, R. M. �Ambiguity as Strategy inOrganizational Communication ,� Communica-tion Monographs (51:3), 1984, pp. 227-242.

El-Shinnawy, M., and Markus, M. L. �The Povertyof Media Richness Theory: Explaining People�sChoice of Electronic Mail vs. Voice Mail,�International Journal of Human-ComputerStudies (46:4), 1997, pp. 443-467.

El-Shinnawy, M., and Vinze, A. S. �Polarizationand Persuasive Argumentation: A Study ofDecision Making in Group Settings,� MISQuarterly (22:2), 1998, pp. 165-198.



Espinoza, J. A., and Garza, R. T. �Social GroupSalience and Inter-Ethnic Cooperation,� Journalof Experimental Social Psychology (21:4),1985, pp. 380-392.

Fairclough, N. Discourse and Social Change,Polity Press, Cambridge, MA, 1992.

Fish, R. S., Kraut, R. E., and Root, R. W. �Videoas a Technology for Informal Communication,�Communications of the ACM (36:1), 1993, pp.48-61.

Flores, F. "Information Technology and the Insti-tution of Identity,� Information Technology &People (11:4), 1998, pp. 351-372.

French, J. C. �DIRE: An Approach to ImprovingInformal Scientific Communication,� Informationand Decision Technologies (19), 1994, pp. 527-541.

Frost, P. J. �Power, Politics, and Influence,� inHandbook of Organizational Communication:An Interdisciplinary Perspective, F. M. Jablin, L.L. Putnam, K. H. Roberts, and L. W. Porter(eds.), Sage Publications, Newbury, CA., 1987,pp. 503-547.

Fulk, J. �Social Construction of CommunicationTechnology,� Academy of Management Journal(36:5), 1993, pp. 921-950.

Fulk, J., and Boyd, B. �Emerging Theories ofCommunication in Organizations,� Journal ofManagement (17:2), 1991, pp. 407-446.

Fulk J., and Collins-Jarvis, L. "Wired Meetings:Technological Mediation of OrganizationalGatherings,� in New Handbook of Organiza-tional Communication, F. M. Jablin and L. L.Putnam (eds.), Sage, Newbury Park, CA, 2000,pp. 624-663.

Fulk, J., and DeSanctis, G. �Electronic Com-munication and Changing OrganizationalForms,� Organization Science (6:4), 1995, pp.337-349.

Fulk, J., and Ryu, D. �Perceiving Electronic MailSystems: A Partial Test of the Social Informa-tion Processing,� Paper presented to a meetingof the International Communication Asso-ciation, Dublin, 1990.

Fulk, J., Schmitz, J. A., and Schwarz, D. �TheDynamics of Context-Behavior Interactions inComputer-Mediated Communication,� in Con-texts of Computer-Mediated Communication,M. Lea (ed.), Harvester Wheatsheaf, London,1991, pp. 1-29.

Fulk, J., Schmitz, J. A., and Steinfield, C. �ASocial Influence Model of Technology Use,� inOrganizations and Communication Technology,J. Fulk and C. Steinfield (eds.), Sage Publi-cations, Newbury Park, CA, 1990, pp. 117-140.

Fussell, S. R., and Benimoff, I. �Social and Cog-nitive Processes in Interpersonal Communi-cation: Implications for Advanced Telecom-munications Technologies,� Human Factors(37:2), 1995, 228-250.

Gabarro, J. J. �The Development of TrustInfluence and Expectations,� in Inter-PersonalBehavior: Communication and Understandingin Relationships, A. G. Athos and J. J. Gabarro(eds.), Prentice Hall, Englewood Cliffs, NJ,1978, pp. 290-303,

Gabarro, J. J. �The Development of WorkingRelationships,� in Intellectual Teamwork:Social and Technical Bases of CollaborativeWork, J. Galegher, R. E. Kraut, and C. Egido(eds.), Erlbaum, Hillsdale, NJ, 1990, pp. 79-110.

Galbraith, J. R. Organization Design, Addison-Wesley, Reading, MA, 1977.

Gale, S. �Human Aspects of Interactive Multi-media Communication,� Interacting with Com-puters (2), 1990, pp. 175-189.

Gallupe, R. B., Cooper, W. H., Grise, M., and Bas-tianutti, L. �Blocking Electronic Brainstorms,�Journal of Applied Psychology (79:1), 1994, pp.77-86.

Gefen, D., and Straub, D. W. �Gender Diffe-rences in the Perception and Use of E-mail: AnExtension to the Technology AcceptanceModel,� MIS Quarterly (21:4), 1997, pp. 389-400.

Gelfand, M. J., and Christakopoulou, S. �Cultureand Negotiator Cognition: Judgement Accu-racy and Negotiation Processes in Indivi-dualistic and Collectivistic Cultures,� Organiza-tional Behavior and Human Decision Processes(79:3), 1999, pp. 248-269.

Ginsburg, M., and Kambil, A. �Annotate! A Web-based Knowledge Management Support Sys-tem for Document Collections,� Working Paper#IS-98-19, Stern School of Business, New YorkUniversity, 1998.

Goffman, E. Forms of Talk, University of Pennsyl-vania Press, Philadelphia, PA, 1981.

Goldberg, D. T. �The Social Formation of RacistDiscourse� in Anatomy of Racism, D. T.



Goldberg (ed.), University of Minnesota,Minneapolis, 1990, pp. 295-318.

Green, R. S., and Cliff, N. �MultidimensionalComparisons of Structures of Vocally andFacially Expressed Emotions,� Perceptions andPsychophysics (17:5), 1975, pp. 429-438.

Greenberg, S. �Peepholes: Low Cost Awarenessof One�s Community,� in Conference Com-panion of CHI �96, ACM Press, New York,1996, pp. 206-207.

Greene, J. O., and Lindsey, A. E. �Encoding Pro-cesses in the Production of Multiple-Goal Mes-sages,� Human Communication Research(16:1), 1989, pp. 195-231.

Gudykunst, W. B �Individualistic and CollectivisticPerspectives on Communication: An Introduc-tion� International Journal of InterculturalRelations (22:2), 1998, pp. 107-134.

Gudykunst, W. B., and Matsumoto, Y. �Cross-Cultural Variability of Communication in Per-sonal Relationships,� in Communication in Per-sonal Relationships across Cultures, W. B.Gudykunst, S. Ting-Toomey, and T. Nishida(eds.), Sage Publications, Newbury Park, CA,1996, pp. 19-55.

Gudykunst, W. B., and Shapiro, R. B. �Communi-cation in Everyday Interpersonal and IntergroupEncounters� International Journal of Inter-cultural Relations (20:1), 1996, pp. 19-45.

Gudykunst, W. B., and Ting-Toomey, S. Cultureand Interpersonal Communication, Sage Publi-cations, Newbury Park, CA, 1988.

Gudykunst, W., Yoon, Y., and Nishida, T. �TheInfluence of Individualism-Collectivism on Per-ceptions of Communication in In-group andOut-group Relationships,� CommunicationMonographs (54:3), 1987, pp. 295-306.

Gumperz, J. J. Discourse Strategies, CambridgeUniversity Press, Cambridge, England, 1982.

Habermas, J. The Theory of CommunicativeAction: Reason and Rationalization of Society,Volume 1, Beacon Press, Boston, 1984.

Habermas, J. The Theory of CommunicativeAction: Lifeworld and Social System, Volume 2,Beacon Press, Boston, 1987.

Habermas, J. On the Pragmatics of Communi-cation (edited by M. Cooke), The MIT Press,Cambridge, MA, 1998.

Halliday, M. A. K., and Hasan, R. Language,Context and Text: Aspects of Language in a

Social-Semiotic Perspective, Oxford UniversityPress, Oxford, England, 1989.

Hall, E. T. Beyond Culture, Doubleday, New York,1976.

Handy, C. �Trust and the Virtual Organization,�Harvard Business Review (73:3), 1995, pp. 40-50.

Hart, P., and Saunders, C., �Power and Trust:Critical Factors in the Adoption and Use ofElectronic Data Interchange,� OrganizationScience (8:1), 1997, pp. 23-42.

Haworth, D. A., and Savage, G.T ., �A Channel-Ratio Model of Intercultural Communication:The Trains Won�t Sell, Fix Them Please,� Jour-nal of Business Communication (26:3), 1989,pp. 231-254.

Hedlund, J., Ilgen, D. R., and Hollenbeck, J. R.�Decision Accuracy in Computer-Mediatedversus Face-to-Face Decision-Making Teams,�Organizational Behavior and Human DecisionProcesses (76:1), 1998, pp. 30-47.

Herring, S. �Interactional Coherence in CMC,�Journal of Computer Mediated Communication(4:4), 1999 (online publication, 26 pages).

Hesse, B. W., Werner, C. M., and Altman, I."Temporal Aspects of Computer MediatedComunication,� Computers in Human Behavior(4:4), 1988, pp. 147-65.

Hightower, R. T., and Sayeed, L. �The Impact ofComputer-Mediated Communication Systemson Biased Group Discussion,� Computers inHuman Behavior (11:1), 1995, pp. 33-44.

Hightower, R. T., and Sayeed, L. "Effects ofCommunication Mode and Pre-discussion Infor-mation Distribution Characteristics on Informa-tion Exchange in Groups,� Information SystemsResearch (7:4), 1996, pp. 451-465.

Hiltz, S. R., Johnson, K., and Turoff, M. �Experi-ments in Group Decision Making, 1: Communi-cation Process and Outcome in Face-to-Faceversus Computerized Conferences,� HumanCommunication Research (13:2), 1986, pp.225-252.

Hofstede, G. Cultures and Organizations: Soft-ware of the Mind, McGraw-Hill, New York,1991.

Hollingshead, A. B. �Information Suppression andStatus Persistence in Group Decision Making:The Effects of Communication Media,� HumanCommunication Research (23:2), 1996, pp.193-219.



Hollingshead, A. B., and McGrath, J. E. �TheWhole Is Less than the Sum of its Parts: ACritical Review of Research on Computer-Assisted Groups,� in Teams Decisions andTeam Performance in Organizations, R. A.Guzzo and E. Salas (eds.), Jossey-Bass, SanFrancisco, 1995, pp. 46-78.

Hollingshead, A. B., McGrath, J. E., andO�Connor, K. M. �Group Task Performanceand Communication Technology: A Longitu-dinal Study of Computer Mediated versus Face-to-Face Work Groups,� Small Group Research(24:3), 1993, pp. 307-333.

Huang, W., Watson, R. T., and Wei, K. K. �Can aLean E-mail Medium be Used for Rich Com-munication? A Psychological Perspective,�European Journal of Information Systems (7),1998, pp. 269-274.

Huber, G. P. "Cognitive Style as a Basis for MISand DSS Designs: Much Ado About Nothing?,�Management Science (29:5), 1983, pp. 567-579.

Huber, G. P. "A Theory of the Effects of Advan-ced Information Technologies on Organiza-tional Design, Intelligence, and DecisionMaking,� Academy of Management Review(15:1), 1990, pp. 47-71.

Huff, C., Sproull, L., and Kiesler, S. "ComputerCommunication and Organizational Commit-ment: Tracing the Relationship in a CityGovernment,� Journal of Applied SocialPsychology (19:16), 1989, pp. 1371-1391.

Hutchins, E. �The Social Organization of Distri-buted Cognition,� Chapter 13 in Perspective onSocially Shared Cognition, L. B. Resnick, J. M.Levine, and S. D. Teasley (eds.), AmericanPsychological Association Press, Washington,DC, 1991, 283-307.

Hwang, M. I. �Did Task Type Matter in the Use ofDecision Room GSS? A Critical Review and aMeta-Analysis,� Omega International Journal ofManagement Science (20:1), 1998, pp. 1-15.

Iedema, R. "Formalizing Organizational Meaning,�Discourse and Society (10:1), 1999, pp. 49-65.

Irvine, J. "Formality and Informality in Communi-cative Events,� American Anthropologist (81:4),1979, pp. 773-790.

Ishii, H., Kobayashi, M., and Grudin, J. �Inte-gration of Inter-Personal Space and SharedWorkspace: ClearBoard Design and Experi-

ments,� Proceedings of CSCW�92� ComputerSupported Collaborative Work, ACM Press,New York, 1992, pp. 33-42.

Janis, I. L. Crucial Decisions, The Free Press,New York, 1989.

Janson, M., Brown, A., and Taillieu, T. �Colruyt:An Organization Committed to Communi-cation,� Information Systems Journal (9), 1999,pp. 175-199.

Janson, M. A., and Woo, C. C. �A Speech ActLexicon: An Alternative Use of Speech ActTheory in Information Systems,� InformationSystems Journal (6), 1996, pp. 301-329.

Jarvenpaa S. L., and Leidner D. E. "Communica-tion and Trust in Global Virtual Teams,� Organ-ization Science (10:6), 1999, pp. 791-815.

Jones, J. W., Saunders, C., and McLeod, R."Media Usage and Velocity in Executive Infor-mation Acquisition: A Exploratory Study,�European Journal of Information Systems (2:4),1993, pp. 260-272.

Jones, J. W., Saunders, C., and McLeod, R."Information Acquisition During Decision-Making Processes: An Exploratory Study ofDecision Rules in Media Selection,� IEEETransactions on Engineering Management(41:1), 1994, pp. 41-49.

Jordan, J. M �Executive Cognitive Control inCommunication,� Human CommunicationResearch (25:1), 1998, pp. 5-38.

Kasher, A. �Foundations of Philosophical Prag-matics,� in Basic Problems in Methodology andLinguistics, R. Butts and J. Hintikka (eds.),Reidel, Dordrecht, Netherlands, 1977.

Kasper, G., and Kellerman, E. (eds.). Communi-cation Strategies: Psycholinguistic and Socio-linguistic Perspectives, Addison WesleyLongman, London, 1997.

Kellermann, K., and Reynolds, R. "WhenIgnorance is Bliss: The Role of Motivation toReduce Uncertainty in Uncertainty ReductionTheory,� Human Communication Research(17:1), 1990, pp. 5-75.

Kelly, G. The Psychology of Personal ConstructsVolume1, Norton and Company, New York,1955.

Kennedy, M., Te�eni, D., and Treleavan, J.�Impacts of Decision Task, Data and Display onStrategies for Extracting Information,� Inter-national Journal of Human Computer Studies(48:2), 1998, pp. 159-180.



Kettinger, W. J., and Grover, V. "The Use ofComputer-Mediated Communication in an Inter-Organizational Context,� Decision Sciences(28:3), 1997, pp. 513-555.

Kiesler, S. �Think Ahead: The Hidden Message inComputer Network,� Harvard Business Review(64), 1986, pp. 46-59.

Kiesler, S., Zubrow, D., Moses, A. M., and Geller,V. �Affect in Computer Mediated Communica-tion: An Experiment in Synchronous Terminal-to-Terminal Discussion,� Human ComputerInteraction (1:1), 1985, pp. 77-104.

King, R. C., and Xia, W. "Media Appropriateness:Effects of Experience on Communication MediaChoice,� Decision Sciences (28:4), 1997, pp.877-910.

Kintsch, W. �The Role of Knowledge in DiscourseComprehension: A Construction-IntegrationModel,� Psychological Review (95:2), 1988, pp.163-182.

Klien, K. J., Dansereau, F., and Hall, R. J. "LevelsIssues in Theory Development, Data Collection,and Analysis,� Academy of ManagementReview (19:2), 1994, pp. 195-229.

Kock, N. Knowledge Transfer Using Expert Sup-port Systems: A Study on the Role of Analysisand Explanation Facilities in ESS for GroupDecision Making, Unpublished Ph.D. Disser-tation, University of Waikato, New Zealand,1998.

Krauss, R. M., and Fussell, S. R. �ConstructingShared Communicative Environments,�Chapter 9 in Perspectives on Socially SharedCognition, L. B. Resnick, J. M. Levine, and S.D. Teasley (eds.), American PsychologicalAssociation Press, Washington, DC, 1991a, pp.172-200.

Krauss, R. M., and Fussell, S. R. �Perspective-Taking in Communication: Representations ofOthers' Knowledge in Reference,� Social Cog-nition (9:1), 1991b, pp. 2-24.

Kraut, R. E., and Higgins, E. T. �Communicationand Social Cognition,� in Handbook of SocialCognition, R. S. Wyer Jr. and T. K. Srull (eds.),Lawrence Erlbaum Associates, Hillsdale, NJ,1984 pp. 87-127.

Kraut, R. E., Rice, R. E., Cool, C., and Fish, R. S.�Varieties of Social Influence: The Role of Utilityand Norms in the Success of a New Com-munication Medium,� Organization Science(9:4), 1998, pp. 437-453.

Lantz, A. �Heavy Users of Electronic Mail,� Inter-national Journal of Human-Computer Inter-action (10:4), 1998, pp. 361-379.

Larkey, L. K. �Toward a Theory of CommunicativeInteractions in Culturally Diverse Workgroups,�Academy of Management Review (21:2), 1996,pp. 463-491.

Latour, B. "The Force and the Reason of Experi-ment,� in Experimental Inquiries, H. E. LeGrand(ed.), Kluwer Academic Publishers, Dordrecht,Netherlands, 1990, pp. 49-80.

Lea, M. �Rationalist Assumptions in Cross-MediaComparisons of Computer-Mediated Communi-cation,� Behaviour and Information Technology(10:2), 1991, pp. 153-172.

Lea, M., and Spears, R. �Computer-MediatedCommunication, De-individuation and GroupDecision-Making,� International Journal of Man-Machine Studies (34:2), 1991, pp. 283-301.

Lee, A. S. �Electronic Mail as a Medium for RichCommunication: An Empirical InvestigationUsing Hermeneutic Interpretation,� MISQuarterly (18:2), 1994 pp. 143-158.

Lee, J., and Heath, R. L. "Managerial MediaSelection and Information Evaluation from theReceiver's Perspective in Decision-MakingContexts,� Management CommunicationQuarterly (13:1), 1999, pp. 76-99.

Lee, J., and Jablin, F. M. �Maintenance Communi-cation in Superior-Subordinate Work Relation-ships,� Human Communication Research(22:2), 1995, pp. 220-257.

Mackay, W. E. �Diversity in the Use of ElectronicMail: A Preliminary Inquiry,� ACM Transactionson Information Systems (6:4), 1988, pp. 380-397.

Mao, J., Benbasat, I., and Dhaliwal, J. S.�Enhancing Explanations in Knowledge-BasedSystems with Hypertext,� Journal of Organiza-tional Computing and Electronic Commerce(6:3), 1996, pp. 239-268.

Markus, H. R., and Kitayama S. "Culture and theSelf: Implications for Cognition, Emotion andMotivation,� Psychological Review (98:2), 1991,pp. 224-253.

Markus, M. L. �Towards a 'Critical Mass' Theoryof Interactive Media,� in Organizations andCommunication Technology, J. Fulk and C. W.Steinfield (eds.), Sage Publications, NewburyPark, CA, 1990, pp. 194-218.



Markus, M. L. �Electronic Mail as the Medium ofManagerial Choice,� Organization Science(5:4), 1994a, pp. 502-527.

Markus, M. L. �Finding a Happy Medium:Explaining the Negative Effects of ElectronicCommunication on Social Life at Work,� ACMTransactions on Information Systems (12:2),1994b, pp. 119-149.

Markus, M. L., Bikson, T. K., El-Shinnawy, M., andSoe, L .L . �Fragments of Your Communication:Email, Vmail, and Fax,� Information Society(8:4), 1992, pp. 207-266.

Matsumoto, D., Wallbott, H. G., and Scherer, K.R. �Emotions in Intercultural Communication,�in Handbook of International and InterculturalCommunication, M. K. Asotte and W. B. Gudy-kunst (eds.), Sage Publications, Newbury Park,CA, 1989, pp. 225-246.

Mayer, R. E. �Structural Analysis of ScienceProse: Can We Increase Problem-Solving Per-formance?� in Understanding Expository Text:A Theoretical and Practical Handbook forAnalyzing Explanatory Text, B. K. Britton and J.B. Black (eds.), Lawrence Elbraum Associates,Hillsdale, NJ, 1985, pp. 65-86.

McDaniel, S. E., and Brinck, T. �Awareness inCollaborative Systems,� SIGCHI Bulletin (29:4),1997, pp. 68-71.

McGrath, J. E. Groups: Interaction and Perfor-mance, Prentice-Hall, Englewood Cliffs, NJ,1984.

McGuire, T. W., Kiesler, S., and Siegel, J. �Groupand Computer Mediated Discussion Effects inRisk Decision Making,� Journal of Personalityand Social Psychology (52:5), 1987, pp. 917-930.

McKenna, K. Y. A., and Bargh, J .A. "Plan 9 fromCyberspace: The Implications of the Internetfor Personality and Social Psychology,� Per-sonality and Social Psychology Review (4:1),2000, 1, pp. 57-75.

McLaughlin, M. L. Conversation: How Talk IsOrganized, Sage Publications, Beverly Hills,CA, 1984.

McLeod, P. L. Personal Communication, CaseWestern Reserve University, 1999.

Melcher, A. J., and Beller, R. "Toward a Theory ofOrganization Communication Consideration inChannel Selection,� Academy of ManagementJournal (10), 1967, pp. 39-52.

Miller, G. R. (ed.). Explorations in InterpersonalCommunication, Sage Publications, NewburyPark, CA, 1976.

Mingers, J. C. "Information and Meaning: Foun-dations for an Intersubjective Account,� Infor-mation Systems Journal (5:4), 1995, pp. 285-306.

Mintzberg, H. The Nature of Managerial Work,Harper and Row, New York, 1973.

Mischel, W. �Toward a Cognitive Social Learning:Re-conceptualization of Personality,� Psycho-logical Review (80:4), 1973, pp. 252-283.

Moore, S. A. �Categorizing Automated Mess-ages,� Decision Support Systems (22:3), 1998,pp. 213-241.

Morand, D. A. �The Role of Behavioral Formalityand Informality in the Enhancement of Bureau-cratic versus Organic Organizations,� Academyof Management Review (20:4), 1995, pp. 831-872.

Motiwalla, L .F. �An Intelligent Agent forPrioritizing E-mail Messages,� Information Re-sources Management Journal (8:2), 1995, pp.16-24.

Muhlbach, L., Bocker, M., and Prussog, A.�Telepresence in Video Communications: AStudy on Stereoscopy and Individual Eye Con-tact,� Human Factors (37:2), 1995, pp. 290-305.

Muhlfelder, M., Klein, U., Simon, S., and Luczak,H. "Teams Without Trust? Investigations inthe Influence of Video-Mediated Communi-cation on the Origin of Trust Among Cooper-ating Persons,� Behaviour and InformationTechnology (18:5), 1999, pp. 349-360.

Mynatt, E. D., Igarashi, T., Edwards, W. K., andLaMarca, A. �Flatland: New Dimensions inOffice Whiteboards,� in Proceedings of CHI�99,ACM Press, New York, 1999, pp. 346-353.

Nelson, K. M., and Cooprider, J. G. �The Contri-bution of Shared Knowledge to IS GroupPerformance,� MIS Quarterly (20:4), 1996, pp.409-432.

Neuliep, J. W., and Hazleton, V., Jr. �A Cross-Cultural Comparison of Japanese and Ameri-can Persuasive Strategy Selection,� Interna-tional Journal of Intercultural Relations (9:4),1985, pp. 389-404.

Ngwenyama, O. K., and Lee, A. S. �Communi-cation Richness in E-mail: Critical Social



Theory and the Contextuality of Meaning,� MISQuarterly (21:2), 1997 pp. 145-167.

Ngwenyama, O. K., and Lyytinen K. J. �Group-ware Environments as Action ConstitutiveResources: A Social Action Framework forAnalyzing Groupware Technologies,� ComputerSupported Cooperative Work: The Journal ofCollaborative Computing (6:1), 1997, pp. 71-93.

Norman, D. A. The Design of Everyday Things,Doubleday, New York, 1990.

Ocker, R., Fjermestad, J., Hiltz, S. R., andJohnson, K. �Effects of Four Modes of GroupCommunication on the Outcomes of SoftwareRequirements Determination,� Journal ofManagement Information Systems (15:1), 1998,pp. 99-118.

O�Connaill, B., Whittaker, S., and Wilbur, S.�Conversations Over Videoconferences: AnEvaluation of Video-Mediated Interaction,�Human-Computer Interaction (8:4), 1993, pp.389-428.

Oetzel, J. G. �Explaining Individual Communi-cation Processes in Homogeneous and Hetero-geneous Groups Through Individualism-Collec-tivism and Self-Construal,� Human Communi-cation Research (25:2), 1998, pp. 202-224.

Ohbuchi, K., and Takahashi, Y. �Cultural Styles ofConflict Management in Japanese and Ameri-cans: Passivity, Covertness, and Effectivenessof Strategies,� Journal of Applied SocialPsychology (24:15), 1994, pp. 1345-1366.

O�Keefe, B. J. �The Logic of Message Design:Individual Differences in Reasoning AboutCommunication,� Communication Monographs(55:1), 1988, pp. 80-103.

O�Reilly, C. A., Chatman, J. A., and Anderson, J.C. �Message Flow and Decision Making,� inHandbook of Organizational Communication:An Interdisciplinary Perspective, F. M. Jablin,L. L. Putnam, K. H. Roberts, and L. W. Porter(eds.), Sage Publications, Newbury, CA. 1987,pp. 600-623.

O�Reilly, C. A., and Pondy, L. R. �OrganizationalCommunication,� in Organizational Behavior,S. Kerr (ed.), Grid, Columbus, OH, 1979. pp.119-150.

O�Reilly, C. A., and Roberts, K. H. �InformationFiltration in Organizations: Three Experiments,�Organizational Behavior and Human Perfor-mance (11:2), 1974, pp. 253-265.

Orlikowski, W. J. "The Duality of Technology:Rethinking the Concept of Technology inOrganizations,� Organization Science (6:3),1992, pp. 398-427.

Orlikowski, W. J., and Yates, J. "GenresRepertoire: Examining the Structuring Com-munication Practices in Organizations,� Admin-istrative Science Quarterly (39:4), 1994, pp.541-574.

Osgood, C. E. �On Whys and Wherefores of E, P,and A,� Journal of Personality and SocialPsychology (12:3), 1969, pp. 194-199.

Palme, J. �Survey of Computer-Based MessageSystems,� in Proceedings of the Conferenceon Human-Computer Interaction (INTERACT�84), B. Shackel (ed), Elsevier North-Holland,Amsterdam, 1985, pp. 923-928.

Patrick, A. S. �The Human Factors of MboneVideoconferences: Recommendations forImproving Sessions and Software,� Journal ofComputer Mediated Communication (4:3), 1999(online publication).

Payne, J. W., Bettman, J. R., and Johnson, E. J.The Adaptive Decision Maker, CambridgeUniversity Press, Cambridge, England, 1993.

Pettigrew, T. F., and Martin, J. �OrganizationalInclusion of Minority Groups: A Social Psycho-logical Analysis,� in Ethnic Minorities: SocialPsychological Perspectives, J. P. Van Ouden-hoven and T. M. Willemnsen (eds.), SwetsNorth America, Berwyn, PA, 1989, pp. 169-200.

Petty, R. E., and Cacioppo, J. T. Communicationand Persuasion: Central and PeripheralRoutes to Attitude Change, Springer-Verlag,New York, 1986.

Phillips, S. R., and Eisenberg, E. M. �StrategicUses of Electronic Mail in Organizations,�Electronic Journal of Communication (3:2),1993 (online publication).

Piaget, J. Success and Understanding, HarvardUniversity Press, Cambridge, MA, 1978.

Pinsonneault, A., Barki, H., Gallupe, R. B., andHoppen, N. �Electronic Brainstorming: TheIllusion of Productivity,� Information SystemsResearch (10:2), 1999, pp. 110-133.

Poole, M. S., and Hirokawa, R. Y. �Communi-cation and Group Decision Making,� in Com-munication and Group Decision Making, R. Y.Hirokawa and M. S. Poole (eds.), Sage Publi-cation, Thousand Oaks, CA, 1996, pp. 3-18.



Poole, M. S., and Jackson, M. H. �Communi-cation Theory and Group Support Systems,� inGroup Support Systems: New Perspectives, L.M. Jessup and J. S. Valacich (eds.), Macmillan,New York, 1993, pp. 281-293.

Quinn, J. B. Intelligent Enterprise: A Knowledgeand Service Based Paradigm for Industry, TheFree Press, New York, 1992.

Rasmussen, J. Information Processing andHuman-Machine Interaction: An Approach toCognitive Engineering, Elsevier, Amsterdam,1986.

Reinsch, N. L., and Beswick, R. W. �Voice Mailversus Conventional Channels: A Cost Minimi-zation Analysis of Individuals' Preferences,�Academy of Management Journal (33:4), 1990,pp. 801-816.

Rice, R. E. �Task Analyzability, Use of NewMedia, and Effectiveness: A Multi-Site Explora-tion of Media Richness,� Organization Science(3:4), 1992, pp. 475-500.

Rice, R. E., and Aydin, C. �Attitudes Toward NewOrganizational Technologies,� AdministrativeSciences Quarterly (36:2), 1991, pp. 219-244.

Rice, R. E., D�Ambra J., and More, E. �Cross-Cultural Comparison of Organizational MediaEvaluation and Choice,� Journal of Communi-cation (48:3), 1998, pp. 3-26.

Rice, R. E., Hughes, G., and Love, G. �Usageand Outcomes of Electronic Messaging at aRand Organization: Situational Constraints, IobLevel, and Media Awareness,� Office, Tech-nology and People (5:2), 1989, pp. 141-161.

Rice, R. E., and Love, G. �Electronic Emotion:Socioemotional Content in a Computer Medi-ated Network,� Communication Research(14:1), 1987, pp. 85-108.

Rice, R. E., and Shook, D. E. "Relationship of JobCategories and Organizational Levels of theUse of Communication Channels, IncludingElectronic Mail: A Meta-Analysis and Exten-sion,� Journal of Management Studies (27:2),1990, pp. 195-229.

Rich, C., and Sidner, C. �Collagen: A Collabora-tion Manager for a Collaborative InterfaceAgent,� User Modeling and User AdaptedInteraction (8:3/4), Fall 1998, pp. 315-350.

Robinson, M., Kovalainen, M., and Auramaki, E."Diary as Dialogue in Papermill ProcessControl,� Communications of the ACM (43:1),2000, pp. 65-70.

Rocco, E. �Trust Breaks Down in ElectronicContext,� Proceedings of CHI�98, ACM Press,New York, 1998, pp. 496-502.

Rogers, L. E., and Farace, R. V. �Analysis ofRelational Communication in Dyads: New Mea-surements Procedures,� Human Communi-cation Research (1), 1975, pp. 222-239.

Romm, C. T., and Pliskin, N. �Electronic Mail asa Coalition-Building Information Technology,�ACM Transactions on Information Systems(16:1), 1998, pp. 82-100.

Rowe, F., and Struck, D. "Cultural Values, MediaRichness and Telecommunication Use in anOrganization,� Accounting, Management andInformation Technologies (9:3), 1999, pp. 161-192.

Rudy, I. A. �A Critical Review of Research onElectronic Mail,� European Journal of Informa-tion Systems (4:4), 1996, pp. 198-213.

Russ, G. S., Daft, R. L., and Lengel, R. H. �MediaSelection and Managerial Characteristics inOrganizational Communication,� ManagementCommunication Quarterly (4:2), 1990, pp. 151-175.

Rutter, D. R., and Stephenson, G. M. �The Roleof Visual Communication in SynchronisingConversation,� European Journal of SocialPsychology (7), 1977, pp. 29-37.

Salazar, A. J. �Understanding the SynergisticEffects of Communication in Small Groups,�Small Group Research (26:2), 1995, pp. 169-199.

Sarbaugh-Thompson, M., and Feldman, M. S.�Electronic Mail and Organizational Communi-cation: Does Saying �Hi� Really Matter?�Organization Science (9:6), 1998, pp. 685-698.

Saunders, C., and Jones, J. �Temporal Se-quences in Information Acquisition for DecisionMaking: A Focus on Source and Medium,�Academy of Management Review (15:1), 1990,pp. 29-46.

Saunders, C., and Miranda, S. "Information Ac-quisition in Group Decision Making,� Infor-mation & Management (34:1), 1998, pp. 55-74.

Schein, E. H. �Three Cultures of Management:The Key to Organizational Learning,� SloanManagement Review (38:1), Fall 1996, pp. 9-20.

Schlosberg, H. �The Description of FacialExpressions in Terms of Two Dimensions,�Journal of Experimental Psychology (44), 1952,pp. 229-237.



Schober, M. F. �Spatial Perspective-Taking inConversation,� Cognition (47:1), 1993, pp. 1-24.

Schoop, M. �Habermas and Searle in Hospital: ADescription Language for Cooperative Docu-mentation Systems in Healthcare,� in Com-munication Modeling: The Language/ ActionPerspective, F. Dignum, and J. Dietz (eds.),Eindhoven University, Eindhoven, Netherlands,1997, pp. 117-132.

Schroder, H. M., Driver, M. J., and Streufert, S.Human Information Procession, Holt, Rinehartand Winston, New York, 1967.

Schultze, U., and Vandenbosch, B. �InformationOverload in a Groupware Environment: NowYou See It, Now You Don�t,� Journal of Organi-zational Computing and Electronic Commerce(8:2), 1998, pp. 127-148.

Schwartz, D., and Te'eni, D. �Tying Knowledge toAction with kMail,� IEEE Intelligent Systems(15:3), 2000, pp. 33-39.

Schwarz, N. �Feeling as Information: Informa-tional and Motivational Functions of AffectiveStates,� in Handbook of Motivation and Cogni-tion: Foundation of Social Behaviour, Volume2, R. M. Sorrentino (ed.), Guilford Press, NewYork, 1990, pp. 527-561.

Scollon, R., and Scollon, S. W. InterculturalCommunication: A Discourse Approach, Black-well, Oxford, England, 1995.

Searle, J R. Speech Acts: An Essay in thePhilosophy of Language, Cambridge UniversityPress, Cambridge, England, 1969.

Sharrock, W., and Button, G. �On the Relevanceof Habermas� Theory of Communicative Actionfor CSCW,� Computer Supported CooperativeWork: The Journal of Collaborative Computing(6:4), 1997, pp. 369-389.

Shirani, A., Tafti, M. H. A., and Affisco, J. F. "Taskand Technology Fit: A Comparison of TwoTechnologies for Synchronous and Asynch-ronous Group Communication,� Information &Management (23:3), 1999, pp. 139-150.

Short, J., Williams, E., and Christie, B. The SocialPsychology of Telecommunications, Wiley,London, 1976.

Siegel, J., Dubrovsky, V., Kiesler, S., andMcGuire, T. W. "Group Processes in Compu-ter-Mediated Communication,� OrganizationalBehavior and Human Decision Processes(37:2), 1986, pp. 157-187.

Sillince, J. A. A. �A Media-Attributes and Design-Choices Theory of the Information Technology-Organizations Relation,� Journal of Organi-zational Computing and Electronic Commerce(7:4), 1997, pp. 279-303.

Simons, R. �Strategic Orientation and TopManagement Attention to Control Systems,�Strategic Management Journal (12:1), 1991,pp. 49-62.

Singelis, T., and Brown, W. �Culture, Self, andCollectivist Communication: Linking Culture toIndividual Behavior,� Human CommunicationResearch (21:3), 1995, pp. 354-389.

Sitkin, S. B., Sutcliffe, K. M., and Barrios-Cholin, J.R. �A Dual Capacity Model of CommunicationMedia Choice in Organizations,� Human Com-munication Research (18:4), 1992, pp .563-598.

Smith, J., and Vanecek, M. �Dispersed-GroupDecision Making Using Non-simultaneousConferencing: A Report of Research,� Journalof Management Information Systems (7:2),1990, pp. 71-92.

Spears, R., Lea, M., and Lee, S. "De-individuationand Group Polarization in Computer MediatedCommunication,� British Journal of SocialPsychology (29:2), 1990, pp. 121-134.

Sproull, L., and Kiesler, S. "Connections: NewWays of Working in the Networked Organi-zation,� MIT Press, Cambridge, MA, 1992.

Srinivasan, A., and Te'eni, D. "Modeling as Con-strained Problem Solving: An Empirical Studyof the Data Modeling Process,� ManagementScience (41:3), 1995, pp. 419-434.

Stasser, G., and Titus, W. "Effects of InformationLoad and Percentage of Shared Information onthe Dissemination of Unshared InformationDuring Group Discussion,� Journal of Per-sonality and Social Psychology (53:1), 1987,pp. 81-93.

Steinfield, C. W., and Fulk, J. "The Theory Imper-ative,� in Organizations and CommunicationTechnology, J. Fulk and C. Steinfield (eds.),Sage Publications, London, 1990, pp. 13-25.

Stephan, W. G., and Stephan, C. W. �IntergroupAnxiety,� Journal of Social Issues (41:2), 1985,pp. 157-166.

Stohl, C., and Redding, W. C. �Messages andMessage Exchange Processes,� in Handbookof Organizational Communication: An Inter-disciplinary Perspective, F. M. Jablin, L. L.



Putnam, K. H. Roberts, and L. W. Porter (eds.),Sage Publications, Newbury, CA., 1987, pp.451-502.

Straub, D. �The Effect of Culture on IT Diffusion:E-mail and Fax in Japan and the U.S,�Information Systems Research (4:3), 1994, pp.23-47.

Straub, D., and Karahanna, E. �KnowledgeWorker Communications and Recipient Avail-ability: Toward a Task Closure Explanation ofMedia Choice,� Organization Science (9:2),1998, pp. 160-175.

Straus, S. G., and McGrath, J. E. �Does theMedium Matter? The Interaction of Task Typeand Technology on Group Performance andMember Reactions,� Journal of AppliedPsychology (79:1), 1994, pp. 87-97.

Street, R. L., and Capella, J. H. (eds.). Sequenceand Pattern in Communication Behaviour,Edward Arnold, London, 1985

Sussman, S. W., and Sproull, L. �Straight Talk:Delivering Bad News Through ElectronicCommunication,� Information SystemsResearch (10:2), 1999, pp. 150-166.

Suzuki, S. �Cultural Transmission in InternationalOrganizations: Impact of Interpersonal Com-munication Patterns in Intergroup Context,�Human Communication Research (24:1), 1997,pp. 147-180.

Tan, B. C. Y., Wei, K. K., Sia, C., and Raman, K.S. �A Partial Test of the Task-Medium FitProposition in a Group Support SystemEnvironment,� ACM Transactions on Compu-ter-Human Interaction (6:1), 1999a, pp. 47-66.

Tan, B. C. Y., Wei, K .K., and Watson, R. T. �TheEqualizing Impact of a Group Support Systemon Status Differentials,� ACM Transactions onInformation Systems (17:1), 1999b, pp. 77-100.

Tan, B. C. Y., Wei, K. K., Watson, R. T., Clapper,D. L., and McLean, E. R. �Computer-MediatedCommunication and Majority Influence:Assessing the Impact in an Individualistic anda Collectivistic Culture,� Management Science(44;9), 1998a, pp. 1263-1278.

Tan, B. C. Y., Wei, K. K., Watson, R. T., andWalczuch, R. M. �Reducing Status Effects withComputer-Mediated Communication: Evidencefrom Two Distinct National Cultures,� Journal ofManagement Information Systems (15:1),1998b, pp. 119-141.

Tang, J. C., and Isaacs, E. A. �Why Do Users LikeVideo? Studies of Multimedia-Supported Col-laboration,� Computer Supported CooperativeWork (1), 1992, pp. 163-193.

Tatar, D. G., Foster, G., and Bobrow, D. G.�Design for Conversation: Lessons from Cog-noter,� International Journal of Man-MachineStudies (34;2), 1991, pp. 185-209.

Te�eni, D. �Analysis and Design of Process Feed-back in Information Systems: Old and NewWine in New Bottles,� Accounting, Manage-ment and Information Technology (2:1), 1992,pp. 1-18.

Te�eni, D., Sagie, A., Zaidman, N., Schwartz, D.,and Hamburger, Y. �Organizational Communi-cation: A Model and Field Study,� IEEE Trans-actions on Professional Communications(44:1), March 2001.

Te'eni, D., and Schwartz, D. G. �Contextualizationin Computer-Mediated Communication,� Infor-mation Systems Review (1), 2000, pp. 59-75.

Thompson, J. D. Organizations in Action: SocialScience Bases of Administrative Theory,McGraw-Hill, New York, 1967.

Trevino, L. K., Daft, R. L., and Lengel, R. H.�Understanding Media Choices: A SymbolicInteractionist Perspective,� in Organizationsand Communication Technology, J. Fulk and C.W. Steinfield (eds.), Sage Publications,Newbury Park, CA, 1990, pp. 71-94.

Trevino, L. K., Lengel, R. H., and Daft, R. L.�Media Symbolism, Media Richness, and MediaChoice in Organizations: A Symbolic Interac-tionist Perspective,� Communication Research(14:5), 1987, pp. 553-574.

Triandis, H. C. Individualism and Collectivism,Westview, Boulder, CO, 1995.

Triandis, H. C. �Cross-Cultural Studies of Indivi-dualism and Collectivism,� in Nebraska Sympo-sium on Motivation, J. J. German (ed.), Univer-sity of Nebraska Press, Lincoln, NE, 1989, pp.41-133.

Trompenaars F. Riding the Waves of Culture:Understanding Cultural Diversity in GlobalBusiness, McGraw Hill, New York, 1998.

Turkle, S. The Second Self, Simon and Schuster,New York, 1984.

Tyre, M.J ., and von Hippel, E. �The SituatedNature of Adaptive Learning in Organizations,�Organization Science (8:1), 1997, pp. 71-83.



Valacich, J. S., Dennis, A., and Connolly, T. �IdeaGeneration in Computer-Based Groups: A NewEnding to an Old Story,� OrganizationalBehavior and Human Decision Processes(57:3), 1994, pp. 448-467.

Valacich, J. S., Dennis, A., and Nunamaker, J. F.,Jr. �Electronic Meeting Support: TheGroupSystems Concept,� International Journalof Man-Machine Studies (34:2), 1991, pp. 261-282.

Valacich, J. S., Paranka, D., George, J. F., andNunamaker, J. F. "Communication Concur-rency and the New Media: A New Dimensionfor Media Richness,� Communication Research(20:2), 1993, pp. 249-276.

Vallacher, R. R., and Wegner D. M. �What DoPeople Think They're Doing? Action Identi-fication and Human Behavior,� PsychologicalReview (94:1), 1987, pp. 3-15.

Van de Ven, A. H., Delbecq, A., and Koenig, R.�Determinants of Coordination Modes WithinOrganizations,� American Sociological Review(41:3), 1976, pp. 322-338.

Van Dijk, T. A., and Kintsch, W. Strategies ofDiscourse Comprehension, Academic Press,New York, 1983.

Venkatesh, A., and Vitalari, N. P. �An EmergingDistributed W ork Arrangement: AnInvestigation of Computer-Based SupplementalWork at Home,� Management Science (38:12),1992, pp. 1687-1706.

Waldron, V. R., and Applegate, J. L. �Inter-personal Construct Differentiation and Con-versational Planning: An Examination of theTwo Accounts for the Production of CompetentVerbal Disagreement Tactics,� Human Com-munication Research (21), 1994, pp. 3-35.

Wallbott, H. G., and Scherer, K. R. �Cues andChannels in Emotion Recognition,� Journal ofPersonality and Social Psychology (51:4),1986, pp. 690-699.

Walther, J. B. �Interpersonal Effects in Computer-Mediated Interaction: A Relational Perspec-tive,� Communication Research (19:1), 1992,pp. 52-90.

Walther, J. B. �Relational Aspects of Computer-Mediated Communication: Experimental Obser-vations Over Time,� Organization Science (6:2),1995, pp. 186-203.

Walther, J. B., and Burgoon, J. K. �RelationalCommunication in Computer Mediated Inter-action,� Human Communication Research(19:1), 1992, pp. 50-88.

Warkentin, M. E., Sayeed L., and Hightower, R. T."Virtual Teams versus Face-to-Face Teams: AnExploratory Study of a Web-Based ConferenceSystem,� Decision Sciences (28:4), 1997, pp.975-996.

Watson, R. T., Ho, T. H., and Raman, K. S.�Culture: A Fourth Dimension of GroupSupport Systems,� Communications of theACM (37:10), 1994, pp. 44-55.

Webster, J. �Desktop Videoconferencing: Exper-iences of Complete Users, Wary Users, andNon-users,� MIS Quarterly (22:3), 1998, pp.257-286.

Webster, J., and Trevino, L. K. �Rational andSocial Theories as Complementary Explana-tions of Communication Media Choices: TwoPolicy-Capturing Studies,� Academy ofManageent Journal (38;6), 1995, pp. 1544-1572.

Weedman, J. �Task and No-Task Functions of aComputer Conference Used in ProfessionalEducation: A Measure of Flexibility,� Inter-national Journal of Man-Machine Studies(34:2), 1991, pp. 303-318.

Weick, K. E. The Social Psychology of Orga-nizing, Addison Wesley, Reading, MA, 1979.

Westmyer, S. A., DiCoccio, R. L., and Rubin, R.B. �Appropriateness and Effectiveness of Com-munication Channels in Competent Inter-personal Communication,� Journal of Com-munication (48:3), 1998, pp. 27-48.

Whitener, E., Brout, S. E., Korgaurd, M. A., andWerner, J. M. �Managers as Initiators of Trust:An Exchange Relationship Framework forUnderstanding Managerial TrustworthyBehavior,� Academy of Management Review(23:3), 1998, pp. 330-349.

Whittaker, S. "Rethinking Video as a Technologyfor Interpersonal Communications,� Interna-tional Journal of Human-Computer Studies(42:5), 1995, pp. 501-529.

Wijayanayake, J., and Higa, K. �CommunicationMedia Choice by Workers in Distributed Envi-ronment,� Information & Management (36:6),1999, pp. 329-338.

Williamson, O. E. Markets and Hierarchies, TheFree Press, New York, 1975.



Wilson, S. R. Coordinating Compliance and FaceGoals Within Persuasive Messages, Unpub-lished Doctoral Dissertation, Purdue University,1989.

Winograd, T., and Flores, F. UnderstandingComputers and Cognition: A New Foundationfor Design, Addison Wesley, Reading, MA,1986.

Wood, R. E. �Task Complexity: Definition of theConstruct,� Organizational Behavior andHuman Decision Processes (37:1), 1986, pp.60-82.

Yates, J., Orlikowski, W. J., and Okamura, K."Explicit and Implicit Structuring of Genres inElectronic Communication: Reinforcement andChange of Social Interaction,� OrganizationScience (10:1), 1999, pp. 83-103.

Zack, M. H. �Interactivity and CommunicationMode Choice in Ongoing ManagementGroups,� Information Systems Research (4:3),1993, pp. 207-239.

Zigurs, I., and Buckland, B. K. �A Theory ofTask/Technology Fit and Group Support Sys-tems Effectiveness,� MIS Quarterly (22:3),1998, pp. 313-334.

Zmud, R. W., Lind, M. R., and Young, F. W. �AnAttribute Space for Organizational Communi-cation Channels,� Information SystemsResearch (1:4), 1990, pp. 440-457.

About the Author

Dov Te�eni is on the faculty at Bar-Ilan University,Israel, where he serves as director of the Univer-sity�s e-learning center and as the foundingdirector of the interdisciplinary center for GlobalKnowledge Management. Dov studies severalrelated areas of information systems in the organi-zational context: human-computer interaction,computer support for decision making and sys-tems design. In addition, he is interested in infor-mation systems for non-profit organizations. Hisresearch usually combines model building, labora-tory and field experiments and the development ofprototype systems such as Spider and kMail. Heis currently working on the design of adaptiveWeb Articles and the use of information tech-nology for supporting communication within multi-nationals corporations.

Documents

Review: A Cognitive-Affective Model of Organizational ... · Te™eni/Organizational Communication and IT 252 MIS Quarterly Vol. 25 No. 2/June 2001 edge of the mechanisms that guide