beyond internal and external: a dyadic theory of ...€¦ · the dyadic theory of relational...

BEYOND INTERNAL AND EXTERNAL: A DYADICTHEORY OF RELATIONAL ATTRIBUTIONS

MARION B. EBERLYUniversity of Washington, Tacoma

ERICA C. HOLLEYMICHAEL D. JOHNSONTERENCE R. MITCHELL

University of Washington, Seattle

Attribution theory argues that people assess the locus of causality of achievement-relevant events as either internal or external. Given the frequency of interpersonalinteractions in organizations, we posit that a third category—relational attributions—may be used. Drawing on relational perspectives, we lay the conceptual foundationand develop a dyadic theory of relational attributions, proposing their antecedentsand linking them to relationship-focused behaviors, which influence the quality ofinterpersonal links within organizations.

According to attribution theory, people havean innate tendency to make sense of their sur-roundings by acting as naive psychologists(Heider, 1958). When confronted with certainevents, people seek to determine their causes.For example, in the organizational context onemight ask, “Why did I get passed over for pro-motion?” or “Why did my boss criticize me formy work on this project?” (e.g., Martinko, Doug-las, & Harvey, 2006). Through the use of attribu-tions, people attempt to (re)establish controlover their lives and improve their ability to pre-dict future events (Kelley, 1971; Thibaut &Walker, 1975). Attribution theory suggests that inanswering “why” questions, people primarilydistinguish between internal (self) and external(outside of self) explanations, thereby determin-ing the locus of causality for an event (Allport,1979/1954; Heider, 1958; Kelley, 1967). Whetherthe cause of an event is seen as internal orexternal systematically influences people’s sub-sequent behaviors, motivations, cognitions, andaffect (Weiner, 1985). For instance, an employeewho is wondering why he did not get assigned

the lead position on a new project might attri-bute it to his lack of skills (internal attribution)or his supervisor’s lack of material support (ex-ternal attribution). The employee may be morelikely to participate in skills training if hemakes an internal attribution for a failure to bepromoted, or, alternatively, he might decide toquit or ask for a job transfer if he makes anexternal attribution (Martinko et al., 2006).

Identifying the locus of causality has been atthe core of attribution theory since its inceptionand has generated an extensive researchstream in the field of organizational behavior(see Martinko, 1995, 2004, and Martinko et al.,2006, for comprehensive reviews). But the ques-tion emerges of whether the “internal” and “ex-ternal” categories capture the entire conceptualspace of this phenomenon. In recent years thefield of organizational behavior has greatly ben-efited from a consideration of levels of analysis(e.g., dyads, teams, business units; House, Rous-seau, & Thomas-Hunt, 1995; Yammarino & Dan-sereau, 2009). In modeling the complexities in-volved in bridging levels of analysis, interestingand compelling theoretical insights may begained (e.g., Contractor, Wasserman, & Faust,2006; Felps, Mitchell, Hekman, Lee, Holtom, &Harman, 2009). Attribution theory has mainly fo-cused on the individual level of analysis andidentified the antecedents and consequences ofattributions to either the self or someone/something outside the self, neglecting any po-

The first author was a Ph.D. candidate at the University ofWashington, Seattle when this research was conducted. Wethank Mark Martinko for valuable comments and insights onan earlier version of the manuscript. We also thank associ-ate editor Gerardo Okhuysen and two anonymous reviewersfor their guidance and encouragement throughout the re-view process.

� Academy of Management Review2011, Vol. 36, No. 4, 731–753.http://dx.doi.org/10.5465/amr.2009.0371

731Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyrightholder’s express written permission. Users may print, download, or email articles for individual use only.

tential relational aspects of performance. Webelieve that examining locus of causality attri-butions at higher levels of analysis has the po-tential to provide unique insights regarding theconsequences of attributions while simultane-ously expanding the scope of attribution theory.In this article we demonstrate this potential byexamining attributions made to the dyad.

Relationships, teams, and groups are activeresearch domains and central components oforganizational life (e.g., Dutton & Ragins, 2006;Ferris et al., 2009; Ilgen, Hollenbeck, Johnson, &Jundt, 2005). Because being embedded within asocial context can make it difficult (if not impos-sible) to view the causes of many events assolely internal or external, we posit that individ-uals draw from a third set of locus attributionsthat we call “relational attributions.” Relationalattributions are those explanations made by afocal individual that locate the cause of anevent within the relationship the individual haswith another person. They are not merely com-binations of an internal and an external attribu-tion but, rather, are attributions uniquelygrounded in the interaction between two part-ners. In other words, relational attributionsare not reducible to the actions of either partneralone. For example, an employee may attributethe failure to meet a project deadline to a lack ofclear communication with her supervisor. Thisemployee does not solely blame her own abili-ties and skills for the missed deadline, nor doesshe attribute blame solely to her supervisor. In-stead, she attributes the failure to the poor in-teraction she had with her supervisor—a featureof their relationship.

By moving beyond the internal/external dis-tinction, this construct advances attribution the-ory and provides a more complete picture of theloci of causality, including their antecedentsand consequences. The introduction of rela-tional attributions and their integration withnew relational perspectives has the potential tobroaden our understanding of attribution theoryand how it predicts organizational behaviors.Figure 1 contrasts relational with internal andexternal attributions and provides specific ex-amples of all three types of attributions, with anemphasis on explanations for negative events.Figure 2 depicts the relationships between thethree locus of causality categories and their an-tecedents and outcomes. The dyadic theory ofrelational attributions—our contribution to attri-

bution theory—is depicted in the gray-shadedarea. We refer to these figures throughout thearticle.

By extending the reach of attribution theory,we also hope to push its boundaries so that itmay “emerge as a major theory of motivation”within the field of organization science (Mar-tinko et al., 2006: 129). Recently, Martinko,Harvey, and Dasborough pointed out that attri-bution processes have generally been “under-utilized in the organizational sciences, yet havetremendous potential to explain a wide range ofworkplace behaviors” (2011: 144). This omissioncan be partially explained by attribution theo-ry’s lack of focus on the relational aspects thatare endemic to organizational life. Thus, con-ceptualizing relational attributions is a signifi-cant step toward maximizing attribution theo-ry’s potential within the organizationalsciences.

Focusing on negative achievement-relatedevents in leader-follower relationships, we pro-pose that relational attributions often trigger re-lationship-focused behaviors (which we labelrelationship work), an argument not present inthe current scope of attribution theory. Specifi-cally, we explain how relational attributions,depending on their specific content, can predicttask- or person-focused voice and citizenshipbehaviors. Hence, relational attributions may becritical in the development of positive social tieswithin organizations and should inform re-search in leadership, teams, social networks,and other topics focusing on interpersonal inter-actions.

The article is structured as follows. First, weestablish how relational attributions differ frominternal and external attributions, and we pro-vide examples of common relational attribu-tions within organizational settings. Second, wedraw on relational self theory to outline the the-oretical rationale for distinguishing relationalattributions from their internal and externalcounterparts. Third, following Kelley’s covaria-tion model (1967, 1973), we describe how rela-tional attributions are formed and propose per-sonal and situational characteristics asantecedents. Fourth, building on Weiner’s (1985)original framework, we develop a theory thatidentifies the general cognitive, affective, andbehavioral consequences associated with rela-tional attributions in response to negativeevents, emphasizing those links not previously

732 OctoberAcademy of Management Review

captured by attribution theory. Fifth, we discussthe theoretical and practical implications of ourconceptualization, including ways in which re-lational attributions may advance other rela-tionship-oriented theories in organizational be-havior at the dyadic as well as higher levels ofanalysis. Finally, we close with a discussion ofthe limitations of our analysis and propose fu-ture research that might address these issues.

BACKGROUND AND DEFINITION

Attribution Theory

Attributions are the causal explanations thatindividuals use to interpret the world aroundthem and adapt to their environment, especiallywhen reacting to events viewed as important,

novel, unexpected, and negative (Martinko, Har-vey, & Douglas, 2007; Weiner, 1990). FollowingHeider’s (1958) initial work, the most influentiallines of attribution research originated from Kel-ley and Weiner. Kelley (1967, 1973) focused onhow individuals determine the cause of a be-havior or event by considering information re-garding the consensus, consistency, and distinc-tiveness of the behavior or event. Kelley’s modelexplores the dimensions people use to locate thecausality of a behavior or event, which can in-volve oneself or others. For example, if a studentreceives a failing grade on an exam (or observesthat a fellow student received a failing grade),she may ask, “Did everyone fail?” (consensus—how shared the behavior/event is), “How didI/they perform on previous exams?” (consisten-

FIGURE 1Contrasting Internal and External Attributions with Relational Attributions Within a Dyad in

Response to Negative Achievement-Related Events

Internal attribution

Self in relation to other

Relational attribution External attribution

Other person/situation

“I did not get a positiveperformance review,because . . .

“I was not chosen as theteam leader, because . . .

“I did not meet the project’sdeadline, because . . .

“My boss always monitorsme closely, because . . .

. . . I once was caught surfingthe web at work.”

. . . I did not ask foradditional help soon

enough.”

. . . I have poorcommunication skills.”

. . . I did not put in enougheffort over the past few

weeks.”

. . . my boss and I don’t havea positive relationship.”

. . . my boss and I do notcommunicate well with each

other.”

. . . my coworker and I didnot give each other frequent

enough updates.”

. . . we dislike each other andhe is looking for a reason to

fire me.”

. . . my boss is incompetent.”

. . . it was my coworker’sturn—people are selected

based on a policy of rotatingresponsibility.”

. . . I had to redo all the workmy coworker turned in.”

. . . he is a control freak.”

Self

2011 733Eberly, Holley, Johnson, and Mitchell

cy—how consistent a behavior/event is acrosssimilar contexts and times), and “How did I/theyperform on other assignments such as writtenpapers?” (distinctiveness—how unique the be-havior/event is to the particular situation). Thepattern of information gleaned could be used toinfer an internal attribution (e.g., ability) and/orexternal attribution (e.g., it was a very toughtest) for failing the exam.

In contrast, Weiner and colleagues focused onthe consequences of four different types ofcausal judgments that people make for eventsregarding their performance (Weiner et al., 1971;see also Martinko, Moss, Douglas, & Borkowski,2007). Specifically, they argued that an individ-ual’s expectations, emotions, and behaviorscould be predicted by understanding whetherthe event’s cause was believed to be (1) internalor external, (2) stable or unstable, (3) controlla-ble or uncontrollable, and (4) global or specific.Our article expands Weiner’s first dimension byadding a relational locus of causality and iden-

tifies how it can be predicted using Kelley’smodel.

Together, Kelley’s (1967, 1973) and Weiner’s(Weiner et al., 1971) models reveal that the attri-butional process is fairly complex and can becognitively taxing. Within a dyad, our focallevel of analysis, the historical interactions be-tween the two partners all serve as potentialcues in the attributional process and require therecollection of and reflection about the dynamicprocesses inherent in any relationship. Becauseof the ambiguity and complexity involved in theattributional process, individuals may makemultiple attributions at once and generate im-plicit confidence levels for each attribution. Forexample, a salesperson may believe that hedid not close a deal because he was not aggres-sive enough (internal attribution) and becausehis company did not allow him to offer addi-tional discounts (external attribution). Thissalesperson may feel very confident that theprice made a significant difference to the cus-

FIGURE 2A Dyadic Theory of Relational Attributions in Achievement-Related Situations

Attribution outcomesAttributionAttribution formation

Consensus Distinctiveness Consistency

Relational characteristics of theattributer

Relational characteristics of thesituation

Internal

External

Relational

Change relationship

Change other

Relational uncertainty andanxiety

Change self

Task focused

Person focused

e.g., relationship work suchas remedial voice and ICBs

Task

Person

Self in relation toother

e.g., relational self, relationalidentification

e.g., interdependence

Self

Other person orsituation

e.g., learn new skills, exertmore effort

e.g., switch project team, quitjob

Low

Low

Low

Low

Low

High High

High

High

High

High

High


tomer but may also be moderately confident thathis own sales strategy contributed to the failure.In addition, he may also believe that he and thecustomer had a weak relationship bond due toinfrequent communication (relational attribu-tion). The salesperson’s subsequent attitudes,motivations, and behaviors may be a function ofthe attribution he feels most confident about,and we believe this last type of explanation (aweak relationship bond) has been absent fromthe literature.

Scope of the Article

Before examining relational attributions inmore detail, we specify some key boundary con-ditions. Obviously, many events that providepeople performance feedback do not involve re-lationships (e.g., scores on a standardized test, amonthly report on the number of widgets pro-duced). These lie outside the boundaries of ouranalysis and are likely covered by the internal/external distinction. Relational attributions pri-marily apply to events that involve two people,and we focus on the dyadic level to improve thepredictability and precision of our arguments.This boundary condition follows Ferris and col-leagues (2009: 1380), who acknowledged theneed for focusing on “dyadic entities” by devel-oping a multidimensional conceptualization ofwork relationships between two actors. Typicaldaily dyadic interactions for employees involvepeers, subordinates, customers, and suppliers.One of the most salient involves the relationshipwith one’s supervisor (Ferris et al., 2009; Graen,1976), and we use it as the primary examplethroughout our discussion.

We also focus on achievement situations thatreflect negatively on performance. Many em-ployees consider performance-related events vi-tal to their future standing within an organiza-tion, and these events can consist of unexpectedand negative episodes. In line with the generalnotion that “bad is stronger than good”(Baumeister, Bratslavsky, Finkenauer, & Vohs,2001), negative events trigger more extensive at-tributional processes because they threatengoal accomplishment and motivate people tofind underlying causes so they can avoid simi-lar events in the future (Weiner, 1990), whereaspositive feedback does not generally motivatesuch changes in behavior. In addition, negativerelationships may have a greater impact on crit-

ical organizational outcomes than positive rela-tionships (Labianca & Brass, 2006). Thus, nega-tive relational achievement situations providethe most likely ground for relational attributionsto occur. Finally, we focus our attention solelyon the locus of causality dimension of the attri-bution process. For our purposes, it was impor-tant to limit ourselves to describing the con-struct of relational attributions and identifyingits nomological network without considering theeffects of interactions with other attribution di-mensions. We will return to these definitionalconstraints in the discussion section.

Relational Perspectives in Attribution Research

Some previous attribution research has beenconducted in explicitly relational contexts andis therefore important for our analysis.Sedikides, Campbell, Reeder, and Elliot (1998)found that the self-serving bias (internal attribu-tions for success, external ones for failure; Zuck-erman, 1979) was less pronounced in a relationalcontext. By including the self as part of a rela-tionship, completely externalizing negativefeedback becomes less likely. Another stream ofresearch has shown that individuals are morelikely to help another person in distress if theyattribute that distress to external causes (Ru-dolph, Roesch, Greitemeyer, & Weiner, 2004;Weiner, 1995). Also, in applying attribution the-ory to the leadership context, Green and Mitch-ell’s (1979) theory emphasizes how leader attri-butions about a subordinate’s performanceinfluence leader-member relations. Martinkoand Gardner (1987) added subordinate attribu-tions and behaviors to Green and Mitchell’smodel, acknowledging that both partners in theleader-member relationship and their interac-tions matter to the attribution process. Martinko,Moss, Douglas, and Borkowski (2007) expandedthis research to demonstrate how leaders’ andmembers’ attributional styles interactively pre-dict their relationship quality. Finally, Anderson(1991) coded participants’ attributions followingfailure and success along thirteen dimensions.Interestingly, interpersonalness, defined as theextent to which the cause of the event reflectedon the attributer’s relationships with other peo-ple, emerged as the strongest dimension. Whilewe view the relationship as a third locus ofcausality as opposed to a separate dimension,as did Anderson (1991), his research demon-


strates that individuals do use relationships tointerpret and make sense of performance-related events.

While all of these perspectives examined at-tributions within relational contexts, theydid not consider that the attribution itself maybe relational and therefore lead to conse-quences not currently predicted by internal andexternal attributions. For example, Martinko,Moss, Douglas, and Borkowski (2007) describedhow clashing attributions by subordinates andleaders over the same event may lead to a de-terioration of the relationship. Their model, how-ever, does not discuss the possibility that theleader-member relationship might prime a rela-tional mindset that could override common at-tributional biases (such as the self-serving bias)and direct the leader and/or the subordinate toascribe attributions to the relationship itself.Martinko et al. (2006), however, argued thatemerging work on leader-member relationshipsis trending toward a more dynamic understand-ing of the leader-member interaction process.This reciprocal interaction between individualsand their relationship partners is what we areattempting to capture with the relational attri-bution concept.

Defining Relational Attributions

The classic distinction between internal andexternal attributions assumes that individualscan clearly distinguish between these two cate-gories. Classic attribution research would arguethat locating the cause of an event in one’s abil-ities (“I was not promoted to a management po-sition because I am not good at strategic think-ing”) is an internal attribution. Alternately, if thecause of an event is ascribed to economic trends(“I was fired because my position was out-sourced”), it would be classified as an externalattribution.

The distinction between internal and external,however, is not always so obvious. Research hasestablished that individuals are often closelylinked to a variety of constituents (peers, cus-tomers, supervisors, etc.) and these ties haveunique consequences. For example, employeeswith many interpersonal links through their jobsare less likely to voluntarily quit (Lee, Mitchell,Sablynski, Burton, & Holtom, 2004). Individualswho are embedded in relationships may gainincreased access to career and promotion oppor-

tunities (Seibert, Kraimer, & Liden, 2001). Em-ployees who fit well with the organization andtheir coworkers (i.e., in terms of personality andvalues) are more likely to stay, perform better,and develop more favorable attitudes towardthe organization and their job (Kristof-Brown,Zimmerman, & Johnson, 2005). These studies,among others, illustrate that employees oftenhave close relationships at work that theymay not be clearly separated from and thatelicit unique motivations. Extending this think-ing to attribution theory, we posit that peopleoften make relational attributions, which we de-fine as those explanations made by a focal indi-vidual that locate the cause of an event withinthe relationship that the individual has with an-other person. Relational attributions reflect anexplanation by one individual (e.g., the subordi-nate) within the relationship and may ormay not differ from the relationship partner’s(e.g., the supervisor) explanation. Figure 1 pro-vides a sample list of possible internal, exter-nal, and relational attributions in response tonegative feedback situations commonly foundin the workplace.

Relational attributions capture features of re-lationships that can be either task focused orperson focused. Organizational science has atradition of distinguishing between task andperson orientations (e.g., initiating structure ver-sus consideration [Stogdill & Coons, 1957], taskversus relationship conflict [De Dreu & Wein-gart, 2003]). Individuals who make relationaltask attributions identify the cause of the eventwithin those relational performance elementsresulting in successful task completion, such ascoordination, exchange of information in atimely manner, and provision of constructivefeedback. Individuals who make relational per-son attributions identify the cause of the eventwithin personal issues that are not directly re-lated to job performance, such as differing val-ues, interpersonal styles, or preferences.

Relational attributions differ from internaland external attributions in that they have twopotential agents of change. In the case of inter-nal attributions, attributers have some controlover ensuring that an event happens differentlyin the future by changing the self (e.g., exertingmore effort or learning new skills). For externalattributions, attributers may have little or nocontrol over the other person or the situation.With relational attributions, the partners share


responsibility for the event and therefore maybe motivated to take steps to fix or improve therelationship. However, compared with internalattributions, one’s efforts must be recognizedand reciprocated by one’s partner in order to besuccessful. A similar idea (one that is part ofleader-member exchange [LMX] theory) arguesthat either partner in the leader-member rela-tionship may initiate the relationship develop-ment process but that the process may not ad-vance without reciprocation (Maslyn & Uhl-Bien,2001). In addition, the relationship conflict liter-ature argues that “to be successful, the repairprocess demands the efforts of both the offend-ing and offended parties, since both play a crit-ical role in maintaining the expressive order”(Ren & Gray, 2009: 107). While relationship con-flict does not necessarily lead to relational at-tributions, this perspective suggests that posi-tive developments can occur only when bothpartners and their interactions are considered(Dirks, Lewicki, & Zaheer, 2009).

THEORETICAL FOUNDATION FORRELATIONAL ATTRIBUTIONS

To provide a theoretical rationale for rela-tional attributions, we focus on the theory of therelational self (Brewer & Gardner, 1996), becauseit is broad in scope and explains the underlyingpsychological mechanisms that guide the inter-dependencies between individuals in dyadic re-lationships. The theory of the relational self sug-gests that people are attuned to the relationalaspects of their environments and developunique motivations based on seeing themselvesin relation to and interdependent with others.

Early social psychological theories empha-sized an individual’s sense of self as beingunique and differentiated from others (Brewer &Gardner, 1996). In contrast, cross-cultural re-search on the self and research on individuals’social identities have revealed that the con-struction of the self depends not only on one’sunique attributes but also on the relationshipsand groups in which one takes part (Markus &Kitayama, 1991; Tajfel, 1982). Based on this lineof research, the self literature and the identityliterature have come to recognize multiple lev-els of self-definition. Specifically, Brewer andGardner (1996) have differentiated between per-sonal, relational, and collective representationsof the self.

The relational self is derived from an individ-ual’s connections and role relationships withsignificant others. It reflects the ways in whichone thinks about oneself within a particular re-lationship (Chen, Boucher, & Tapias, 2006), andit has implications for self-definition, self-evaluation, self-regulation, and behaviors ex-pressed in relation to significant others (Ander-sen & Chen, 2002; Baldwin, 1992) and, hence, forthe attributional process. Because the relationalself is composed of ties with significant others,the interactions and events that involve it canactivate one’s relational self (Chen et al., 2006).When activated, the relational self prompts peo-ple to focus on themselves in relation to theirinterpersonal context, and it influences the in-formation that people attend to during interac-tions. In their search for causal understanding,individuals are directed by their relational selfto explore the relational components of theirinteractions and focus their attention on howthey relate to (or fit with) their partners. Accord-ingly, relational self theory suggests that indi-viduals find not only internal and external ex-planations but also relational explanations forachievement-oriented interpersonal events.Based on the above discussion, we propose thefollowing.

Proposition 1: Relational attributionsrepresent a third category of the locusof causality dimension of attributionsand are conceptually different frominternal and external attributions.

Formation of Relational Attributions

Kelley’s covariation principle (1967, 1973) ex-plains how people process information frommultiple observations in order to make causalattributions: “An effect is attributed to the one ofits possible causes with which, over time, it co-varies” (1973: 108). The principle suggests thatindividuals determine the cause of events byconsidering information related to the consen-sus, consistency, and distinctiveness of theevents. Consensus information indicates to whatextent the behavior or event is widely shared: ifthe majority of people experience an event orbehavior, then consensus is high. Consistencyinformation refers to the extent to which anevent or action is consistent across a similarcontext or time: if someone behaves similarly


during comparable events, then consistency ishigh. Finally, distinctiveness information sug-gests to what extent a given response is uniqueto a particular event or person: if an event orbehavior transcends different situations, thendistinctiveness is low. In combination, consen-sus, consistency, and distinctiveness informa-tion provide the basis on which individualsmake an internal or external attribution. For ex-ample, an employee who is told by his boss thathe did not receive a pay raise based on meritmay ascertain (1) that the supervisor has givenmerit increases to other employees (low consen-sus), (2) that he has never received a merit in-crease from this supervisor (high consistency),and (3) that he has never received a merit in-crease from other supervisors in the past (lowdistinctiveness) and is therefore likely to con-clude that he probably does not deserve a meritincrease (internal attribution).

A key question, then, is under which circum-stances do individuals make relational attribu-tions? We suggest that relational attributionsare most likely when consensus is low and bothdistinctiveness and consistency are high. Forexample, another employee observes that (1)others have received merit increases (low con-sensus), (2) she has never received one from herboss (high consistency), but (c) she has fre-quently received merit raises for her work fromother supervisors (high distinctiveness). Be-cause other supervisors have provided meritraises to her, this employee would not make asolely internal attribution (“I’m not a meritoriousperson”) so easily, and because her current su-pervisor provided merit increases to her col-leagues, neither would she make a solely exter-nal attribution (“My supervisor is a jerk andnever gives anyone a merit raise”). The combi-nation of low consensus with high distinctive-ness and consistency suggests that the eventrests on the interaction between the partners(“My supervisor and I do not have a positiverelationship”). We therefore propose the follow-ing.

Proposition 2: Within a given relation-ship, when individuals perceive a re-lationship event as being low in con-sensus but high in distinctiveness andconsistency, they are more likely tomake a relational attribution than aninternal or external attribution.

Of course, attributers’ perceptions of the con-sensus, distinctiveness, and consistency ofevents are subject to interpretation and, thus,may not mirror objective reality. Kelley statesthat the attribution process, then, is necessarily“incomplete, subject to bias, ready to proceed onincomplete evidence, and so on” (1973: 109).Moreover, other combinations of the three Kelleydimensions do not map so clearly onto internal,external, or relational attributions. In the exam-ple above, what happens when the employeeobserves that (1) others have received merit in-creases (low consensus), (2) she has receivedmerit increases before from the current boss(low consistency), and (c) she has received meritincreases from past supervisors (low distinctive-ness)? Multiple attributions could be made inthis instance. Which attribution prevails and isperceived as most probable and therefore likelydrives cognitive, affective, and behavioral reac-tions is likely a function of other factors, such asthe context or the individual’s traits and states.Therefore, we now address some personal andsituational characteristics that may influenceindividuals’ interpretative processes in thesemore ambiguous situations and, thus, may elicitrelational attributions.

Personal Antecedents toRelational Attributions

Individuals are often primed to think of them-selves as connected with others, which is whenthe relational self is activated (Baldwin, 1992;Brewer & Gardner, 1996). Past interpersonal ex-periences form cognitive maps that include im-ages of the self and others, as well as scriptsbased on expected interaction patterns—toolsthat help us navigate the social world (Baldwin,1992). The different roles employees hold at work(e.g., role of subordinate) are necessarily rela-tional since their meaning depends on the com-plementary role (e.g., role of supervisor) withinthe role relationship, and the resulting rela-tional identity is a function of both individuals’expectations and goals (Sluss & Ashforth, 2007).While individuals may have all the tools tothink “relationally,” the chronic tendency to doso likely varies across people.

For example, people whose personal self isprimarily dominant see themselves as differen-tiated from others and may therefore be morelikely to see a clear separation between them-


selves and a relationship partner. For these in-dividuals the distinction of internal versus ex-ternal attributions will likely explain theirattribution processes. In contrast, people whoserelational self is most often dominant find theirfocus shifted away from the individual and to-ward the relationship (what Markus and Ki-tayama call “information about the self in rela-tion to another person” [1991: 230]). An employeewith a dispositionally active relational self mayquickly recognize that his supervisor is treatinghim differently from his coworkers and, thus,identify low consensus in the supervisor’s be-havior.

In addition to dispositional differences in theactivation of the relational self, the level of re-lational identification individuals experience intheir supervisor-subordinate relationship mayelicit relational attributions. Sluss and Ashforthdefine relational identification as “the extent towhich one defines oneself in terms of a givenrole-relationship” (2007: 11). When the relation-ship with the supervisor has significant mean-ing, individuals may extend their self-definitionto include the role relationship, which makes itmore difficult for them to distinguish or differen-tiate between the self and those aspects of therelationship partner that are pertinent to therole relationship (Aron & McLaughlin-Volpe,2001), thus making it potentially more difficult toassign causes of events to either the self or thepartner. In this situation individuals are morelikely to consider the relationship as a wholeand evaluate each event from the perspective ofwhat it means for the relationship.

Proposition 3: Individuals with highlyactivated relational selves are morelikely to form relational attributionsthan internal or external attributions.

Proposition 4: Individuals with highlevels of relational identification inthe subordinate-supervisor role rela-tionship are more likely to form rela-tional attributions than internal or ex-ternal attributions.

Situational Antecedents toRelational Attributions

Although some individuals chronically thinkof themselves as interconnected (e.g., throughculture and upbringing), characteristics of the

situation can also activate the relational self(Markus & Wurf, 1987). An increased level ofinteraction and interdependence among em-ployees is an aspect that characterizes today’sorganizational environment (Ferris et al., 2009).Specific relational cues in the work context in-clude work tasks, performance feedback, co-worker interactions, and leadership behaviors(Johnson, Selenta, & Lord, 2006). Obviously, insome work situations interdependence is rela-tively low and relationships are not focal (e.g., aconsultant is hired to provide technical exper-tise that no one else has). Such settings will notnecessarily prime employees’ relational self-concepts, and attribution processes more likelywill be captured by the internal and externalloci of causality.

Oftentimes, however, performance tasks arestructured such that employees depend on oth-ers for their personal outcomes. Minimally, em-ployees depend on their supervisors for re-sources like information, financial support, andrewards or assignments. The level and type ofinterdependence may vary in each relationship,depending on the type of work to be completed(Rusbult & Van Lange, 2003, 2008). Employeesmay have to share information or resources inorder to complete a task (input interdepen-dence), or they may share performance out-comes with group members or their supervisors(output interdependence). They may need eachother (reciprocal interdependence, such aswhere a supervisor relies on an area specialist’sexpertise), or one party may be dependent on theother in an unfolding sequence of interactions(serial interdependence). Interdependence isparticularly salient within the employee-super-visor relationship and is often emphasized infeedback-related situations like performanceevaluations. We propose that employees andsupervisors who are highly interdependent aremore likely to identify the cause of a negativeevent as grounded within their relationships.

In support of this idea, Sedikides and col-leagues (1998) tested the hypothesis that closerelationships place limits on individuals’ self-enhancement tendencies, such as the self-serving bias (Zuckerman, 1979). Specifically,they tested the extent to which the self-servingbias exists when two people who are close toeach other collaborate on a task. In a set of twoexperiments, members of either distant or closedyads worked together on an interdependent


outcome task. Following the task, members re-ceived false feedback at the dyadic level re-garding the performance of both partners. Studyparticipants were then asked to attribute thedyad’s performance to either the self or the part-ner. In dyads where the partners did not knoweach other, they demonstrated a pattern associ-ated with the self-serving bias, assuming moreresponsibility for the dyad’s success than for itsfailure. In dyads where the partners were close,however, each attributed failure and successequally to both partners. These results suggestthat in situations of interdependence with a pairthat is close (such as supervisor and coworkers),employees may be less motivated by self-esteem maintenance and take some responsibil-ity for the outcome, even if it is negative. Thus,relational attributions may occur more oftenthan the basic premise of a self-serving biasmight suggest.

Proposition 5: Situations high in inter-dependence are more likely to resultin relational attributions than internaland external attributions.

Consequences of Relational Attributions

Past research has shown that attributions canhave predictable and wide-ranging conse-quences for an attributer’s motivation, emotions,and behaviors. Weiner’s (1985) attributional the-ory of motivation and emotion is represented asa temporal sequence in which an event or per-formance feedback initiates appraisal pro-cesses that, in turn, influence attributions. Attri-butions can impact cognitive and affectivereactions, which directly trigger behavioral re-sponses. Thus, we now turn to identifying theunique cognitive and affective reactions thatfollow relational attributions, and we explainhow relational attributions can trigger a set ofrelationship-oriented behaviors that we labelrelationship work.

General Affective and CognitiveConsequences of Relational Attributions

Relational attributions are complex becauseboth people in the relationship are agentic andcan initiate changes in the relationship at anytime. Individuals need to simultaneously con-sider their own actions, their partners’ reactions,

and potential behaviors beyond those. Contem-plating actions in response to relational attribu-tions may also involve retrospection to evaluatepast behavior, recalling not just one’s actionsbut the reactions of the other person and theinteractive effects on the relationship, raisingsuch questions as “When I tried to influence mysupervisor in the past, how did she react andwhat did I do that seemed to work/not work?”

This complexity often translates into uncer-tainty regarding the “correct response” when arelational attribution is made. Since the cause ofthe event is seen as occurring within the rela-tionship, actions taken to improve an outcomealso affect the relationship. Thus, when explor-ing which action steps to take to remedy a situ-ation, the attributer needs to consider the rela-tionship partner’s possible reactions to anysteps. The attributer is therefore likely to engagein extensive thought trials injected with uncer-tainty, since one can never be completely cer-tain about another’s reaction. For instance, if anemployee wants to be more available for con-versations with her supervisor, she must takehis reactions into account. What will he thinkabout her increased availability? Will he per-ceive her efforts as manipulative or sincere?Could such an action possibly make mattersworse?

Granted, reactions associated with internaland external attributions may also be hard topredict. In response to negative feedback, oneemployee might consider leaving her job, whileanother might decide to improve her skill setthrough advanced training. Both options containuncertainty: the one cannot be sure she will finda job elsewhere, and the other cannot be surethat training will be useful. Thus, all three lociof attributions can generate a certain level ofuncertainty. However, we propose here that re-lational attributions lead to a specific type ofuncertainty: relational uncertainty.

Based on uncertainty reduction theory (Berger& Calabrese, 1975), relationship communicationresearchers distinguish between three types ofrelational uncertainty: uncertainty about theself, uncertainty about the partner, and uncer-tainty about the relationship itself. Self-uncer-tainty refers to doubts about being involved in arelationship and the ability to execute desiredactions. Partner uncertainty emerges from aninability to predict the partner’s attitudes, val-ues, and behaviors. Finally, relationship uncer-


tainty focuses on the dyad as a whole and en-compasses the ambiguity people experienceregarding the status of the relationship andtheir perceptions of it (Knobloch & Knobloch-Fedders, 2010; Knobloch & Solomon, 1999).

With relational attributions, individuals mustcope with all three levels of uncertainty andinclude them in their thought trials. An em-ployee may attribute failure to meet a deadlineto a lack of communication with his boss. Toimprove the communication pattern, he mightconsider sending her an email every morninginforming her about the status of the project.Rising doubts and questions likely span allthree uncertainty levels. First, the employeemight have doubts about finding the time tosend the emails (self-uncertainty). Second, hemay be unclear about whether his boss will findthe time to read the emails every morning (part-ner uncertainty). Finally, he may wonderwhether such an action will initiate an ex-change where his supervisor will provide regu-lar updates and feedback and, ultimately, im-prove their communication (relationshipuncertainty).

In addition to perceptions of uncertainty, rela-tional attributions in response to negativeevents are also likely to induce feelings of anx-iety, for two reasons. First, they threaten thestability of the relationship. Uncertainty aboutthe nature of an interaction with a significantother can threaten assumptions about people’sability to predict and control their own lives(Thibaut & Walker, 1975), as well as their needfor certainty in their relationships and their en-vironment (Van den Bos & Lind, 2002). This threatto one’s values and needs is an aversive expe-rience, which can often induce stress and anxi-ety (Lazarus, 1991) and, hence, can threaten theattributer’s psychological well-being (e.g.,Wright & Bonett, 2007). Uncertainty is a well-known workplace stressor and a common causeof anxiety (Garst, Frese, & Molenaar, 2000;O’Driscoll & Beehr, 1994), and relational uncer-tainty in particular has been shown to be re-lated to negative emotions such as sadness andfear (Knobloch & Solomon, 2002).

Second, the threat of a malfunctioning or de-teriorating relationship violates humans’ innateneed for belongingness. Baumeister and Leary(1995: 497) have provided compelling theoreticaland empirical evidence that humans have a“pervasive drive” to form and maintain interper-

sonal bonds and experience emotional distressand anxiety at the prospect of a threatened re-lationship (Leary & Baumeister, 2000). Making arelational attribution in response to a negativeevent makes salient the possibility of not beingable to maintain the relationship, and since therelationship with one’s supervisor is not easilysubstitutable (people may replace old friendswith new friends but may not easily choose adifferent supervisor) and critical to one’s goalachievement (e.g., with regards to merit in-creases and promotions), the relational attribu-tion likely leads to feelings of anxiety over howto improve the relationship and continue to ful-fill one’s needs and goals. As Baumeister andLeary note, “People feel anxious at the prospectof losing important relationships” (1995: 506).Based on this discussion, we propose the follow-ing.

Proposition 6: In contemplating possi-ble actions in response to making re-lational attributions, attributers expe-rience (a) relational uncertainty and(b) anxiety.

Relationship Work As a BehavioralConsequence of Relational Attributions

In early attribution research Weiner (1985)posited that attributions play a significant rolein shaping an individual’s expectation of suc-cess and, therefore, the individual’s desire toexpend effort on goal-directed activities. Most ofWeiner’s predictions were targeted at the direc-tion or amount, not the content, of the effort.Relational attributions offer a unique opportu-nity not only to predict motivation levels butalso to identify the specific behaviors relationalattributers are likely to engage in. Generally,we propose that individuals are more likely toseek to proactively repair the relationshipthrough relationship work when they make re-lational attributions than when they make inter-nal or external attributions. Relationship work isfocused on addressing the true underlyingcauses of relationship events in order to repairor strengthen the relational processes, as op-posed to merely changing the façade of the re-lationship through superficial and less endur-ing strategies as impression management(Rosenfeld, Giacalone, & Riordan, 2001). Becauseof their interdependent nature and basic need


for belonging, people have a basic interest increating and maintaining well-balanced rela-tionships in every aspect of their lives, includingthe workplace. The anxiety triggered by rela-tional attributions motivates actions targeted atcultivating and avoiding breaks in existing re-lationships, as well as at regaining and restor-ing relational value (Baumeister & Leary, 1995;Leary & Baumeister, 2000). Relational attribu-tions direct employees’ attention toward theirrelationships, making a commitment to relation-ship improvement more likely and improvingthe odds that employees will initiate the im-provement process.

Just as internal attributions in a dyad triggerchanges in the self (e.g., expending more effort)while external attributions trigger attempts tochange the other or the situation (e.g., asking formore resources for a given project), relationalattributions are likely to motivate changing therelationship in order to achieve desired out-comes. In making relational attributions, attrib-uters will pay attention to the features of therelationship that could be enhanced (Brewer &Gardner, 1996; Markus & Kitayama, 1991). Indi-viduals who make relational attributions aremore likely to attend to, remember, and act oninformation that is relevant to their relation-ships (Gabriel & Gardner, 1999) and to expendeffort toward developing a high-quality or work-able relationship (Maslyn & Uhl-Bien, 2001). Bymaking relational attributions, individuals rec-ognize the need to fix something on a relationallevel, and, as a consequence, multiple facets ofthe relationship can be improved. Relational at-tributions are unique in that they are associatedwith relationship schema, and, hence, attribut-ers are made more aware of potential behaviorsassociated with improving or maintaining them(Brewer & Gardner, 1996).

Proposition 7: Relational attributionsare more likely to result in relation-ship work than are internal or exter-nal attributions.

Relationship work may take two primary be-havioral forms, and its exact nature depends onwhether attributers perceive an event as arisingfrom relational task or relational person con-cerns. First, individuals may approach their re-lationship partners, discuss their relevant taskor person problems openly, and hope for thedevelopment of a mutually agreed-upon solu-

tion. This response follows Goffman’s (1967) re-lationship restoration process, which opens witha challenge phase where “the offended partycalls attention to the offender’s misconduct”(Ren & Gray, 2009: 108–109). When an employeemakes an attribution to the relationship with hisor her supervisor, the employee cannot be surethat the supervisor will make the same attribu-tion and may therefore choose to explicitly voicehis or her concerns to the supervisor to identifywhether the supervisor has made a similar at-tribution. The content of this initial conversationis likely going to reflect the specific task or per-son explanations inherent in the relational at-tribution.

The procedural choice literature suggests thatthe initiation of an informal discussion with theoffending party—called “remedial voice”—is acommon first step toward conflict management(Peirce, Pruitt, & Czaja, 1993). While remedialvoice is more commonly studied in response tointerpersonal mistreatment, such as verbal ag-gression and humiliation, its change-orientedand relational nature makes it applicable to ourtheory and plausible to assume that a conversa-tion with the relationship partner to discuss andremedy the situation is a step commonly taken.

Proposition 7a: Relational task and re-lational person attributions may leadto task- or person-focused remedialvoice behaviors, such as talking aboutthe issue with the supervisor.

Second, individuals may choose not to overtlyexpress their attributions but may instead at-tempt to change the relationship by changingaspects of their own behavior, with the hope thatthe relationship partner will reciprocate. Since arelational attribution necessarily implicates as-pects of the self, such a strategy without involv-ing voice may be successful. While changingone’s behavior is a common response to internalattributions, this same behavior in response to arelational attribution is intended to initiate areciprocal exchange. This is similar to the rela-tionship development process described by LMXtheory where leaders provide followers certainlevels of latitude, await follower reactions, andsubsequently restrict or expand the latitude.Through ongoing exchanges the individuals inthe relationship “test” one another; if the part-ner’s response is positive, exchanges continueand the relationship develops into mutual trust,


affection, and loyalty (Dienesch & Liden, 1986;Liden, Sparrowe, & Wayne, 1997). In the aboveemail example, the employee might decide tosend a daily email to his supervisor summariz-ing the project’s progress. If the supervisor readsthe emails and responds with feedback andtime-sensitive information important to the proj-ect, the exchange will likely continue and ulti-mately improve coordination and communica-tion between the partners.

A specific type of behavior that may occur toinitiate the relationship repair process is inter-personal citizenship behaviors (ICBs)—coopera-tive behaviors targeted at close others, such ascoworkers and supervisors. ICBs take the form ofproviding assistance beyond one’s job require-ments, something that directly or indirectly en-hances individual, team, and organizationalperformance (Bowler & Brass, 2006). Such behav-iors notably occur in the context of interpersonalrelationships (Settoon & Mossholder, 2002) andmay be independent of organizational citizen-ship behaviors directed at the organization(Bowler & Brass, 2006). ICBs have been shown tohave a stronger effect on such key organiza-tional outcomes as performance, morale, andresource availability than other forms of citizen-ship behaviors (Podsakoff, MacKenzie, Paine, &Bachrach, 2000). Because of their positive contri-butions to coordination and the establishment ofa favorable work climate, it seems likely thatICBs also lead to the development of high-quality interpersonal relationships in the work-place.

Task-focused ICBs involve “the resolution ofwork-related problems of a less personal na-ture,” such as supplying information, offeringadvice, helping out with the completion of aconcrete task, assuming responsibility for solv-ing problems, and making improvement sugges-tions (Settoon & Mossholder, 2002: 256). Person-focused ICBs deal with problems associatedwith one’s affiliations and social support sys-tems at work and include listening, being avail-able for emotional support, counseling, anddemonstrating concern and respect for others.We suggest that individuals will engage in thecitizenship behaviors that match the content ofthe relational attribution.

Proposition 7b: Relational task and re-lational person attributions may leadto changes to the self (e.g., task- and

person-focused ICBs) in an effort to ini-tiate a reciprocal exchange leading toa mutual relationship developmentprocess.

Relational attributions may not only lead tothe types of positive and sincere relationshipimprovement efforts described here but mayalso trigger more negative behaviors, such aswithdrawal and counterproductive work behav-iors, for example, when the attributer believesnothing can be done to remedy the relationship.We have chosen to focus on relationship workbecause it is an outcome not previously cap-tured by attribution theory and unique to rela-tional attributions.

IMPLICATIONS OFRELATIONAL ATTRIBUTIONS

General Implications

Today’s organizations encompass increas-ingly interconnected networks built on numer-ous relationships. Management researchers ac-knowledge the increased complexity of work lifeby modeling how relationships develop, theways information flows through networks, howinfluence is a function of network position andties, and how relationships may best be concep-tualized (Ferris et al., 2009; Graen & Uhl-Bien,1995; Sparrowe & Liden, 1997, 2005; Young & Per-rewé, 2000). Attribution theory to date is missinga systematic relational perspective. While someattempts have been made to consider the con-sequences of attributions in relationships, theyhave relied on the traditional internal and ex-ternal distinction and, thus, have not been ableto fully explain how attributions inform relation-ship dynamics in organizations. We hope thisarticle pushes the boundaries of attribution the-ory by modeling how relational attributions canpredict behaviors that are essential in forminghigh-quality relationships.

Adopting relational attributions as a third lo-cus of causality has numerous theoretical impli-cations for organizational research. By drawingon existing relational perspectives, we have es-tablished a theoretical rationale for the exis-tence of relational attributions by differentiat-ing them from internal and external attributions.As a result, this article can provide a more com-plete picture of the attribution process at work,as well as afford a better understanding of the


complexity with which attributions are made. Byrecognizing that individuals are motivated bytheir relationships with others and that theserelationships influence attributional processes,we hope to broaden the applicability of attribu-tion theory to current and future research ques-tions, as well as enhance its predictive validity.In predicting relationship work such as voiceand ICBs, relational attributions point to orga-nizationally relevant outcomes that have not yetbeen considered in attribution theory. Moreover,they provide additional explanations for rela-tionship-oriented behaviors in organizationsand may aid in identifying specific conditionsunder which such behaviors flourish.

These contributions suggest an importantmanagerial implication. For managers to be mo-tivated to enhance the quality of their relation-ships with subordinates, they must first ac-knowledge that relational performance-impeding elements exist and that theseelements may be proactively addressed throughrelationship work. The proposed existence of re-lational attributions suggests that performance-related feedback should focus on the relation-ships managers develop with their employeesand the other relationships in which employeesare embedded. When managers make relationalattributions and would like their employees tomake the same attributions, they may benefitfrom emphasizing employees’ embeddednesswithin relationships (thus activating their rela-tional selves) or from guiding their employeesthrough Kelley’s information dimensions. If bothparties decide that the relationship needs to beimproved, they can jointly engage in relation-ship work. Also, to ensure that employees feelcomfortable using remedial voice to addresstheir own relational attributions, managers mayexplicitly state and continuously emphasizetheir open-door policy and willingness to getinput and feedback on any issue.

The primary contribution of this article is theinclusion of relational aspects in attribution the-ory. We believe that this addition holds impor-tant implications for the broader relational liter-ature, which can draw on attribution theory andelements such as the loci of causality and Kel-ley’s information categories to develop uniqueexplanations for relationship work and develop-ment. For example, research on ICBs could ben-efit from an attributional perspective. Relationalattributions may be an antecedent of ICBs, sug-

gesting that employees exhibit helping behav-iors even in the face of negative events. ICBs areoften conceptualized within a social exchangeperspective, where employees perform helpingbehaviors as a way of reciprocating positivetreatment they have received (Bowler & Brass,2006). Our analysis suggests that ICBs may alsobe performed as a way to initiate relationshipimprovement in a negative situation, ratherthan to reciprocate what has been given. Addingrelational attributions to the nomological net-work of ICBs could shed additional light on howorganizations develop and sustain a positivesocial support climate.

In general, relational attributions have the po-tential to explain why some individuals are mo-tivated to improve a bad or less than ideal rela-tionship. Often, research focuses on howindividuals can improve relationships (e.g.,Tomlinson & Mayer, 2009); attribution theory isnow formulated to describe when people aremotivated to do so. Because relationships arecritical to organizational functioning, under-standing when individuals are willing to en-gage in relationship work is an important impli-cation of our analysis.

Implications of Relational Attributions forResearch Involving Dyads

To illustrate the wide-ranging potential of ourtheory to contribute to other literature, we dis-cuss the theoretical implications for three dis-tinct topics in the management field that weidentify as part of the broader relational litera-ture: leadership, teams, and social networks.

Implications for leadership research. Increas-ingly, leadership researchers recognize the in-teractive nature of the leader-follower relation-ship by acknowledging that followers activelyshape what constitutes leadership (Avolio, 2007;Avolio, Walumbwa, & Weber, 2009). LMX theorywas among the first to acknowledge the rela-tional basis of leadership, positing that leader-ship and influence emerge not merely as a re-sult of certain leader or follower characteristicsand behaviors but as a result of these individu-als’ interactions within their relationship (Dan-sereau, 1995; Graen & Uhl-Bien, 1995). Followers’relationships with their leaders are potent pre-dictors for many important outcomes (Gerstner& Day, 1997; Graen & Uhl-Bien, 1995). It is notsurprising that followers are highly attuned to


their relationships with leaders, since they maygain acceptance and access such benefits asplum assignments, pay raises, and promotionsif the relationship is characterized by mutualtrust, loyalty, respect, and extensive communi-cation (Graen & Cashman, 1975; Graen & Scan-dura, 1987). Followers’ attention is often directednot only toward their leaders but also towardtheir relationships with their leaders, underscor-ing our argument that individuals often assignthe cause of an event to the relationship itself.

To date, however, LMX research has mainlyexamined the dynamics of the leader-followerrelationship in its initial stages, where interac-tions are particularly important in determiningwhether the follower will be a part of the in-group or outgroup. It has long been assumedthat LMX quality is quite stable over time oncethe relationship is established, but this assump-tion has not been supported by empirical re-search (Robert C. Liden, personal communica-tion). For those studies where LMX quality wasmeasured multiple times, stability in LMX rela-tions was only moderate. For example, Bauerand Green (1996) measured LMX quality twice,once twelve weeks after follower entry and thenagain thirty-four weeks after entry, finding thatthe ratings correlated .54. Similarly, with a timespan of one year, Epitropaki and Martin (2005)found a .64 correlation between LMX quality rat-ings over time. These results suggest what re-searchers in the communication literature havefound—that relationships can grow and deteri-orate over time (e.g., Hinde, 1997)—and the lead-ership literature would benefit greatly from un-derstanding these fluctuations.

We believe that the concept of relational at-tributions may aid in developing new theoryregarding relationship change. As indicatedabove, relational attributions help to identifythe circumstances in which individuals wouldbe motivated to improve an existing relation-ship and, because they trigger relationshipwork, may help explain how LMX quality maybe sustained and how followers may potentiallymove from the outgroup to the ingroup (or viceversa). For example, relational attributions maycontribute to the development of high-qualityLMX relationships through their influence onmember performance. Research has shown thatmembers’ performance becomes a key predictorof relationship quality once leaders and mem-bers have had time to interact (Nahrgang,

Morgeson, & Ilies, 2009). When employees makerelational attributions, they are more likely tofocus on the relational aspects of performance,possibly engaging in ICBs, which, in turn, influ-ence performance ratings (Whiting, Podsakoff, &Pierce, 2008), enhancing the relationship quality.Thus, relational attributions may provide theo-retical leverage for exploring how LMX qualitycan improve as a result of relational attribu-tions.

In addition, understanding followers’ attribu-tions for success and failure may provide lead-ers with valuable insights about follower moti-vation and actions. While relationship workmay be initiated solely by followers, the successof such improvement attempts is partially deter-mined by leaders’ willingness and ability to re-ciprocate the behavior and to initiate a dyadicproblem-solving process by communicatingopenly and engaging in conflict resolution (Mas-lyn & Uhl-Bien, 2001). Leaders must be involvedin any attempt at relationship repair. Relationalattributions may explain the circumstances un-der which leaders are willing to engage in rela-tionship work with followers (such as when theymake relational attributions themselves). Iden-tifying the antecedents of relational attributionsis important to the leadership literature becauseit offers theoretical mechanisms through whichleaders can influence followers’ attributionalprocesses (e.g., during a performance evalua-tion session).

Implications for team research. Although ourdiscussion of relational attributions focused ondyads, this phenomenon may be relevant forresearch on teams as well. Most important, re-lational attributions may be invoked to explainintrateam causes of poor performance. Researchexamining the attributions made by teams inresponse to feedback found that teams gener-ally share credit for good performance but sin-gle out individuals as the cause of poor perfor-mance (Dorfman & Stephan, 1984; Naquin &Tynan, 2003). Research on the latter case oftenfocuses on reactions to the poorest-performingmember of the team (Jackson & LePine, 2003;LePine & Van Dyne, 2001; Taggar & Neubert,2004).

We suggest that this research could be en-hanced by considering the interrelations be-tween team members in seeking the cause of ateam’s poor performance, rather than viewingindividuals as atomistic members of the team.


Much literature has examined the effects ofteam processes as the causes of team perfor-mance (for a review see Ilgen et al., 2005), someof which directly implicated relationships be-tween team members (e.g., relationship conflict[Jehn, 1995]; psychological safety, “a shared be-lief that the team is safe for interpersonal risktaking” [Edmondson, 1999: 354]). However, theseteam processes have mostly been examined atthe team level (e.g., examining the overall levelof conflict within a team), neglecting to acknowl-edge that problems within teams may arise fromspecific dyadic relationships existing in theteam. Because of the interdependence within ateam, a dyad-level conflict may “poison” allteam interactions and may ultimately result inpoorer team performance. For example, a teamconsisting of diverse experts may fail to meet itsperformance goals because there is animositybetween the team leader and one team member,who therefore fail to share critical informationwith each other. Such a coordination loss be-tween the two individuals results in increasedcoordination demands for the others and, ulti-mately, an overall decrease in performance.

Therefore, instead of examining the team as awhole or focusing on individual team membersas the causes of poor team performance, ouranalysis recommends examining whether teammembers make an attribution to a specific rela-tionship within their team, which may generaterelationship work from the team members in-volved in the dyad and also from the team mem-bers outside the dyad (e.g., voicing concerns to-ward the involved team members or attemptingto mediate the conflict). Thus, relational attribu-tions may provide a theoretical mechanism toexplain when team members are motivated tofocus on relational performance elements likecoordination with other team members and mayallow team researchers to identify the condi-tions under which constructive team process im-provement takes place.

Implications for social network research. So-cial network theory suggests that individuals’success at work depends on a unique set offormal and informal relationships with otherconstituents (beyond the relationships with theirboss or team members) of the organization (e.g.,Brass, 1985; Granovetter, 1973). While strong tiescharacterized by emotional closeness and fre-quent interaction and reciprocation provide so-cial support and better access to influence (Fer-

ris et al., 2009), weak ties ensure access tounique, diverse, and nonredundant informationand can therefore increase domain-relevantknowledge and creativity (Granovetter, 1973;Perry-Smith, 2006).

Because some ties and nodes are more criticalthan others in determining a network’s success,people within a given network may assignblame to a specific relationship (e.g., a weak tiebetween two nodes). For example, networkmembers may argue that they failed to obtain anew client because the network member with aunique tie to the client’s industry (i.e., the bound-ary spanner) did not obtain timely informationowing to deteriorating communication betweenthe member and the contact. While network the-ory and analysis generally focus on an overallset of relationships within either complete net-works or egocentric networks, attributions forevents may actually be made to very specificties, which may then shed light on specific op-portunities for social network improvements. Inthe above example, understanding the precisereason for losing a potential new client can mo-tivate relationship work between the boundaryspanner and his or her contact but also possiblybetween other boundary spanners and theirconnections in an effort to prevent potential re-lational troubles.

The above sections have provided details onhow dyadic relational attributions are essentialfor understanding important organizational be-haviors. Without providing the same level ofdetail, we would suggest that topics such asmentoring, socialization, and managing conflictinvolve relationship activity and may thus alsobenefit from the new relational perspectivegained within attribution theory. A better under-standing of how relational attributional pro-cesses impact affect, cognition, and behaviorswill be helpful to many aspects of our discipline.

Implications for Levels of AnalysisBeyond the Dyad: Reflections onOne-to-Many Relationships

So far, we have focused only on relationshipsbetween two individuals. But, of course, rela-tionships also exist at other levels of analysis,and to begin to illustrate how our theory may beapplied to these other levels, we would like tohighlight the implications of our theory for rela-


tionships where one partner is not an individ-ual.

While individuals often derive significantmeaning from the relationships they have withother individuals (e.g., supervisor, coworkers),their relationships to higher-level entities, suchas the groups and communities they are a partof, may also critically influence their cognitions,affect, behaviors, and attributions (Brewer &Gardner, 1996). Moreover, individuals may an-thropomorphize these units at higher levels(Sluss & Ashforth, 2007); by viewing the organi-zations they belong to as unified entities, indi-viduals may be even more likely to view theirinteractions with them in terms of relationships.For example, person-group fit research wouldsuggest that employees may attribute problemswithin a group to their own lack of fit with thegroup’s culture or value system (Kristof, 1996).Rather than considering the group’s values andbeliefs as “wrong” (and, thus, invoking an exter-nal attribution), individuals may simply per-ceive that their values and the group’s valuesare inconsistent. As such, the attribution for lackof fit would be to the relationship between theindividual and the group, rather than simply tothe individual (internal) or the group (external).Although individuals may seek an opportunityto exit the group as a consequence of this rela-tional attribution, they may also pursue rela-tionship work, if they perceive that their effortsmay bring about conciliation between their andthe group’s values and beliefs. Moreover, thegroup may make the same relational attributionand decide to voice concerns toward the groupmember in an effort to initiate a resolution pro-cess.

Individuals’ relationships to collectives mayalso be interpreted through the lens of socialnetwork theory. For example, a person’s positionwithin a given network can be meaningfullydescribed by his or her network centrality,broadly defined as “individuals’ degree of ac-cess to others within emergent intraorganiza-tional networks” (Settoon & Mossholder, 2002).Central individuals gain more access to infor-mation and control resources like informationmore easily (Brass & Burkhardt, 1993). They aremore creative (Perry-Smith, 2006), perceive theirjob as similar to objective reality (Dean & Brass,1985), are perceived as having higher status(Ibarra, 1992), and receive more influence andpromotions (Brass, 1985). Thus, position within a

social network can be seen as a contributingfactor to career success or failure. For instance,employees may blame being passed over for apromotion on a lack of connections with keyconstituents in the organization or on low net-work centrality, both of which reduce access tovaluable information. The attribution refers tothe employee’s position within a social networkand his or her embeddedness within a set ofrelationships, and we would therefore define itas a relational attribution.

The relational attribution here can again helpexplain when individuals are motivated tochange their network position. A relational at-tribution with regard to network position couldmotivate employees to actively engage in net-working behaviors to build a network suitablefor goal achievement. Social network theorymay draw on this extended attribution theory inorder to identify the circumstances under whichindividuals are motivated to improve their so-cial capital. Explicitly engaging in a conversa-tion about social capital with employees givesleaders the chance to provide feedback aboutrelational aspects of performance and to high-light potential performance improvement oppor-tunities. Doing so provides the feedback-rich en-vironment that is important for employeedevelopment (Ilgen, Fisher, & Taylor, 1979).

These descriptions suggest that relational at-tributions may contribute to research examiningrelationships at higher levels of analysis. How-ever, our suggestions remain largely specula-tive in nature, and we would like to offer somecaveats regarding our theory’s extension tohigher levels of analysis. Our theory highlightsthe often reciprocal nature of relationship workwhere—to be successful— both relationshippartners need to contribute. Considering agroup or network as a relationship partnerposes the question of to what extent collectivesare actually able to engage in relationshipwork. As suggested by groups researchers, col-lectives of individuals can have personalitiesand feelings and exhibit behaviors toward oth-ers (including group members) through the es-tablishment of norms, routines, and habits (e.g.,Cole, Walter, & Bruch, 2008; George, 1990; Hof-mann & Jones, 2005; Pearsall & Ellis, 2011). Forexample, in their discussion of the emergence ofcollective constructs, Morgeson and Hofmannposit that “as interaction occurs within largergroups of individuals, a structure of collective


action emerges that transcends the individualswho constitute the collective” (1999: 252). How-ever, the processes with which groups actmay not be isomorphic to individuals’ behav-iors. Similarly, the mechanisms with which re-lational attributions are formed and shapegroup dynamics may be different from the onesput forth in this article. For example, Kelley’sdimensions of consistency, distinctiveness, andconsensus could be interpreted differently athigher levels of analysis when consideringgroup-level outcomes (e.g., consensus may referto the extent to which other groups have expe-rienced a similar event). An extension of ourdyadic theory to higher levels of analysis shouldtherefore only proceed with caution and shouldtake into consideration our field’s rich knowl-edge of multilevel theorizing (e.g., Mathieu &Chen, 2011; Morgeson & Hofmann, 1999).

LIMITATIONS AND FUTURE RESEARCH

We introduced boundary conditions at the be-ginning of our theoretical analysis to restrict ourfocus, and these constraints require some com-ments. First, we recognize that people receiveboth positive and negative feedback. While thelatter is more likely to result in attributionalanalysis because of the desire to avoid similarsituations in the future, it is also true that peoplemay analyze positive events, especially if theywere unexpected. Positive goal discrepanciescan have motivating forces (e.g., Ilies & Judge,2005), but the mechanisms by which they oper-ate may be different. Although relational attri-butions for positive events are not as likely toimpact an attributer’s behaviors as strongly asrelational attributions for negative events, wesuspect that unique relationship-focused behav-iors may still emerge. For example, if an em-ployee receives positive feedback from his su-pervisor, engages in the attribution process, andmakes a relational attribution, he may subse-quently engage in relationship work to maintainthe relationship with his supervisor. Realizingthe benefits of positive relationships as a resultof the relational attribution, he may also engagein relationship work targeted at other importantrelationships, such as his relationship with hissupervisor’s manager. Moreover, when a rela-tional attribution occurs in response to a promo-tion, the employee may decide to “pay-it-forward” and engage in relationship work with

other employees, realizing how critical it is totheir subsequent success. These ideas are spec-ulative, and we encourage future work to exam-ine relational attributions in response to bothpositive and negative events.

Second, we focused on one’s supervisor as thebearer of bad news. However, one can receiveperformance-related feedback from peers, sub-ordinates, and other organizational members.We suspect that the power of the other partyinfluences the attributional process, but we areunsure exactly how that influence affects rela-tional attributions.

Third, we focused on the attributional processwhere feedback involves interpersonal rela-tions. Many people work in contexts or on taskswith machinery or technology where the taskitself provides feedback. Other people have sol-itary jobs; many people now work at home awayfrom coworkers. Certainly, relational attribu-tions are less likely to occur in these contextswith few interpersonal interactions.

Fourth, we primarily discussed situationswhere individuals make only one attribution.People often have multiple attributions for per-formance-related events, however, and feelmore or less confident about any given one.Thus, the overall frequency and strength of re-lational attributions compared to and/or com-bined with internal and external attributionsneed to be examined further.

Fifth, our theory proposes a relationship be-tween relational attributions and relationshipwork, without going on to explain the circum-stances under which relationship work is moreor less likely to successfully improve the rela-tionship. For example, relationship work may beless successful if it requires employees to pres-ent themselves in unusual, new, and unfamiliarways. Under such circumstances, self-presenta-tion requires more self-regulatory effort, whichmay interfere with subsequent and continuousrelationship efforts (Vohs, Baumeister, & Cia-rocco, 2005). Alternatively, relationship worksuccess may depend on individuals’ level of po-litical skill—the ability to “adjust their behaviorto different and changing situational demandsin a manner that appears to be sincere, inspiressupport and trust, and effectively influences andcontrols the responses of others” (Ferris et al.,2009: 291–292). If relationship work is not per-ceived as sincere because of low political skill,


employees’ efforts may backfire and furtherthreaten their relationship.

Finally, we chose to focus solely on the locusof causality dimension, without systematicallyintegrating relational attributions with the othercausality dimensions. Weiner and colleagues(1971) drew on the various attributional dimen-sions to model how the consequences of an at-tribution may only be understood fully whenexamining different dimensions in combination.We chose to focus on the general consequencesof relational attributions but acknowledge po-tential interactions with perceptions of control-lability, stability, and globality, as well as withthe other loci of causality (internal and externalattributions).

We see the issues mentioned above as possi-ble avenues for future research. However, webelieve that because of the newness of the rela-tional attributions construct, empirical valida-tion research is initially necessary to both mea-sure relational attributions and test thepropositions put forth in this article. The great-est challenge is to demonstrate how peoplefreely and spontaneously make relational attri-butions without being prompted by guidedquestions or scaled items. We would suggestthat empirical research begin with qualitativestudy designs where people are asked to makeattributions for situations in which they receivenegative feedback from a supervisor/leader. Fol-low-up studies could more precisely test propo-sitions about the causes of relational attribu-tions by manipulating aspects of the event (e.g.,consensus, consistency, distinctiveness) in sce-narios or in an experimental context, as well aspersonal attributes (e.g., relational self) and con-textual variables (e.g., interdependence). Attri-butions about the locus of causality (internal,external, relational) could be assessed by ratingopen-ended responses to questions aboutcauses, or by using traditional attributionalmeasurement techniques such as scales (Elig &Frieze, 1979; Kent & Martinko, 1995). Researcherscould manipulate the attribution or provide at-tributions in scenarios to measure study partic-ipants’ attitudinal and behavioral responses.

CONCLUSION

In a recent review of the applications of attri-bution theory to industrial and organizationalpsychology, Martinko and colleagues concluded

that the majority of recent attribution researchattempts to apply the traditional attribution par-adigm to new phenomena “without a concurrenteffort to enhance or make modifications to attri-bution theory per se” (2006: 173). With this articlewe have attempted to break from that pattern byexamining attribution theory from a relationalperspective. Following the trend in manage-ment literature examining the effects of employ-ees’ relationships at work (e.g., Chiaburu & Har-rison, 2008; Johns, 2006), we extend theconceptual domain of attributions by introduc-ing a new category, relational attributions, tothe locus of causality dimension. With the addi-tion of a third category, attribution theory mayaccount for more variance in important em-ployee outcomes and predict previously ne-glected but organizationally critical outcomes,such as relationship work. Since employees areoften motivated by relationships with others, thetheoretical refinement of attribution theorycould elevate it to a major motivational theory inorganizational behavior (Martinko et al., 2006).

REFERENCES

Allport, G. 1979. (First published in 1954.) The nature of prej-udice. New York: Doubleday Anchor.

Andersen, S. M., & Chen, S. 2002. The relational self: Aninterpersonal social-cognitive theory. Psychological Re-view, 109: 619–645.

Anderson, C. A. 1991. How people think about causes: Exam-ination of the typical phenomenal organization of attri-butions for success and failure. Social Cognition, 9: 295–329.

Aron, A., & McLaughlin-Volpe, T. 2001. Including others in theself: Extensions to own and partner’s group member-ship. In C. Sedikides & M. B. Brewer (Eds.), Individualself, relational self, collective self: 89–108. Philadelphia:Psychology Press.

Avolio, B. J. 2007. Promoting more integrative strategies forleadership theory-building. American Psychologist, 62:25–33.

Avolio, B. J., Walumbwa, F. O., & Weber, T. J. 2009. Leader-ship: Current theories, research, and future directions.Annual Review of Psychology, 60: 8.1–8.29.

Baldwin, M. W. 1992. Relational schemas and the processingof social information. Psychological Bulletin, 112: 461–484.

Bauer, T. N., & Green, S. G. 1996. The development of leader-member exchange: A longitudinal test. Academy ofManagement Journal, 39: 1538–1567.

Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs,K. D. 2001. Bad is stronger than good. Review of GeneralPsychology, 5: 323–370.


Baumeister, R. F., & Leary, M. R. 1995. The need to belong:Desire for interpersonal attachments as a fundamentalhuman motivation. Psychological Bulletin, 117: 497–529.

Berger, C. R., & Calabrese, R. J. 1975. Some exploration ininitial interaction and beyond: Toward a developmentaltheory of communication. Human Communication Re-search, 1: 99–112.

Bowler, W. M., & Brass, D. J. 2006. Relational correlates ofinterpersonal citizenship behavior: A social networkperspective. Journal of Applied Psychology, 91: 70–82.

Brass, D. J. 1985. Men’s and women’s networks: A study ofinteraction patterns and influence in an organization.Academy of Management Journal, 28: 327–343.

Brass, D. J., & Burkhardt, M. E. 1993. Potential power andpower use: An investigation of structure and behavior.Academy of Management Journal, 36: 440–470.

Brewer, M. B., & Gardner, W. 1996. Who is this “we”? Levelsof collective identity and self representations. Journal ofPersonality and Social Psychology, 71: 83–93.

Chen, S., Boucher, H. C., & Tapias, M. P. 2006. The relationalself revealed: Integrative conceptualization and impli-cations for interpersonal life. Psychological Bulletin, 132:151–179.

Chiaburu, D. S., & Harrison, D. A. 2008. Do peers make theplace? Conceptual synthesis and meta-analysis of co-worker effects on perceptions, attitudes, OCBs, and per-formance. Journal of Applied Psychology, 93: 1082–1103.

Cole, M. S., Walter, F., & Bruch, H. 2008. Affective mecha-nisms linking dysfunctional behavior to performance inwork teams: A moderated mediation study. Journal ofApplied Psychology, 93: 945–958.

Contractor, N. S., Wasserman, S., & Faust, K. 2006. Testingmultitheoretical, multilevel hypotheses about organiza-tional networks: An analytic framework and empiricalexample. Academy of Management Review, 31: 681–703.

Dansereau, F. 1995. A dyadic approach to leadership: Creat-ing and nurturing this approach under fire. LeadershipQuarterly, 6: 479–490.

Dean, J. W., & Brass, D. J., Jr. 1985. Social interaction and theperception of job characteristics in an organization. Hu-man Relations, 38: 571–582.

De Dreu, C. K. W., & Weingart, L. R. 2003. Task versus rela-tionship conflict, team performance, and team membersatisfaction: A meta-analysis. Journal of Applied Psy-chology, 88: 741–749.

Dienesch, R. M., & Liden, R. C. 1986. Leader-member ex-change model of leadership: A critique and further de-velopment. Academy of Management Review, 11: 618–634.

Dirks, K. T., Lewicki, R. J., & Zaheer, A. 2009. Repairing rela-tionships within and between organizations: Building aconceptual foundation. Academy of Management Re-view, 34: 68–84.

Dorfman, P. W., & Stephan, W. G. 1984. The effects of groupperformance on cognitions, satisfaction, and behavior:A process model. Journal of Management, 10: 173–192.

Dutton, J. E., & Ragins, B. R. (Eds.). 2006. Exploring positiverelationships at work: Building a theoretical and re-search foundation. Mahwah, NJ: Lawrence Erlbaum As-sociates.

Edmondson, A. C. 1999. Psychological safety and learningbehavior in work teams. Administrative Science Quar-terly, 44: 350–383.

Elig, T. W., & Frieze, I. H. 1979. Measuring causal attributionsfor success and failure. Journal of Personality and SocialPsychology, 37: 621–634.

Epitropaki, O., & Martin, R. 2005. From ideal to real: A longi-tudinal study of implicit leadership theories, leader-member exchanges and employee outcomes. Journal ofApplied Psychology, 90: 659–676.

Felps, W., Mitchell, T. R., Hekman, D. R., Lee, T. W., Holtom,B. C., & Harman, W. S. 2009. Turnover contagion: Howcoworkers’ job embeddedness and job search behaviorsinfluence quitting. Academy of Management Journal, 52:545–561.

Ferris, G. R., Liden, R. C., Munyon, T. P., Summers, J. K., Basik,K. J., & Buckley, M. R. 2009. Relationships at work: To-ward a multidimensional conceptualization of dyadicwork relationships. Journal of Management, 35: 379–403.

Gabriel, S., & Gardner, W. L. 1999. Are there “his” and “hers”types of interdependence? The implications of genderdifferences in collective versus relational interdepen-dence for affect, behavior, and cognition. Journal of Per-sonality and Social Psychology, 77: 642–655.

Garst, H., Frese, M., & Molenaar, P. C. M. 2000. The temporalfactor of change in stressor-strain relationships: Agrowth curve model on a longitudinal study in EastGermany. Journal of Applied Psychology, 85: 417–438.

George, J. M. 1990. Personality, affect, and behavior ingroups. Journal of Applied Psychology, 75: 107–116.

Gerstner, C. R., & Day, D. V. 1997. Meta-analytic review ofleader-member exchange theory: Correlates and con-struct issues. Journal of Applied Psychology, 82: 827–844.

Goffman, E. 1967. Interaction ritual: Essays on face-to-facebehavior. Garden City, NY: Anchor Books.

Graen, G. 1976. Role-making processes of leadership devel-opment. In M. D. Dunnette (Ed.), Handbook of industrialand organizational psychology: 1201–1245. Chicago:Rand McNally.

Graen, G., & Cashman, J. F. 1975. A role-making model ofleadership in formal organizations: A developmentalapproach. In J. G. Hunt & L. L. Larson (Eds.), Leadershipfrontiers: 143–166. Kent, OH: Kent State University Press.

Graen, G. B., & Scandura, T. A. 1987. Toward a psychology ofdyadic organizing. Research in Organizational Behav-ior, 9: 175–208.

Graen, G. B., & Uhl-Bien, M. 1995. Relationship-based ap-proach to leadership: Development of leader-memberexchange (LMX) theory of leadership over 25 years: Ap-plying a multi-level multi-domain perspective. Leader-ship Quarterly, 6: 219–247.

Granovetter, M. 1973. The strength of weak ties. AmericanJournal of Sociology, 78: 1360–1380.


Green, S. G., & Mitchell, T. R. 1979. Attributional processes ofleaders in leader-member interactions. OrganizationalBehavior and Human Performance, 23: 429–458.

Heider, F. 1958. The psychology of interpersonal relations.Hillsdale, NJ: Lawrence Erlbaum Associates.

Hinde, R. A. 1997. Relationships: A dialectical perspective.Hove, UK: Psychology Press.

Hofmann, D. A., & Jones, L. M. 2005. Leadership, collectivepersonality, and performance. Journal of Applied Psy-chology, 90: 509–522.

House, R., Rousseau, D. M., & Thomas-Hunt, M. 1995. Themeso paradigm: A framework for the integration of mi-cro- and macroorganizational behavior. Research in Or-ganizational Behavior, 17: 71–114.

Ibarra, H. 1992. Homophily and differential returns: Sex dif-ferences in network structure and access in an advertis-ing firm. Administrative Science Quarterly, 37: 422–447.

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

Ilgen, D. R., Hollenbeck, J. R., Johnson, M., & Jundt, D. 2005.Teams in organizations: From input-process-outputmodels to IMOI models. Annual Review of Psychology,56: 517–544.

Ilies, R., & Judge, T. A. 2005. Goal regulation across time: Theeffects of feedback and affect. Journal of Applied Psy-chology, 90: 453–467.

Jackson, C. L., & LePine, J. A. 2003. Peer responses to a team’sweakest link: A test and extension of LePine and VanDyne’s model. Journal of Applied Psychology, 88: 459–475.

Jehn, K. A. 1995. A multimethod examination of the benefitsand detriments of intragroup conflict. AdministrativeScience Quarterly, 40: 256–282.

Johns, G. 2006. The essential impact of context on organiza-tional behavior. Academy of Management Review, 31:386–408.

Johnson, R. E., Selenta, C., & Lord, R. G. 2006. When organi-zational justice and the self-concept meet. Organiza-tional Behavior and Human Decision Processes, 99: 175–201.

Kelley, H. H. 1967. Attribution theory in social psychology.Nebraska Symposium on Motivation, 15: 192–240.

Kelley, H. H. 1971. Attributions in social interactions. Morris-town, NJ: General Learning Publishing Group.

Kelley, H. H. 1973. The processes of causal attribution. Amer-ican Psychologist, 28: 107–128.

Kent, R. L., & Martinko, M. J.. 1995. The measurement ofattributions in organizational research. In M. J. Martinko(Ed.), Attribution theory: An organizational perspective:17–34. Delray Beach, FL: St. Lucie Press.

Knobloch, L. K., & Knobloch-Fedders, L. M. 2010. The role ofrelational uncertainty in depressive symptoms and re-lationship quality: An actor partner interdependencemodel. Journal of Social and Personal Relationships, 27:137–159.

Knobloch, L. K., & Solomon, D. H. 1999. Measuring the sourcesand content of relational uncertainty. CommunicationStudies, 50: 261–278.

Knobloch, L. K., & Solomon, D. H. 2002. Intimacy and themagnitude and experience of episodic relational uncer-tainty within romantic relationships. Personal Relation-ships, 9: 457–478

Kristof, A. L. 1996. Person-organization fit: An integrativereview of its conceptualizations, measurement, and im-plications. Personnel Psychology, 49: 1–49

Kristof-Brown, A. L., Zimmerman, R. D., & Johnson, E. C. 2005.Consequences of individuals’ fit at work: A meta-analysis of person-job, person-organization, person-group, and person-supervisor fit. Personnel Psychology,58: 281–342

Labianca, G., & Brass, D. J. 2006. Exploring the social ledger:Negative relationships and negative asymmetry in so-cial networks in organizations. Academy of Manage-ment Review, 31: 596–614.

Lazarus, R. S. 1991. Progress on a cognitive-motivational-relational theory of emotion. American Psychologist, 46:819–834

Leary, M. R., & Baumeister, R. F. 2000. The nature and func-tion of self-esteem: Sociometer theory. Advances in Ex-perimental Social Psychology, 32: 1–62.

Lee, T. W., Mitchell, T. R., Sablynski, C. J., Burton, J. P., &Holtom, B. C. 2004. The effects of job embeddedness onorganizational citizenship, job performance, volitionalabsences, and voluntary turnover. Academy of Manage-ment Journal, 47: 711–722

LePine, J. A., & Van Dyne, L. 2001. Peer responses to lowperformers: An attributional model of helping in thecontext of groups. Academy of Management Review, 26:67–84

Liden, R. C., Sparrowe, R. T., & Wayne, S. J. 1997. Leader-member exchange theory: The past and potential for thefuture. Research in Personnel and Human ResourcesManagement, 15: 47–119.

Markus, H., & Wurf, E. 1987. The dynamic self-concept: Asocial psychological perspective. Annual Review of Psy-chology, 38: 299–337.

Markus, H. R., & Kitayama, S. 1991. Culture and the self:Implications for cognition, emotion, and motivation. Psy-chological Review, 98: 224–253

Martinko, M. J. (Ed.). 1995. Attribution theory: An organiza-tional perspective. Delray Beach, FL: St. Lucie Press.

Martinko, M. J. 2004. Parting thoughts: Current issues andfuture directions. In M. J. Martinko (Ed.), Attribution the-ory in the organizational sciences: Theoretical and em-pirical contributions: 297–305. Greenwich, CT: Informa-tion Age.

Martinko, M. J., Douglas, S. C., & Harvey, P. 2006. Attributiontheory in industrial and organizational psychology: Areview. International Review of Industrial and Organi-zational Psychology, 21: 127–187.


Martinko, M. J., & Gardner, W. L. 1987. The leader memberattribution process. Academy of Management Review,12: 235–249.

Martinko, M. J., Harvey, P., & Dasborough, M. T. 2011. Attri-bution theory in the organizational sciences: A case ofunrealized potential. Journal of Organizational Behav-ior, 32: 144–149.

Martinko, M. J., Harvey, P., & Douglas, S. C. 2007. The role,function, and contribution of attribution theory to lead-ership: A review. Leadership Quarterly, 18: 561–585.

Martinko, M. J., Moss, S. E., Douglas, S. C., & Borkowski, N.2007. Anticipating the inevitable: When leader andmember attribution styles clash. Organizational Behav-ior and Human Decision Processes, 104: 158–174.

Maslyn, J. M., & Uhl-Bien, M. 2001. Leader-member exchangeand its dimensions: Effects of self-effort and other’s ef-fort on relationship quality. Journal of Applied Psychol-ogy, 86: 697–708.

Mathieu, J. E., & Chen, G. 2011. The etiology of the multilevelparadigm in management research. Journal of Manage-ment, 37: 610–641.

Morgeson, F. P., & Hofmann, D. A. 1999. The structure andfunction of collective constructs: Implications for multi-level research and theory development. Academy ofManagement Review, 24: 249–265.

Nahrgang, J. D., Morgeson, F. P., & Ilies, R. 2009. The devel-opment of leader-member exchanges: Exploring howpersonality and performance influence leader andmember relationships over time. Organizational Behav-ior and Human Decision Processes, 108: 256–266.

Naquin, C., & Tynan, R. 2003. The team halo effect: Whyteams are not blamed for their failures. Journal of Ap-plied Psychology, 88: 332–340.

O’Driscoll, M. P., & Beehr, T. A. 1994. Supervisor behaviors,role stressors and uncertainty as predictors of personaloutcomes for subordinates. Journal of OrganizationalBehavior, 15: 141–155.

Pearsall, M. J., & Ellis, A. P. J. 2011. Thick as thieves: Theeffects of ethical orientation and psychological safetyon unethical team behavior. Journal of Applied Psychol-ogy, 96: 401–411.

Peirce, R. S., Pruitt, D. G., & Czaja, S. J. 1993. Complainant-respondent differences in procedural choice. Interna-tional Journal of Conflict Management, 4: 199–222.

Perry-Smith, J. E. 2006. Social yet creative: The role of socialrelationships in facilitating individual creativity. Acad-emy of Management Journal, 49: 85–101.

Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach,D. G. 2000. Organizational citizenship behaviors: A crit-ical review of the theoretical and empirical literatureand suggestions for future research. Journal of Manage-ment, 26: 513–563.

Ren, H., & Gray, B. 2009. Repairing relationship conflict: Howviolation types and culture influence the effectivenessof restoration rituals. Academy of Management Review,34: 105–126.

Rosenfeld, P., Giacalone, R., & Riordan, C. 2001. Impressionmanagement: Building and enhancing reputations atwork. New York: International Thompson BusinessPress.

Rudolph, U., Roesch, S. C., Greitemeyer, T., & Weiner, B. 2004.A meta-analytic review of help giving and aggressionfrom an attributional perspective: Contributions to ageneral theory of motivation. Cognition and Emotion, 18:815–848.

Rusbult, C. E., & Van Lange, P. 2003. Interdependence, inter-action, and relationships. Annual Review of Psychology,54: 351–375.

Rusbult, C. E., & Van Lange, P. 2008. Why we need interde-pendence theory. Social and Personality PsychologyCompass, 2: 2049–2070.

Sedikides, C., Campbell, W. K., Reeder, G. D., & Elliot, A. J.1998. The self-serving bias in relational context. Journalof Personality and Social Psychology, 74: 378–386.

Seibert, S. E., Kraimer, M. L., & Liden, R. C. 2001. A socialcapital theory of career success. Academy of Manage-ment Journal, 44: 219–237.

Settoon, R. P., & Mossholder, K. W. 2002. Relationship qualityand relationship context as antecedents of person- andtask-focused interpersonal citizenship behavior. Journalof Applied Psychology, 87: 255–267.

Sluss, D. M., & Ashforth, B. E. 2007. Relational identity andidentification: Defining ourselves through work rela-tionships. Academy of Management Review, 32: 9–32.

Sparrowe, R. T., & Liden, R. C. 1997. Process and structure inleader-member exchange. Academy of Management Re-view, 22: 522–552.

Sparrowe, R. T., & Liden, R. C. 2005. Two routes to influence:Integrating leader-member exchange and social net-work perspectives. Administrative Science Quarterly,50: 505–535.

Stogdill, R. M., & Coons, A. E. 1957. Leader behavior: Itsdescription and measurement. Columbus: Ohio StateUniversity Press for Bureau of Business Research.

Taggar, S., & Neubert, M. 2004. The impact of poor performerson team outcomes: An empirical examination of attribu-tion theory. Personnel Psychology, 57: 935–968.

Tajfel, H. 1982. Social identity and intergroup relations. Cam-bridge: Cambridge University Press.

Thibaut, J., & Walker, L. 1975. Procedural justice: A psycho-logical analysis. Hillsdale, NJ: Lawrence Erlbaum Asso-ciates.

Tomlinson, E. C., & Mayer, R. C. 2009. The role of causalattribution dimensions in trust repair. Academy of Man-agement Review, 34: 85–104.

Van den Bos, K., & Lind, E. A. 2002. Uncertainty managementby means of fairness judgments. Advances in Experi-mental Social Psychology, 34: 1–60.

Vohs, K. D., Baumeister, R. F., & Ciarocco, N. J. 2005. Self-regulation and self-presentation: Regulatory resourcedepletion impairs impression management and effortfulself-presentation depletes regulatory resources. Journalof Personality and Social Psychology, 88: 632–657.


Weiner, B. 1985. An attributional theory of achievement mo-tivation and emotion. Psychological Review, 92: 548–573.

Weiner, B. 1990. Attribution in personality psychology. InL. A. Perrin (Ed.), Handbook of personality: Theory andresearch: 465–485. New York: Guilford Press.

Weiner, B. 1995. Judgments of responsibility: A foundation fora theory of social conduct. New York: Guilford Press.

Weiner, B., Frieze, I., Kukla, A., Reed, L., Rest, S., & Rosenbaum,R. M. 1971. Perceiving the causes of success and failure.Morristown, NJ: General Learning Publishing Group.

Whiting, S. W., Podsakoff, P. M., & Pierce, J. R. 2008. Effects oftask performance, helping, voice, and organizationalloyalty on performance appraisal ratings. Journal of Ap-plied Psychology, 93: 125–139.

Wright, T. A., & Bonett, D. G. 2007. Job satisfaction andpsychological well-being as nonadditive predictors ofworkplace turnover. Journal of Management, 33: 141–160.

Yammarino, F. J., & Dansereau, F. 2009. Multi-level issues inorganizational behavior and leadership, vol. 8. Bingley,UK: Emerald Group.

Young, A. M., & Perrewé, P. L. 2000. The exchange relation-ship between mentors and protégés: The developmentof a framework. Human Resource Management Review,10: 177–209.

Zuckerman, M. 1979. Attribution of success and failure revis-ited, or the motivational bias is alive and well in attri-bution theory. Journal of Personality, 47: 245–287.

Marion B. Eberly ([email protected]) is an assistant professor in the Milgard School ofBusiness, University of Washington, Tacoma. She received her Ph.D. in organizationalbehavior and human resources from the University of Washington, Seattle. Her re-search focuses on the influence of emotions and attributions within leader-followerrelationships and the impact of leadership on organizational entry and exit.

Erica C. Holley ([email protected]) is a doctoral candidate in organizational behaviorand human resources at the Foster School of Business, University of Washington. Herresearch focuses on the management of challenging interpersonal events that havethe potential to create conflict within dyadic and/or larger group/team relationships.

Michael D. Johnson ([email protected]) is an assistant professor of management andorganization in the Foster School of Business, University of Washington. He receivedhis Ph.D. in organizational behavior and human resource management from MichiganState University. His research focuses primarily on groups and teams, identity, andaffect.

Terence R. Mitchell ([email protected]) is the Edward Carlson Professor of Business Ad-ministration at the University of Washington. He received his Ph.D. in organizationalpsychology from the University of Illinois. His research focuses on motivation, lead-ership, turnover, and decision making.


Copyright of Academy of Management Review is the property of Academy of Management and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

beyond internal and external: a dyadic theory of ...€¦ · the dyadic theory of relational...

Documents