The Importance of Conceptual Clarity and Methodological Diversity for Studying Confidential Gossip – a Response and Addition to Fan et al. (2021)

When introducing a new concept such as confidential gossip, researchers are essentially treading on new ground (cf. Podsakoff et al., 2016). Therefore, it is important to acknowledge that developing a definition for a new concept, which confidential gossip is, is not an easy task. However, although the part of the concept that refers to ‘gossip’ is not new to the literature, Fan et al.’s essay also lacks an explicit, unambiguous definition of gossip. That is, they write: ‘But the information shared is not very definite, is not official and is based on hearsay. So it’s gossip.’ (p.1652), ‘“formal gossip” would be an oxymoron’, and ‘It is the relational dimensions of gossip that enable us to determine whether it is gossip or not’ (p.1653). These are fragments of what could eventually become a definition of gossip, yet, an explicit definition is missing.

This is problematic because conceptual clarity is essential for scientific progress. For quantitative and qualitative studies alike, without conceptual clarity, it becomes difficult or even impossible to compare findings across studies, for studies to build onto one another, to run meta-analyses or meta-studies, or to integrate findings in theoretical models and across disciplines (Locke & Greenberg, 2003).

Unfortunately, this lack of conceptual clarity is rather characteristic of studies on gossip, and as such, Fan et al.’s essay is no exception. Dores Cruz, Nieper et al. (2021) systematically reviewed 6114 scientific articles on gossip and identified 324 articles that included a definition of gossip. Their review showed that in the gossip literature, vastly different definitions of gossip have been used (as an example, compare Bosson, Johnson, Niederhoffer & Swan’s (2006) definition of gossip as: ‘an exchange of personal information about absent third parties that can be either evaluatively positive or negative’ to Kurland and Pelled’s (2000) definition of gossip as: ‘information about intense third-party social relationships’). Moreover, many articles on gossip did not define the construct at all.

It is obviously problematic if researchers in a scientific field do not agree on how to define the central construct that they are examining. That is, if researchers speak to one another about gossip, but are actually speaking about different things entirely, this is likely to result in a Babylonic confusion of tongues within which it becomes impossible to compare or integrate research findings. For example, one study may show that gossip has a negative impact on organizational processes whereas another study demonstrates that gossip positively affects the same processes, but the studies may have conceptualized gossip in an entirely different way, with the former study conceptualizing it as negative evaluative talk and the latter conceptualizing it as positive as well as negative talk (for details, see Dores Cruz, Nieper, et al., 2021). Such comparisons of “apples” and “oranges” interfere with building a coherent knowledge base on the impact gossip has on organizational processes, and result in research that fails to provide practitioners (e.g., managers, team coaches, organizational employees) with clear insights on the effects of gossip.

To contribute to a solution for this undesirable situation, Dores Cruz, Nieper et al. (2021) distilled, from all the available definitions they found in their systematic review, those elements on which the most consensus exists in the gossip literature. That is, they screened the 324 definitions of gossip they found. This revealed that there is high consensus regarding two essential elements of the gossip definition: (1) the definition refers, implicitly or explicitly, to the gossip triad, comprised of a sender, a receiver, and a target (i.e., the person(s) the information is about), (2) the target of the gossip is absent (i.e., physically absent or not able to access the communicated content at the time of communication. Thus, constructing an integrative definition based on all the definitions available in the scientific literature, Dores Cruz, Nieper et al. (2021) concluded that there is consensus in the gossip literature to define gossip as: ‘a sender communicating to a receiver about a target who is absent or unaware of the content.’ (Dores Cruz, Nieper, et al., 2021, p. 256). It is important to note that this does not imply that this is “the correct” definition of gossip. Rather, it simply is the definition for which there was the most “common ground” within the gossip literature on the basis of the state of that literature in 2021, when the systematic review was conducted.

This definition excludes two elements that are frequently mentioned as relevant to gossip: evaluative valence (i.e., gossip contains ‘positive’ or ‘negative’ information about the target) and informality (i.e., gossip is communication that takes place in an informal rather than formal context). Instead of including these elements in the definition, Dores Cruz, Nieper et al. (2021) propose a dimensional scaling framework, allowing gossip to vary on these dimensions (such that gossip valence can vary from positive to negative and the context in which it occurs can vary from informal to formal). This enables researchers to systematically examine the impact of variation on these dimensions.

An example of how this approach can be applied to the evaluative dimension is the study by Grosser, Lopez-Kidwell and Labianca (2010) in which they examined how relationship ties between employees relate to their engagement in positive and negative gossip. Their results showed that whereas positive gossip is exchanged both in instrumental and in friendship relationships, negative gossip tends to flow only between individuals who share a friendship tie and not between those who are only involved in a work-required instrumental relationship. These findings shed light on differences between gossip with positive and negative valence, and as such form a good basis for developing theory on what these differences could look like.

The ethnographic study by Hallett, Harger and Eden (2009) on gossip in formal school meetings is an example of this approach for exploring the formal versus informal dimension of gossip. Based on their findings, they propose that the greater the formality of the context, the lower the likelihood of gossip, and the higher the likelihood that gossip statements are more ambiguous and indirect. Just as in Grosser et al.’s (2010) study, by proposing how variance on a specified dimension (in this case formality) affects gossip processes and/or the effects of gossip, their findings offer building blocks for developing theory on gossip.

These examples show that allowing for variation, rather than narrowly placing gossip on a particular point on the dimensions of evaluative content and formality, creates the necessary ‘thinking space’ for understanding the different forms that gossip can take and understanding their effects on organizational outcomes.

To connect the confidentiality dimension that Fan et al. (2021) propose with gossip research that is based on a postpositivist perspective, we advocate a similar approach. That is, to systematically examine whether confidential gossip plays a particular role within the communicative constitution of organizations, and whether this role is different from the role of non-confidential gossip or non-gossip communication, a good starting point could be the broad, integrative definition of gossip (i.e., a sender communicating to a receiver about a target who is absent or unaware of the content; Dores Cruz, Nieper, et al., 2021). From there, studies could systematically examine whether gossip that varies on the confidentiality dimension (i.e., gossip ranging from non-confidential to confidential) differentially shapes gossip processes and outcomes, thereby showing that confidentially indeed matters. It is through such systematic comparisons that the claim that confidential gossip has fundamentally different effects than non-confidential gossip could be empirically substantiated. This brings us to the next issue, which is defining confidentiality clearly.

In the previous paragraph, we discussed the importance of clarity about what gossip is. The same applies to clarity about what confidentiality is. Fan et al. (p.1654) state that their ‘(…) fundamental definition of confidential gossip is that it is an activity where gossip and secrecy overlap’, and ‘gossip becomes confidential when it involves marked boundaries’ (p.1655). They also discuss what it is not: ‘confidential gossip is not necessarily or even often about uncovering ’juicy facts’ or the hidden darkness of organization’ (p.1661).

These descriptions of confidential gossip lack clarity because they do not specify whether they pertain to attributes of the gossip sender, the receiver(s), or both parties. That is, if gossip senders aim to keep gossip confidential, is this enough to define gossip as confidential, even if the receiver does not perceive the sender’s intentions? Or is confidentiality an attribute of gossip that is purely ‘in the eye of the beholder’? Furthermore, it should be clarified from whom the content of gossip is to be kept confidential – only from the target or also from others? For the purpose of Fan et al.’s broad phenomenological analyses, these mutable meanings for participants may be adequate, but for future studies based on a postpositivist perspective, to build on Fan et al.’s work, including such qualifiers is essential to formulate clear theoretical propositions and testable predictions regarding the effects of confidential gossip in organizations.

It is noteworthy that some of Fan’s other work provides a more specific definition of confidential gossip, defining it as: ‘A particular form of gossip involving a few selective individuals who engage in highly sensitive and confidential evaluations about absent parties, events, or issues that are important to conceal and keep secret from particular others in their potential to influence outlier thinking and meaning making’ (Fan & Dawson, 2022, pp. 152–153, see also Fan & Grey, 2021). This definition, however, still does not specify whether senders and/or receiver(s) should perceive gossip as confidential for it to qualify as such.

Conceptual clarity regarding what we can and cannot label as confidential gossip as such is important in light of the impact that Fan et al. (2021) ascribe to confidential gossip. For instance, they state that confidential gossip is ‘qualitatively different from gossip in general. Put another way, confidentiality meaningfully changes the character of the content of gossip. The cloaking of secrecy makes it more meaningful, intense, important or otherwise consequential’. However, in this statement, the authors confound the content of gossip with its confidential character, which creates confusion in understanding the role of confidential gossip in organizations. The authors’ statements would, in principle, lend themselves to be translated into clear (and testable) propositions, but to do so, it is essential to make very clear what ‘confidentiality’ means.

To define confidentiality and thereby provide researchers with possibilities to operationalize it in their studies, it would be useful to start by distinguishing gossip senders’ motives for sending confidential gossip and gossip receivers’ perceptions of confidential gossip and their reactions to it, which are discussed interchangeably in Fan et al.’s (2021) article. To start with the first, consider the statement: ‘We might expect confidential gossip to be especially prevalent in particular organizational contexts where open communication is for some reason difficult or discouraged’ (p.1657). This is essentially a proposition regarding the effects of organizational context dimensions on gossip senders’ motivation to engage in confidential gossip. Whereas research questions regarding senders’ motives to engage in confidential gossip are interesting to study, they are very different from the question of how confidential gossip affects ‘group and individual identity recognition’ (a construct Fan et al., 2021, introduce on p. 1657), as this question pertains to gossip receivers’ interpretations of confidentiality and the effects these interpretations have on them. Examining senders’ motivation for sharing confidential gossip as well as examining receivers’ reactions to gossip perceived as confidential are interesting directions for future research to explore, but it is important to distinguish them from one another. This is especially important because sometimes, gossip senders may aim to keep gossip confidential, but this may be unclear to the receiver, or vice versa: gossip intended to be non-confidential by the sender may be understood as confidential by the receiver. Clearly distinguishing senders’ motives from receivers’ perceptions and reactions helps to gain insight into what happens in such situations.

It is equally important that research distinguishes situations in which gossip is confidential with regards to the targeted individual(s) and/or with regards to third parties in the organization. Research has shown that gossip senders tend to send gossip in ways that protect them against possible retaliation from gossip targets (Dores Cruz, Thielmann, et al., 2021). Whether the meaningfulness ascribed to gossip by receivers differs between such gossip and gossip that is kept confidential for non-targets as well, is an interesting research question.

To connect Fan et al.’s ideas on the role of confidential gossip in organizations with research grounded in a postpositivist perspective, we suggest specifying clear propositions regarding: 1) the senders’ motives for keeping gossip confidential and the factors that cause them, 2) the receivers’ perceptions of gossip confidentiality and how these impact their reactions, and 3) whether it matters for whom gossip is kept secret. Such propositions can be based on existing theoretical frameworks, see, for example, Giardini and Wittek’s (2019) theory paper on when people refrain from engaging in gossip, Beersma and Van Kleef’s framework regarding motives driving gossip (Beersma & Van Kleef, 2012; Dores Cruz et al., 2019) and Wu and colleagues’ framework regarding when gossip is likely to be honest or not (Wu et al., 2021). A particularly relevant theoretical framework is the Social Identity Theory of Information Access Regulation (SITIAR, Bingley et al., 2021), which proposes that information access regulation (as it occurs for example via engaging in confidential gossip) shapes shared social identity, explaining why people who share access to information feel a sense of togetherness, and a sense of separation from those who do not have similar access. Testing specific propositions on the impact of confidentiality that build on these frameworks can help to develop theory regarding which aspects of confidentiality shape the effects of gossip, and as such clarify the role confidential gossip plays in organizational life and help dissipate the ‘mystique’ of secrecy. This could help build a connection between Fan et al’s ideas on confidential gossip and research grounded in a postpositivist perspective.

When discussing empirical research on confidential gossip, Fan et al. (2021) draw attention to the ‘formidable methodological barriers’ this entails (p. 1658) due to its elusive character: ‘It emerges, submerges and re-emerges in between social connections and tensions’ (p. 1658). Arguing that the secret character of confidential gossip makes it difficult to observe by outsiders, they suggest that ethnography or participant observation is the most feasible method to study confidential gossip, as it ‘’gets inside’ these interactions, whereas methods relying on retrospective accounts (e.g., interviews) or diary-keeping may fail because of the ephemeral nature of the phenomenon’ (p. 1660-1661).

We certainly agree that ethnographic methods offer an important means to increase understanding of (confidential) gossip. At the same time, however, we feel that Fan et al. (2021) provide a somewhat limited view of the research methods that can be employed to study gossip. In their overview of possible research methods that have been used to study gossip, they mention several methods in passing (i.e. archive studies, diary studies, experimental research, questionnaire studies, online studies, and interview studies, see page 1658), but then they quickly move on to discuss participant observation, quoting Hannerz (1967, p. 45) who stated: ‘probably there is no other way of acquiring knowledge on gossip’.

With regards to developing research methods for gossip research, we would like to extend the perspective Fan et al. (2021) present by pointing to some important developments. First, diary studies have evolved to include experience sampling, where participants answer questions about gossip activities several times per day (Dores Cruz, Thielmann, et al., 2021) and studies have even used wearable Electronically Activated Recorders to obtain naturalistic gossip observations (Robbins & Karan, 2020). These approaches enable fine-grained observations of gossip without employing participant observation and could be especially suitable for capturing data about ephemeral gossip episodes, because they reduce the time between the event and reporting and or even directly record the event. Second, experiments based on hypothetical scenarios (Dores Cruz et al., 2019; Dores Cruz, van der Lee, et al., 2021; Farley, 2011; McAndrew et al., 2007) and behavioral measures (Beersma & Van Kleef, 2011; Dores Cruz et al., 2023; Feinberg et al., 2012; Fonseca & Peters, 2018; Peters & Fonseca, 2020; Samu et al., 2020; Wu et al., 2016a, 2016b) enable testing hypotheses on the causes and consequences of gossip in a controlled setting, demonstrating that gossip plays an important role in organizational processes (i.e., promoting prosocial behavior, building social bonds, exerting power).

Of course, experimental designs cannot completely represent all the conditions that apply to ‘real-life’ human interactions, and therefore they are always to some extent artificial (Berkowitz & Donnerstein, 1982; Ilgen, 1986). Unfortunately, we have all too often experienced that concerns about this artificiality of experimental studies, and the allegedly limited external validity of their results associated with it, can lead researchers to dismiss experimental methods altogether, and instead rely on qualitative methods that are supposed to lead to results that one can generalize to ‘the real world’.

This is unfortunate for two reasons. First, although one might find experiments to ‘score low’ on external validity, they ‘score high’ on internal validity (Ilgen, 1986). That is, experiments do one thing exceptionally well; they provide confidence that a cause-and-effect relationship established in a study cannot be explained by other factors (Cook & Campbell, 1979). Thus, experiments, to a larger or smaller extent, sacrifice mundane realism (the extent to which materials and procedures are similar to events that occur in the real world; Aronson & Carlsmith, 1968), for control. Arguably, in order to fully understand a phenomenon (be it confidential gossip or any other phenomenon), both studies that answer questions related to causality (i.e., what causes the phenomenon to occur and how does it affect other phenomena) as well as studies that answer questions related to the contexts in which the phenomenon occurs and how it takes form in these contexts, are needed. As such, experiments and ethnographic studies can complement one another; both methods allow us to complete different parts of the puzzle.

Second, the idea that qualitative studies by default result in more generalizable knowledge than quantitative studies is incorrect. Whether findings from a study obtained at a certain point in time, in one context, and with a specific sample, generalize to other times, contexts, and samples, is an empirical question, and this applies to findings derived from both qualitative and quantitative studies (Mook, 1983).

In addition, experimental studies that create an “artificial” setting in which participants can engage in (confidential) gossip might help to address some of the ethical complexities that Fan et al. point to (see p. 1659). Whereas real-life confidential gossip is, in its nature, sensitive, laboratory situations allow creating scenarios in which it might be easier to evoke gossip, as being identified as a (confidential) gossiper is less undesirable within an artificial scenario (see, for example, Dores Cruz et al., 2023; Peters & Fonseca, 2020). As such, an experimental approach to studying gossip may overcome ethical difficulties that exist in qualitative approaches.

Therefore, we would advocate a broader methodological perspective on studying confidential gossip than Fan et al. (2021) advise. Many of the propositions that implicitly appear in their essay, such as: ‘confidentiality meaningfully changes the character of the content of the gossip’ (p.1656), ‘the various things gossip ’does’ in organizations are likely to be inflected differently and perhaps heightened when confidentiality is added to the mix’ (p.1656) lend themselves very well to be examined in experience sampling studies, studies in which gossip is recorded using wearable audio recorders, and/or experimental research designs. Systematic testing of the implicit causal mechanisms in the above propositions can add to participant observation studies of confidential gossip; experiments allow building theory on the causal relations that connect confidential gossip to organizational processes and outcomes whereas qualitative, participative research allows for understanding confidential gossip as embedded in the context in which it occurs. A sole focus on either one of these methods would lead to an incomplete understanding of the phenomenon.

With their essay, Fan et al. (2021) put confidential gossip on the “research-map” of organizational scientists and they should be commended for doing so. Yet, their claimed contributions, i.e., that the essay shows that confidential gossip is a distinctive sub-category of gossip that has specific implications for organizations and that it offers a methodological platform for future empirical studies of confidential gossip, deserve some adaptations and nuance when viewed from a postpositivist perspective.

In this essay, we provided suggestions for offering this nuance and adaptations by looking at Fan et al’s arguments from a postpositivist perspective. Specifically, we argued for the importance of clearly defining gossip as: ‘a sender communicating to a receiver about a target who is absent or unaware of the content.’ (Dores Cruz, Nieper, et al., 2021, p. 256). In line with Dores Cruz and colleagues’ recommendations to use a dimensional scaling approach to studying gossip, we further argued that clear predictions on how confidentiality might shape gossip processes and outcomes can be developed by distinguishing between sender motives for, and receiver perceptions of, confidentiality, and by distinguishing whether gossip is kept confidential from just the gossip target or also from other parties. Proposing and testing such predictions can provide confidence that Fan et al.’s conclusion that confidential gossip is a distinctive sub-category of gossip that impacts organizations in different ways than non-confidential gossip does, is warranted.

Additionally, we argued that employing a broader range of research methods than Fan et al. (2021) discussed would greatly add to understanding confidential gossip and its impact on organizations. We hope that researchers interested in confidential gossip will use the recommendations provided here along with Fan and colleagues’ thought-provoking ideas in order to achieve a better comprehension of the role that confidential gossip plays in organizations, and that our paper will spark constructive debate on (confidential) gossip.

Drafted the article: BB

Revised the article critically for important intellectual content: BB, EM, MT, TDC, AN

Final approval of the version to be published: BB, EM, MT, TDC, AN

This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 771391).

The authors declare that they have no competing interests.