Peeling Away the Layers of Workplace Gossip: A Framework, Review, and Future Research Agenda to Study Workplace Gossip as a Dynamic and Complex Behavior

: Workplace gossip, a ubiquitous organizational behavior broadly deﬁned as talking about an absent target in an evaluative way, has received increasing scientiﬁc attention over the past few years. The complexity and dynamism of the workplace gossip phenomenon create challenging research conditions such that studies focusing on the allegedly same type of workplace gossip can produce differential ﬁndings. To address this problem and better align theory and methodology, our manuscript ﬁrst proposes a framework of workplace gossip that accounts for the interdependencies of the context-embeddedness and dynamism of workplace gossip. Guided by this framework, we systematically evaluate extant workplace gossip research, spanning a total of N = 104 empirical research articles. Highlighting current methodological challenges that indicate a neglect of the dynamic nuances and contexts of workplace gossip, we argue that previous organizational research painted an overly simplistic picture of workplace gossip. By looking beyond traditional organizational research designs, we derive recommendations to advance research on workplace gossip and, ultimately, to establish it as a complex and dynamic social interaction behavior.


Introduction
Workplace gossip-evaluative talk between two or more members of an organization about an absent third party of the same organization-is a key component of organizational life [1][2][3][4]. Depending on job rank, gossip accounts for up to 20% of all work emails [5], and more than 90% of the workforce engages in gossip [6]. As a specific form of informal communication, gossip is a vessel for fast information exchange, and for making sense of ambiguous situations, establishing or maintaining social norms, creating or increasing social bonds, or coping with stress [7][8][9]. In light of the new ways of working, these important social functions of gossip gain relevance, as they can help people to navigate and deal with increasingly high degrees of volatility, uncertainty, complexity, and ambiguity (VUCA). The sudden global shift toward remote work due to the Covid-19 pandemic [10] has further highlighted the importance of effective informal communication channels and their use to maintain a sense of belongingness [11][12][13]. This focus has underscored workplace gossip as an important organizational behavior. Indeed, workplace gossip has become a popular research topic among scientific scholars and practitioners, with a substantial increase over the past few years in empirical studies (e.g., [14][15][16][17]).
Notwithstanding its social functions and benefits, previous empirical work has also linked workplace gossip to a range of detrimental consequences for individuals, teams, and organizations as a whole (e.g., decreased well-being, knowledge-hiding behavior, or reduced organizational citizenship behavior; [15,18,19]). Reflecting these differential empirical findings, recent efforts to distill defining features of workplace gossip (e.g., [3,20,21]) and integrate empirical findings (e.g., [22,23]) have illustrated that workplace gossip is a complex construct to study. As a dynamic behavioral process involving at least three people (i.e., the sender, the receiver, and the target of gossip), gossip is highly influenced by context, varies in content (e.g., positive or negative valence), and can thus be conceptualized and assessed in myriad ways (e.g., [22,23]). Recent integrative works in this area [20,22,23] have provided first critical insights into the definitional features, antecedents, functions, and consequences of workplace gossip and revealed critical research gaps. However, we argue that the scholarly conceptualization and empirical understanding of workplace gossip have so far only scratched the surface of the complexity of this social interaction phenomenon.
Our current paper augments knowledge and refines the future research agenda for workplace gossip along three key themes. First, workplace gossip research has always been multidisciplinary, as indicated by Sun et al. [22], who reviewed the gossip literature published in both organizational and psychological journals. However, the workplace gossip literature is also deeply rooted in other research disciplines, such as communication studies (e.g., [24]) or ethnographic studies (e.g., [25]). We believe that, to advance workplace gossip research, mutual understanding, sharing, and combining research approaches across various disciplines are crucial. Thus, in our work, we do not exclude empirical studies based on their research area but rather review and integrate workplace gossip studies across a greater variety of research streams to identify new possibilities for future research.
Second, as a behavioral process, workplace gossip is inherently influenced by the context in which it occurs [26]. An important distinction in this regard concerns dyadic or group interactions. Although recent integrative works have adopted a multilevel perspective of workplace gossip and its correlating variables (i.e., individual, interpersonal, and organizational levels; [22,23]), they have not differentiated between dyadic and group settings. As dyadic and group contexts can differ greatly [27], accounting for the group context is important in reviewing and conducting workplace gossip research. Moreover, our conceptual framework includes the most proximal context surrounding the gossip phenomenon-the conversational context. Since workplace gossip is in its essence a communication behavior, conversational factors, such as the communication channel or setting, can shape the workplace gossip that is exchanged, such as its subtlety or content [24].
Third, a theoretical account of the workplace gossip phenomenon and its subsequent operationalization in empirical work requires a consideration of the temporal features of workplace gossip. Differentiating between various degrees of temporal granularity in measuring workplace gossip may influence our understanding of the workplace gossip phenomenon. Specifically, we propose a theoretically derived conceptual framework of workplace gossip that: (1) accounts for the interplay of contextual conditions at several organizational levels (including a differentiation between dyadic and group contexts); (2) adds the conversational context as an important additional level to account for contextual factors during a gossip conversation; and (3) considers time as a crucial element to capture and differentiate between highly granular behavioral nuances and more coarse accumulations of workplace gossip. We consider these distinctions in a two-dimensional conceptual model that reflects the complexity and dynamism of the workplace gossip phenomenon and allows us to evaluate the alignment of workplace gossip conceptualization and operationalization in previous research.
To ensure that theoretical contributions are advanced and backed up by a strong empirical basis, it is paramount that operationalizations align with the conceptualization of workplace gossip. While [22] provided an extensive summary of workplace gossip measures, a critical and thorough evaluation of how they align with the conceptualization in empirical studies on workplace gossip of any discipline has been missing to date. To this end, we use our conceptual framework of workplace gossip as a theoretical lens to systematically evaluate how the methodologies utilized in the extant research align with the complexity and dynamism of workplace gossip. Synthesizing previous multidisciplinary findings allows us to uncover two current methodological challenges that indicate a neglect of the complexity and dynamism of the behavioral processes of workplace gossip. We provide recommendations to address these challenges and advance innovative and interdisciplinary future research. Ultimately, we aim to encourage scholars to conceptualize and empirically capture workplace gossip as a complex and dynamic organizational behavior embedded in social interactions.

Conceptualizing Workplace Gossip as a Multilayer Social Phenomenon
Workplace gossip is a distinct organizational construct [1,22]. In their review of gossip more broadly (i.e., not specifically workplace gossip), Dores Cruz et al. [20] proposed four defining features of gossip: (1) an interpersonal phenomenon involving at least three people (i.e., the sender, receiver, and target); (2) an absent or unaware target; (3) a positive, negative, or neutral valence of the gossip content; and (4) informality. These four features naturally extend to gossip in the workplace as well. However, at its very core, gossip is defined as talk (e.g., [1,3]) and as "a type of conversation that occurs as a function of social dynamics requiring social interactions among individuals" [28] (p. 76). Thus, an additional fifth defining feature of workplace gossip is that it is, in its essence, an interactive communication and organizational behavior with people actively gossiping [22]. As such, workplace gossip is shaped not only by context [26] but also by temporal factors and can be assessed at different temporal scopes.
To further carve out the complexity and dynamism of workplace gossip, we introduce the granularity and layers of workplace gossip as conceptual dimensions that have important implications for the measurement of workplace gossip. Integrating granularity and layers and their intricate interdependencies, we propose a dynamic, multilayer conceptual framework of workplace gossip (see Figure 1). In this model, we consider workplace gossip along different degrees of granularity: gossip events, gossip episodes, gossip conversations, and accumulated gossip conversations. Furthermore, we view workplace gossip as embedded in interrelated layers: the conversational layer, the individual layer, the interpersonal layer (i.e., further differentiated between a dyadic layer and a group layer), and the organizational layer. Note that there may be more layers of context beyond the organizational boundaries, such as societal culture or the historical time period. Consistent with the purpose of this paper, however, we focus on workplace gossip within an organization and accordingly assume that workplace gossip is embedded within four interconnected layers of context.

Granularity of Workplace Gossip
With the fifth defining feature, we point to the dynamic behavioral dimension of gossip-or rather, the communicative activity of gossiping. If we want to understand gossip as a behavior and the social exchange that informs gossip, we believe it is crucial to dis-

Granularity of Workplace Gossip
With the fifth defining feature, we point to the dynamic behavioral dimension of gossip-or rather, the communicative activity of gossiping. If we want to understand gossip as a behavior and the social exchange that informs gossip, we believe it is crucial to distinguish the degree of granularity at which researchers locate gossip. By granularity, we mean the level of resolution or the thickness of the (microscopic to macroscopic) lens used to investigate workplace gossip. Workplace gossip research has the potential to uncover different behavioral dynamics and interrelations, depending on the degree of granularity. As with other ubiquitous organizational behaviors (e.g., leader-follower interactions; [29]), investigating workplace gossip at a microscopic level (i.e., single statements) provides different insights from investigations of workplace gossip at a macroscopic level (i.e., workplace gossip extent over the past months). Thus, it is crucial to choose a method that provides an appropriate degree of resolution (see Klonek et al. [30] for an extensive review of high-resolution methods).

Gossip Events
Adopting a social interaction lens, the most fine-grained observations of workplace gossip with the highest level of resolution concern gossip events, in which the smallest unit is a single statement that represents a complete thought (cf. sense units [31]). Gossip events may alternate in valence, content, and functions and are essentially the behavioral expressions of gossip, hence gossiping. As an example of a gossip event, consider a situation in which Sarah talks to her colleague Charlie about their absent colleague Ian: "I can't believe that Ian is late again for our meeting. This is so disrespectful". Assessing workplace gossip at the event level could, for example, provide insights into the emergence of gossip behavior, fine-grained sequential patterns of workplace gossip, or workplace gossip emergence and its relation to physiological arousal.

Gossip Episodes
Compared to gossip events, a slightly less granular consideration of workplace gossip concerns gossip episodes [32], which we define as temporal sequences of gossip events during a conversation. Based on their attributional process model of workplace gossip, Lee and Barnes [32] argued that, within a gossip episode, the gossip may be reciprocated by the gossip receiver. Thus, the gossip sender may change within a gossip episode, as the conversational partners take turns gossiping. For example, Charlie might respond to Sarah's gossip statement with: "I know. I'm also kind of annoyed. But I heard that Ian is a single parent and has a lot on his plate right now with finding a long-term babysitter, so I wouldn't be too hard on him". This back-and-forth gossip would be considered a gossip episode. Assessing gossip at the episodic level could, for example, provide insights into how workplace gossip during a conversation may relate to emotional contagion processes in a group conversation.

Gossip Conversations
Gossip events and episodes occur within the larger context of a gossip conversation. During a conversation, many different topics can be discussed. In our example, Charlie and Sarah may switch from topic to topic. After ending this particular gossip episode, they change the topic and talk about their work tasks ("So anyway, what's on your agenda for today?" "Oh, the usual grind, and I have to talk to HR today about my request for time off."), but they may come back to gossiping throughout the conversation about Ian or about other members of their organization ("Speaking of time off, did you hear that Janet was in Finland over the holidays? She went on a dog sledding tour. So cool!").
Investigating gossip conversations can, for example, provide insights into the effects of workplace gossip on the status of the individuals engaging in gossip [33]. Evidently, assessing workplace gossip at the conversational level comes with some information loss. Since gossip can quickly shift between topics, valences, and other communication behaviors, a whole conversation comprising only negative gossip is likely rather uncommon. Although the degree of granularity decreases from gossip events over gossip episodes to gossip conversations, at each of these three levels, a specific gossip situation can be assessed in situ (e.g., by direct observation [25]), retrospectively (e.g., by applying critical incident techniques [34]), or fictionally (e.g., using scenario-based experiments [35]).

Accumulated Gossip
Many studies have assessed gossip as something that manifests or accumulates over time (e.g., over the past six months [36]). In the example above, Sarah and Charlie gossip about their colleagues Ian and Janet. Over time, they may start to regularly discuss their colleagues. At first, these discussions only happen during their private meetings behind closed doors. However, over time, they start to gossip in larger group settings as well (e.g., over lunch) with other colleagues and about a variety of other people (both positively and negatively). Thus, over time, gossip behavior accumulates and yields general states of workplace gossip that may contribute to an overall gossip climate in an organization. As these gossip states are an accumulation of many different gossip situations, they could only be assessed retrospectively (e.g., in the past six months) or in general (e.g., without giving any reference point of time). Accumulated gossip conversations, when combined with the suitable methodology (e.g., longitudinal studies), could provide insights into how a gossip climate develops over time in a team and whether there are key factors that may inhibit the emergence of a gossip climate [22].

Layers of Workplace Gossip
Because it is embedded in the flow of communication between at least two conversational partners, we understand gossip as a dynamic communication behavior. Like other communication activities, gossip is highly influenced by context [21,26,37]. The group communication literature argues that "context is not a single influence, but multiple influences at various levels of abstraction that jointly work together to create influence" [38] (p. 34). Accordingly, previous theoretical and empirical work on gossip has also argued that there is more than one level of context-or what we call gossip layers-that interrelates with workplace gossip (e.g., [3,32,39,40]), meaning that there are multidirectional relationships between the different layers and workplace gossip. Previous integrative work on gossip research has distinguished among organizational, interpersonal, and individual layers (e.g., [22,23]). We further differentiate the interpersonal layer to account for differences between dyadic and group settings. Moreover, we propose an additional layer that reflects the conversational nature of the gossip phenomenon.

Organizational Layer
The organizational layer includes variables such as organizational justice [18], organizational trust [41], or organizational change [26]. In our example, all colleagues work in an organization with many opportunities for informal communication. Since the organization has only recently undergone a change in team constellations and management, shared work norms and values (e.g., punctuality) still need to be formed and maintained. With the regular negative gossip about meeting lateness, the organizational norm of meeting punctuality is slowly established. This relationship can be multidirectional, meaning that workplace gossip can also affect the organizational context at large. Such an outcome would be the case when gossiping becomes a regular daily occurrence across the organization and shapes a "gossipy" climate [23]. This climate could become harmful for the organization at large, for example, by potentially decreasing organizational performance via increasing employee turnover intentions.

Interpersonal Layer-Dyads and Groups
Previous work did not differentiate between interpersonal dyadic contexts and interpersonal group contexts. This omission is problematic because the two contexts (i.e., dyadic vs. group) are distinct and entail major differences [27]. Thus, in our framework, we further subdivide the interpersonal layer into an interpersonal dyadic and an interpersonal group layer. The interpersonal dyadic layer includes phenomena such as relationships among two members of the gossip triad (e.g., friendships between gossipers [42]) or interpersonal behaviors (e.g., receiver-helping behavior toward the sender [35]). The interpersonal group layer includes variables such as group cohesion (e.g., [43]), group norms (e.g., [44]), or group characteristics (e.g., tenure diversity [28]). In our example, Sarah and Charlie are not only colleagues but also workplace friends. Sarah and Ian, however, do not have a close relationship. Sarah trusts Charlie and knows that gossiping negatively about Ian with Charlie does not pose a reputational concern for her, even though negative gossiping is against the existing norms in their work group. She trusts that Charlie will not tell Ian about her gossiping behavior. After regularly engaging in negative gossip together, the workplace friendship between Sarah and Charlie grows stronger, while the overall group norm that negative gossip is socially unacceptable diminishes.

Individual Layer
The individual layer includes individual characteristics, such as personality traits (e.g., trait mindfulness [45]), individual beliefs (e.g., just world beliefs [46]), or gossip motives (e.g., information exchange or negative influence [47]). For example, Sarah may have a higher sensitivity for injustice, which is why she perceived Ian's constant meeting lateness as disrespectful and unjust. She was irritated and wanted to vent her negative emotions.
After the gossip episode about Ian, she may have felt short-term emotional relief.

Adding Another Layer: The Conversational Layer
Extending the previous perspective of gossip as a context-embedded phenomenon (e.g., [23]), we add the conversational layer as the most acute and proximal social context, including information about conversational characteristics. Research has shown that gossip conversations are critically influenced by who and how many people are present in the conversation (e.g., a dyad or a group [6,21]) or by what the mode of communication is (e.g., face to face or over the phone [24]). This suggests that the conversational layer is necessary for understanding gossip and its associations at other layers of the model. To return to the specific gossip example above, in this particular conversation, only Sarah and Charlie were present (i.e., dyadic gossip setting), and they talked face to face during their private meeting (i.e., face-to-face gossip mode).

The Core: Gossiping Characteristics
Embedded in these layers, gossiping happens. Here, we include the characteristics of the gossip that is exchanged, such as valence (e.g., positive or negative), content (e.g., job-related or personal topics), functions of gossip (e.g., emotion venting), and other characteristics, such as gossip veracity. Again, gossip is both shaped by and shapes the various layers of gossip in our model. In our example, even though Sarah and Charlie work in a group with persisting norms against gossiping, they, on the other hand, agree with the organizational value of punctuality, they both share a trusted relationship, and they happen to be in a dyadic, face-to-face setting. Thus, Sarah uses this moment to vent her negative emotions by negatively gossiping about Ian. The negative gossip, as opposed to positive gossip, may strengthen the bond between Sarah and Ian even further.

Methods
Using our proposed conceptual multilayer framework, we conducted a preregistered systematic review of the extant workplace gossip research following the PRISMA guidelines [48] (https://osf.io/ksmqf, accessed on 1 January 2023). We do not intend to summarize and integrate antecedents, functions, and consequences of workplace gossip (for integrative works, see Sun et al. [22] and Wax et al. [23]), nor is it our intention to build consensus toward a broader definition of gossip (see Dores Cruz et al. [20]). Rather, build-ing on and extending these previous works, we seek to gain insights into the interplay of the conceptualization and operationalization of workplace gossip. By contrasting different methodological approaches and the insights gained regarding the layers and granularity of gossip, we aim to identify opportunities for future research and theory development.
First, we conducted a comprehensive search in online literature databases (Web of Science, PsychINFO, and EBSCO) for peer-reviewed journal articles published through 11 January 2023. We searched for the combined keywords related to gossip (i.e., "gossip*", "informal talk", "malicious talk", "personal talk", "rumor", "rumour", "chit*chat", "hearsay", "slander") and work ("work*", "job*", "organization*", "office", "business"). Next, we reviewed the reference lists (backward search) and citing literature (forward search) of the influential articles in the field (articles on workplace gossip or gossip with more than 50 citations based on Web of Science; e.g., [1]) and of recent integrative work (e.g., [22,23]). This search yielded 3069 unique articles. Each article was initially evaluated based on its title and abstract.
We examined the full texts of 342 remaining articles to check whether they met our inclusion criteria. To be included in our review, articles needed to: (1) be empirical (i.e., quantitative or qualitative studies); (2) be published in English; (3) focus on workplace gossip (i.e., investigating workplace gossip as a standalone construct and variable of interest); and (4) investigate samples consisting of adults with work experience or adult student samples in a simulated work context. Further, because there is some conceptual overlap between workplace gossip and related constructs (e.g., workplace rumor; for an overview, see Sun et al. [22]) the articles also needed to meet our working definition of workplace gossip: an exchange of evaluative information at work about colleagues and/or clients who are not present [1].
Ultimately, 104 journal articles comprising a total of 139 individual empirical studies met our inclusion criteria and formed the basis for our review (see Supplementary Material Figure S1 for a flow diagram of the selection process). Note that the number of articles and studies varies due to several multi-study papers, of which some studies from the same papers may have met our inclusion criteria, while others did not. In the following, when stating relative frequencies, we refer to the single studies instead of the articles. Two coders then thoroughly extracted information about the conceptualization and operationalization of workplace gossip and organized the final studies based on our proposed conceptual framework ( Figure 1). Table 1 summarizes the frequencies of the methodologies across the studies. Table 2 provides a summary of all included articles and the applied analytical approach, the assessed degree(s) of granularity of gossip, and the assessed layer(s) of gossip for each included article.  [57,63] Note. WG = workplace gossip, IV = independent variable, DV = dependent variable, LMX = leader-member exchange. Note that the percentages within each category (e.g., time frame) do not necessarily add up, as more than one code (e.g., past and general) could be coded. Note that, for calculating the percentages, we referred to the single studies (N = 139) instead of the (sometimes multi-study) articles.  As shown in Table 1, more than half of the reviewed studies explored the consequences of workplace gossip (e.g., [66,127]), and only about a quarter of the studies examined antecedents of workplace gossip (e.g., [44,56]). Most studies used either time-lagged (e.g., [71,103]) or cross-sectional designs (e.g., [69,115]). Experimental study designs that allowed for causal inferences were scarce (e.g., [17,73]). The few longitudinal studies were mostly qualitative field studies (e.g., [26,60]). Strikingly, the vast majority of the studies relied on surveys to assess workplace gossip (e.g., [106]).
Turning to our proposed conceptual framework, we identified two overarching methodological challenges in existing workplace gossip research. First, the degree of granularity with which workplace gossip has been measured has been overwhelmingly coarse. Second, the understanding of workplace gossip and its associations within and across the different layers has greatly varied.

Captured Granularity of Workplace Gossip in Previous Studies
In terms of the degrees of granularity of workplace gossip, studies that explored specific gossip behaviors remain scarce (see Table 1). Only two studies investigated specific gossip events emerging within a conversation [25,80], and only one study additionally investigated gossip episodes [25]. About one-quarter of the reviewed studies investigated specific gossip conversations (e.g., [17]).
These studies mostly concerned past conversations (e.g., by using the critical incident technique [46]) or fictional conversations (e.g., by using scenario-based experiments [73]). Only rarely did research directly observe conversations in situ (e.g., ethnographic observations [60]). As presented in Table 1, the majority of the studies included in our review, however, measured unspecific gossip behaviors by investigating either accumulated gossip states over a period of time in the past (e.g., [75,99]) or inquiring about general experiences of gossip at work (e.g., [55,59]). While informative, this high level of aggregation and generalization of workplace gossip neglects specific situational differences and transient variations in valence, content, and functions of workplace gossip behavior [26,80]. Relatedly, current research has mostly focused on single valences only, mostly negative gossip, thereby neglecting potential differential effects of different valences. Further, a fifth of the studies did not differentiate at all between positive and negative gossip (e.g., [42]), even though in the majority of the studies, workplace gossip has been conceptualized as both positive and negative (e.g., [1]). Since valence substantially influences whether gossip has beneficial or detrimental effects [14], the high level of aggregation and the focus on single valences may also explain why research has produced contradictory findings (e.g., 'negative-only' gossip leading to higher vs. lower performance [15,54]). Another problem with this approach is that the process of emergence and discontinuation of workplace gossip in conversations has been largely overlooked. We know little about how gossip partners may alternate between different valences and how this alternation is shaped by temporal and contextual factors within and across conversations.

Captured Layers of Workplace Gossip in Previous Studies
In our conceptual framework, we construed workplace gossip as a dynamic social interaction phenomenon that is embedded in multiple layers, with gossip characteristics, such as valence, content, and functions, building the core. As shown in Table 1, almost half of the included studies investigated some gossip characteristics, most of which considered its different valences (e.g., [33]). Few studies captured some content characteristics of workplace gossip. Simpler variations of content assessments are represented by studies that differentiated between job-and non-job-related content (e.g., [85]). In contrast, studies following a qualitative approach often captured gossip content in more complex forms. For example, in their extensive interview study, Vaidyanathan [57] explored the different gossip topics in academic workplaces, with a particular focus on the interplay of ethical violations with the power of the wrongdoers. More commonly investigated workplace gossip characteristics were gossip functions (for a recent summary, see Sun et al. [22]). Few studies went beyond these well-established characteristics of workplace gossip and explored other nuances of gossip. For example, in their two-study paper, Outlaw and Baer [35] investigated the differential effects of gossip interestingness and truthfulness on the receiver's state happiness, perceived relationship with the sender, and helping behavior toward the sender.
Turning to the conversational layer of gossip, as shown in Tables 1 and 2, the majority of past research did not account for differences in conversational contexts. That is, in most studies, it is unclear whether the gossip took place in dyadic, group settings, or both. We further noted a lack of information concerning the communication mode (e.g., whether employees gossiped face to face or in writing), as the majority of the studies included in our review did not clearly state which communication mode was used to gossip. The conversational setting and communication mode, however, can significantly shape the extent, characteristics, and consequences of gossip. Larger group settings may reduce the perceived intimacy and informality, thus affecting the characteristics, extent, and consequences of the gossip exchanged [21]. Moreover, reduced social cues in virtual communication compared to face-to-face communication [129] can affect gossiping behavior. For example, gossip via telephone involves specific linguistic features that demonstrate the purpose and confidentiality of the gossip [24].
The individual layer of workplace gossip has been investigated in more detail, with the majority of studies exploring how workplace gossip is related to individual variables, such as affect (e.g., negative mood [58]), behavior (e.g., performance [70]), behavioral intentions (e.g., motives to gossip [47]), personality traits (e.g., extraversion [118]), or cognition (e.g., self-construal [52]; see Tables 1 and 2). Individual characteristics have been explored for the sender, receiver, and target of the gossip (for an extensive summary, see Sun et al. [22] and Wax et al. [23]).
As workplace gossip is an interpersonal phenomenon, the past, current, and future relationships and interactions between the gossip triad and other involved persons affect the quantity and quality of workplace gossip. For instance, as shown in Table 1, more than half of the included studies accounted for the interpersonal layer. Specifically, a third of the studies captured the dyadic layer of gossip, for example, by measuring the psychological proximity between the sender and the target (e.g., [62]). Further, more than a third of the studies captured the interpersonal group layer, for example, by capturing group cohesion (e.g., [43]) or perceived group norms (e.g., [44]). We also coded network studies as capturing the interpersonal group layer because the results of these network studies are inherently informative of the group, with individual-level values being interpretable only in relation to the group. Notably, in comparison to most studies focusing exclusively on the sender, target, or more recently, the receiver, network studies include the whole gossip triad, thereby capturing the interpersonal dynamics of workplace gossip.
In terms of the organizational layer, previous studies have measured organizational contextual variables, such as job insecurity [55] or organizational tenure diversity [28] (see Table 1). Notably, studies that assessed interpersonal or organizational variables often used rather static, single-source, self-reported constructs (e.g., affiliation intentions [105]) that reflected the individual's perception of the interpersonal or organizational variables [23].
Overall, our findings indicate that most studies have tended to overlook the complex and dynamic behavioral nature of the workplace gossip phenomenon. Notably, as shown in Table 2, the few studies that captured the multiple layers of workplace gossip are mostly studies that moved beyond "boxes-and-arrows" methodological approaches and adopted qualitative (e.g., [60]), mixed-methods (e.g., [35]), or network designs (e.g., [68]). Especially qualitative studies from research fields other than organizational behavior, such as ethnography or communication studies, have depicted a more holistic picture of workplace gossip that captures its complexity by considering multiple layers, as well as its dynamism by considering multiple degrees of granularity.

Future Research Agenda
Overall, our extensive cross-disciplinary review showed that previous research has mostly focused on gossip instead of gossiping, relying on unidirectional "boxes-and-arrows" research models that barely reflect the complexity and dynamism of workplace gossip. Despite past calls for more diverse methodologies and nuanced research (e.g., [39,126]), the extant research is dominated by self-report measures that focus on perceptions of unspecific and accumulated workplace gossip states instead of workplace gossip in actual conversations. This focus is unfortunate considering that workplace gossip is at its core an organizational behavior that occurs in dynamic social interactions [28] and is embedded in different, interdependent layers of context and individual characteristics. In the following, we provide several recommendations for capturing a more complex and dynamic picture of workplace gossip. In keeping with our conceptual framework, we specifically focus on aspects of granularity and layers of workplace gossip.

Rethinking the Sender-Receiver Perspective
Most previous research on workplace gossip has followed the traditional senderreceiver model (especially in the organization and management literature; e.g., [46,98]). The classic sender-receiver model assumes that information is transmitted from a sender to a receiver [130]. From this theoretical stance, communication merely serves as a vessel or carrier to exchange information. In the broader communication literature, this perspective has been criticized as being overly simplistic and mechanical [131]. Aggerholm and Thomsen [132] (p. 175) discuss communication as a complex and dynamic process, whereby organizations are "constituted by complex webs of sense-making activities between groups and individuals whose understandings intersect, clash and interfere with each other." In line with constructivist views, communication is not only a neutral tool to pass on information from one person to another, but it is also through communication that conversational partners make sense of their surroundings and co-construct a shared cognitive understanding (e.g., [133]). Applied to the workplace gossip phenomenon, this perspective suggests that the assumed dichotomy between the gossip sender and the gossip receiver is too simplistic. During a conversation, the roles of sender and receiver shift and likely blur and overlap, as all involved parties might engage in some gossip. Thus, instead of an individual act, research should start to investigate workplace gossip as a joint activity. Instead of differentiating between the sender and the receiver, a great starting point could be to conceptualize them as gossip participants (e.g., [23,89]).

Contextualizing Workplace Gossiping
As individual factors have been explored in much more detail, we encourage future studies to place a stronger emphasis on the other three layers of workplace gossip: the conversational, the interpersonal (dyadic and group), and the organizational layers.

Distinguishing Different Conversational Contexts
Although gossip is a communication behavior that is embedded in a conversation between two or more people, our results indicate that most research has neglected the conversational layer of workplace gossip, including the conversational setting (i.e., dyadic or group) and the conversational mode (e.g., face to face, written, virtual, telephone). An easy fix for this omission, even in traditional self-report studies, would be to give more specific and varying descriptions of reference for gossip situations. For example, when applying a critical incident technique, researchers could instruct participants to think about a specific gossip setting or mode, such as a situation in which they gossiped in a virtual meeting with one other colleague present.

Recognizing (Remote) Work Trends and Its Influence on Workplace Gossip
Several studies have reported that up to one-third of workplace gossip happens via digital technology (e.g., via phone [74]). The global shift toward working from home due to the Covid-19 pandemic and being dependent on remote communication will likely only accelerate this trend. Remote communication provides a special conversational mode. Compared to face-to-face gossip, remote gossip can sometimes require more obvious purpose statements [24], and it can also have more amplified effects than face-to-face gossip due to the technological opportunities to "record" gossip [134]. For example, screenshots of chats or emails can serve as evidence to further support the veracity of the gossip. Likewise, these abilities can also impose a higher risk for the gossiper of the gossip being leaked. Thus, future studies should distinguish between different modes of workplace gossip (e.g., face to face vs. virtual) to unravel their possibly differential effects and account for global trends in modern workplace communication.
Methodologically, tracing or recording virtual conversations can provide rich datasets. At the same time, ethical considerations and privacy concerns regarding both the act of engaging in online gossip and the act of recording actual virtual behavior pose significant barriers to accessing these kinds of data sources. One possibility to avoid this could be to employ group experiments in which participants have to communicate online instead of accessing field data that are already "out there" (see Liu et al. [135] for an application example of collecting and analyzing virtual written team communication data). In doing so, participants can give their informed consent prior to their data being recorded and analyzed. Another possibility would be to analyze publicly available datasets or archival data. For example, Mitra and Gilbert [5] applied natural language processing and content analysis to analyze the use and flow of gossip in work emails using the publicly available Enron email dataset. Taken together, new technology both changes the way we communicate informally and also offers new methodologies to access richer datasets and capture workplace gossip as a complex and dynamic construct.

Operationalizing Workplace Gossip as an Interpersonal Behavior
Future studies should also assess the interpersonal context using multiple sources to obtain a more holistic understanding of workplace gossip. For example, instead of only relying on the individual perception of the relationship by one person involved, studies could capture the perceptions of the relationship of all actors involved by conducting a network study. Network studies offer the opportunity to move beyond dyadic interpersonal relations and provide insights into interpersonal group dynamics, as well as organizational dynamics. Combined with temporal and longitudinal data of a varying degree of granularity, network studies could provide insights into the process through which single gossip statements between close dyads result in a gossip climate at the organizational level. In their network study of the effects of receiving gossip from multiple senders about several targets on the evolution of friendship ties, Estévez et al. [68] offered first inspirations for combining more than one data point of network data and applying a complex contagion model.

Addressing Interdependencies between Gossip Characteristics and Its Layers
Our review revealed that the knowledge about the complex interdependencies between the different gossip layers and gossip characteristics is still limited. For example, previous studies have suggested that gossip dynamics are impacted by the status of the gossip target [57,69,93]-but this effect may be different in different contexts. In a more formal setting, such as a virtual meeting between a group of people (conversational layer) of varying status (individual layer) and with different dyadic relationships with each other (interpersonal layer), this effect may lead to more covert gossiping (gossip characteristics) in that situation, especially when the overall organizational climate is perceived as psychologically unsafe to share one's own opinions (organizational layer).
In another case, varying individual status of gossip participants may not impact gossip behavior over a longer period of time when the relationship between gossip participants is close; the setting is informal, face to face, and dyadic; and the organization has a "gossipy" atmosphere, meaning that even overt, negative gossip is not uncommon. Evidently, gossip layers are interdependent. Although previous studies have already investigated isolated boundary conditions and mechanisms of gossip at different levels, more complex study designs are needed that are able to capture how the type of gossip, the gossip layers, and the temporal dynamics of gossip interact. Qualitative, longitudinal, multi-source, and/or multi-level research designs allow us to capture multiple layers and degrees of granularity of gossip and how they interact. Ethnographic and qualitative studies provide inspiration for how to conduct a longitudinal, qualitative study to gain such a rich dataset (e.g., [26,81]). Although survey studies have made significant contributions to our understanding of workplace gossip, we argue that more direct measures of gossiping are needed for gossip research to move forward. The layperson's understanding of gossip as malicious talk [136] is likely not reflective of what gossip actually looks like in most organizational settings. Within a conversation, a particular gossip statement can be fast and fleeting, is often ambiguous, and is supported by a series of nonverbal behaviors (e.g., gaze direction; [118]). There can also be a mismatch among verbal, nonverbal, and paralinguistic behaviors, implying more delicate and ambiguous nuances of the valence of workplace gossip (e.g., when using sarcasm [80]).
Instead of reflecting open hostility and crude statements, such as "Colleague X is totally incompetent," gossip can often be much more subtle [64]. For example, in a larger and more formal setting (e.g., a meeting), a negative gossip statement might only reveal itself as gossip in combination with a specific accompanying facial expression. Study designs that rely on self-reported gossip are not able to capture these intricate, fine-grained, and dynamic nuances.
Future research should pay more attention to the question of how gossip is encoded and unfolds within a conversation, using methodologies that are able to identify the specific, fine-grained (combination of) verbal, paraverbal, and/or nonverbal behaviors that are reflective of gossip. To access such subtleties in interaction, more diverse study designs and data sources are needed [39]. Looking beyond the boundaries of traditional organizational research and adopting a more diverse approach can be beneficial to capturing more content and context data (for a comprehensive overview of promising qualitative methodologies, see Nyein et al. [137]).
Although many scholars have recurrently voiced concerns about social desirability effects and argued that workplace gossip can only be accurately assessed using self-reports (e.g., [28,46,65]), a look beyond traditional organizational studies toward discourse studies (e.g., [138]) or ethnographic studies (e.g., [25]) proves these assumptions wrong. In these qualitative studies, researchers were able to directly observe and measure gossiping behavior (e.g., [25,60,138]). Notably, the participants still engaged in a large amount of negative workplace gossip, despite knowing that they were being observed or even recorded (e.g., [25]).
Further, research on related constructs, such as small talk, can also offer inspiration for how to directly assess the behavior of interest (e.g., see Van De Mieroop [139] for an analysis of recorded small talk instances in medical settings). For behavioral observation data, several approaches lend themselves to structuring and analyzing the behavioral data. We see great potential for future research using: (1) quantitative and qualitative interaction coding; and (2) automated text analysis.

Applying Interaction Coding to Uncover Granular Processes of Gossiping
Previous research has mostly focused on aggregated gossip of either negative or positive valence [22]. In reality, though, people can have opposite opinions about a person or change their opinion about a person altogether. In the course of a conversation, the topic and gossip target may change rather quickly (e.g., [25]). Consequently, people may shift back and forth between different valences. Indeed, scholars have argued that positive and negative gossip can co-occur (e.g., [28,64]). More granular observational data could provide opportunities to uncover the dynamic nuances, interplay, and processes of the emergence of gossip in a conversation. Again, a look beyond organizational studies toward communication studies could provide rewarding inspiration. For example, Aslan [138] provided nuanced insights into the sequential organization of gossip in conversations by qualitatively analyzing recorded conversations among residents in a retirement home.
Although the first studies have already applied quantitative behavioral coding systems for gossip behavior (e.g., [140]), future research is needed to develop and test valid observer ratings of workplace gossip over a variety of different contexts and samples. Exploratory multi-source multi-method studies could provide detailed initial insights into the complexity and dynamism of gossiping behavior. A reliable coding system could then be used to detect the content (e.g., valence and work-relatedness) or functions (e.g., emotion venting or group norms) of workplace gossip throughout a conversation. Doing so will allow for a fine-grained analysis of the quantity, interplay, and temporal embeddedness of specific gossip statements.

Leveraging (Semi-)Automated Analytical Methods
Automated text analysis based on transcribed conversations that include gossip statements can be used to investigate the emotional tone of gossip conversations or other lexical features that characterize gossip, such as a specific combination of words (e.g., "I do not want to talk negatively about XY, but . . . ") that demonstrate the beginning of a gossip episode (e.g., [5], see also Banks et al. [141] and Hickman et al. [142] for best practices for text preprocessing and analysis). Following this point, we see potential for machine learning approaches to (semi)-automatically detect and code gossip. With newer technology and more computing power at hand, future research may now be able to answer the call for a more process-oriented approach to study workplace gossip more easily [39]. To develop a reliable automated detection system, using insights from in-depth exploratory studies of diverse samples and contexts is indispensable. Doing so would also emphasize the importance and benefits of conducting interdisciplinary research and utilizing insights on workplace gossip that exist outside of one's own discipline.

Conclusions
Workplace gossip research has increased substantially over the past few years. To contribute to a more nuanced understanding of this ubiquitous behavior in organizations, we propose a dynamic, multilayer framework of workplace gossip that integrates the different degrees of granularity of gossip (i.e., gossip events, episodes, conversations, and accumulated states), different layers of workplace gossip (i.e., conversational, individual, interpersonal, and organizational), and different gossip characteristics at the core. We used this framework as a theoretical lens to systematically evaluate the extant empirical workplace gossip research. We found that studies mostly focused on accumulated gossip states and used overly simplistic, static views of gossip. By drawing on multiple disciplines and study approaches beyond traditional organizational research methodologies, we highlight different directions for future research that address the dynamic, complex, and context-embedded nature of workplace gossip as an interpersonal behavior.