Psychological and Hierarchical Closeness as Opposing Factors in Whistleblowing: A Meta-Analysis

Although employees are an important means of detecting and preventing misconducts through whistleblowing, many witnesses choose to remain silent. One reason to remain silent is the discomfort of reporting a colleague. Intuitively, employees should be less likely to report a close or trusted colleague, but a previous review suggests that the opposite may actually be true. However, later studies have shown mixed effects of social closeness on whistleblowing. To gain a better understanding of how social closeness affects whistleblowing, we meta-analyzed 22 experimental studies on intentions to blow the whistle. Overall, the studies show no effect of social closeness on whistleblowing intentions, d =  − 0.21, p = .05. However, when separating the studies by type of closeness, we find that psychological closeness has a negative effect, d =  − 0.46, p < .001, while hierarchical closeness has a positive effect, d = .34, p < .001 on whistleblowing intentions. This means that employees are most likely to report misconduct if the perpetrator is at the same hierarchical level in the organization and not a close or trusted friend. Since close psychological bonds are more likely to develop between employees at the same hierarchical level, the two types of closeness may counteract each other. This dilemma could be part of the explanation why so many witnesses choose to remain silent.


Introduction
Unethical behavior has become a major concern for public and private organizations, with its trickling-down effects amounting to trillions of dollars in losses each year. The Association of Certified Fraud Examiners estimates the total annual fraud losses to be more than $4.5 trillion globally (Association of Certified Fraud Examiners (ACFE, 2020). Over the decades, there have been a number of high-profile scandals such as Enron, WorldCom, Volkswagen, and many other organizations, that have revealed the indispensable role whistleblowers play in exposing wrongdoings. Therefore, it is intuitive to assume that whistleblowers have become an effective tool in uncovering wrongdoings. Indeed, tips provided by employees are the single most effective method to prevent or detect unethical behavior (ACFE, 2020). Despite the fact that employees constitute an important mechanism that can help prevent misconducts through whistleblowing, many of those witnesses choose to remain silent and not to report wrongdoings. According to the 2018 Global Business Ethics Survey, nearly 30% of employees who observed a misconduct in the workplace did not report it (Ethics & Compliance Initiative, 2018).
Organizations are entities in which individuals engage in interpersonal relationships. Workplaces are places where people spend most of their daily lives, interact with others, do things together, and inevitably form friendships with coworkers and supervisors among others (Berman et al., 2002). One would assume that interpersonal closeness is negatively related to the willingness to report a wrongdoing because of the trust and loyalty developed between colleagues (Greenberger et al., 1987). Indeed, individuals are less willing to report a friend (Hess et al., 2019;King, 1997;Waytz et al., 2013). Within organizations, individuals are members of formal (e.g., team, workgroup) and informal (e.g., demographics) groups (Ashforth & Johnson, 2001). Sharing a group membership can create psychological closeness between colleagues and as a result show ingroup favoritism toward the groups of which they are members (Tajfel & Turner, 1979). In addition, workplaces are hierarchical in structure and organized through social relations of power. On the contrary, power affects interpersonal relationships by increasing social distance (Lammers et al., 2012). Power dynamics, which are present in almost every human social relationship (Guinote, 2017) including those formed within work settings, may be an impediment to reporting. Previous research has shown that individuals are less willing to report a high-status wrongdoer than a low-status wrongdoer (Rehg et al., 2008).
Although there is a growing body of literature on whistleblowing, among the most understudied factors is the relationship between the whistleblower and the wrongdoer (Anvari et al., 2019;Bergemann & Aven, 2020;Hess et al., 2019). Given that whistleblowing involves an individual who observes someone else performing an unethical act, there are reasons to believe that the whistleblowing decision is likely to depend on how close or distant the former is to the latter. In a previous meta-analysis (Mesmer-Magnus & Viswesvaran, 2005), and to the authors' surprise, a strong positive correlation was found between closeness to the wrongdoer and whistleblowing intentions. Their findings suggested that the closer the would-be whistleblower is to the wrongdoer, the greater their intention to blow the whistle. They attributed this finding to either the interpersonal closeness or closeness with respect to organizational structure. This finding served as a springboard for our meta-analysis. To elucidate this intriguing finding about the relationship between closeness and whistleblowing intentions, we focused on two different dimensions of closeness that could have different effects: psychological closeness and hierarchical closeness. This distinction is important in whistleblowing research as it might drive reporting intentions in opposite directions.
The aforementioned arguments suggest a need to focus on the closeness between the whistleblower and the wrongdoer to better understand the whistleblowing decision. Therefore, our meta-analysis approaches the phenomenon from the perspective of the relationships resulting from either the psychological closeness, contingent on an intimate relationship or any form of similarity (Gino & Galinsky, 2012), or the hierarchical power distance which is formalized in a hierarchy. In this meta-analytic review, power distance is termed "hierarchical closeness" as power results from the extent to which the whistleblower is hierarchically close to the wrongdoer within the organization (Taylor & Curtis, 2013). In addition, we use social closeness, or inversely put social distance, as a catch-all term to describe how close the whistleblower is to the wrongdoer, either in terms of psychological closeness or in terms of hierarchical closeness. To do so, we focus on experimental studies to allow for causal inferences about the role of social distance/closeness on whistleblowing intentions. Finally, compared to the previous meta-analysis by (Mesmer-Magnus & Viswesvaran, 2005), we identify 20 articles.
The remainder of this research is organized as follows. In the next section, we briefly develop the theoretical arguments regarding the social context of whistleblowing and present the hypotheses to be tested. Following this, we describe our meta-analytic procedures, including the literature retrieval, application of selection criteria, coding procedure, and effect size computation. Third, we present our findings. Finally, we discuss the main findings as well as provide suggestions for future research.

The Social Nature of Whistleblowing
The subject of whistleblowing has received much attention by scholars since the 1980s (Vandekerckhove & Lewis, 2012), making it one of the most debated issues of business ethics literature (Teo & Caspersz, 2011). Whistleblowing is typically defined as "the disclosure by organization members (former or current) of illegal, immoral, or illegitimate practices under the control of their employers, to persons or organizations that may be able to effect action (Near & Miceli, 1985, p. 4). Whistleblowing is a dynamic process that requires at least three social actors: (a) the observer of wrongdoing and would-be whistleblower, (b) the perceived wrongdoer who commits the wrongdoing, and (c) the recipient of the report on wrongdoing. Each of these actors takes actions in response to others (Near & Miceli, 1996). In the following, we focus more closely on the psychological closeness and the hierarchical power distance as the sources of social distance/closeness.

Psychological Closeness
Humans, as social primates, have evolved to live in social groups, spending most of their life socializing with other fellow humans. The strength of social relationships is characterized by interpersonal closeness (Marsden & Campbell, 1984) stemming from feelings of connection between individuals (Gino & Galinsky, 2012), either because they belong to the same social group or because they share even subtle forms of social categorization such as in a minimal group paradigm (Tajfel et al., 1971). To capture the various levels of manifestation of an interpersonal relationship, we conceptually define it as psychological closeness that is "feelings of attachment and perceived connection toward another person or people" (Gino & Galinsky, 2012, p.16).
A central feature of human behavior is people's tendency to view themselves and others in terms of group memberships, through the lens of ingroups (groups they belong) and outgroups (groups they do not belong) (Tajfel & Turner, 1979;Tajfel et al., 1971). A behavioral implication of this social categorization is ingroup favoritism. It is argued that people tend to favor ingroups over outgroups. Group affiliation was particularly important during our ancestral past as it provided immense survival and reproductive benefits and, thus, an evolutionary explanation of ingroup favoritism is plausible (McDonald et al., 2012). Behaviors affected by the ingroup/outgroup categorization have been discerned in many contexts such as preferential treatment (Levine et al., 2005), preferential allocation of resources (Fowler & Kam, 2007), and cooperation (Yamagishi & Mifune, 2009). In addition, people tend to justify (Gino & Galinsky, 2012) or judge less harshly (van der Toorn et al., 2015) negative actions of those they feel close to.
Organizations can be viewed as large social groups where individuals interact and form personal and informal relationships with their colleagues beyond the formal or professional ones. People at work continuously engage in social relationships that vary in depth and quality (e.g., friendships, close relationships, romantic relationships), with their coworkers, supervisors, and others (Campbell & Campbell, 2012). Given how much time individuals spend at work, workplace relationships are considered to be an important part of the organizational life and can influence employees' attitudes and behaviors. The frequent interaction with colleagues helps in the development of trust and social relationships at work. Often, these peer relationships are governed by an unwritten code of loyalty called "wall of silence" that allows for solidarity and protection (Rothwell & Baldwin, 2007), and the disclosure of a colleague's malfeasance can be viewed as an act of betrayal and a breach of trust (Greenberger et al., 1987;Trevino & Victor, 1992) that paves the way for adverse outcomes such as retaliation (Curtis et al., 2021;Mesmer-Magnus & Viswesvaran, 2005) and stigmatization (Pershing, 2003).
Since trust and loyalty are critical components of human relationships, would-be whistleblowers may refrain from reporting colleagues, especially the closest ones, as a response to the fear of being retaliated or stigmatized. In essence, our argument is that psychological closeness is negatively related to whistleblowing.

Power and Status 1
Another factor that influences the decision to blow the whistle is the power distance between the observer of wrongdoing and the wrongdoer. Power is the most fundamental element in the development of any society and an important form of social influence (Russell, 1938) and stems from asymmetries in social relations (Hershcovis et al., 2017). In an organizational context, the distribution of power is formalized in the hierarchical structure (Hofstede, 2001). Whether it stems from a position in an organization, an interpersonal relationship, or an individual characteristic (Ragins & Sundstrom, 1989), power tends to increase social distance affecting interpersonal relationships (Smith & Magee, 2015) including employees' relationships with their supervisors (Clugston et al., 2000;Erdogan & Liden, 2002).
As power distance reflects an understanding of hierarchy and positional authority, we consider power distance as the relative hierarchical distance between two parties. Specifically, we use the definition provided by Taylor and Curtis (2013). The authors termed the hierarchical distance between the wrongdoer and the would-be whistleblower as "power distance" (p.23). Given that hierarchical distance is the degree of closeness within the organizational hierarchy, we, in turn, define it as hierarchical closeness.
Power relations between whistleblowers and wrongdoers are important when investigating the whistleblowing decision process. Prior research suggests that whistleblowing is less likely when it involves wrongdoers who are at higher levels of an organization because of the costs that come with blowing the whistle against these individuals (Dozier & Miceli, 1985;Miceli et al., 1991;Near & Miceli, 1987). For instance, the power of the wrongdoer is crucial in the retaliation process. Individuals who blow the whistle on high-ranking wrongdoers are more likely to be retaliated against (Cortina & Magley, 2003;J.-Y. Lee et al., 2004;Rehg et al., 2008).
The power and status of the wrongdoers also influence whether they are protected or not by the organization. Organizations are less likely to take action against powerful or high-level wrongdoers (Bergman et al., 2002;Near & Miceli, 1995). Their reluctance to sanction powerful employees can be due to their dependence on the wrongdoing or the wrongdoer (Near & Miceli, 1995;Rehg et al., 2008) as some wrongdoings are often committed by corporate executives on behalf of the organization. However, organizations may also be more willing to take action against high-status perpetrators because of the potential organizational damage coming from a high-profile incident (Bergman et al., 2002).
Social influence theories may explain whistleblowing when the perpetrator is a superior. Prior research has shown that obedience pressure coming from people in positions of authority has an adverse effect on individuals' judgments and decisions (Lord & DeZoort, 2001) and it is highly likely to explain whistleblowing decisions (Taylor & Curtis, 2013).
Taking the above into consideration, we suggest that the lesser the hierarchical closeness, the higher the intentions of the observer to blow the whistle.

Methods
Our meta-analysis incorporates experimental studies that investigate how the social closeness between the wrongdoer and observer of wrongdoing influences the latter's willingness to blow the whistle. To this end, the following steps were taken: (a) identification of articles through database searching, (b) definition and application of inclusion criteria to screen out non-eligible studies, (c) coding of study characteristics, (d) calculating effect sizes, (e) conducting meta-analysis, (f) conducting moderator analysis, and (g) estimating the publication bias.

Literature Search
In order to identify as many relevant studies as possible, five complementary search strategies were employed. First, we examined seven electronic databases: (1) Web of Science, (2) Scopus, (3) Business Source Premier, (4) ABI/INFORM Global, (5) Proquest, (6) PsychInfo, and (7) Google Scholar, using search terms related to (1) whistleblowing and (2) closeness. We applied a search strategy using the terms (whistleblow* OR whistle-blow* OR "blow* the whistle" OR "peer report*") AND (closeness OR cohes* OR relation* OR friend* OR likeab* OR distance). Google Scholar results were retrieved using the Publish or Perish software (Harzing, 2007). Second, we reviewed the references cited in relevant reviews (Culiberg & Mihelič, 2017;L. Gao & Brink, 2017;G. Lee & Xiao, 2018) and meta-analysis (Mesmer-Magnus & Viswesvaran, 2005). Third, we emailed prominent scholars in the field of whistleblowing asking them if they were aware of any unpublished studies. Fourth, we emailed requests to professional organization listservs (e.g., Society for Judgment and Decision Making, European Association for Decision Making) seeking for any unpublished manuscripts. Finally, using a two-way "snowballing" technique, we manually searched all references cited in the retrieved studies (backward searching) as well as all articles citing the original studies using Google Scholar and Web of Science (forward searching). These five search strategies yielded a final sample of 20 articles that resulted in 22 studies and 23 effect sizes. The publication window ranged from 1972 to May 2022. We chose 1972 as a starting point for the literature search because all literature on whistleblowing traces whistleblowing in an organizational context to 1972 (Vandekerckhove, 2006, p.8).
Our literature search resulted in 4322 articles, which was then narrowed to 2590 after deduplication. After reviewing each article's title, abstract, and keywords, 2538 articles were excluded and 56 underwent full-text review. A total of 20 articles (16 published, 4 unpublished) comprising a total of 4316 participants met the minimum quality criteria and were examined in detail. These articles are marked with an asterisk in the reference section. An overview of the study selection process is summarized in the PRISMA (Moher et al., 2009) flow chart in Fig. 1. Table 1 lists the studies that were included in this meta-analysis, their effect sizes, the conceptualization and operationalization of closeness in each study, study location, and a brief description of the whistleblowing context.

Study Selection Criteria
All three authors followed three inclusion criteria to determine the relevance of retrieved studies for research purposes. First, to ensure causality, we only included experimental studies, in which participants were randomly assigned to a "close" condition and compared to a "distant" condition. Non-experimental studies (e.g., correlational or qualitative) were excluded. Second, the studies must have provided sufficient statistical information to compute effect sizes of interest. Third, the studies must have been written in English. Given that our research question concerns the relationship between whistleblower and wrongdoer, we excluded studies on whistleblower-organization relationship.

Coding Procedure
The first two authors coded the means, standard deviations, and sample sizes of both experimental conditions for each observation to calculate Cohen's d. For each study, the coders also coded the following information: (a) form of closeness (psychological closeness or hierarchical closeness), (b) type of research participants (students or professionals), (c) publication status (published or unpublished) (d) cultural country scores. The agreement between coders was very good as indicated by the high intraclass correlation coefficient (ICC > 0.81) for the continuous variables and the high Cohen's kappa (κ = 1) for the categorical variables. Any discrepancies were resolved by reaching a consensus through discussion.

Effect Size Calculation
All statistical analyses were performed using the opensource statistical software (R Core Team, 2020). If authors did not provide Cohen's d, reported data (means and standard deviations, F-values, or beta coefficients) extracted from the selected papers was transformed to Cohen's d with the R package "compute.es" (Del Re, 2013). Experimental group sample sizes were calculated from reported sample sizes. If group sample sizes were not reported, we divided the total sample into equal groups, even if this resulted in fractional sample sizes. When a single study provided multiple measures of the dependent variable, we implemented the Borenstein et al., (2009, Eq. (24.4)) method to calculate the aggregate dependent effect sizes prior to meta-analysis, so that each study contributes only a single effect size to the meta-analysis. This aggregation method is recommended as it was found to be the least biased and most precise (Hoyt & Del Re, 2018).

Weighting of Effect Sizes
For the meta-analysis of the selected studies, we used a random-effects model, which we deemed to be more appropriate than the alternative, a fixed-effect model. The fixed-effect model assumes that all studies included in the meta-analysis share a common true effect size and differences between observed effect sizes are due to sampling error, an unrealistic assumption. A random-effects model accounts for the heterogeneity of effect sizes and hence, results allow for more generalizable conclusions (Field, 2001;Hedges & Vevea, 1998;Hunter & Schmidt, 2000). Research has shown that the Hedges-Vevea method (Hedges & Vevea, 1998) works better for standardized mean differences (Marín-Martínez & Sánchez-Meca, 2010). Therefore, we employed the Hedges-Vevea method in which the study-level effect sizes are weighted by inverse variances. Meta-analysis was conducted using the "metafor" R-package (Viechtbauer, 2010).

Culture as a Moderator
Considering that cross-cultural differences in moral values can be a source of difference in ethical judgment and decision-making (Graham et al., 2011;Haidt et al., 1993), we further chose to examine national culture as an additional moderator of the relationship between closeness and whistleblowing. People from collectivist cultures might be less willing to blow the whistle because loyalty, a core value that binds people together, is deemed important in collectivist societies (Dungan et al., 2015). Indeed, cross-cultural differences in perceptions of whistleblowing exist, with people from collectivist cultures viewing whistleblowing less favorably than people from individualist cultures (e.g., Chiu, 2003;Tavakoli et al., 2003). For the purpose of this meta-analysis, we extracted cultural country scores from the Hofstede Insights Country Comparison database (Hofstede Insights, 2022). When a study did not provide information about the country, we asked the authors for additional information, or if this was not possible, we used the first author's institutional affiliation to determine the country (Kling et al., 1999). Due to the number of moderator analyses (2), a Bonferroni correction was applied to correct the p-values. After the Bonferroni correction, a p-value of < 0.025 was considered statistically significant.

Publication Bias
A common concern for any meta-analysis is the publication bias, often referred to as the "file drawer problem" (Rosenthal, 1979). Publication bias is present when the probability that a study is submitted to journals or accepted for publication is contingent to the magnitude, direction, or significance of the study's results. For example, studies with statistically significant findings are more likely to be submitted or published than studies with non-significant results and thus, these studies would provide a biased view of the actual effect size. Publication bias was evaluated using the contour-enhanced funnel plot (Peters et al., 2008) and the Egger's test (Egger et al., 1997). Publication bias was further assessed by conducting the p-curve analysis (Simonsohn et al., 2015(Simonsohn et al., , 2014a. If publication bias was found, the trim-and-fill method (Duval & Tweedie, 2000) was applied, providing effect sizes adjusted for publication bias.

Results
Our meta-analysis followed a hierarchical breakdown analysis (Hunter & Schmidt, 2004). First, a global model of all studies used in this meta-analysis was analyzed, followed by a hierarchical moderator analysis with closeness as a moderator. The results are presented in Table 2. The results of the global model revealed a significant, overall effect size, d = − 0.21. This finding suggests that the closer to the wrongdoer one is, the less likely one is to blow the whistle. In addition, there was a substantial heterogeneity between studies, I 2 = 91.2%. Breaking down by the moderator variable slightly reduced the heterogeneity for the studies on psychological closeness, I 2 = 86.0%, and completely for those on hierarchical closeness, I 2 = 0%. At a moderator level, effect sizes were significantly different between the two subgroups of studies. The hierarchical breakdown analysis revealed a significant effect for psychological closeness, such that feeling closer to the wrongdoer makes one less willing to report him/her (d = − 0.46), and a significant effect for hierarchical closeness, such that being hierarchically closer to the wrongdoer increases reporting intentions (d = 0.34). A forest plot for each subgroup is displayed in Fig. 2. The results of a meta-regression analysis revealed no moderating effect of culture (p = 0.39), indicating that the cultural background (individualist vs. collectivist) does not influence the relationship between closeness and whistleblowing decision. Publication bias was assessed using a contour enhanced funnel plot, the Egger's test, and a p-curve analysis. A visual assessment of the contour funnel plot revealed no asymmetry indicating that there is no publication bias (Fig. 3). The Egger's test confirmed the absence of publication bias (p = 0.80). As a final robustness check we conducted a p-curve analysis. The p-curve results based on the significant right-skewness test (z = − 9.365, p < 0.001) and the non-significant flatness test (z = 6.367, p > 0.999) suggested that the included studies reflected a true effect rather than publication bias or selective reporting (e.g., p-hacking) (Fig. 4). Taken the results of all tests together, there was no evidence of publication bias and therefore no further adjustment of effect size was needed.

Discussion
The closeness of the wrongdoer to the whistleblower might be an exhibition of the personal power of the former over the latter . The view of (Mesmer-Magnus & Viswesvaran, 2005, p.291) who stated that "the closer the observer was to the wrongdoer, whether interpersonally or with respect to organizational structure, the more likely it was that he or she intended to blow the whistle", served as a springboard for our meta-analysis. In order to understand the direction and magnitude of the effect of closeness on whistleblowing, it is important to further explore it, as the way it is operationalized can potentially drive reporting intentions in opposite directions. People are less likely to report a wrongdoer when he or she is a close other compared with a distant other. On the other hand, the relative hierarchical position between the whistleblower and the wrongdoer may affect reporting, such that people are more likely to blow the whistle when the perpetrator is a peer than when the perpetrator is a superior. In this meta-analytic review, we focused on the notion of social distance/closeness that stems either from psychological closeness or hierarchical closeness.
Using a random-effects model, we examined 20 articles and 22 studies that provided 23 effect sizes. We followed a hierarchical breakdown strategy, in which we first analyzed all studies before breaking them down to a moderator level. Specifically, our main meta-analysis revealed a significant overall effect, d = 0.21, and a high level of heterogeneity, 91.2%. Further decomposition of closeness into psychological and hierarchical closeness subgroups reduced the heterogeneity for both subgroups. As expected, our results revealed the two dimensions of closeness to be related to whistleblowing with their effects working in opposite directions. In particular, when people feel psychologically close to the perpetrator (e.g., friends), the less likely they are to blow the whistle. Conversely, the closer they are to the perpetrator with respect to the organizational hierarchy (e.g., peer), the more likely to report him/her. Additionally, to account for any culture differences, we used Hofstede's cultural dimension of individualism and collectivism. Culture, however, did not moderate the effect of closeness on whistleblowing.

Implications
The findings from this study have implications for both practice and theory. With respect to practice, it is crucial for organizations and managements to foster an organizational climate that eases the decision to report observed wrongdoing committed by individuals at a higher level in the organizational hierarchy. This may be achieved by establishing safeguards that protect whistleblowers against retaliation. One way to protect those employees who decide to disclose a wrongdoing is by developing an anti-retaliation policy that punishes those who retaliate against whistleblowers. Moreover, once concerns are raised, organizations should make sure that the whistleblower's identity is kept confidential. Anonymous reporting channels could help towards this direction. Regarding violations committed by close others such as friends, it is difficult for organizations to formally intervene besides encouraging reporting and protecting the whistleblower. Additionally, given that wrongdoings committed by superiors are less likely to be questioned and therefore to be reported, organizations' ethics training programs should increase individuals' awareness of violations. It is important for employees to understand that reporting superior wrongdoing is equally or more important as reporting peer wrongdoing given that fraud is strongly correlated with the level of authority of the wrongdoer (ACFE, 2020).

Limitations
In this meta-analysis we conducted a comprehensive literature search, spanning from 1972 to 2022, following a Fig. 2 Forest plot for the effects of each subgroup. Error bars represent the 95% confidence intervals detailed search strategy. Moreover, the inclusion of unpublished research aimed to reduce the problem of publication bias, whereas the inclusion of experimental studies allowed for a causal inference concerning the relationship between closeness and whistleblowing. By focusing our interest on experimental research and not on research that uses surveys such as the Merit Systems Protection Board survey in the USA, we eliminated any possibility of non-independence of the samples. However, the last two represent both a strength and limitation for a meta-analysis. The decision to include unpublished studies can be problematic, as they have not been peer-reviewed and therefore may be of lower quality than published research. The exclusion of non-experimental studies may have not identified some relevant studies that could generate a larger body of evidence. One limitation of this study is the small sample size. Therefore, we hope that our findings will encourage future research to further investigate the role of closeness between the observer of the wrongdoing and the wrongdoer. Finally, our study may be subjected to language bias as we only included articles published in English. However, English language bias is accepted in meta-analysis (Eisend, 2019).

Future Research
Subsequent research could examine the role of other forms of closeness such as spatial closeness or psychological closeness on the basis of group memberships and associated social identities (for the latter see (Anvari et al., 2019). In this review, only one study (Kaplan et al., 2009) attempted to investigate whistleblowing by focusing on gender as a group. Given that individuals can be members of many formal groups within the organization (e.g., workgroup, departments) (Ashforth & Johnson, 2001), future work should explore how work-related group memberships influence the decision to report a colleague. This is particularly interesting, as organizational group boundaries may create intergroup biases (Granitz & Ward, 2001), which may influence reporting decision. Future research should also examine hierarchical relationships in which the whistleblower is equally or more powerful than the wrongdoer (e.g., superiorto-superior and superior-to-subordinate), something that has not been investigated so far. We suggest that future research investigates the complexity of relationships within organizations by focusing on how the interplay of various personal and hierarchical relationships influences the decision to blow the whistle. For example, situations where the two types of closeness may counteract each other could be an interesting topic for future research. Following what we mentioned above about the unexplored relationship in which the wouldbe whistleblower is more powerful than the wrongdoer, a topic of interest could be a close relationship (e.g., friendship) between a supervisor and a subordinate. Future studies may also look for additional moderators of the relationship between closeness and whistleblowing. Although ingroup favoritism is quite robust as a phenomenon, it is not unconditional. Prior research has shown that individuals evaluate unlikable ingroup members more negatively than unlikable outgroup members, a phenomenon called the black sheep effect (Marques et al., 1988). As such, the decision to blow the whistle could be a manifestation of ingroup favoritism or the black sheep effect. The emergence of one or the other depends on several factors such as the characteristics of the misconduct, the characteristics of the wrongdoer, and the characteristics of the person who evaluates the misconduct (Otten & Gordijn, 2014). Thus, the identification of moderating variables favoring one or the other phenomenon may represent a new direction for future research.
Another potential area for future research could be identifying possible mediators to gain an understanding of the "why" factor behind our findings. The breach of loyalty may act as a possible mediator explaining why individuals do not blow the whistle, as retaliation is a significant factor in predicting the whistleblowing decision (Mesmer-Magnus & Viswesvaran, 2005). Additionally, the act of reporting is capable of eliciting stigma when loyalty norms are violated. Derogatory labels such as "snitch" or "rat" can convey a stigmatizing view of the whistleblowers being disloyal. Aside from the stigma of being disloyal, individuals are tainted by the stigma by association due to their association with "bad apples" (Pontikes et al., 2010). As such, stigma by association may compel individuals to refrain from blowing the whistle in order to protect their own reputation (Hussinger & Pellens, 2019). Understanding the mechanisms behind our findings may help organizations develop policies to encourage reporting wrongdoing.

Conclusion
This meta-analysis aims at assessing how the closeness to the wrongdoer influences the decision to report unethical behavior by disentangling the effects of "psychological closeness" versus "hierarchical closeness" on whistleblowing, two dimensions of closeness that their effects work in opposite directions. The results of our research show that the operationalization affects the link between closeness and whistleblowing behavior. In particular, a personal connection to the deviant person decreases the willingness to engage in whistleblowing behavior. On the contrary, results indicate that there is a tendency for closeness, as reflected in the organizational hierarchy, to increase whistleblowing behavior.
Acknowledgements The authors thank Jacob Lund Orquin for his valuable comments and helpful suggestions that improved the manuscript.

Data Availability
The data and code that support the findings of this study are available on the Open Science Framework (OSF) at: https:// osf. io/ f7aj4/? view_ only= 11eb7 c401e 2b4fc c9d0e 15f8e 1b4de 42.

Declarations
Ethics Approval This article does not contain any studies with human participants or animals performed by the authors.

Conflict of Interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.