Articulo del mes

* Acad&my of ManagemeTit Review200D. Vol. 25, No. 2. 428-438.

NOTE

PASSING THE WORD: TOWARD A MODEL OFGOSSIP AND POWER IN THE WORKPLACE

NANCY B. KURLANDLISA HOPE PELLED

University of Southern California

Although gossip is widespread, seldom has it been a topic of management research.Here we build a conceptual model oi workplace gossip and its eifects on the power oiemployees who initiate it. After defining and distinguishing among different kinds o(workplace gossip, we develop propositions about Ihe effect of that gossip on gosslp-ers' expert, referent, reward, and coercive power. We then suggest how moderatorsmay shape those effects and discuss implications of the model.

As early as the Hawthorne Studies (Roethlis-berger & Dickson, 1943), management scholarsrecognized the existence of the informal organi-zation. Unlike the formal organization, whichappears in organization charts and reflects pre-scribed patterns for officially sanctioned mes-sages, the informal organization consists ofspontaneous, emergent patterns that result fromindividuals' discretionary choices (Stohl, 1995:65). This informal network, also called thegrapevine (e.g., Baird, 1977; Daniels, Spiker, &Papa, 1997), has received considerable attentionin the years since its discovery (e.g., Davis, 1953;Katz & Kahn, 1978; Krackhardt & Hanson, 1993;Podolny & Baron, 1997). Still, there is a need forcloser examination of its specific components—for example, rumor, "catching up," and gossip(Goldsmith & Baxter, 1996). Accordingly, in thisarticle we explore one such component: work-place gossip.

Although psychologists (e.g.. Fine & Rosnow,1978), sociologists (e.g., Eder & Enke, 1991), andanthropologists (e.g., Dunbar, 1996) have exam-ined the nature and role of gossip in larger so-ciety, scholars have yet to develop a conceptualmodel of workplace gossip—or even agree on

We are grateful for comments Chris Earley and threeanonymous reviewers provided. We also thank Tom Cum-mings, Janet Fulk, Bill Gartner, Mike Kamins, Peter Kim,Peter Monge, Nandini Rajagopalan, Kathleen Reardon, andPatti Riley for their helpful comments and suggestions.

Both authors contributed equally; our names appear inalphabetical order.

its definltion^despite Noon and Delbridge's(1993) call for research on the topic. Thus, it isimportant to begin redressing this gap. In thisarticle we draw on writings from multiple disci-plines to offer a definition and theoretical modelof workplace gossip and its consequences.

Models of general communication typicallyhave been of two kinds. The first, most commonkind is the linear model (e.g., Berlo, 1960;Osgood, Suci, & Tannenbaum, 1957; Shannon &Weaver, 1949), in which the researcher treatscommunication as a "left-to-right, one-way" pro-cess (Rogers & Kincaid, 1981: 33). Key compo-nents of linear models are the source (personwho initiates communication), message (contentof the communication), channel (transmissionmedium), and receiver (person receiving themessage; Ruch, 1989). Communication is viewedas a process by which a message is transferredfrom an active source, through a channel, to apassive receiver.

The second kind of general communicationmodel is the convergence model (Rogers & Kin-caid, 1981). In convergence models (e.g., Kincaid,1979; Pearce, Figgins, & Golen, 1984) researcherstreat communication as a two-way process. Sug-gesting that participants in the communicationprocess are simultaneously sending and receiv-ing messages, researchers developing thesemodels make less distinction between senderand receiver. Instead, they delve into the rela-tionships among communication participants,the larger social networks in which those rela-tionships exist, and the dynamic nature of com-

42S

2000 Kurland and Felled 429

munication (e.g., how communication changesits participants).

To ensure practical value in communicationmodels, researchers may need to balance thesimplicity of linear models with the complexityof convergence models. As Smeltzer and Leo-nard have suggested, a communication modelshould "contain enough elements so that userscan relate their personal experiences and train-ing to the model. But it must not become socomplex that practitioners find it impossible tounderstand" (1994: 32). Thus, our model lies be-tween the linear and convergence categories.Like linear models, its primary emphasis is onthe flow of a message (gossip) from source (gos-siper) to receiver (gossip recipient).' However,with our model we improve on traditional linearmodels by paying greater attention to the com-munication context—specifically, the culture inwhich gossip occurs. Also, the receiver in ourmodel has a more active role than in strict linearmodels: we consider the interplay betweensource and receiver—that is, how the relation-ship between gossiper and recipient moderatesthe effects we propose. Additionally, we incor-porate the receiver's reaction to the message inour model's dependent variable: the source'spower over the receiver. The receiver's interpre-tation of the gossip largely determines howmuch power the source gains.

Power is the dependent variable in our modelfor several reasons. First, social scientists (e.g.,Berger, 1994; Giddens, 1984; Mumby, 1988) havesuggested that communication in general tendsto shape power structures in organizations aswell as society. Second, in extant writings ongossip, scholars have hinted at linkages topower (e.g., Emler, 1994). Third, power is a mul-tidimensional construct (French & Raven, 1959;Hinkin & Schriescheim, 1989); as such, it hassufficient breadth to capture a variety of work-place gossip effects. Finally, power is often acritical asset to employees (Pfeffer, 1992).

Although the focus of our model is the gossiper-recipient dyad, it is important to keep in mindthat such dyads are embedded in social net-works. Mutual friends and acquaintances of thegossiper and recipient can influence the prolif-

' The source may be either a supervisor, subordinate, orpeer of the recipient. That is, the direction of gossip mayeither be upward, doivnwfaid, or lateral.

eration and impact of gossip (Burt & Knez, 1996;Jaeger, Skelder, & Rosnow, 1998). Indeed, re-searchers (Martin, Feldman, Hatch, & Sitkin,1983; Martin & Siehl, 1983) have observed thateven an ostensibly minor story about one em-ployee can ultimately transform a corporate cul-ture, if that story is shared by many organiza-tional members.

A complete network analysis of gossip is be-yond the scope of our model, for as Burt andKnez note, even "a minimal assumption of activethird parties creates enormous complexity fortheoretical analysis" (1996: 72). Nevertheless, atseveral points in this article, we touch on howsuch networks play a role in gossip-power link-ages.

KEY CONCEPTS IN THE PROPOSED MODEL

Definition and Types of Gossip

As prior researchers have noted (Jaeger et al.,1998; Schein, 1994), gossip traditionally has beendefined as idle chatter, chitchat, or the eviJtongue. These negative connotations largelyarose from religious writings (e.g., Exod. 23 : 1;Lev. 19 :16; Prov. 25 :18). Many authors (e.g., Bok,1984) continue to treat gossip as improper andoverly subjective. Some, however, recently haveoffered neutral definitions, such as "evaluativetalk about a person who is not present" (Eder 8EEnke, 1991: 494) and "the process of informallycommunicating value-laden information aboutmembers of a social setting" (Noon & Delbridge,1993: 25). Unlike their negative counterparts,these more even-handed definitions allow forgossip's functional as well as dysfunctional side(e.g., Dunbar, 1996; Tebbutt, 1995). Here, we drawupon and adapt these neutral conceptualiza-tions, defining workplace gossip as iniormaland evaJuative talk in an organization, usuallyamong no more than a few individuals, aJboufanother member of that organization who is notpresent.

Although laypersons and academics (e.g.,Ayim, 1994) occasionally may suggest that gos-sip encompasses informal communicationabout objects or events—not just people—ourtreatment focuses on talk about other persons.We delimit our definition in this manner for tworeasons. First, in scholarly writings on gossip inlarger society (e.g., Eder & Enke, 1991; Harris,1993; Rosnow & Fine, 1976), researchers predom-

430 Academy oi Management Review April

inantly treat the concept as communicationabout people. Second, the American Manage-ment Association (AMA) recently asserted thatthe grapevine may include a wide range of in-formal communication, whereas gossip focusessolely on information about people (Smith, 1996).

lust as there are distinctions between gossipand other forms of informal communication,there are important distinctions among differentkinds of gossip. A review of relevant literaturepoints to three dimensions useful for makingthese distinctions: sign, credibility, and work-relatedness. Following writings on feedback(e.g., Ilgen, Fisher, & Taylor, 1979), we definesign as the positivity or negativity of the infor-mation being related. When gossip consists offavorable news about others—for example, stat-ing that "Mary received a raise"—its sign ispositive. When gossip consists of unfavorablenews about others, its sign is negative.^

Credibility is the extent to which the gossip isbelievable—that is, it is seemingly accurate andtruthful. Message credibility has been the sub-ject of considerable research in the fields ofcommunication, marketing, and social psychol-ogy (e.g., Boehm, 1994; McCroskey, 1969; Slattery8f Tiedge, 1992). A recent review attests to itsimportance as a communication feature (Self,1996).

Consistent with prior literature (e.g.. Morrow,1981; Tushman, 1979) in which authors have dis-tinguished between work-related and non-work-related communication, we distinguish amongwork-related (professional) and non-work-related (social) gossip. We define worfc-reJated-ness as the degree to which gossip is focused ona subject's work life, such as job performance,career progress, relationships with other organ-izational members, and general behavior in theworkplace.

Definition and Types of Power

Also essential to our model is the concept ofpower. Pfeffer has described power as "the po-

tential ability to influence behavior, to changethe course of events, to overcome resistance,and to get people to do things that they wouldotherwise not do" (1992: 30). Finkelstein has re-ferred to power as "the capacity of individualactors to exert their will" (1992: 507). Based onthese writings and the writings of others (French& Raven, 1969; House, 1988; Shackleton, 1995), wedefine power here as the ability to exert one'swill, influencing others to do things that theywould not otherwise do. In the model we specif-ically focus on the gossiper's power over gossiprecipients.

The multidimensionality of power is well rec-ognized. French and Raven (1959) advanced atypology of power that remains popular (e.g.,Atwater, 1995; Davis & Schoorman, 1997; Hinkin& Schriesheim, 1994), distinguishing among fivekinds of power that one individual (whom wecall Person A) can have over another individual(whom we call Person B): coercive power, rewardpower, legitimate power, expert power, and ref-erent power.^ Although organizational scholarshave offered other power typologies (e.g.,Finkelstein, 1992; Yukl & Falbe, 1991), Frenchand Raven's original classification is the mostwidely accepted and adopted. Their typology isparticularly useful for describing individual-level power, which is the focus of our model.Hence, our propositions pertain to four of thesepower types (coercive, reward, expert, and refer-ent) that we expect gossip to influence. (We donot consider legitimate power as an outcomebecause it is largely based on one's position—that is, hierarchical rank—rather than on socialprocesses.) Our predictions refer to the Frenchand Raven dimensions, but we draw from arange of power and influence writings to de-velop those predictions.

^ Within the categories of positive gossip and negativegossip, it is possible to make additional distinctions. Forexample, gossip can be negative if it describes an unfortu-nate event that befell someone (e.g., a broken leg), but it canalso be negative if it describes unethical behavior. Here, weinterpret gossip as negative when it constitutes a "smear"that could detract from a subject's reputation. Positive gos-sip, however, tends to enhance a subject's reputation.

^ Coercive power is the power that emerges from PersonB's belief that Person A has the ability to punish him or her.flewaid power is the power that emerges from Person B'sbelief that Person A can provide him or her with desiredoutcomes. Legitimate power is the power that emerges fromPerson B's perception that Person A has a legitimate right,based on position in the organization, to influence him orher. Expert power is the power that emerges from Person B'sbelief that Person A has special knowledge or expertise thatPerson B needs. Finally, referent power is the power thatemerges from Person B's attraction for and desire to beassociated with Person A.

2000 Kuiland and Pelled 431

THEORETICAL BACKGROUND ANDHYPOTHESES

Figure 1 presents our model. In the followingsections we develop propositions about the il-lustrated linkages.

Linkages Between Gossip and Power

One main effect of negative gossip may beenhanced coercive power. When the gossiper

nGURE 1Proposed Model of Gossip and Power

Gossip features:- credibility- work-relatedness

Positivegossip

Rewardpower

Coercivepower

Negativegossip

s\

\\

t

Context- relationsh- organizati

factors:ip qualityonal culture

Expertpower

Referentpower

relates negative news about a third party, recip-ients may infer that the gossiper also couldspread negative information about them (Yerk-ovich, 1977). Because such information can dam-age reputations and/or careers (Emler, 1994;Fine, 1977; Glazer & Ras, 1994; Tebbutt, 1995),negative gossip may constitute an implicitthreat by the gossiper. French and Raven (1959)proposed that when Person B perceives that Per-son A can administer punishments. Person Ahas coercive power over Person B. Along thesame lines, other researchers (e.g.. Hunt &Nevin, 1974; Tedeschi, 1972) have advanced thenotion that implicit and explicit threats can en-hance power and influence.* Those who feelthreatened may comply in order to avoid retri-bution (Kipnis, Schmidt, & Wilkinson, 1980).Thus, negative gossip may give the gossipercoercive power over recipients.

Proposifion i; 7n a wort seffing, nega-tive gossip will enhance the gossiper'scoercive power over gossip recipients.

Positive gossip, in contrast, is likely to affectreward power. When a gossiper shares positivenews about another worker, recipients may inferthat the gossiper also could spread positive in-formation about them. Because such informationcan strengthen reputations and/or careers, pos-itive gossip shows the ability to distribute (al-beit indirectly) desired outcomes. French andRaven (1959) suggested that when Person B per-ceives that Person A has control over valuedoutcomes. Person A has reward power over Per-son B. Along the same lines, Etzioni (1961) pro-posed that control of material and symbolic re-wards are a basis for power. Additionally,Emerson asserted that power "resides in control

'' There may be limits to the effectiveness ol implicitthreats (e.g., the threat of spreading negative information) inattempts to gain power. First, if the gossiper has few con-nections to others, recipients may be less concerned aboutthe gossiper's ability to spread dark secrets. Second, somenews—for example, information that is hard to remember—may be especially difficult to spread (Zimbardo & Leippe,1991), Third, as our Proposition 7 suggests, recipients may beless afraid of the gossiper when they have a good relation-ship with him or her. Moreover, those vrho do feel threatenedmay strive to decrease their dependence on the person mak-ing the threat (Bacharach & Lawler, 1980; Tjosvold, 1995). AsBacharach and Lawler have noted, coercion "should be mosteffective when the target is highly dependent on the user"{1980: 177).

432 Academy of Management Review April

over the things [another person] values Inshort, power resides implicitly in the other's de-pendency" (1962: 32). Resource dependence the-orists (e.g., Salancik & Pfeffer, 1977), too, haveadvanced the notion that power comes from thecontrol of relevant resources—resources that areimportant to others. Hence, by revealing the gos-siper's ability to control an important resource(reputation), positive gossip may give the gos-siper reward power over recipients.

Proposition 2: In a work setting, posi-tive gossip will enhance the gossiper'sreward power over gossip recipients.

Gossip in general, whether positive or nega-tive, is apt to influence expert power, for it canfacilitate an exchange of data and help build aknowledge base (e.g.. Code, 1994; Dunbar, 1996).When a gossiper shares information about oth-ers, the recipient may learn more about the or-ganization's values. As Heath (1994) has ob-served, stories shared by coworkers can helpemployees understand principles by which theirorganization operates. Additionally, gossip canreveal that the gossiper has relevant knowledgeabout persons in the work environment. As thegossiper demonstrates such knowledge (an abil-ity that depends, in part, on the gossiper's net-work centrality), the recipient may come to viewthe gossiper as a source of useful information,and the gossiper may thereby gain expertpower.

Proposifion 3: In a work setting, gossipwill enhance the gossiper's expertpower over gossip recipients.

In the case of referent power, we expect gossipto have competing effects. One possibility isthat gossip reduces referent power, for gossipmay be seen as a small or petty activity. Asmentioned earlier, in religious writings andother sources of guidance and education, gossipis often denounced as idle, immoral, or improper(Levin & Arluke, 1987). Socialized by such teach-ings, many persons perceive gossip as repre-hensible, and they look down on those who en-gage in the behavior. Gossip, therefore, maydetract from the referent power of the gossiper.

This effect is likely to be particularly pro-nounced when gossip is negative. As describedearlier, positive gossip can enhance the reputa-tion of its subjects, whereas negative gossiptends to destroy subjects' reputations. Hence,

those who condemn gossip from an ethicalstandpoint will be especially hard pressed tofind anything redeeming about negative gossip.

Proposition 4a: In a work setting, gos-sip will reduce the gossiper's referentpower over recipients. This effect willbe stronger for negative gossip thanfor positive gossip.

The competing argument is that gossip en-hances referent power. As gossipers sharenews, they draw recipients into their social cir-cles (e.g., Dunbar, 1996; Eder & Enke, 1991). Theserecipients, in turn, may appreciate being in-cluded. Moreover, through gossip, recipientsmight realize that the gossiper is on the inside ofa social network. This realization is apt to makerecipients more interested in knowing and be-ing liked by the gossiper. Consistent with thisnotion, impression management scholars havefound that people can enhance their image bymanaging information about others with whomthey are associated (Gardner & Martinko, 1988).

If gossip enhances referent power, this effectis apt to taper off at very high levels—that is, asthe frequency of the gossip and the pool of re-cipients increase. As Levin and Arluke have ob-served, a person who gossips too much "maybecome defined as a 'big mouth' or a 'yenta' whowill 'talk to anyone about anything,' as a personwho cannot be trusted to keep a secret or to bediscreet with 'privileged information'" (1987; 16).Moreover, when gossipers talk incessantlyabout others, they may become resented for us-ing so much of recipients' time. Thus, we offerthe following.

Proposition 4b: In a work setting, gos-sip will have a curvilinear effect onthe gossiper's referent power over re-cipients; it will enhance referentpower until it reaches a very highlevel, at which point it will detractfrom referent power.

Moderators of Linkages Between Gossip andPower

The strength of the above linkages may beinfluenced by characteristics of the gossip andby contextual factors, including organizationalculture and the relationship between gossiperand recipient.

2000 Kmland and Pelled 433

Features of the gossip. As described earlier,one particularly relevant characteristic of gos-sip is its credibility. Upon reviewing a variety ofempirical findings and conducting their ownstudy. Slater and Rouner (1998) concluded thatmessage credibility has considerable influenceon judgments of source credibility.^ Thus, gossipthat lacks credibility can lead a recipient toview the gossiper as a noncredible source. Evenif the recipient's view of the gossiper is notwidely held, he or she may assume that othersshare this view, for a common cognitive bias isthe faJse consensus effect: the tendency to over-estimate the prevalence of one's own opinionsor experiences (Kelley, 1967; Whitley, 1998). Ac-cording to Fiske and Taylor, "Researchers con-sistently find that consensus information (i.e.,the opinions or experience of others) is rela-tively underutilized in the judgment process"(1991: 93). They explain that those "who agreewith us are more likely to come to mind when weattempt to infer what others will believe" (1991:75). Recipients, therefore, may infer that the gos-siper also lacks credibility with others and willnot be believed when sharing negative or posi-tive gossip. Hence, when gossip credibility islow, recipients are less likely to view the gos-siper as someone with coercive or rewardpower.

In addition, credibility may affect the relation-ship between gossip and expert power. If recip-ients believe that a gossiper's information isinaccurate, they may begin to question or doubtany future information the gossiper relays. As aresult, that gossip will contribute less to, andmay detract from, the gossiper's expert power. Inline with this reasoning, Krackhardt (1990) hasfound that employees with more accurate infor-mation about the informal network have higherreputational power than those whose informa-tion is less accurate.

Lack of credibility also may diminish any pos-itive link, and enhance any negative link, be-tween gossip and referent power. Recipientsmay resent the gossiper who seems to relatefar-fetched or incorrect information, for they may

^ Although message and source credibility are conceptu-ally distinct, they are often closely related. Indeed, credibil-ity is a complex feature, and that complexity may make itmore challenging to measure, compared to other features ofgossip. Those who test the proposed model should keep thiscaveat in mind.

perceive that the gossiper is attempting to mis-lead them. As Zucker (1986) has suggested, indi-viduals perceived as providing accurate infor-mation are more trusted than those who shareinaccurate knowledge.

Proposifion 5; The effects of gossip oncoercive, reward, expert, and referentpower will be moderated by gossipcredibility. Any tendency for gossip foenhance the four power types will bestronger when credibility is high thanwhen it is low. Any tendency for gos-sip to reduce referent power will beweaker when credibility is high thanwhen it is low.

Like credibility, the work-relatedness of gos-sip may play a moderating role. Rewards (e.g.,high performance ratings and promotions) andpunishments in the organization (e.g., demo-tions and firings) are based largely on an em-ployee's work-related behavior. It is, in fact, il-legal to take many personal events (topics ofsocial gossip), such as marriage, a major illness,or a change of housing, into account when de-termining such rewards and punishments (Holl-witz, Goodman, & Bolte, 1995; Madison & Knud-son-Fields, 1987). Although some managers stillconsider those personal factors when allocatingresources, legislation (and the possibility ofcostly lawsuits) constrains their ability to do so.Thus, the employee who engages in work-related gossip has a greater ability to influencerewards and punishments in the workplace thandoes an employee who engages in gossip aboutother topics.

Work-related gossip is also particularly likelyto shape expert power. Fiske and Taylor (1991)have pointed out that a given context can en-courage us to attend to some information morethan other information. Being in the workplacemakes employees particularly attuned to work-related information. When the recipient is in awork context, "professional" topics such as aperson's salary, promotion, and recognition gen-erally have more relevance than do divorce,plastic surgery, or other "social" topics. Thus, agossiper who provides work-related informationabout others is especially likely to be used as aninformation source and seen as an expert in theworkplace.

In addition, the work-relatedness of gossipmay diminish any negative link between gossip

434 Academy ol Management fleview April

and referent power. Recipients are less likely toperceive the gossiper as wasting their time atthe office when the gossip is relevant to thatsetting. Hence, they will be less resentful of thegossiper when the work-relatedness of gossip ishigh.

Proposition 6: The effects of gossip oncoercive, reward, expert, and referentpower will be moderated by the work-relatedness of the gossip. Any ten-dency for gossip to enhance coercive,reward, and expert power will bestronger when work-relatedness ishigh. Any tendency for gossip to re-duce referent power will be weakerwhen work-relatedness is high.

Gossiper-recipient relationship quality. Likethe nature of the gossip, the context of that gos-sip—specifically, the quality of the relationshipbetween gossiper and recipient—may act as amoderator. Relationship quality is the degree towhich a relationship is characterized by mutualsupport, informal influence, trust, and frequentinformation exchange (Lee, 1998). Employeeswho have a habit of gossiping with each other,for example, can be characterized as having ahigh relationship quality. Much of the literatureon relationship quality pertains to supervisor-subordinate dyads or leader-member exchangetheory (Dansereau, Graen, & Haga, 1975), butone can also characterize peer relationships interms of relationship quality (Kram & Isabella,1985).

Negative gossip is less likely to enhance co-ercive power when relationship quality is high.If a recipient trusts a gossiper, that recipientmay believe the gossiper will avoid harminghim or her. Even if the gossiper is spreadingnegative news about others, the recipient maybe confident that his or her own dark secrets willnot be revealed by that gossiper.

Positive gossip, however, is more likely to en-hance reward power when relationship qualityis high. A recipient who is a close friend of agossiper may believe that gossiper will try tohelp him or her when possible. Thus, if thatgossiper is spreading positive news about oth-ers, the recipient is especially likely to think thegossiper will do the same for him or her.

Proposition 7: The effect of gossip oncoercive and reward power will be

moderated by gossiper-recipient rela-tionship quality. Any tendency tornegative gossip to enhance coercivepower will be weaker when relation-ship quality is high. Any tendency forpositive gossip to enhance rewardpower will be stronger when relation-ship quality is high.

Relationship quality also may shape gossipeffects on referent power. Gossip is more likelyto enhance referent pow^er when the quality of arelationship is high. If the gossiper and recipi-ent have a close and trusting relationship, therecipient is apt to view such gossip as appropri-ate, for informal communication is characteris-tic of high-quality relationships (Fairhurst, 1993;Lee & Jablin, 1995). Consistent with this logic is"halo effect" research (Nisbett & Wilson, 1977),which has revealed "a tendency to evaluate allcomponents of a target person in the same wayonce a general evaluation, positive or negative,is formed" (Fiske & Taylor, 1991: 256). Thus, in thecontext of a strong relationship, any positivelink between gossip and referent power will bestronger. Also, when relationship quality ishigh, recipients who frown upon gossip in gen-eral may be more forgiving of the gossiper.Hence, any negative link between gossip andreferent power will be weaker.

Proposition 8: The effect of gossip onreferent power will be moderated bygossiper-recipient relationship qual-ity. Any tendency for gossip to en-hance such power will be strongerwhen relationship quality is high. Anytendency for gossip to reduce suchpower will be weaJcer when reiafion-ship quality is high.

Organizational culture. Another moderatingcontextual factor may be organizational cuJfure:the "system of shared values (that define what isimportant) and norms that define appropriateattitudes and behaviors for organizationalmembers (how to feel and behave)" (O'Reilly &Chatman, 1996: 160). In some organizations theculture advocates considerable formal commu-nication, while discouraging informal communi-cation (Smeltzer & Leonard, 1994). If there is acultural injunction against informal communi-cation, then employees will be constrained intheir use of gossip to spread news about others.

2000 Kurland and Pelled 435

Gossip recipients may recognize these con-straints and conclude that gossipers have fewopportunities to help or harm reputations. Theeffect of gossip on reward and coercive power,therefore, will be weaker.

Also, when culture encourages formal commu-nication and discourages informal communica-tion, organizational members may not look togossip as a source of information. Evidence hasshown that individuals refrain from an informa-tion-seeking strategy if they expect the strategyto have high social costs (Miller & Jablin, 1991).In an antigossip culture, seeking informationfrom a gossiper may have such costs. Conse-quently, it may be difficult for the gossiper togain expert power via gossip.

The link between gossip and referent power,too, may be shaped by culture. An antigossipculture may reinforce a recipient's belief thatgossip is wrong or immoral. Thus, any tendencyfor gossip to reduce referent power will be stron-ger when the culture discourages such informalcommunication.^

Proposition 9: The effects of gossip oncoercive, reward, expert, and referentpower will be moderated by organiza-tional culture. Any tendency for gossipto enhance coercive, reward, and ex-pert power will be weaker when theculture discourages informal commu-nication. Any tendency for gossip toreduce referent power will be strongerwhen the culture discourages infor-mal communication.

CONCLUDING REMARKS

The proposed model contributes to both man-agement research and practice. On the aca-demic side, it is—to the authors' knowledge—the first theoretical model of workplace gossipand its consequences. Noon and Delbridge (1993)took a significant step with their thought-provoking discussion of gossip in organizationsand their call for research on the topic. Ourmodel takes their work a step further, offering a

^ It is possible that some employees will reject the valuesof the dominant culture and appreciate the individual whogoes against it (e.g., by gossiping in an antigossip culture).These employees may respect that gossiper ior taking sucha risk.

refined conceptualization and specific predic-tions about the phenomenon. On the practitionerside, the proposed framework illustrates that,contrary to the adage "small people talk aboutother people," gossip can make a person quite"large" in an organization. At the same time, themodel shows conditions under which gossipmay backfire. An understanding of such dynam-ics of gossip is likely to help organizations andtheir members capitalize on this widespreadgenre of informal communication.

REFERENCES

Atwater, L. 1995. The relationship between supervisorypower and organizational characteristics. Group andOrganization Management, 20: 460-485.

Ayim, M. 1994. Knowledge through the grapevine: Gossip asinquiry. In R. F. Goodman 8t A. Ben-Ze'ev (Eds.), Goodgossip: 85-99. Lawrence: University oi Kansas Press.

Bacharach, S, B., & Lawler, E. J. 1980, Power and politics inorgranizafions. San Francisco: Jossey-Bass.

Baird, I. E. 1977. The dynamics o/ organizafionai communi-cation. New York: Harper & Row.

Berger, C. R. 1994. Power, dominance, and social interaction.In M. L. Knapp & G. R, Miller (Eds.), Handbook of inter-personal comniunication: 450-507. Thousand Oaks, CA:Sage.

Berlo, D. K. 1960. lilie process of communication: An introduc-tion to theory and practice. New York: Holt, Rinehart, StWinston.

Boehm, L. E. 1994. The validity effect: A search for mediatingvariables. Personality and Social Psychology Bulletin,20: 285-293.

Bok, S. 1934. Secrets: On tJie ethics of concealment and rev-elation. New York: Vintage.

Burt, R. S., & Knez, M. 1996. Trust and third-party gossip, InR. M. Kramer & T. R. Tyler (Eds.), Trust in organizations:68-89. Thousand Oaks, CA: Sage.

Code, L. 1934. Gossip, or in praise oi chaos, ln R. F. Goodman& A. Ben-Ze'ev (Eds.), Good gossip: 100-106. Lawrence:University of Kansas Press.

Daniels, T., Spiker, B., & Papa, M. 1997. Perspectives on or-ganizational communicaiion (3rd ed.). Madison, WI:Brown and Benchmark.

Dansereau, F., Graen, G., & Haga, W. J. 1975. A vertical dyadapproach to leadership within formal organizations.OrgranizafionaJ Beliavior and Human Performance, 13:46-78.

Davis, J. H., & Schoorman, F. D. 1997. Toward a stewardshiptheory of management. Academy of Management Re-view, 22: 20-47.

Davis, K. 1953. A method of studying communication patternsin organizations. Personnel PsychoJogy, 6: 301-312.

436 Academy of Management Review April

Dunbar, R. 1S96. Grooming, gossip, and the evolution of lan-guage. Cambridge, MA; Harvard University Press.

Eder. D., & Enke, J. L. 1991. The structure of gossip: Opportu-nities and constraints on collective expression amongadolescents. American Sociological fleview, 56: 494-508.

Emerson, R. M. 1962. Power-dependence relations. AmericanSociological Review, 27: 31-41.

Emler, N. 1994. Grossip, reputation, and social adaptation. InR. F. Goodman & A. Ben-Ze'ev (Eds.), Good gossip: 117-138. Lawrence: University of Kansas Press.

Etzionl, A. 1961. A comparative analysis of complex organi-zations: On power, involvement, and their correlates.New York: Free Press.

Fairhurst, G. T. 1993. The leader-member exchange patternsol women leaders in industry: A discourse analysis.Communication Monographs, 60: 321-351.

Fine, G. A. 1977. Social components oi children's gossip.Journal of Communication, 27(1): 181-18S.

Fine, G. A., & Rosnow, R. L. 1978. Gossip, gossipers, gossip-ing. PersonaJit]r and Social Psychology Bulletin, 4:161-168.

Finkelstein, S. 1992. Power in top management teams:Dimensions, measurement, and validation. Academy ofManagemenl Journal 35: 505-538.

Fiske, S. T., & Taylor, S. E. 1991. Social cognifion. New York:McGraw-Hill.

French, J. R. P., & Raven, B. 1959. The bases oi social power.In D. Cartwright (Ed.), Studies in social power: 150-167.Ann Arbor: University of Michigan Institute for SocialResearch.

Gardner, W. L., & Martinko, M. J. 1988. Impression managementin organizations. Journal of Management, 14: 321-339.

Giddens, A. 1984. The constitution of society: Outline of thetheory of structuration. Cambridge: Polity Press.

Glazer, I. M, & Ras, W. A, 1994.On aggression,human rights,and hegemonic discourse: The case of a murder iorfamily honor in Israel. Sex Roles, 30: 269-288.

Goldsmith, D. J., & Baxter, L. A. 1996. Constituting relation-ships in talk: A taxonomy of speech events in social andpersonal relationships. Human Communication fle-search, 23: 87-114.

Harris, T. E. 1993. Applied organizational communication:Perspectives, principles, and pragmatics. Hillsdale, NJ:Lawrence Erlbaum Associates.

Heath, R. L. 1994. Management of corporate communication:From interpersonal contacts to external affairs. Hills-dale, NJ: Lawrence Erlbaum Associates.

Hinkin, T. R., & Schriesheim, C. A. 1989. Development andapplication of new scales to measure the French andRaven (1959) bases of social power. Journal of AppliedPsychology, 74: 561-567.

Hinkin, T. R., & Schriesheim, C. A. 1994. An examination ofsubordinate-perceived relationships between leader re-ward and punishment behavior and leader bases ofpower. Human Relations, 47: 779-800.

Hollwitz, J., Goodman, D. F., & Bolte, D. 1995. Complying with

the Americans with Disabilities Act: Assessing the costsoi reasonable accommodation. Public Personnel Man-agement, 24: 149-157.

House, R. J. 1988. Power and personality in complex organi-zations. In B. M. Staw & L. L. Cummings (Eds.), Aesearchin organizational behavior, vol. 10: 305-357. Greenwich,CT: JAI Press.

Hunt, S. D., 8E Nevin, J. R. 1974. Power in a channel of distri-bution: Sources and consequences. Journal of MarketingResearch, 11: 186-193.

Ilgen, D. R., Fisher, C. D., & Taylor, M. S. 1979. Consequencesof individual feedback on behavior in organizations.Journal of Applied Psychology, 64: 349-371.

Jaeger, M. E., Skelder, A. A., & Rosnow, R. L. 1998. Who's up onthe low down: Gossip in interpersonal relations. In B. H.Spitzberg & W. R. Cupach (Eds.), The dark side of closerelationships: 103-117. Mahwah, NJ: Lawrence ErlbaumAssociates.

Katz, D., & Kahn, R. L. 1978. The social psychology of organi-zations. New York: Wiley.

Kelley, H. H. 1967. Attribution theory in social psychology. InD. Levine (Ed.), Nebraska symposium oa motivation, vol.15: 192-240. Lincoln: University of Nebraska Press.

Kincaid, D. L. 1979. The convergence model of communica-tion. Paper 18. Honolulu: East-West Communication In-stitute.

Kipnis, D., Schmidt, S. M., & Wilkinson, L 1980. Intraorgani-zational influence tactics: Explorations in getting one'sway. /ournai of Applied Psychology, 65: 440-452.

Krackhardt, D. 1990. Assessing the political landscape:Structure, cognition, and power in organizations. Ad-ministrative Science Quarterly, 35: 342-369.

Krackhardt, D., & Hanson. J. R. 1993. Informal networks: Thecompany behind the chart. Harvard Business Review,71(4): 104-111.

Kram. K, E., & Isabella, L. A. 1985. Mentoring alternatives:The role of peer relationships in career development.Academy of Management Journal, 28: 110-132.

Lee, J. 1998. Maintenance communication in superior-subordinate relationships: An exploratory investigationof group social context and the "Pelz effect." SouthernCommunication/ournai, 63: 144-159.

Lee, J., 8t Jablin, F. M. 1995. Maintenance communication insuperior-subordinate work relationships. Human Com-munications Researcli, 22: 220-257.

Levin, J., & Arluke, A. 1987. Gossip: The inside scoop. NewYork: Plenum.

Madson, R. B., & Knudson-Fields, B. 1987. The law and em-ployee-employer relationships: The hiring process.Management Solutions, 32(2): 12-20.

Martin, J., Feldman, M. S., Hatch, M. J., & Sitkin, S. B. 1983. Theuniqueness paradox in organizational stories. Adminis-trative Science Quarterly, 28: 438-453.

Martin, J., & Siehl, C. 1983. Organizational culture and coun-terculture: An uneasy symbiosis. Organizational Dy-namics, 12(2): 52-64.

2000 Kudand and Pelled 437

McCroskey, ]. C. 1969. A survey oi experimental research onthe effects oi evidence in persuasive communication.Speech Monographs, 55: 169-176.

Miller, V. D., & Jablin, F. M. 1991. Information seeking duringorganizational entry: Influences tactics, and a model ofthe process. Academy ol Management Review, \B: 92-120.

Morrow, P. C. 1981. Work related communication, environ-mental uncertainty, and subunit effectiveness: A secondlook at the iniormation processing approach to subunitcommunication. Academy of Management Journal 24:351-858.

Mumby, D. K. 1988. Communicafion and power in organiza-tions: Discourse, ideology, and domination. Norwood, NJ:Ablex Publishing.

Nisbett, R. E., & Wilson, T. D. 1977. The halo effect: Evidencefor unconscious alteration of judgments. Journal ol Per-sonality and Social Psychology, 35: 250-256.

Noon, M., & Delbridge, R. 1993. News from behind my hand:Gossip in organizations. Organization Studies, 14: 23-36.

O'Reilly, C. A., Ill, & Chatman, I. A. 1996. Culture as socialcontrol: Corporations, cults, and commitment. In B. M.Staw & L. L. Cummings (Eds.), flesearch in organization-al behavior, vol. 13: 157-200. Greenwich, CT: JAI Press.

Osgood, C. E., Suci, G. J., & Tannenbaura, P. H. 1957. Themeasurement ol meaning. Urbana: University of IllinoisPress.

Pearce, C, Figgins, R., & Golen, S. 1984. Principles ot busi-ness communication.- Theory, appJicaHon, and technoJ-ogy. New York: Wiley.

Pieffer, J. 1992. Managing with power: Politics and influencein organizations. Boston: Harvard Business School Press.

Podolny, J. M., & Baron, J. N. 1997. Resources and relation-ships: Social networks and mobility in the workplace.American SocioJogicaJ Review, 62: 673-693.

Roethlisberger, F. J., & Dickson, W. J. 1943. Management and(he worier. Cambridge, MA: Harvard University Press.

Rogers, E. M., & Kincaid, D. L. 1981. Comrounication networks.New York: Free Press.

Rosnow, R. I., & Fine, G. A. 1976. flumor and gossip: Thesocial psychology oi hearsay. New York: Elsevier.

Ruch, W. V. 1989. JntemaHonaJ handhoolr of corporate com-munication. Jefierson, NC: McFarland.

Salancik, G. R., & Piefier, J. 1977. Who gets power—and howthey hold on to it: A strategic contingency model ofpower. Organizational Dynamics, 5(3): 2-21.

Schein, S. 1994. Used and abused: Gossip in medieval soci-ety. In R. F. Goodman & A. Ben-Ze'ev (Eds.), Gfood gossip:139-153. Lawrence: University oi Kansas Press.

Self, C. C. 1996. Credibility. In M. B. Salwen & D. W. Stacks(Eds.), An integrated approach to communication theoryand research: 421-442. Hillsdale, NJ: Lawrence ErlbaumAssociates.

Shackleton, V. 1995. Business leadership. London: Routledge.

Shannon, C. E., & Weaver, W. 1949. The mathematicaJ theoryol communication. Urbana: University of Itlinois Press.

Slater, M. D., Si Rouner, D. 199B. How message evaluation andsource attributes may influence credibility assessmentand belief change. Journalism and Mass CommunicationQuarterJy, 73: 974-991.

Slattery, K., & Tiedge, J. T. 1992. The effect oi labeling stagedvideo on the credibility of TV news stories. Journal olBroadcasting and Electronic Media, 36: 279-286.

Smeltzer, L. R., & Leonard, D. J. 1994. Managerial communi'cation.- Strategies and applications. Boston: Irwin.

Smith, B. 1996. Care and ieeding of the oifice grapevine.Management fleview, 85: S.

Stohl, C. 1995. OrganirationaJ communication: Connectod-ness in action. Thousand Oaks, CA: Sage.

Tabbutt, M. 1995. Women's talk? A social history ol "gossip"in working class neighborhoods, 1830-19B0. Brookiieid,VT: Ashgate.

Tedeschi, J. T. 1972. The social influence process. Chicago:Aldine-Atherton.

Tjosvold, D. 1995. Effects of power to reward and punish incooperative and competitive contexts. Journal ol SocialPsychology, 135: 723-736.

Tushman, M. L. 1979. Impacts oi perceived environmentalvariability on patterns of work related communication.Academy oi Management Journal 22: 482-500.

Whitley, B. E. 1998. False consensus on sexual behavioramong college women: Comparison oi iour theoreticalexplanations. Journal ol Sex Research, 35: 206-214.

Yerkovich, S. 1977. Gossiping as a way of speaking. Journalol Communication, 27(1): 192-196,

Yukl, G., & Falbe, C. M. 1991. Importance oi different powersources in downward and lateral relations. Journal oiApplied Psychology, 76: 416-423.

Zimbardo, P. G., 8r Leippe, M. R. 1991. The psychology oiattitude change and social influence. New York:McGraw-Hill.

Zucker, L. G. 1986, Production of trust: Institutional sources ofeconomic structure, 1840-1920. In B. M. Staw & L. L.Cummings (Eds.), flesearch in organizational behavior,vol. 3: 53-111. Greenwich, CT: JAI Press.

Nancy B. Kurland is an assistant professor oi management and organization in theMarshall School of Business, University of Southern California, where she teachescourses in organizational behavior, business ethics, and leadership. She received herPh.D. from the University of Pittsburgh. She researches gossip, telecommuting, genderissues, the social impact of technology, and ethics and incentives.

Academy of Management Review April

Lisa Hope Pelled is an assistant professor of management and organization in theMarshall School of Business, University of Southern California, where she teachescourses in organizational behavior, managing interpersonal relations, and multicul-tural management. She received her Ph.D. from Stanford University. Her researchareas include organizational demography, workplace emotions, communication, andconflict.