no employee an island: workplace loneliness and job performance€¦ · no employee an island:...

r Academy of Management Journal2018, Vol. 61, No. 6, 2343–2366.https://doi.org/10.5465/amj.2015.1066

NO EMPLOYEE AN ISLAND: WORKPLACE LONELINESS ANDJOB PERFORMANCE

HAKAN OZCELIKCalifornia State University, Sacramento

SIGAL G. BARSADEUniversity of Pennsylvania

This research investigates the link between workplace loneliness and job performance.Integrating the regulatory loop model of loneliness and the affect theory of social ex-change, we develop a model of workplace loneliness. We focus on the central role ofaffiliation in explaining the loneliness–performance relationship, predicting that de-spite lonelier employees’ desire to connect with others, being lonelier is associated withlower job performance because of a lack of affiliation at work. Through a time-laggedfield study of 672 employees and their 114 supervisors in two organizations, we findsupport that greater workplace loneliness is related to lower job performance; the me-diators of this relationship are lonelier employees’ lower approachability and lesseraffective commitment to their organizations. We also examine the moderating roles ofthe emotional cultures of companionate love and anger, as well as of the loneliness ofother coworkers in the work group. Features of this affective affiliative context moderatesome of the relationships between loneliness and the mediating variables; we also findsupport for the full moderated mediation model. This study highlights the importance ofrecognizing the pernicious power of workplace loneliness over both lonelier employeesand their organizations. We offer implications for future research and practice.

“No man is an island, entire of itself; every man isa piece of the continent, a part of the main.”

JohnDonne, 2012 [1624]

Loneliness—“a complex set of feelings that occurswhen intimate and social needs are not adequatelymet” (Cacioppo et al., 2006: 1055)—is an aversivepsychological state. Psychologists have long studiedthepainful experience of loneliness and its outcomes(see Cacioppo & Patrick, 2008; Heinrich & Gullone,2006 for reviews). Surprisingly, though, there hasbeen very little examination of the processes andoutcomes of loneliness in the workplace, eventhough most people spend a large part of their livesat work. Better understanding loneliness at work isimportant given the myriad pernicious emotional,cognitive, attitudinal, and behavioral outcomes thathave been found as a result of being lonely (Cacioppo

& Patrick, 2008; Hawkley & Cacioppo, 2010;Heinrich & Gullone, 2006; Masi, Chen, Hawkley, &Cacioppo, 2011). These negative outcomes occurbecause people have an innate, primary drive toform social bonds and mutual caring commitments(Murray, 1938; Schachter, 1959), and they sufferwhen these social bonds do not meet their expecta-tions or their need to belong (Baumeister & Leary,1995), including at work (Barsade & O’Neill, 2014;Chiaburu & Harrison, 2008; Dutton & Heaphy, 2003;Wright, 2012).

This inquiry into the processes and outcomes ofloneliness in the workplace is also important as loneli-ness has been characterized by scholars, as well as bytheU.S. SurgeonGeneral (Murthy, 2017), as “amodernepidemic” in need of treatment (Killeen, 1998). Lone-liness is experienced by adults of all ages (Masi et al.,2011),withnosign that loneliness levelsareabating (seeQualter et al., 2015 for a review).

Because loneliness is a relational construct (Weiss,1973), it influences not only how lonelier people feelabout themselves, but also how they feel about andbehave toward others (Cacioppo & Hawkley, 2009;Jones, 1982; Jones & Hebb, 2003), and, importantly,how others feel about and behave toward them

Thank you to Paul Dickey, Lauren Rhodewalt, andKevin Sweeney for research assistance. We also thank theCSUSResearch and Creative Activity Grant Program at theCalifornia State University, Sacramento, the Robert KatzFund for EmotionResearch at theWharton School, and theWharton Center for Leadership and Change Managementfor their support of this study.

2343

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

https://doi.org/10.5465/amj.2015.1066

(Heinrich & Gullone, 2006). That broader impactmakes loneliness particularly relevant to examine atwork, since connectionwithothershas been found tobe an inherent part of employee motivation and sat-isfaction (Dutton & Heaphy, 2003). Indeed, we knowfromsomeof the earliest field studies inmanagementthat employees are driven not only by economicneeds but also by the need to establish relationshipsand social attachments (Gouldner, 1954;Mayo, 1949).Given the pervasive and pernicious influence ofloneliness in other life domains, and givenhowmuchtime employees spend at work with each other, theinvestigation of workplace loneliness is important.

To better understand the relationship betweenworkplace loneliness and employee attitudes, be-havior, and performance, and to build our model ofworkplace loneliness, we draw from two theoreticaltraditions. The first comes from thenearly 40 years ofpsychological research tying greater loneliness tolowered affiliation (e.g., Cacioppo & Patrick, 2008;Heinrich & Gullone, 2006 for reviews), specificallyCacioppo and Hawkley’s (2009) regulatory loopmodel of loneliness, which describes the psycho-logical mechanisms within lonelier people. Thesecond is the affect theory of social exchange(Lawler, 2001, 2006; Lawler & Thye, 2007), a socio-logical theory that takes into account the role of otherpeople’s thoughts about the lonely person. We in-tegrate these theories to build and empirically testa model of workplace loneliness that helps explainwhy lonelier employees are less affiliative—andwhythat lackofaffiliationrelates topoorer jobperformance.Through this process,we aim to extendour knowledgein three domains. First, by improving our knowledgeabout loneliness in organizational settings, we aim tocontribute to a more nuanced understanding of howaffect is related to job performance (Elfenbein, 2007),andaddress thecall for greater studyof theoutcomesofdiscrete affect at work (Barsade & Gibson, 2007). Sec-ond, because of the relational nature of loneliness, weadd to the emerging theorizing about relational sys-tems at work (Dutton & Heaphy, 2003; Grant & Parker,2009) by better understanding how an employee’s re-lational environments can include feelings of lack ofaffiliationanddisconnection.Last,wecontribute to theloneliness literatures in psychology and sociology byexpanding our understanding of this powerful socialemotion to the workplace context.

THE NATURE OF LONELINESS

We take an affective prototype approach to lone-liness (Heinrich & Gullone, 2006; Horowitz, French,

& Anderson, 1982). According to affective prototypetheory, each type of discrete affect is determined bya cluster, or prototype, of features that distinguish it.These features include all aspects of the affectiveexperience: feeling states, physiological markers,and cognitive and behavioral categories (Shaver,Schwartz, Kirson, & O’Connor, 1987). The affectiveprototype model is broad enough to encompass af-fect as a state (including emotions andmoods), a trait(dispositional affect), or a sentiment (evaluative af-fective responses to social objects)—all dependingon the specific affect’s intensity, duration, specific-ity, and evaluative focus (Barsade & Gibson, 2007).Loneliness, similar to other types of affect,1 has beencharacterized as a state and a sentiment, but is gen-erally not thought of as a trait (although it can bechronic) (Peplau & Perlman, 1982; Spitzberg & Hurt,1989). The prototype model of loneliness also com-ports with the most cited definition of loneliness,which requires three key components: (1) an un-pleasant and aversive feeling, (2) generated from asubjective negative assessment of one’s overall re-lationships in a particular social domain, and (3)a belief that these social relationships are deficient(Peplau & Perlman, 1982). Defining lonelinessthrough this affective prototype lens, we take intoconsideration the feelings loneliness arouses, thecognitions that comprise it, and the behaviors itevokes.

Because loneliness is a subjective experience, anemployee does not have to be alone to feel lonely,and lonely employees can be lonely even wheninteracting frequently with many others if these in-teractions do not provide lonelier employees withtheir desired level of closeness (Fischer & Phillips,1982). Whether employees feel lonely depends onthe level of closeness, security, and support they seekin their interpersonal relationships (Jones & Hebb,2003). Thus, the same work environment could ful-fill the interpersonal needs of some employeeswhileleaving other employees lonely.

Distinguishing Loneliness from Related Constructs

We differentiate workplace loneliness from re-lated constructs such asworkplace ostracism, whichis an employee’s perception that she or he is beingintentionally excluded or ignored by others at work

1 We intentionally use the term “affect” as an “umbrellaterm encompassing a broad range of feelings that in-dividuals experience,” including discrete emotions andgeneralized mood (Barsade & Gibson, 2007: 37).

2344 DecemberAcademy of Management Journal

(Ferris, Brown, Berry, & Lian, 2008). Ostracism dif-fers from loneliness, the subjective nature of whichmakes it possible for two equally ostracized em-ployees to feel different levels of loneliness—and foremployees to feel lonely without being ostracized atall. Empirical research examining these constructstogether has foundmoderated relationships betweenthe two (Wesselmann, Wirth, Mroczek, & Williams,2012), and ostracism by others has been discussed asan antecedent to loneliness (Leary, 1990), but lone-liness and ostracism have not been found eithertheoretically or empirically to be the same construct(Wesselmann et al., 2012; Williams, 2007).

Loneliness is also neither a form of depression nora personality trait. Cacioppo and Patrick (2008) of-fered compelling evidence that although loneliness isrelated to these constructs (and loneliness predictsdepression [Nolen-Hoeksema & Ahrens, 2002]), con-ceptions of loneliness as a form of depression, shy-ness, or poor social skills are inaccurate (Anderson& Harvey, 1988; Hawkley, Burleson, Berntson, &Cacioppo, 2003; Hawkley & Cacioppo, 2010). Thereis also much evidence across myriad studies thatloneliness is not a disposition, and differs from per-sonality traits such as negative affectivity, introver-sion, and disagreeableness (e.g., Cacioppo et al., 2006;Russell, Peplau, & Cutrona, 1980). Last, lonelinessdiffers from solitude (Heinrich & Gullone, 2006),which, contrary to loneliness, is a pleasant, desirableand freely chosen state (Derlega & Margulis, 1982).

WORKPLACE LONELINESS: ACONCEPTUAL MODEL

Loneliness researchers have found that peopleexperience different levels of loneliness in differentdomains of their lives, such as family life, the ro-mantic realm, and the social domain (DiTommaso &Spinner, 1993). In line with this research, we focusspecifically on the workplace domain, and concep-tualize degree ofworkplace loneliness as employees’subjective affective evaluations of, and feelingsabout, whether their affiliation needs are being metby the people they work with and the organizationthey work for.

Affiliation, a central influence on human behav-ior (Hill, 1987), is the degree to which people haveclose interpersonal bonds, harmonious relationships,and a sense of communion with others (Depue &Morrone-Strupinsky, 2005; Mehrabian & Ksionzky,1974; Murray, 1938). Motivated by the desire forsocial contact and belongingness (Murray, 1938;Schachter, 1959), affiliation is a key component of

social interactions (Hess, 2006;Mehrabian&Ksionzky,1974) and largely explains how loneliness influ-ences life outcomes (Weiss, 1973). Affiliation mani-fests in two major ways: through attitudes andbehaviors. Affiliative attitudes reflect how close andattached people feel to their social environments(Freeman, 1992), which, within the work context,we operationalize as employees’ affective commit-ment to the organization (Allen & Meyer, 1990).Affiliative behaviors manifest in people’s behav-ioral expressions of attachment and involvement insocial interactions, particularly those expressionsthat indicate closeness (Mehrabian & Ksionzky,1974). Inherent to the behavioral manifestation ofaffiliation is having others perceive you as being ap-proachable, both verbally and nonverbally (Mehrabian& Ksionzky, 1974; Wiemann, 1977). Therefore, in thework context, we operationalize affiliative behav-iors as an employee’s affiliative approachability to-ward coworkers (which we refer to as “employeeapproachability” for conciseness).

We hypothesize that the workplace loneliness–affiliation relationship is related to work outcomes.Because loneliness arises from a person’s basicneed to belong to a social environment in which heor she feels psychologically secure and protected(Baumeister & Leary, 1995), it can be adaptive byincreasing the lonely person’s motivation to build orrebuild affiliations—but only if experienced as a short-lived emotion (Cacioppo et al., 2006; Qualter et al.,2015), such as feeling lonely for an afternoon, or inthe first days of starting a new job. However, one ofthe most robust findings in the loneliness literature(e.g., Cacioppo & Patrick, 2008; Heinrich & Gullone,2006; Jones & Hebb, 2003; Qualter et al., 2015) isthat when loneliness becomes a more establishedsentiment—that is, a valenced assessment and com-prising feelings about a particular group or in-terpersonal setting (Frijda, 1994)—it actually has theopposite effect, impeding the satisfaction of belong-ingness needs “through faulty or dysfunctional cogni-tions, emotions, and behaviors” (Heinrich & Gullone,2006: 698). This occurs when people come to believethat the meaningful social connections they want arenot available; their fear, hypervigilance, and subjectivefeelings of rejection cause them to withdraw and giveup on building the very interpersonal relationshipsthey crave (Masi et al., 2011).

Theoretical Framework

By combining these two theories relevant toour questions about workplace loneliness—the

2018 2345Ozcelik and Barsade

psychologically oriented regulatory loop model(Cacioppo & Hawkley, 2009) and the interpersonallyand contextually oriented affect theory of social ex-change (Lawler, 2001)—we build a model of work-place loneliness showing that greater workplaceloneliness predicts lower job performance, as medi-ated by lower employee affiliation (both attitudinaland behavioral). In addition, we predict that two sit-uational factors moderate the relationship betweenworkplace loneliness and lowered affiliation: theemotional culture of the employee’s work group(culture of anger and culture of companionate love)and the aggregate loneliness of other employees in theemployee’s work group. We elaborate upon ourmodel of workplace loneliness below (see Figure 1).

Regulatory loop model of loneliness. In theirregulatory loop model of loneliness, Cacioppo andHawkley (2009) showed that once people havemadethe evaluation that their relational needs are notbeing met, and that a particular context makesthem feel lonely, they develop an acute need to feelpsychologically protected and secure. This makeslonelier people more vigilant and defensive about

interpersonal relationships, prompting them to con-tinually appraise situations to see whether theserelationships can meet their belongingness needsand alleviate their loneliness (Weiss, 1973). Be-cause loneliness isoftenaccompaniedbya“perceptionthatone is sociallyon theedgeand isolated fromothers”(Cacioppo, Grippo, London, Goessens, & Cacioppo,2015: 243), lonelier people are overly vigilant to socialthreats, although not to other types of threat (Cacioppo,Balogh, & Cacioppo, 2015; Cacioppo, Bangee, Balogh,Cardenas-Iniguez, Qualter, & Cacioppo, 2016). This so-cial hypervigilance leads them to fall prey to sociallybased attentional, confirmatory, andmemory biases, allof which induce them to view their “social world asthreatening and punitive” (Cacioppo &Hawkley, 2009:451). As a result, lonelier people become less secure insocial interactions (Hawkley et al., 2003) and moreanxious about being negatively evaluated by others(Cacioppo et al., 2006; Cacioppo & Patrick, 2008;Jones, 1982). This socially anxious state leadslonelier people to be more likely to engage in in-appropriate self-disclosure patterns (Jones, Hobbs,& Hockenbury, 1982) and to show other deficits in

FIGURE 1Conceptual Model of Workplace Loneliness

Workplace Loneliness

(Self-Report)

Employee Affective Commitment

to the Organization(Self-Report)

Employee Approachability (Coworker Ratings)

Culture of Companionate

Love (Work Group)

Job Performance (Supervisor Ratings)

Loneliness of Employee’s Coworkers (Coworker Ratings)

Culture of Anger

(Work Group)


social skills (Bell, 1985; Russell et al., 1980) that un-dermine their social interactions by eliciting more neg-ative feelings, displays, and behaviors on the part ofothers (see Heinrich & Gullone, 2006 for a review).

In short, loneliness can create a hard-to-break cycleof negative social interactions, whereby greater lone-liness causes more intrapsychic negative attributionsabout others, which prompt awkward or negative be-haviors that actually breed less affiliation (Masi et al.,2011), which in turn gets reciprocated in kind(Mehrabian & Ksionzky, 1974; Perlman & Peplau,1981).Thus, thepeoplewhomostneed tobuild securerelationships have themost trouble doing so (Rokach,1989).2 The memoirist Emily White (2010: 161) de-scribed thebitter ironyof this loneliness cycle atwork:

Out of everyone [in the group], I was probably the onewho needed sociability the most. . . Yet when con-fronted with [others] who wanted to get close to me,I retreated into a stiff and stammering version of my-self, and became oddly resentful of the people whowere trying to befriend me. . . [M]y loneliness was al-tering my behavior and my perceptions of others.

Although the regulatory loop model of lonelinessdoes an excellent job of explaining the maladaptivepsychological motivations of lonely employees, thismodel alone is not enough to predict the effects ofworkplace loneliness on job performance. This isbecause it does not focus on the interpersonal natureof work—that is, the expectations that employeeshave of each other in their workplace interactions.The regulatory loopmodel also does not explain howthe lack of affiliation employees feel toward theirwork group extends to a lowered commitment to thelarger organization. To fill these gaps, we turn to theaffect theory of social exchange, which providesa more interpersonally and contextually orientedlens through which to understand the affiliative andperformance outcomes of loneliness within the so-cial context of organizations.

Affect theory of social exchange.Lawler’s (2001)affect theory of social exchange “explains how andwhen emotions produced by social exchange gen-erate stronger or weaker ties to relations, groups,or networks” (321). The theory posits that social

exchanges between people produce positive ornegative feelings (Lawler & Thye, 1999), which in-dividuals in the exchange consider either intrinsi-cally rewarding or punishing (Lawler, 2001; Lawler&Thye, 2007). The global feelings arising from thesesocial exchanges then trigger cognitive attributionalefforts to understand the sources or causes of thesefeelings, and these attributions influence how peo-ple assess their relations with specific others (suchas members of their work team), as well as withtheir broader group (their organization as a whole)(Lawler, Thye, & Yoon, 2008). When negative feel-ings predominate, therefore, people decide thatcertain relationships are not worth the effort, andthis decision fans out to ever-widening circles ofpotential connections.

All of this occurs through the process of affiliation(or disaffiliation), and affiliation in turn predictsprosocial behavior. Positive feelings experienced bya person from interactions with others within thegroup will be related to stronger affiliative attach-ments to the individuals in the group and organiza-tion, leading to a greater exchange of affective andhelping resources (Lawler, Thye, & Yoon, 2014).Conversely, negative feelings, such as greater lone-liness, will be related to weaker attachments, lessaffiliation (Lawler, 2006), and less exchange of af-fective and other help. The affect theory of socialexchange shares with the regulatory loop model ofloneliness the notion that affiliation (or its absence)is the key process resulting from an evaluation ofone’s social connections, and thus influences thequality of a lonely person’s relationships. However,the affect theory of social exchange goes further byanalyzing these feelings in a broader interpersonalcontext. This broader context is especially impor-tant in building a workplace model of lonelinessbecause workplaces depend heavily on interper-sonal exchange relationships. By considering therole of a lonely worker’s coworkers, the affect theoryof social exchange helps us make better predictionsabout the relationship between loneliness and jobperformance. In particular, this theory reinforces therole of affiliation, which also helps to explain whycoworkers notice and care that their lonelier col-leagues are less approachable, and why less ap-proachability relates to lonelier employees’ loweredjob performance. Furthermore, by considering attri-butional processes, the affect model of social ex-change (unlike the regulatory loop model) helpsexplainwhy the lower affiliation lonelier employeesfeel toward their immediate coworkers extends tolower affective commitment to the organization as

2 Research has found that the reason why lonely peoplecannot get themselves out of this negative regulatory loophas todowith their attributions. They feel unable to changetheir situation because they attribute their loneliness toperceived lack of control, which leads to lower expecta-tions of success and thus lower motivation to engage (seeHeinrich & Gullone, 2006 for a review of this literature).


a whole. Below, we elaborate on how each of theprocesses above is uniquely explained by the affecttheory of social exchange, by the regulatory loopmodel, and by the complementary predictions madeby both models.

Greater Employee Workplace Loneliness Relatedto Lower Employee Approachability

The affect theory of social exchange argues thatdepending on their expected outcomes in a possibleexchange relationship, people make decisions aboutwhether to exchange, with whom, and under whatconditions (Lawler, 2001). As the regulatory loopmodel of loneliness (Cacioppo & Hawkley, 2009)highlights at the psychological level, employees ex-periencing greater work loneliness will likely eval-uate their previous social exchanges with coworkersas negative (“When we get together, I feel bad”);therefore, lonelier employees will tend to withdrawfrom existing relationship opportunities. In addi-tion, lonelier employees will expect the worst in thefuture, causing them to continue to withdraw—atendency exacerbated by deficits in their social skillscaused by the loneliness itself (Bell, 1985; Russellet al., 1980). For example, lonelier people have beenfound to have more problems in taking part ingroups, being friendly, introducing themselves, andmaking friends with others (Horowitz & French,1979). This is despite the fact that lonelier people donothave inferior social skills to beginwith (Cacioppo& Hawkley, 2005). Rather, as they become lonelier,their preoccupation with their own feelings can leadto deficits in empathy for others (Jones et al., 1982),impeding their capacity to connect successfully andto be perceived as affiliatively approachable (Bell,1985; Mehrabian & Ksionzky, 1974). For example,lonelier people have been found to respond moreslowly to conversationpartners, ask fewer questions,and focus more on themselves compared to lesslonely people (Jones et al., 1982).

Because of the importance of affiliation as a majorresource exchanged in social relationships (Foa &Foa, 1974), people are quite accurate in perceiving,recording, and recalling other people’s affiliativepatterns in their social environment (Freeman,1992). For example, people continually evaluatethe level of affiliation they maintain with others intheir social interactions by paying close attention tothe verbal and nonverbal cues they receive (Miles,2009). Therefore, our model predicts that whenlonelier employees show a lack of affiliative behav-iors, their coworkers take notice. Specifically, they

perceive lonelier employees as less affiliativelyapproachable—less available for close interpersonalwork bonds, personal communion with others, andharmonious relationships.

Hypothesis 1. Employees who experience higherlevels of workplace loneliness will be less affiliativelyapproachable toward their coworkers (employeeapproachability).

Greater Workplace Loneliness and ReducedAffective Commitment to the Organization

Affective commitment refers to an employee’s af-filiation with, emotional attachment to, identifica-tionwith, and involvement in his or her organizationthrough feelings (such as belongingness, affection,and warmth) that lead to a rewarding work experi-ence (Meyer & Allen, 1997). We predict that em-ployees who experience loneliness in their workgroupswill reduce their affective commitment to theorganization for both psychological and in-terpersonal reasons.

As predicted by the regulatory loop model ofloneliness, the social hypervigilance of lonelierpeople (Cacioppo & Hawkley, 2009) leads to theiraffective and attitudinal withdrawal from thegroups of which they are a part. The attributionalprocesses described by the affect theory of socialexchange explain how this group-level reduction inaffective commitment leads to a reduction in af-fective commitment to the organization as a whole.Specifically, people generalize from their smallerdyadic or group-level exchanges to the broadergroup (Lawler, 2001; Lawler & Yoon, 1996). Be-cause the feelings that arise from social exchangestrigger cognitive efforts to understand the sourcesor causes of these feelings, people tend to interpretand explain these feelings with reference to thelarger social units in which they are embedded,such as an organization (Lawler, 2006). As a resultof being lonely in their work group, lonelier em-ployees assume that the probability of a positiveexchangewith others across the organization is alsolower, and will project the responsibility for theircurrent loneliness onto the broader organization.This attribution leads them to be less willing to in-vest themselves emotionally, and thus to lowertheir affective commitment to the organization asa whole (Lawler, 2001).

Hypothesis 2. Employees who experience higherlevels of workplace loneliness will be less affectivelycommitted to their organization.


Greater Workplace Loneliness and ReducedEmployee Performance: The Mediating Role ofEmployee Approachability and Employee AffectiveCommitment to the Organization

We predict that lower employee approachabilitywill negatively relate to the employee’s job perfor-mance. A growing stream of research has shownthat employee job performance is significantly tiedto an employee’s ability to build and maintain a re-lational support system and interpersonal networks(Chiaburu & Harrison, 2008; Grant & Parker, 2009;Kahn, 2007). The regulatory loop model of loneli-ness predicts that lonelier employees are less ap-proachable; the affect theory of social exchangepredicts that coworkers will notice this loweredlevel of approachability as a deficit in the exchangerelationship, and that they will withdraw in re-sponse (Lawler et al., 2014). Indeed, lonely behaviorevokes negative social responses in others (Perlman& Peplau, 1981) and less interest in future inter-actions from others (Bell, 1985), who view lonelierpeople as harder to get to know. This reciprocalprocess impairs the normal development of socialrelationships (Solano, Batten, & Parish, 1982). Yetthe exchange of interpersonal resources in workteams (Seers, 1989) has been found to positively in-fluence performance outcomes by enabling em-ployees to receive more guidance and emotionalsupport (Liden, Wayne, & Sparrowe, 2000), greaterhelpwith thework itself (Kamdar&VanDyne, 2007),engagement in better communication, coordination,and balance of team member contributions (Hoegl &Gemuenden, 2001), greater usage of teammates’ re-sources and task information (Farh, Lanaj, & Ilies,2017), and better opportunities to develop creativeideas (Grant & Berry, 2011). Therefore, we predictthat lonelier employees’ lower approachability re-duces the ability of these employees to receive re-lational and task resources from their coworkers,thus harming their job performance.

Hypothesis 3. Employees who are lonelier at work willhave poorer job performance, as partially mediated bytheir lower approachability towards their coworkers.

We also hypothesize that lonelier employees willbe less likely to perform effectively because of theirreduced affective commitment to the organization.As suggested by the regulatory loop model of lone-liness (Cacioppo & Hawkley, 2009), the cycle ofnegative social interactions experienced by lonelieremployees will fuel their sense that their work en-vironment is notmeeting their relational needs. This

appraisal, according to the affect theory of socialexchange (Lawler et al., 2008), will lead lonelieremployees to attribute their negative feelings to theiroverall organization; thus, the impaired social ex-change relationship between an employee and hisor her organization relates to a withdrawal of sup-port and effort from the employee (VanDyne, Graham,& Dienesch, 1994). Employees with less affectivecommitment to the organization have been shownto undertake fewer extra-role duties and to havegreater absenteeism and turnover (Hackett, Bycio, &Hausdorf, 1994; Meyer, Stanley, Herscovitch, &Topolnytsky, 2002). Especially important for ourprediction, they have also been found to put inless effort and to perform at lower levels (Meyer,Paunonen, Gellatly, Goffin, & Jackson, 1989). Indeed,in a meta-analytic review of the influence of com-mitment on work outcomes, Meyer et al. (2002) de-termined that of all types of commitment, affectivecommitment to the organization had the most con-sistent and significant relationship with employees’job performance.

Hypothesis 4. Employees who experience higherlevels of workplace loneliness will have poorer jobperformance, as partially mediated by their lowerlevels of affective commitment to the organization.

The Importance of the Broader Affective AffiliativeContext as a Moderator of the WorkplaceLoneliness–Affiliation Relationship

A basic premise of the affect theory of social ex-change is that the interpersonal context and socialnorms in which the relationship is embedded areimportant to how people evaluate their exchangerelationships and subsequent affiliative responses(Lawler et al., 2008). Complementing this view atthe psychological level, the regulatory loop modelof loneliness (Cacioppo & Hawkley, 2009) suggeststhat lonelier people frequently appraise their en-vironment for potential social threats (Cacioppo &Hawkley, 2009; Cacioppo et al., 2016; Jones, 1982;Weiss, 1973) and appraise other people’s potentialfor making the needed relationships available(Cacioppo et al., 2006; Heinrich & Gullone, 2006;Russell et al., 1980). Therefore, both theorieswouldpredict that the affiliative context of the lonelyemployees is an important factor in explaininghow affiliative they will ultimately feel. Supportingthis, in one of the few examinations of loneliness atwork, Wright (2015) directly theorized that the or-ganizational context of lonelier employees would


influence their subsequent work responses. Draw-ing on these ideas, we examine the employee’saffiliative context, which we operationalize as theemotional culture of the work group and the degreeof loneliness of other coworkers in the employee’swork group.

Work group emotional culture as a moderatorof the employee workplace loneliness–affiliationrelationship. We focus on the role of emotionalculture as amoderator of workplace loneliness andaffiliative attitudes and behaviors—specificallythe emotional culture of companionate love andthe emotional culture of anger. Because bothcompanionate love and anger are other-focusedand affiliation-oriented, both these social emo-tions are relevant to loneliness and its affiliativeoutcomes (Barsade & O’Neill, 2014; Smith &Lazarus, 1993).

Emotional culture of companionate love. Barsadeand O’Neill (2014) defined an emotional culture ofcompanionate love as:

the behavioral norms, artifacts, and underlyingvalues and assumptions reflecting the actual ex-pression or suppression of affection, caring, com-passion, and tenderness, and the degree of perceivedappropriateness of these emotions, transmittedthrough feeling and normative mechanisms withina social unit. (558)

A stronger emotional culture of companionatelove encourages harmonious relationships, a senseof community, and greater affiliative behaviors(Barsade & O’Neill, 2014), which should increaseorganizationalmembers’ capacity to respond to eachother’s relational needs. The fact that workplaceloneliness can exist even in a strong emotional cul-ture of companionate love shows that such a cultureis not a perfect antidote to loneliness. However, thevalues, norms, and caring behaviors in a strong ver-sus weak emotional culture of companionate lovewould offer the expectation of a greater level ofcompassion and tenderness being expressed by em-ployees, thus mitigating, rather than strengthening,the negative cycle predicted by the regulatorymodelof loneliness. There is recent empirical support forthis buffering influence of an emotional culture ofcompanionate love. In a study examining firefighterscopingwith intensework–family conflict (a differenttype of emotional vulnerability), a stronger emo-tional culture of companionate love buffered againstthe ill effects of firefighters suppressing their emo-tions about this conflict and subsequent healthproblems (O’Neill &Rothbard, 2017). In addition, the

normative mechanisms of an emotional culture ofcompanionate love could encourage lonelier em-ployees to appear approachable to better fit into theculture.

We also predict that because employees ina stronger emotional culture of companionate lovebelieve that their colleagues are trying harder toconnect (Barsade & O’Neill, 2014), lonelier em-ployees will have fewer negative attributions abouttheir work group. Under these circumstances, theaffect theory of social exchange (Lawler, Thye, &Yoon, 2008) suggests a weaker relationship betweengreater workplace loneliness and employees’ low-ered affective commitment to the organization asa whole. We predict the opposite for a weaker emo-tional culture of companionate love—a culturecharacterized by indifference and callousnessamongemployees (Barsade&O’Neill, 2014): becauseof the heightened self-protective tendencies andvigilance evoked by a weaker emotional culture ofcompanionate love, there will be a stronger negativerelationship between greater workplace lonelinessand affiliative outcomes.

Hypothesis 5. The emotional culture of compan-ionate love in an employee’s work group will mod-erate the negative relationship between employeeworkplace loneliness and (a) employee approach-ability and (b) employee affective commitment tothe organization. Specifically, a stronger emotionalculture of companionate love will lessen the nega-tive relationship between workplace loneliness andthese affiliative outcomes, whereas a weaker cultureof companionate love will increase the negative re-lationship between workplace loneliness and theseaffiliative outcomes.

Emotional culture of anger. The emotional cul-ture of anger is transmitted in the same way asa culture of companionate love—through feelingand normative mechanisms within a social unit—but involves the expression of irritation, annoy-ance, anger, grumpiness, frustration, and hostility(adapted from Barsade & O’Neill, 2014; see OnlineAppendix C). In a stronger culture of anger—whereexpressing angry feelings is more acceptable thanin a weaker culture of anger—the social vigilance(Jones, 1982), heavily self-protective attitudes, anda negative view of others (Heinrich & Gullone,2006) that are so problematic for lonelier em-ployees will likely increase. We also predict sucha culture to heighten employees’ sensitivity to re-jection (Cutrona, 1982), self-consciousness (Jones,1982), expectations that others will evaluate them


negatively (Jones & Hebb, 2003), and a view ofothers as less trustworthy (Rotenberg, 1994). Allof this will increase the negative cycle present inthe regulatory model of loneliness (Cacioppo &Hawkley, 2009), leading to lower approachability.In linewith the affect theory of social exchange, wepredict that these heightened feelings of a lack ofaffiliation and the attributions that go with thesefeelings will cause employees to feel lower affec-tive commitment to the organization as a whole.We predict that an emotional culture of anger willincrease the negative relationship between greaterworkplace loneliness and lower affiliative out-comes, and that a weaker emotional culture of an-ger will reduce the relationship between thosevariables.

Hypothesis 6. The emotional culture of anger in anemployee’s work group will moderate the negativerelationship between employee workplace lonelinessand (a) employee approachability and (b) employeeaffective commitment to the organization. Specifi-cally, a stronger emotional culture of anger will in-crease the negative relationship between workplaceloneliness and these affiliative outcomes, whereasa weaker culture of anger will reduce the negativerelationship between workplace loneliness and theseaffiliative outcomes.

Coworkers’ loneliness as a moderator of the em-ployeeworkplace loneliness–affiliationrelationship.Another important part of employees’ affiliativeworkplace context is the aggregate loneliness levelof their coworkers. What effect will being sur-rounded by lonelier coworkers have on lonelieremployees? Although one might expect lonelierpeople to have a stronger tendency to approachothers to fulfill their relational needs, the empiri-cal evidence for the regulatory loopmodel suggeststhe opposite: that because lonelier people tend tobehave in less trusting and less positively rein-forcing ways (Rotenberg, 1994), they have a hardertime providing the closeness that interaction part-ners crave. Indeed, two or more people who allfeel lonely have been found to fail at producingmutually satisfying interactions (Jerrome, 1983;Weiss, 1973). Furthermore, as predicted by theaffect theory of social exchange, because of thegreater need to understand and attribute theseaversive feelings, employees surrounded by lonelycoworkers receiving less affiliation will show lessaffective commitment to the organization as awhole. As a result, for employees not getting theiraffiliative needs met because they are surrounded

by lonelier coworkers—coworkers who are alsolikely to be less responsive to these relational needsbecause of their own loneliness—the negative re-lationship is predicted to be even stronger betweengreater workplace loneliness and lower affiliativeoutcomes. We expect the opposite if an employee’scoworkers are not as lonely, since these coworkerswill be better able to meet the affiliative needs oflonely employees (Cacioppo & Hawkley, 2009).

Hypothesis 7. The workplace loneliness level of theother coworkers in an employee’s work group willmoderate the negative relationship between em-ployee workplace loneliness and (a) employee ap-proachability and (b) employeeaffective commitmentto the organization. Specifically, higher levels ofcoworker loneliness will increase the negative re-lationship between workplace loneliness and theseaffiliative outcomes, and lower levels of coworkerloneliness will lessen the negative relationship be-tween workplace loneliness and these affiliativeoutcomes.

METHOD

We conducted a field study using a multirater,multilevel, time-lagged research design based onmultiple sources of data collected from employ-ees, their coworkers, and supervisors in a for-profit private company (Private Company) anda city government of a West Coast U.S. metropolisof approximately 120,000 people (Public Munic-ipality). At Time 1, participants completed mea-sures of self-reported workplace loneliness, allcontrol variables, and emotional culture. Six weekslater, at Time 2, supervisors rated employees’ perfor-mance and employees reported their affectivecommitment to the organization. In addition,employees responded to a coworker survey to rateeach of the other members of their work groupwith regard to their approachability, from a workgroup list we provided based on organizationalrecords.

Sample

We studied a total of 672 employees across 143work groups, and their 114 supervisors. Thissample consisted of the Public Municipality (n 5477 employees across 99 work groups, and their 81supervisors) and the Private Company (n 5 195employees across 44 work groups, and their 33supervisors). The occupational diversity withinthe PublicMunicipality—we surveyed 88 different


positions as varied as clerks, truck drivers, man-agers, engineers, police officers, and many others—contributes to the generalizability of our results.The private for-profit company is a service andmanufacturing outsourcer with over 41 differentpositions, including project managers, accountingspecialists, supervisors, administrative assistants,and material handlers.3 Across the entire sample,54% of participants were male and 41% female(5% did not specify).4 The average age of the re-spondents was 43.32 years (SD 5 10.44), rangingfrom 18 to 71 years. A total of 16.9% of partici-pants reported that their highest level of educa-tion was graduating from high school; 44.2% hadsome college or an Associate (AA) degree; 30.2%had completed a bachelor’s degree; and 8.7% hadcompleted a graduate degree. The average length ofemployees’ organizational tenure was 8.12 years(SD 5 7.21).

All full-time employees in each organizationwhomet the definition of beingpart of awork groupwereinvited to participate in the study. A work group(such as a shift, department, or work unit) was de-fined as a group of three or more employees whowork together, interact with each other on a dailybasis, and have a shared immediate supervisor.At Time 1, surveys were sent to 673 employeesin the Public Municipality and 293 employees inthe Private Company; 477 (response rate 5 71%)employees in the Public Municipality and 195(response rate 5 67%) employees in the PrivateCompany completed the Time 1 survey. Six weekslater, at Time 2, of the employees who completedthe Time 1 surveys, 391 employees in the PublicMunicipality and 167 employees in the PrivateCompany completed the Time 2 survey, yielding

a Time 2 response rate of 82% and 86%, re-spectively.5 All participants gave informed consentto the study and to matching their surveys fromTime1 toTime2,whichwedid via e-mail addressesfor online surveys, and via names for paper–pencilsurveys. To collect performance data at Time 2, wedistributed supervisor surveys to the supervisors ofall employees who were invited to participate atTime 1. In the Public Municipality, 81 supervisorscompleted 555 out of the 673 job performance rat-ings (82%); in the Private Company, 33 supervisorscompleted 223 out of 293 surveys (76%). Amongthe employees who completed both Time 1 andTime 2 surveys, we received completed supervisorjob performance ratings for 342 out of 391 (87%)employees in the Public Municipality and 139 outof 167 employees (83%) in the Private Company.These high response rates for employers and su-pervisors likely stem from the support of the orga-nizations in survey administration, in raisingemployee interest, and in allowing employees tocomplete the survey on work time.

Measures

For every variable in the section below, unlessnoted otherwise, employees were asked to rate thedegree to which they agreed that each of the state-ment items reflected their feelings about their expe-rience in their job and organization (on a five-pointscale, 15 “strongly disagree” through 55 “stronglyagree”).Allmeasures are available in their entirety inOnline Appendix A.

Employee workplace loneliness measure. Wemeasured employees’ workplace loneliness usinga version of the 20-item UCLA Loneliness Scale(Russell et al., 1980), themostwidelyused lonelinessmeasure in the psychological literature. Because3 To test for a possible response bias between the two

organizations, we compared these organizations with re-spect to their employees’ responses to the independent,mediating, dependent, and control variables in our study.There were significant differences between the two orga-nizations with respect to the education level of employees(t5 6.58, df5 595, p, .01), gender (t5 4.87, df5 625, p,.01), tenure (t 5 7.45, df 5 623, p , .01), trait negativeaffectivity (t 5 22.93, df 5 637, p , .01), and employeeapproachability (t5 3.66,df5 610,p, .01). To account forthese differences, we controlled for them, as well as for theorganization type across all of our analyses.

4 Respondents who did not specify gender were not in-cluded in analyses with gender as a control; however,when we conducted all the analyses with those partici-pants, the pattern and significance of results were thesame.

5 To test for a response bias between those who respon-ded only at Time 1 and those who responded at Times 1and 2, we compared the respondents’ demographic char-acteristics, including their education level, tenure in theorganization, overall job experience, and age.We found nosignificant difference between the two groupswith respectto their education level (t5 1.19, df5 595, n.s.), tenure inthe organization (t520.06, df5 623, n.s.), and overall jobexperience (t51.83,df5 620, n.s.). Therewas a significantdifference between the two groupswith respect to their age(t5 2.26, df5 625, p, .05): thosewho participated only inthe Time 1 survey were younger (mean age 5 40.8 years)than those who participated in Time 1 and Time 2 surveys(mean age5 43.3 years). Thus, we controlled for age acrossall of our analyses.


loneliness can vary across domains (Schmidt &Sermat, 1983), and is traditionally measured differ-ently across domains (DiTommaso & Spinner, 1993),to measure workplace loneliness we adapted thismost frequently used scale to fit the work setting: forexample, we changed “I feel left out” to “I feel left outin this organization.”Other items in our version of thescale included “I lack companionship at my work,”“There isnoone Ican turn to in thisorganization,”and“I am no longer close to anyone in this organization.”Byasking respondents to evaluate how they feel abouttheir job andorganization, this scale takes a sentimentapproach to loneliness. Higher ratings on the scaleindicated greater workplace loneliness. Themean forthis self-report measure of workplace loneliness was2.23 (SD5 0.64), with a Cronbach’s a of 0.94.

Employee job performance measures. To mea-sure job performance, we asked supervisors to rateeach employee on his or her degree of effective in-dividual task performance and team member roleperformance (employee teamwork behaviors neces-sary for the good of the work group). We measuredindividual taskperformancewith the following threeitems on the 1–5 scale described above: “This em-ployee satisfactorily completes assigned duties,”“This employee is a good individual contributor,”and “This employee is an effective performer.” Tomeasure the team member role performance, weused Welbourne, Johnson, and Erez’s (1998) four-item scale. This scale measured employee effective-ness in items such as: “Working as part of a team orwork group.”We averaged these two scales together,and the overall job performance mean was 4.08 (SD5 0.70), Cronbach’s a 5 0.93.

Affiliation mediator measures. Work group mem-bers rated each other’s approachability based ona four-item scale adapted from Wiemann’s (1977) af-filiation measure, with the following items: Thiscoworker “can be easily approached by other em-ployees when they need help with their work-relatedproblems,” “can be easily approached by other em-ployees when they need help with their personalproblems,” “is a likeable person,” and “is distant inhis or her personal relations with other employees”(reverse-coded). For each employee, we calculatedapproachability by taking the average of all the rat-ings the employee received from their other workgroup members. The mean approachability ratingwas 3.77 (SD 5 0.43), Cronbach’s a 5 .83. Beforeaggregating coworkers’ ratings for this measure,we assessed within-group homogeneity within co-worker ratings using three frequently recommendedindices: within-group interrater agreement index

(rwg), interclass correlation coefficient (ICC[1] andICC[2]), and averagedeviation (AD). Themeanswere0.81 (for rwg), 0.57 (ICC[1]), 0.86 (ICC[2]), and 0.20(AD). Because the means for all the indices werewithin the range of generally acceptable norms(LeBreton & Senter, 2008), there was support for ag-gregating each employees’ approachability as ratedby other coworkers to the work group level.

We used Allen and Meyer’s (1990) six-item scale tomeasure employee self-reported affective commitmentto theorganization.This scale includessuch itemsas“Ireally feel as if this organization’s problems are myown” and “This organization has a great deal of per-sonalmeaning forme.”Themean for thescalewas3.50(SD 5 0.75), Cronbach’s a 5 .81. Considering the po-tential overlap between items measuring affectivecommitment and workplace loneliness, we assessedthe validity of these two separate constructs by con-ducting a confirmatory factor analysis. The fit indices,includingcomparative fit index (CFI),normed fit index(NFI), and incremental fit index (IFI), showed that thetwo-factor model was a good fit with the data (x2 51768.75,df5 298,CFI5 .96,NFI5 0.95, IFI5 0.96).Allindicators in the model exhibited significant relation-ships (p, .01)with their intendedlatentvariables in thepredicted direction, providing support for constructvalidity (Anderson&Gerbing, 1988). In addition, a two-factor model with workplace loneliness and affectivecommitment as separate factors had a better fit thana one-factor model where all items were loaded ontoa single latent variable (x2 5 2319.82, df 5 299, CFI 5.95, NFI5 0.94, IFI5 0.95, Dx25 551.07, Ddf5 1, p,.01), offering support that affective commitment andworkplace loneliness are two distinct constructs.

Affiliation moderator measures—affective affili-ative context. We used scales from Barsade andO’Neill (2014) to measure the emotional culture ofcompanionate love and the emotional cultureof angerin an employee’s work group. Employees rated thedegree towhichotherworkgroupmembers expressedcompanionate love and anger at work by respondingto a five-point scale, ranging between15 “never” and5 5 “very often.” The emotional culture of compan-ionate love culture scale consisted of four items (af-fection, caring, compassion, and tenderness) witha mean of 2.91 (SD5 .56) and a Cronbach’s a of 0.84.The emotional culture of anger scale included sixitems (such as anger, frustration, hostility, and irrita-tion) with amean of 2.82 (SD5 .43) and a Cronbach’sa of 0.88. The rwg, ICC[1], ICC[2], and AD indices, inrespective order, were 0.73, 0.26, 0.62, and 0.58 for anemotional culture of companionate love, and 0.76,0.31, 0.67, and 0.56 for an emotional culture of anger.


These indices support aggregating the two emotionalculture variables to the work group level.

To measure the workplace loneliness of each em-ployee’s coworkerswithin his or her work group, weused the employee workplace loneliness measuredescribed above. For each employee, we createda coworker loneliness score that reflected the aggre-gation of each employee’s coworkers’ self-reportsof workplace loneliness across the work group (ex-cluding the focal employee). The mean for thismeasure was 2.23 (SD5 .38). The rwg, ICC[1], ICC[2],and AD indices were 0.72, 0.37, 0.75, and 0.53,which offer support for aggregating coworker lone-liness to the work group level.

Control Variables. The following control vari-ables were included in all analyses, and the meansand standard deviations of all the variables can beseen in Table 1.

With regard to the demographic variables reportedearlier, we controlled for both age and gender, sincethey have been shown to relate to loneliness (Schmitt& Kurdek, 1985), and the degree to which others per-ceive a person as being approachable (Kite, Deaux,& Miele, 1991). In addition, we controlled for em-ployee education, since that could be positively cor-related with employee affiliative behaviors andperformance (Ng & Feldman, 2008).

We controlled for organizational tenure, sinceworking longer in an organization might cause orindicate stronger affiliation with that organization(Meyer et al., 2002).

Because of differences between the two organiza-tions, we included organization type as a controlvariable in all analyses (Public Municipality 5 0,Private Company 5 1).

For personality, we focused on agreeableness, ex-traversion, trait negative affectivity and trait positiveaffectivity as these personality traits have been foundto be related to loneliness (Cacioppo et al., 2006;Teppers, Klimstra, VanDamme, Luyckx, Vanhalst, &Goossens, 2013). Controlling for negative affectivityalso helps to determine whether workplace loneli-ness has useful predictive value beyond the negativeoutcomes of generalized negative affect. We measuredagreeableness and extraversion with four-item scalesdeveloped by Donnellan, Oswald, Baird, and Lucas(2006). Cronbach’s a scores were 0.78 and 0.77, re-spectively.Wemeasured trait negative affectivity (TraitNA) and trait positive affectivity (Trait PA) using thePANAS scale (Watson, Clark, & Tellegen, 1988), ratedfrom1–5 (15 “very slightly” through55 “extremely”).Trait NA and Trait PA consisted of 10 items each, witha Cronbach’s a of 0.88 and 0.87, respectively.

Loneliness is a context-specific emotion that var-ies across different life domains (Schmidt & Sermat,1983). Given the empirical evidence for work–lifespillover (Rothbard, 2001), loneliness experiencedoutside of work may be related to our hypothesizedrelationships. Therefore, we controlled for em-ployees’ private-life loneliness in their family, ro-mantic, and social-life domains. We used five-pointscales developed byDiTommaso and Spinner (1993)for each of these domains. The Cronbach’s a for thefamily life (three items), romantic life (two items),and social life (four items) loneliness scales were0.87, 0.94, and 0.96, respectively.

Analytic Method

Because our data involved a multilevel structurewith individual employeesnestedwithinworkgroups,eachofwhichhadone supervisor rating all employees,we usedmixed-effectsmodels (e.g., hierarchical linearmodeling [HLM] [Raudenbush & Bryk, 2002]) andmultilevel structural equation modeling (MSEM) witha latent variable measurement model (Preacher,Zyphur, & Zhang, 2010) to account for potential non-independence of individuals in the same work groupbeing rated by the same supervisor. In each HLManalysis predicting our outcome variables (level 1) weincluded thework group (level 2) as a random effect tocontrol for additional sources of group-level variance.When testing our moderation hypotheses, followingKreft, de Leew, and Aiken (1995), we grand-meancentered all interaction terms. We used MSEM whentesting our mediation hypotheses; and used non-parametricmultilevel bootstrappinganalysis (Bollen&Stine, 1990) to examine the indirect effect ofworkplaceloneliness on performance through the mediating ef-fects of employee approachability (Hypothesis 3) andemployeeaffectivecommitment (Hypothesis4). Forallmodels, we grand-mean centered noncategorical vari-ables (Hofmann & Gavin, 1998) and produced robuststandard errors, offering greater robustness for certaintypes of misspecification. We also checked thesenoncategorical variables for normality in light of thepreviously recommended thresholds and guidelines(Piovesana & Senior, 2018). Some of our variablesshowed high positive or negative asymmetry; specifi-cally, Trait NA, family-life loneliness, romantic-lifeloneliness, and social-life loneliness had skewnessscores above 1.0, and job performance showed askewness score of 20.99. To normalize their distribu-tions, we log-transformed these variables. Given thecentrality of the workplace loneliness variable in ourmodel, and its moderately high positive asymmetry


TABLE1

Mea

ns,StandardDev

iation

s,an

dZero-Ord

erCorrelation

sof

StudyVariables

(n5

672)

Variable

Mea

nSD

12

34

56

78

910

1112

1314

1516

1718

1.Age

43.32

10.4

2.Educa

tion

a2.29

0.85

–.03

3.Male(1

5

Fem

ale,

25

Male)

1.57

0.49

.01

.04

4.Org.T

enure

(years)

8.12

7.21

.41*

*.05

.10*

5.Priva

teCom

pan

yb

0.32

0.46

–.05*

*227

**–.19*

*–.29*

*

6.Agreeablene

ss3.79

0.63

.12*

.15*

*–.27*

*–.01

.02

7.Extrave

rsion

3.10

0.80

–.15*

*.05

–.03

–.10*

.03

.21*

*8.

Neg

ative

Affectivity

1.55

0.53

–.10*

–.17*

*.05

–.02

.12*

–.14*

*–.10*

9.Positive

Affectivity

3.78

0.57

–.05

.14*

*–.04

–.04

.01

.29*

*.37*

*–.24*

*

10.

Fam

ily-Life

Lon

eliness

1.40

0.66

–.02

–.07

.03

.03

.02

–.14*

*–.09*

.21*

*–.25*

*

11.

Rom

antic-Life

Lone

line

ss1.89

1.23

.01

–.06

–.08

.06

.01

–.01

–.08*

.12*

*–.21*

*.33*

*

12.

Soc

ial-Life

Lon

eliness

1.78

0.88

.06

–.02

.03

.06

.05

–.20*

*–.31*

*.21*

*–.31*

*.38*

*.24*

*

13.

Workp

lace

Lon

eliness

2.23

0.64

–.01

–.08

.08

–.03

–.01

–.35*

*–.31*

*.27*

*–.37*

*.20*

*.13*

*.28*

*

14.

Employe

eApp

roacha

bility

3.77

0.43

.05

.12*

–.03

.03

–.14*

*.18*

*.10*

–.14*

*.15*

*–.17*

*209

*–.12*

*–.31*

*

15.

Affec

tive

Com

mitmen

t3.50

0.75

.16*

*.10*

–.02

.13

–.05

.29*

*.20*

*–.23*

*.33*

*–.15*

*214

**–.14*

*–.66*

*.32*

*

16.

JobPe

rforman

ce4.08

0.70

–.05

.16*

*–.12*

*.06

–.05

.13*

*.01

–.11*

*.18*

*–.11*

–.12*

*–.04

–.28*

*.26*

*.33*

*17

.Culture

ofLo

ve2.91

0.56

.06

.04

–.41*

*–.05

.40*

*.29*

*.06

–.04

.17*

*–.08

–.05

–.06

–.31*

*.30*

*.28*

*.17*

*18

.Culture

ofAnge

r2.82

0.43

–.07

–.13*

*.08*

–.01

.03

–.10*

.06

.12*

*–.07

.02

–.01

.02

.16*

*-.17

**–.16*

*–.14*

*–.26*

*

19.

Cow

orke

rLon

eliness

2.23

0.38

–.08

–.07

.15*

*–.05

–.01

–.12*

*.01

.06

–.10*

–.02

.02

–.01

.14*

*–.21*

*–.18*

*–.20*

*–.51*

*.28*

*

a15

HighSch

ool,25

College,3

5Bac

helor’sDeg

ree,

45

GraduateDeg

ree.

b05

Public

Municipality,

15

Priva

teCom

pan

y.*p,

.05

**p,

.01


(skewness score of 0.73) we also log transformed thisvariable. We conducted all analyses with and withoutcontrol variables, and with and without any trans-formations, and the pattern and significance of resultswere the same as those reported here (results availablefrom the first author).

RESULTS

The means, standard deviations, and correlationsamong all variables included in the study are pre-sented in Table 1. Demographic variables were notfound to be correlated to either workplace lone-liness or loneliness in private-life domains. Theworkplace loneliness measure had significant zero-order correlations with affective commitment, em-ployee approachability, and supervisor-rated jobperformance, offering preliminary evidence thatworkplace loneliness is related to how employeesthink, behave, and perform at work.

Test of Main Effect Hypotheses: Does WorkplaceLoneliness Relate to Employee Approachabilityand Employee Affective Commitment tothe Organization?

To test the first hypothesis, we conducted anHLM analysis to examine the relationship betweenself-reported workplace loneliness and employeeapproachability; the results appear in Table 2. Asseen in Column I, controlling for the demographic,trait personality, and private-life loneliness vari-ables, workplace loneliness was significantly nega-tively related to employee approachability (g 52.62, p , .01), supporting Hypothesis 1.

Our second hypothesis predicts a negative re-lationship between workplace loneliness and em-ployee affective commitment to the organization. Asseen in Table 2, Column II, controlling for the de-mographics, trait personality, and private-life lone-liness variables, workplace loneliness significantlynegatively predicted employee affective commit-ment to the organization (g 5 23.76, p , .01), pro-viding support for Hypothesis 2.

Test of Mediation Hypotheses: Does AffiliationMediate the Negative Relationship BetweenWorkplace Loneliness and Job Performance?

To test the mediation effect of employee ap-proachability and employee affective commitmentpredicted in the Hypotheses 3 and 4, we followedPreacher, Zyphur, and Zhang’s (2010) procedure

using 1–1–1 multilevel mediation analysis as thepredictors. Outcomes and mediators were all mea-sured at level 1 (employee level), with a latent vari-able to account for the clustering of employeeswithin work groups. The results of the mediationanalysis can be seen inTable 3.Workplace lonelinesswas significantly negatively related to supervisor-rated employee performance (g 5 270.38, p , .01)(see Column I), offering support for our predictionthat loneliness negatively relates to performance. Aswe predict in Hypotheses 3 and 4, when employeeapproachability (g 5 13.95, p , .01) and employeeaffective commitment to the organization (g 5 7.19,p , .05) were entered into the equation (see ColumnII), both were significant and there was a decrease inthe significance of workplace loneliness (g5239.30,p , .05), providing support for a partial mediation.To directly test the indirect effect of loneliness vari-able through our mediators, we used a bootstrappinganalysis (Preacher&Hayes, 2008) by generating 1,000bootstrap samples and computing 95% biased-corrected confidence intervals (CI). We conductedthis analysis for each mediator separately, as esti-mating confidence intervals for indirect effects inmodels including both mediators was not computa-tionally possible. In each analysis testing the indirecteffect of a given mediator, we included the othermediator as a control variable. Similar to the resultsabove, our results indicated that the indirect effect ofworkplace loneliness on performance via employeeapproachability toward coworkers was 25.81 (CI 5[–12.34, 20.48]) and via employee affective commit-ment to the organization was 224.73 (CI 5 [–43.49,28.43]). In both cases, the indirect effect CI excludedzero, indicating significant mediation.

Test of Moderation Hypotheses: Does AffectiveAffiliative Context Moderate the RelationshipBetween Workplace Loneliness and EmployeeApproachability and Employee AffectiveCommitment to the Organization?

Hypothesis 5 predicts that the emotional culture ofcompanionate love in anemployee’swork groupwillmoderate the negative relationship between work-place loneliness and employee approachability(Hypothesis 5a) and employee affective commitmentto the organization (Hypothesis 5b). Table 4 presentsthe results of HLM analyses with a cross-level in-teraction testing these hypotheses. With regard toemployee approachability (see Column II), there isno significant interaction between workplace lone-liness and an emotional culture of companionate


love,offeringnosupport forHypothesis5a.Withregardto employee affective commitment, there is a signifi-cant interaction between an emotional culture ofcompanionate love and workplace loneliness on af-fective commitment (g 5 0.80, p , .01, Table 4, Col-umn IV), providing support for Hypothesis 5b. Tointerpret this significant interaction, per Aiken andWest (1991), we interpreted the form of the significantinteraction by plotting and computing slopes at high(one standard deviation above themean) and low (onestandard deviation below the mean) values of themoderatingvariable.The results, illustrated inFigure2,indicate that the predicted negative relationship be-tweenworkplace loneliness and affective commitmentis less negative for employees in work groups withstronger cultures of companionate love (g522.34,p,.01) than for employees in work groups with weakercultures of companionate love (g 5 24.62, p, .01) .

Hypothesis 6 predicts that the emotional culture ofanger in an employee’s work group moderates thenegative relationship between workplace lonelinessand employee approachability (Hypothesis 6a) andemployee affective commitment to the organization(Hypothesis 6b). For the model predicting employeeapproachability (seeTable 4, ColumnVI) therewasnosignificant interaction between workplace lonelinessand emotional culture of anger, and thus no supportfor Hypothesis 6a. There was a negative and signifi-cant interaction betweenworkplace loneliness and an

emotional culture of anger predicting employee af-fective commitment (g521.62, p, .01) (see Table 4,Column VIII). The relationship supports Hypothesis6b, and the simple slope analysis results (see Figure 3)indicate that the negative association between work-place loneliness and affective commitment was morenegative for employees in work groups with strongeremotional cultures of anger (g527.18, p, .01) thanfor those in work groups with weaker anger cultures(g 5 22.05, p, .01).

Hypothesis 7 predicts that workplace lonelinessof coworkers in an employee’s work group moder-ates the relationship between an employee’s work-place loneliness and employee approachabilitytoward coworkers (Hypothesis 7a) and employeeaffective commitment (Hypothesis 7b). In the modelpredicting employee approachability (see Table 4,Column X), there was no significant interaction be-tween an employee’s workplace loneliness and theloneliness of an employee’s coworkers in his or herwork group, offering no support for Hypothesis 7a.With regard to employee affective commitment tothe organization, there was a significant negativeinteraction between an employee’s workplace lone-liness and the loneliness of coworkers in the em-ployee’s work group (g 5 21.70, p , .01, ColumnXII). The simple slope analysis results (see Figure 4)indicate that the predicted negative relationshipbetween self-reported workplace loneliness and

TABLE 2Hierarchical Linear Modeling Analyses Testing the Relationships between EmployeeWorkplace Loneliness, and Employee

Approachability and Employee Affective Commitment to the Organization

Employee Approachability Employee Affective Commitment to the Organization

I (n 5 533) II (n 5 566)

Control VariablesAge –.001 .008**Education –.002 .032Male –.017 .039Organizational Tenure –.026 .086Private Company –.174** –.035Agreeableness .018 .044Extraversion –.007 –.044Trait Negative Affectivity –.212 –.138Trait Positive Affectivity –.011 .170**Family-Life Loneliness –.195 –.063Romantic-Life Loneliness –.032 –.170Social-Life Loneliness .108 .287*Predictor VariableWorkplace Loneliness –.62** 23.76**Pseudo-R2 .16 .41

Note: Unstandardized coefficients.*p , .05

**p , .01.


affective commitment was more negative for em-ployees working in work groups where coworkershad higher levels of workplace loneliness (g 525.51, p , .01) as compared to those who were inwork groups with coworkers with lower levels ofworkplace loneliness (g 5 22.26, p , . 01), sup-porting Hypothesis 7b.

Test of Full Moderated Mediation Model

Last, as a final summary analysis,we examined thefull moderated mediation, and whether the indirecteffect of workplace loneliness on job performancevia employee approachability and employee affec-tive commitment varies as a function of the affectiveaffiliative context as defined by the combination ofthe three moderators: (1) emotional culture of com-panionate love, (2) emotional culture of anger, and(3) work group coworker loneliness. Specifically, weexamined and compared the magnitudes of theconditional indirect effects of workplace lonelinesson performance through employee approachabil-ity toward coworkers and employee affective

commitment across work group contexts. We ana-lyzed whether these indirect effects were stronger in“low affiliation” contexts (i.e., weaker emotionalculture of companionate love, stronger emotionalculture of anger, and greater coworker loneliness), ascompared to “high affiliation” contexts (i.e., strongeremotional culture of companionate love, weakeremotional culture of anger, and less coworker lone-liness). For each of these moderating variables wedetermined the higher and lower categories to be onestandard deviation above themean and one standarddeviation below the mean level of this cumulativeaffiliative context-moderating variable, respectively.

We tested our hypothesized relationships by fit-ting a multilevel structural equation model, wherewe simultaneously entered both mediators and thethree moderators. This allowed us to test for mod-erated mediation (i.e., conditional indirect effects)with moderators at level 2, and predictors, media-tors, and outcomes at level 1, following the methodsuggested by Preacher, Rucker, and Hayes (2007),allowing for analyzing moderated mediation inconjunction with indirect effects. The resultsshowed that in work groups characterized by “lowaffiliation,” the indirect effect of workplace loneli-ness on performance (conditional indirect effect 52.38.5; 95% CI5 [-67.58,29.41]; p, .05) was morenegative as compared to that in work groups char-acterized by “high affiliation” (conditional indirecteffect5218.26; 95% CI5 [–34.29,22.24]; p, .05).The difference between these two conditional in-direct effects (difference 5 20.23) was statisticallysignificant (z5 2.19, p, .05). These results supportthe full moderated mediation model.

DISCUSSION

Loneliness research has left no doubt that loneli-ness is a painful and pernicious emotion, and ourstudy offers evidence that its negative outcomes alsoextend into people’s work lives. We develop andtest a workplace model of loneliness, showing sup-port for our hypotheses that greater employeeworkplace loneliness is related to lower job perfor-mance, and that this loneliness–performance re-lationship is mediated by less approachability andlowered affective commitment on the part of lonelieremployees.

The results are more qualified for our hypothe-sized moderators. We found full support for thehypothesized interaction between workplace lone-liness and all aspects of the affective affiliative con-text (an emotional culture of anger, an emotional

TABLE 3Hierarchical Linear Modeling Analyses Testing Media-tion Hypotheses Predicting the Relationship between

Workplace Loneliness and Job Performance,Mediated byEmployee Approachability and Employee Affective

Commitment to the Organization

JobPerformance—Supervisor

Ratings (n 5 477)

I II

Control VariablesAge –0.37* –0.59**Education 2.68 3.14Male 23.81 25.69Organizational Tenure 5.62 6.99Private Company 21.86 1.29Agreeableness 3.09 4.09Extraversion 26.54** 26.95**Trait Negative Affectivity 29.69 –0.29Trait Positive Affectivity 4.25 6.15Family-Life Loneliness 0.84 3.65Romantic-Life Loneliness 28.71 24.74Social-Life Loneliness 9.17 16.38Predictor VariableWorkplace Loneliness 270.38** 239.30*MediatorsEmployee Approachability 13.95**Affective Commitment 7.19*

Note: Unstandardized coefficients.*p , .05

**p , .01.


TABLE4

Hierarchical

Linea

rMod

elingAnalyses

TestingtheMod

erationof

Affec

tive

Affiliative

Con

text

Variables

andEmploye

eW

orkplace

Lon

elinesson

Employe

eApproac

hab

ilityan

dEmploye

eAffec

tive

Com

mitmen

ttotheOrgan

ization

Mod

erationAnalysis

forCulture

ofCom

pan

ionateLov

eMod

erationAnalysis

forCulture

ofAnge

rMod

erationAnalysis

forCow

orker

Lon

eliness

Employe

eApproac

hab

ility

(n5

533)

Affective

Com

mitmen

t(n

556

6)

Employe

eApproac

hab

ility

(n5

533)

Affective

Com

mitmen

t(n

556

6)

Employe

eApproac

hab

ility

(n5

533)

Affec

tive

Com

mitmen

t(n

556

6)

III

III

IVV

VI

VII

VIII

IXX

XI

XII

Con

trol

Variables

Age

–.001

–.001

.007

**.008

**–.001

–.001

.007

**.008

**–.001

–.001

.007

*.007

*Education

–.009

–.009

.029

.034

–.008

–.006

.031

.037

–.008

–.008

.031

.044

Male(1

5Fem

ale,

25

Male)

.010

.009

.085

.088

–.017

–.022

.039

.041

–.013

–.013

.063

.076

Organ

izational

Ten

ure

–.034

–.029

.075

.087

–.024

–.017

.094

.11

–.028

–.027

.091

.10

Priva

teCom

pan

y–.30*

*–.29*

*–.081

–.073

–.17*

–.16*

–.025

–.018

–.18*

*–.18*

*–.022

–.004

Agreeab

leness

.009

.009

.021

.026

.018

.017

.036

.036

.016

.016

.042

.047

Extrave

rsion

–.006

–.006

–.037

–.034

–.007

–.007

–.036

–.039

–.007

–.006

–.036

–.033

TraitNeg

ativeAffec

tivity

–.25*

–.25*

–.19

–.19

–.23*

–.24*

–.14

–.16

–.24*

–.24*

–.20

–.17

TraitPositiveAffectivity

–.006

–.007

.16*

*.15*

–.006

–.003

.16*

*.18*

*–.006

–.006

.16*

*.16*

*Fam

ily-LifeLon

eliness

–.19

–.18

–.056

.006

–.19

–.188

–.076

–.038

–.20

–.20

–.11

–.030

Rom

antic-LifeLon

eliness

–.019

–.019

–.16

–.16

–.027

–.032

–.18

–.18

–.027

–.027

–.16

–.16

Soc

ial-LifeLon

eliness

.079

.070

.27*

.23

.099

.101

.30*

.27*

.091

.090

.29*

.23

PredictorVariable

Workp

lace

Lon

eliness(W

L)

–.60*

*–.60*

*23.61

**23.61

**–.64*

*–.64*

*23.69

**23.67

**–.55*

*–.55*

*23.41

**23.46

**Mod

erators

Culture

ofLov

e.29*

*.29*

*.14*

*.13*

*Culture

ofAnge

r–.15*

–.15*

–.093

–.11

Cow

orke

rLon

eliness

–.30*

*–.30*

*–.23*

–.18*

Interactions:

WL3

Culture

ofLov

e.30

.80*

*W

L3

Culture

ofAnge

r–.60

21.62

**W

L3

Cow

orke

rLon

eliness

–.076

21.70

**PseudoR2

.23

.23

.44

.43

.23

.23

.43

.43

.24

.24

.45

.45

Note:

Unstan

dardized

regression

coefficien

ts.

*p,

.05

**p,

.01


culture of companionate love, and coworker loneli-ness) as it relates to employee affective commit-ment to the organization. However, we found nosignificant effects of the interaction between work-place loneliness and any aspect of the affectiveaffiliative context on employee approachability. Apossible explanation for these results is the manyobstacles lonelier employees face in skillfully en-gaging in compelling affiliative behaviors (Cacioppo& Hawkley, 2005); these hurdles may make it

difficult for lonelier employees to appear approach-able, despite their intentions to do so. However,given the importance of approachability in garneringthe help that improves job performance, future re-search should examine other moderators of em-ployee approachability. Last, we found support forour overall mediated moderation model.

Our results provide support for the two theoreticalframeworks we integrated. Both models predict awithdrawal at work on the part of lonelier employeesthat negatively relates to job performance. The regu-latory loop model focuses on the psychological per-spective of lonelier employees’ affiliativewithdrawal,while the affect theory of social exchange explainsthe interpersonal and contextual aspects of this phe-nomenon. Thus, in addition to drawing on insightsgained from each theory’s unique contributions, webuild on the strength of using these two theories intandem to fully understand the relationship betweenworkplace loneliness and job performance.

Future Research Directions and ManagerialImplications: Reducing Workplace Loneliness

Our results, coupled with the work of others in thisdomain (Heinrich & Gullone, 2006; Weiss, 1973),stronglysuggest that loneliness isnot just an individualphenomenon but a social phenomenon that we showhere relates to jobperformance.Apractical implicationof our results for managers is to consider lonelinessan organizational problem that needs to be tackledto help employees and improve job performance.

FIGURE 2Moderated Relationship between Workplace

Loneliness and an Emotional Culture ofCompanionate Love on Employee Affective

Commitment


Loneliness and an Emotional Culture of Anger onEmployee Affective Commitment


Loneliness andCoworkers’Loneliness on EmployeeAffective Commitment


This recommendation raises interesting questionsfor future research: How can organizations help em-ployees alleviate their feelings of workplace loneli-ness? How can employees alleviate their ownloneliness? How often and how fast does employeeworkplace loneliness change? Examining possible an-tecedents of workplace loneliness, including whetherdifferent typesofcognitiveattributions lead todifferinglevels and types of loneliness, might suggest ways toprevent it in the first place. For example, interestingrecent work examined whether it is “lonely at thetop,”with initial lab and field evidence indicating thatthis truism is false (Waytz, Chou, Magee, & Galinsky,2015; Wright, 2012).

So, what shouldmanagers do to combat lonelinessat work? If organizations can provide timely and ef-fective support for lonelier employees, they can helpbreak the negative cycle of workplace loneliness. Todo so, organizations can apply loneliness interven-tion programs that have been shown to combatloneliness in other life spheres. These interventionshave included social support, opportunities for so-cial interactions, and programs meant to changethe maladaptive social cognitions to which lonelierpeople are prone. Interestingly, a meta-analysis ofthese common intervention strategies to reduceloneliness (Masi et al., 2011) found that the most ef-fective interventions were those targeting malad-aptive social cognitions, such as lonelier people’snegatively biased perceptions of how they are per-ceived or how trustworthy others are. For example,lonelier people have been shown to benefit signifi-cantly from intervention programs that focus onclarifying participants’ needs in friendship, analyz-ing their current social network, setting friendshipgoals, and developing strategies to achieve thesegoals (Stevens, 2001). Importantly, because loneli-ness inhibits the motivation and skills of lonelierpeople to reach out (Cacioppo & Hawkley, 2005;Spitzberg & Hurt, 1989), any intervention requiresconsiderable effort and followup—it is not enough tojust bring lonelier people together with others(Jerrome, 1983). However, it is also important to keepin mind that, due to structural constraints, work-place interpersonal relations and context are oftennondiscretionary; thus, it may make sense for em-ployees to disaffiliate if their coworkers are trulynot a good fit in terms of meeting their needs(Hess, 2006). In this case, leaving the group or orga-nization and finding more suitable socio-emotionalworkplace surroundingsmay sometimes be themostviable solution for an individual employee’s work-place loneliness (although it can still be a damaging

outcome to an organization that would prefer theemployee not leave).

Study Limitations, Contributions, and Conclusion

The results of our study should be interpreted inlight of its limitations. Although our theoretical modeland the time-lagged design suggest the causalityimplicit in our model, we cannot confirm causality.There is a possibility, for example, that poorer per-formance leads employees to be isolated from theircoworkers, leading to greater loneliness, althoughwe would argue that the preponderance of past the-ory and empirical work operates in the other direc-tion.Additionally, as in other affective and relationalphenomena, a causal reciprocal cycle might be atwork (Hareli & Rafaeli, 2008), whereby lonelinessharms performance and poor performance increasesloneliness.

Additionally, while managerial performance rat-ings are a common performance metric used in or-ganizations, it would be useful to analyze further theloneliness–performance relationship by using ob-jective measures, such as sales volume or output.Last, our study analyzed workplace loneliness onlywithin the United States. Because employees in dif-ferent cultures candiffer in theways they experienceand express their emotions (Matsumoto, 1989), ex-amining workplace loneliness cross-culturally seemslikely to yield additional insights.

Its limitationswithstanding, our study contributesto a variety of research domains. First, we contributeto our understanding of the relationship betweenaffect and organizational performance. Since affecthas become recognized as significant to under-standing organizational life (Barsade & Gibson,2007), a growing body of research has examinedthe influence of generalized positive and negativeaffect on workplace performance (Elfenbein, 2007).However, when it comes to understanding the in-fluence of discrete, or specific, types of affect, orga-nizational behavior is still in its early stages (Barsade& Gibson, 2007). By examining workplace loneli-ness, we contribute to understanding workplace af-fect in a more nuanced way.

Second, these findings support a relational per-spective in understanding the workplace in thatthe social context of work significantly shapes em-ployees’ behaviors (Barsade & O’Neill, 2014; Dutton&Heaphy, 2003; Grant & Parker, 2009). Past researchhas had a strong focus on the outcomes of high-quality relationships (Dutton & Heaphy, 2003), andconsidered these relationships as a motivational


factor (Grant & Parker, 2009) that increases prosocialbehaviors, perspective taking, and creativity (Grant& Berry, 2011). Our study also looks at outcomes ofrelationships, but takes the opposite approach byfocusing on the outcomes of a perceived lack ofmeaningful relationships.

Third, looking outside of our field, we contributeto sociological research by offering new insights forthe affect theory of social exchange (Lawler, 2001).Specifically, our results suggest that workplaceloneliness can create an affective filter for employeesin social exchanges by negatively influencingtheir perceptions of coworkers in these exchanges.Our study also contributes to the field of psychology.For decades, myriad psychological researchers havestudied loneliness in a variety of life contexts(Cacioppo & Patrick, 2008), yet have paid scant at-tention to loneliness within work organizations. Forexample, though loneliness was considered suffi-ciently important to earn its own special issue inPerspectives on Psychological Science (Sbarra,2015), no article in that issue gave mention to work-place loneliness. Given our results, understandingthe nature and outcomes of workplace loneliness isuseful for social scientists across fields.

In conclusion, our study shows that the work-place, where people spend so much of their time, isnot immune from the negative outcomes of loneli-ness. We find that loneliness hurts not only thelonelier employees but also their coworkers andtheir organizations. In an era sometimes called bymedia “the age of loneliness” (Monbiot, 2014),a more complete understanding of workplaceloneliness will not only help organizations and thepeople in them, but perhaps help create a healthiersociety as well.

REFERENCES

Aiken, L. S., & West, S. G. 1991. Multiple regression:Testing and interpreting interactions. NewburyPark, CA: Sage.

Allen, N. J., & Meyer, J. P. 1990. The measurement andantecedents of affective, continuance and normativecommitment to the organization. Journal of Occupa-tional Psychology, 63: 1–18.

Anderson, J. C., & Gerbing, D.W. 1988. Structural equationmodeling in practice: A review and recommendedtwo-step approach. Psychological Bulletin, 103:411–423.

Anderson, C. A., & Harvey, R. J. 1988. Brief report: Dis-criminating between problems in living: An exami-nation ofmeasures of depression, loneliness, shyness,

and social anxiety. Journal of Social and ClinicalPsychology, 6: 482–491.

Barsade, S. G., &Gibson, D. E. 2007.Whydoes affectmatterin organizations? Academy of Management Per-spectives, 21: 36–59.

Barsade, S. G., & O’Neill, O. A. 2014. What’s love got to dowith it? The influence of a culture of companionatelove and employee and client outcomes in a long-termcare setting. Administrative Science Quarterly, 59:551–598.

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

Bell, R. A. 1985. Conversational involvement and loneli-ness. Communication Monographs, 52: 218–235.

Bollen, K. A., & Stine, R. 1990. Direct and indirect effects:Classical and bootstrap estimates of variability. So-ciological Methodology, 20: 115–140.

Cacioppo, J. T., & Hawkley, L. C. 2005. People thinkingabout people: The vicious cycle of being a social out-cast in one’s ownmind. InK.D.Williams, J. P. Forgas&W. von Hippel (Eds.), The social outcast: Ostracism,social exclusion, rejection, and bullying: 91–108.New York, NY: Psychology Press.

Cacioppo, J. T., & Hawkley, L. C. 2009. Perceived socialisolation and cognition. Trends in Cognitive Sci-ences, 13: 447–454.

Cacioppo, J. T., Hawkley, L. C., Ernst, J. M., Burleson, M.,Berntson, G. G., Nouriani, B., & Spiegel, D. 2006.Loneliness within a nomological net: An evolutionaryperspective. Journal of Research in Personality, 40:1054–1085.

Cacioppo, J. T., & Patrick, B. 2008. Loneliness: Humannatureand theneed for social connection.NewYork,NY: W. W. Norton & Company.

Cacioppo, S., Balogh, S., & Cacioppo, J. T. 2015. Implicitattention to negative social, in contrast to nonsocial,words in the Stroop task differs between individualshigh and low in loneliness: Evidence from event-related brain microstates. Cortex, 70: 213–233.

Cacioppo, S., Bangee,M., Balogh, S., Cardenas-Iniguez, C.,Qualter, P., & Cacioppo, J. T. 2016. Loneliness andimplicit attention to social threat: A high-performanceelectrical neuroimaging study. Cognitive Neurosci-ence, 7: 138–159.

Cacioppo, S., Grippo, A. J., London, S., Goossens, L., &Cacioppo, J. T. 2015. Loneliness: Clinical import andinterventions. Perspectives on Psychological Sci-ence, 10: 238–249.

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


performance. Journal of Applied Psychology, 93:1082–1103.

Cutrona, C. E. 1982. Transition to college: Loneliness andthe process of social adjustment. In L. A. Peplau & D.Perlman (Eds.), Loneliness: A sourcebook of currenttheory, research, and therapy: 291–309. New York,NY: Wiley.

Derlega, V. J., & Margulis, S. T. 1982. Why loneliness oc-curs: The interrelationship of social psychological andprivacy concepts. In L. A. Peplau & D. Perlman (Eds.),Loneliness: A sourcebook of current theory, re-search and therapy: 152–164. New York, NY: Wiley.

Depue, R. A., & Morrone-Strupinsky, J. V. 2005. A neuro-biological model of affiliative bonding: Implicationsfor conceptualizing a human trait of affiliation. Be-havioral and Brain Sciences, 28: 313–349.

DiTommaso, E., & Spinner, B. 1993. The development andinitial validation of The Social and Emotional Lone-liness Scale for Adults. Personality and IndividualDifferences, 14: 127–134.

Donne, J. 2012. Devotions upon emergent occasions–17meditation (1624) (“Forwhom the bell tolls”). In I. Bell(Ed.), Collected Poetry: 314–316. New York, NY:Modern Library.

Donnellan, M. B., Oswald, F. L., Baird, B. M., & Lucas, R.2006. The Mini-IPIP Scales: Tiny-yet effective mea-sures of the Big Five Factors of Personality. Psycho-logical Assessment, 18: 192–203.

Dutton, J. E., & Heaphy, E. D. 2003. The power of high-quality relationships at work. In K. S. Cameron, J. E.Dutton & R. E. Quinn (Eds.), Positive organizationalscholarship: 263–278. San Francisco, CA: Berrett-Koehler Publishers.

Elfenbein, H. A. 2007. Emotion in organizations: A reviewand theoretical integration. In A. Brief & J. Walsh(Eds.),Academy ofmanagement annals, 1: 315–386.Amsterdam: Elsevier.

Farh, C. I., Lanaj, K., & Ilies, R. 2017. Resource-based con-tingencies of when team-member exchange helpsmember performance in teams. Academy of Man-agement Journal, 60: 1117–1137.

Ferris, D. L., Brown, D. J., Berry, J. W., & Lian, H. 2008. Thedevelopment and validation of the workplace ostra-cism scale. Journal of Applied Psychology, 93: 1348–1366.

Fischer, C. S., & Phillips, S. L. 1982. Who is alone? Socialcharacteristics of people with small networks. In L. A.Peplau&D. Perlman (Eds.),Loneliness:A sourcebookof current theory, research and therapy: 21–39. NewYork, NY: Wiley.

Foa, U. G., & Foa, E. B. 1974. Societal structures of themind. Springfield, IL: Charles C Thomas.

Freeman, L. C. 1992. Filling in the blanks: A theory ofcognitive categories and the structure of social affili-ation. Social Psychology Quarterly, 55: 118–127.

Frijda, N. H. 1994. Varieties of affect: Emotions and epi-sodes, moods, and sentiments. In P. Ekman & R. J.Davidson (Eds.), The nature of emotion: Fundamentalquestions: 59–67. Oxford, U.K.: OxfordUniversity Press.

Gouldner, A. W. 1954. Patterns of industrial bureau-cracy. New York, NY: The Free Press.

Grant, A. M., & Berry, J. W. 2011. The necessity of others isthe mother of invention: Intrinsic and prosocial mo-tivations, perspective taking, andcreativity.Academyof Management Journal, 54: 73–96.

Grant,A.M., &Parker, S.K. 2009. Redesigningworkdesigntheories: The rise of relational and proactive per-spectives.Academy ofManagement Annals, 3: 317–375.

Hackett, R., Bycio, P., & Hausdorf, P. 1994. Furtherassessments of Meyer and Allen’s (1991) three-component model of organizational commitment. Jour-nal of Applied Psychology, 79: 15–23.

Hareli, S., & Rafaeli, A. 2008. Emotion cycles: On the socialinfluence of emotion in organizations. Research inOrganizational Behavior, 28: 35–59.

Hawkley, L. C., Burleson, M. H., Berntson, G. G., &Cacioppo, J. T. 2003. Loneliness in everyday life:Cardiovascular activity, psychosocial context, andhealth behaviors. Journal of Personality and SocialPsychology, 85: 105–120.

Hawkley, L. C., & Cacioppo, J. T. 2010. Loneliness matters:A theoretical and empirical review of consequencesand mechanisms. Annals of Behavioral Medicine,40: 218–227.

Heinrich, L. M., & Gullone, E. 2006. The clinical signifi-cance of loneliness: A literature review. Clinical Psy-chology Review, 26: 695–718.

Hess, J. 2006. Distancing from problematic coworkers. InJ. Harden Fritz & B. Omdahl (Eds.), Problematic re-lationships in the workplace: 205–232. New York,NY: Peter Lang Publishing Inc.

Hill, C. A. 1987. Affiliation motivation: people who needpeople. . .but in differentways. Journal of Personalityand Social Psychology, 52: 1008–1018.

Hoegl, M., & Gemuenden, H. R. 2001. Teamwork qualityand the success of innovative projects: A theoreticalconcept and empirical evidence. Organization Sci-ence, 12: 435–449.

Hofmann, D. A., & Gavin, M. B. 1998. Centering decisionsin hierarchical linear models: Implications for re-search in organizations. Journal of Management, 23:723–744.


Horowitz, L.M., & French,R. 1979. Interpersonal problemsof people who describe themselves as lonely. Journalof Consulting and Clinical Psychology, 47: 762–764.

Horowitz, L. M., French, R., & Anderson, C. A. 1982. Theprototype of a lonely person. In L. A. Peplau & D.Perlman (Eds.), Loneliness: A sourcebook of currenttheory, research and therapy: 183–205. New York,NY: Wiley.

Jerrome, D. 1983. Lonely women in a friendship club.British JournalofGuidance&Counselling, 11: 10–20.

Jones, W. H. 1982. Loneliness and social behavior. In L. A.Peplau&D. Perlman (Eds.),Loneliness:A sourcebookof current theory, research and therapy: 238–252.New York, NY: John Wiley Sons.

Jones, W. H., & Hebb, L. 2003. The experience of loneli-ness: Objective and subjective factors. InternationalScope Review, 5: 41–68.

Jones, W. H., Hobbs, S. A., & Hockenbury, D. 1982. Lone-liness and social skill deficits. Journal of Personalityand Social Psychology, 42: 682–689.

Kamdar, D., & Van Dyne, L. 2007. The joint effects of per-sonality and workplace social exchange relationshipsin predicting task performance and citizenship per-formance. Journal of Applied Psychology, 92: 1286–1298.

Kahn, W. A. 2007. Meaningful connections: Positive re-lationships and attachments at work. In J. E. Dutton &B. Ragins (Eds.), Exploring positive relationships atwork: Building a theoretical and research founda-tion: 189–206. Mahwah, NJ: Lawrence Erlbaum As-sociates Publishers.

Killeen, C. 1998. Loneliness: An epidemic in modern so-ciety. Journal of Advanced Nursing, 28: 762–770.

Kite, M. E., Deaux, K., & Miele, M. 1991. Stereotypes ofyoung and old: Does age outweigh gender? Psychol-ogy and Aging, 6: 19–27.

Kreft, I. G. G., de Leew, J., & Aiken, L. S. 1995. The effect ofdifferent forms of centering in hierarchical linearmodels.MultivariateBehavioral Research, 30: 1–21.

Lawler, E. J. 2001. An affect theory of social exchange.American Journal of Sociology, 107: 321–352.

Lawler, E. J. 2006. The affect theory of social exchange. InP. J. Burke (Ed.), Contemporary social psychologicaltheories: 244–267. Stanford, CA: Stanford UniversityPress.

Lawler, E. J., & Thye, S. R. 1999. Bringing emotion intosocial exchange theory.Annual Review of Sociology,25: 217–244.

Lawler, E. J., & Thye, S. R. 2007. Social exchange theory ofemotions. In J. E. Stets & J. H. Turner (Eds.),Handbookof the sociology of emotions: 295–320.NewYork,NY:Springer.

Lawler, E. J., Thye, S. R., & Yoon, J. 2008. Social exchangeand micro social order. American Sociological Re-view, 73: 519–542.

Lawler, E. J., Thye, S. R., & Yoon, J. 2014. Emotions andgroup ties in social exchange. In J. E. Stets & J. H.Turner (Eds.), Handbook of the sociology of emo-tions, vol. II: 77–101. New York, NY: Springer.

Lawler, E. J., & Yoon, J. 1996. Commitment in exchangerelations: test of a theory of relational cohesion.American Sociological Review, 61: 89–108.

Leary, M. 1990. Responses to social exclusion: Socialanxiety, jealousy, loneliness, depression, and lowself-esteem. Journal of Social and Clinical Psychol-ogy, 9: 221–229.

LeBreton, J. M., & Senter, J. L. 2008. Answers to 20 ques-tions about interrater reliability and interrater agree-ment. Organizational Research Methods, 11:815–852.

Liden, R. C., Wayne, S. J., & Sparrowe, R. T. 2000. Anexamination of the mediating role of psychologicalempowerment on the relations between the job, in-terpersonal relationships, and work outcomes. Jour-nal of Applied Psychology, 85: 407–416.

Masi, C. M., Chen, H. Y., Hawkley, L. C., & Cacioppo, J. T.2011. A meta-analysis of interventions to reduceloneliness. Personality and Social Psychology Re-view, 15: 219–266.

Matsumoto, D. 1989. Cultural influences on the perceptionof emotion. Journal ofCross-CulturalPsychology, 20:92–105.

Mayo, E. 1949. Hawthorne and the western electriccompany: The social problems of an industrializedcivilization. London, U.K.: Routledge.

Mehrabian, A., & Ksionzky, S. 1974. A theory of affilia-tion. Lexington, MA: Lexington Books.

Meyer, J., & Allen, N. 1997. Commitment in the work-place. Thousand Oaks, CA: Sage Publications.

Meyer, J. P., Paunonen, S. V., Gellatly, I. R., Goffin, R. D., &Jackson, D. N. 1989. Organizational commitment and jobperformance: It’s the nature of the commitment thatcounts. Journal of Applied Psychology, 74: 152–156.

Meyer, J. P., Stanley, D. J., Herscovitch, L., & Topolnytsky,L. 2002. Affective, continuance, and normative com-mitment to the organization: A meta-analysis of ante-cedents, correlates, and consequences. Journal ofVocational Behavior, 61: 20–52.

Miles, L. K. 2009. Who is approachable? Journal of Ex-perimental Social Psychology, 45: 262–266.

Monbiot, G. 2014,The age of loneliness is killing us.Guardian, October 14. Retrieved from https://www.theguardian.com/.


https://www.theguardian.com/

https://www.theguardian.com/

Murray, H. A. 1938. Explorations in personality. Oxford,U.K.: Oxford University Press.

Murthy, V. (2017). Work and the loneliness epidemic: re-ducing isolation at work is good business. HarvardBusiness Review. Retrieved from https://hbr.org/cover-story/2017/09/work-and-the-loneliness-epidemic.

Ng, T. H., & Feldman, D. C. 2008. The relationship of age toten dimensions of job performance. Journal of Ap-plied Psychology, 93: 392–423.

Nolen-Hoeksema, S., & Ahrens, C. 2002. Age differencesand similarities in the correlates of depressive symp-toms. Psychology and Aging, 17: 116–124.

O’Neill,O.A.,&Rothbard,N.P.2017. Is loveall youneed?Theeffects of emotional culture, suppression, and work-family conflict on firefighter risk-taking and health.Academy ofManagement Journal, 60: 78–108.

Peplau, L. A., & Perlman, D. 1982. Loneliness: A source-book of current theory, research and therapy. NewYork, NY: John Wiley and Sons.

Perlman, D., & Peplau, L. A. 1981. Toward a social psy-chology of loneliness. In M. Duck & R. Gilmour (Eds.),Personal relationships in disorder: 31–56. London,U.K.: Academic Press.

Piovesana, A., & Senior, G. 2018. How small is big: Samplesize and skewness. Assessment, 25: 793–800.

Preacher, K. J., & Hayes, A. 2008. Asymptotic and resam-pling strategies for assessing and comparing indirecteffects in multiple mediator models. Behavior Re-search Methods, 40: 879–891.

Preacher, K. J., Rucker, D. D., & Hayes, A. F. 2007.Addressing moderated mediation hypotheses: The-ory, methods, and prescriptions. Multivariate Be-havioral Research, 42: 185–227.

Preacher, K. J., Zyphur, M. J., & Zhang, Z. 2010. A generalmultilevel SEM framework for assessing multilevelmediation. Psychological Methods, 15: 209–233.

Qualter, P., Vanhalst, J., Harris, R., Van Roekel, E., Lodder,G., Bangee, M., Maes, M., & Verhagen, M. 2015.Loneliness across the life span. Perspectives on Psy-chological Science, 10: 250–264.

Raudenbush, S. W., & Bryk, A. S. 2002. Hierarchical lin-earmodels: Applications and data analysis method.Newbury Park, CA: Sage.

Rokach, A. 1989. Antecedents of loneliness: A factorialanalysis. Journal of Psychology, 123: 369–384.

Rotenberg, K. J. 1994. Loneliness and interpersonal trust.Journal of Social and Clinical Psychology, 13:152–173.

Rothbard, N. P. 2001. Enriching or depleting? The dy-namics of engagement in work and family roles. Ad-ministrative Science Quarterly, 46: 655–684.

Russell, D., Peplau, A., & Cutrona, C. E. 1980. Concurrentand discriminant validity evidence for the UCLALoneliness Scale. Journal of Personality and SocialPsychology, 39: 472–480.

Sbarra, D.A. 2015. Taking stock of loneliness: Introductionto the special section. Perspectives on PsychologicalScience, 10: 200–201.

Schachter, S. 1959. The psychology of affiliation: Ex-perimental studies of the sources of gregariousness.Stanford, CA: Stanford University Press.

Schmidt, N., & Sermat, V. 1983. Measuring loneliness indifferent relationships. Journal of Personality andSocial Psychology, 44: 1038–1047.

Schmitt, J. P., & Kurdek, L. A. 1985. Age and gender dif-ferences in and personality correlates of loneliness indifferent relationships. Journal of Personality As-sessment, 49: 485–496.

Seers, A. 1989. Team-member exchange quality: A newconstruct for role-making research. OrganizationalBehavior and Human Decision Processes, 43:118–135.

Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C. 1987.Emotion knowledge: Further exploration of a pro-totype approach. Journal of Personality and SocialPsychology, 52: 1061–1086.

Smith, C. A., & Lazarus, R. S. 1993. Appraisal components,core relational themes, and the emotions. Cognitionand Emotion, 7: 233–269.

Solano, C. H., Batten, P. G., & Parish, E. A. 1982. Lonelinessand patterns of self-disclosure. Journal of Personalityand Social Psychology, 43: 524–531.

Spitzberg, B. H., & Hurt, H. T. 1989. The relationshipof interpersonal competence and skills to re-ported loneliness across time. In M. Hojat &R.Crandall (Eds.),Loneliness: Theory, researchandapplications: 157–172. Newbury Park, CA: SagePublications.

Stevens, N. 2001. Combating loneliness: A friendship en-richment programme for older women. Ageing andSociety, 21: 183–202.

Teppers, E., Klimstra, T. A., Van Damme, C., Luyckx, K.,Vanhalst, J., & Goossens, L. 2013. Personality traits,loneliness, and attitudes toward aloneness in adoles-cence. Journal of Social andPersonal Relationships,30: 1045–1063.

Van Dyne, L., Graham, J. W., & Dienesch, R. M. 1994. Or-ganizational citizenship behavior: Construct re-definition,measurement, and validation.AcademyofManagement Journal, 37: 765–802.

Watson, D., Clark, L. A., & Tellegen, A. 1988. Devel-opment and validation of brief measures of posi-tive and negative affect: The PANAS scales. Journal


https://hbr.org/cover-story/2017/09/work-and-the-loneliness-epidemic

https://hbr.org/cover-story/2017/09/work-and-the-loneliness-epidemic

of Personality and Social Psychology, 54:1063–1070.

Waytz, A., Chou, E. Y.,Magee, J. C., & Galinsky, A. D. 2015.Not so lonely at the top? The relationship betweenpower and loneliness. Organizational Behavior andHuman Decision Processes, 1130: 69–78.

Weiss, R. S. 1973. Loneliness: The experience of emotionaland social isolation. Cambridge, MA: MIT Press.

Welbourne, T.M., Johnson, D. E., & Erez, A. 1998. The role-based performance scale: Validity analysis andtheory-based measure. Academy of ManagementJournal, 41: 540–555.

Wesselmann, E. D., Wirth, J., Mroczek, D. K., & Williams,K. D. 2012. Dial a feeling: Detecting moderation of af-fect decline during ostracism. Personality and Indi-vidual Differences, 53: 580–586.

White, E. 2010. Lonely: A memoir. New York, NY: HarperCollins.

Wiemann, J. M. 1977. Explication and test of a model ofcommunicative competence. Human Communica-tion Research, 3: 195–213.

Williams, K. D. 2007. Ostracism. Annual Review of Psy-chology, 58: 425–452.

Wright, S. 2012. Is it lonely at the top? An empirical studyof managers’ and nonmanagers’ loneliness in organi-zations. Journal of Psychology, 146: 47–60.

Wright, S. 2015. Coping with loneliness at work. In A.Sha’ked & A. Rokach (Eds.), Addressing loneliness:Coping, prevention and clinical interventions: 122–134. New York, NY: Routledge.

Hakan Ozcelik ([email protected]) is a Professor of Man-agement at the California State University, Sacramento. Hereceived his PhD from the University of British Columbia.He studies emotions in organizational life, focusing ontopics such as discrete emotions, surface acting, leadership,emotional climate, decision making, cross-cultural com-munication, organizational neuroscience, and utilizingfilm-making to analyze workplace emotions. He also leadsan interdisciplinary platform to explore organizationalwisdom.

Sigal G. Barsade ([email protected]) is theJoseph Frank Bernstein Professor of Management at theWharton School of Business, University of Pennsylvania.She studies the influence of work emotions on organiza-tional outcomes generally, and on performance, creativity,decision making, and negotiations specifically; emotionalcontagion; emotional intelligence; affective diversity; andimplicit emotions in organizations.


mailto:[email protected]

mailto:[email protected]

Copyright of Academy of Management Journal is the property of Academy of Managementand its content may not be copied or emailed to multiple sites or posted to a listserv withoutthe copyright holder's express written permission. However, users may print, download, oremail articles for individual use.

no employee an island: workplace loneliness and job performance€¦ · no employee an island:...

Documents