Abstract
Despite its omnipresence, quantitative research examining both the nature and the function of adolescent gossip has been limited. The present study aimed to address this limitation in the literature by examining the nature and function of adolescent gossip; in particular, it aimed to explore observed differences between the gossip of those popular and non-popular adolescent dyads. Using both observational lab–based and sociometric school–based methods, the present study assessed the associations between social status and gossip functionality among 130 ninth grade students (92 females). Popular adolescent dyads engaged in more instances of gossip and were more likely to use gossip to foster intimacy, establish norms, and share both information and their opinions. Results also pointed to potential mechanisms by which popular adolescent dyads may gather alliances that support their status within the social hierarchy.
To date, gossip has been an understudied but ubiquitous feature of adolescent social relationships. Developmental researchers have studied gossip within the context of other relationally aggressive behaviors such as excluding, ignoring, and spreading rumors (for a review, see Crick et al., 1999) but have only begun to explore the potential influence of gossip. Recent conceptions of popularity have focused on the high status that adolescents achieve within the peer group by uniquely blending together both strategic social behaviors and prosocial approaches (Closson, 2009; Puckett, Wargo Aikins, & Cillessen, 2008; Wargo Aikins & Litwack, 2011). Gossip may be particularly salient for promoting and maintaining popularity during adolescence by building social capital, establishing norms, and elevating individuals’ self-images, while also building alliances that support youth’s positions. By adolescence, both dyadic intimacy and group acceptance are salient developmental tasks and are supported by adolescents’ growing cognitive abilities that allow them to use socially nuanced behavior in a strategic way. As such, the current study examines the nature and potential function of adolescent gossip; in particular, it aims to understand observed differences between the gossip of popular adolescent dyads versus those who are not.
Much of the research thus far has considered gossip within adult or young adult populations or relies upon ethological or anthropological methods of study (e.g., Besnier, 1989; Dunbar, 2004; Eder & Enke, 1991; Eder & Stanford, 1986; Levin & Arluke, 1987; Rosnow, 2001). However, it is apparent that quite early in development, children begin talking about others and understand both the nature and impact of these gossip interactions (Fine, 1977; Kuttler, Parker, & LeGreca, 2002). These interactions are both frequent and important for adolescents (Gottman & Mettetal, 1986; McDonald, Putallaz, Grimes, Kupersmidt, & Coie, 2007) characterized by discussions regarding other people’s personal characteristics, behaviors, and business. Although gossip is often construed as only pejorative in nature, broader definitions have suggested that gossip may include both positive and negative discussions of any non-present third party (Dunbar, 2004; Eder & Enke, 1991; Foster, 2004; Gottman & Mettetal, 1986; Levin & Arluke, 1987).
Notably, Gottman and colleagues observed that the function of gossip changes over the course of childhood, from promoting group solidarity and asserting group norms in early and middle childhood to providing a forum for solving interpersonal dilemmas (Gottman & Mettetal, 1986; Parker & Gottman, 1989) and promoting positive views of oneself at a cost to others (Eder & Enke, 1991; Eder & Stanford, 1986) in adolescence. In contrast, Besnier (1989) and Fine (1977) suggested that these functions are not mutually exclusive and likely operate simultaneously by adolescence. As such, adolescent gossip may uniquely combine assertive strategic and sociable approaches to build power and popular status through both the denigration of others and the facilitation of interpersonal alliances. These approaches appear consistent with the well-documented dual role of aggression and sociability in contributing to popularity (Cillessen & Rose, 2005).
Burgeoning quantitative research regarding gossip has focused almost exclusively on the impact on dyadic friendships, suggesting that gossip promotes heightened friendship intimacy over time while also contributing to increased friendship difficulties for at least some youth (Banny, Heilbron, Ames, & Prinstein, 2011; Grotpeter & Crick, 1996; Menzer et al., 2012). Positive friendship quality may be one important contributor to status gaining. Adolescents may use these positive relationships to buffer them from the otherwise negative consequences that might result from gossiping about others. However, these gossip behaviors may also be fundamental for achieving status, reflecting the strategic side of status seeking.
Although common conceptions of gossip often center on the use of these approaches by “mean” or “popular” adolescents, limited research to date has actually examined the associations between peer status and gossip characteristics. Based on observations of naturally occurring gossip, Eder and her colleages suggested that only popular adolescents within the peer group were likely to initiate or challenge gossip episodes; whereas all adolescents, regardless of status, provided comments that supported initial gossip overtures (Eder & Enke, 1991; Eder & Stanford, 1986). Empirical work by McDonald and colleagues (2007) found that well-liked early adolescents were both more likely to gossip about their peers and to do so in a more evaluative manner than rejected youth. Although at first blush, these findings may be taken to suggest that popular adolescents are more likely to use this form of social behavior to control social interactions and environments, the groups of adolescents under study in these two papers were not necessarily the same. Eder and colleagues (Eder & Enke, 1991; Eder & Stanford, 1986) work was putatively observing adolescents who were perceived to be popular or high status among their peers; in contrast, McDonald and colleagues (2007) were studying those young adolescents who were well-liked but not necessarily perceived popular. This is an important distinction given that over the last decade, well-liked youth have been shown to be distinct from those regarded as high status or popular (Cillessen & Rose, 2005).
The current study was designed to examine the associations between gossip and peer status, as well as the differences in gossip valence, target, function, and topics between those adolescent dyads who were perceived popular by their peers from those who were not. Building on Eder’s ethological work and our hypothesis that popular adolescents use gossip strategically to both obtain and maintain their social status, it was anticipated that perceived popular adolescents would gossip more than their peers. Furthermore, although positive gossip may create camaraderie and positive emotion between friends, negative gossip may create the sense of superiority over outside peers that is necessary to foster their status. Thus, given the expected goal of gossip for perceived popular adolescents, we also hypothesized that they would gossip in a more valenced way. Gossip’s description as a discussion of others’ characteristics and behaviors are anticipated to be more prototypic of popular adolescent dyads’ gossip as they may use these discussions to denigrate others and to serve as a yardstick against which to compare themselves and thereby assert their superiority. A particular topic of gossip for popular youth is likely to be peers’ associations, as deciphering and analyzing these is likely the grist for the establishment of social grading. The growing importance of romantic relationships during adolescence and the corresponding attention to opposite sex peers likely carries over into peer gossip. In fact, popular adolescents are involved in mixed gender peer groups and date at higher rates (Furman & Collins, 2009). Thus, popular adolescents may not only be more apt to discuss physical attractiveness but may be more likely to discuss male and female peers in that context. In addition, consistent with Besnier (1989) and Fine’s (1977) speculation that adolescent gossip may serve multiple functions during adolescence, we hypothesize that popular adolescent dyads would be particularly adept at both building status and support within their relationship simultaneously. As such, we expected that the function of gossip for popular adolescent dyads would be both to share their opinions and establish group norms in a manner that might be evaluative of others while building intimacy within their own conversation.
Method
Participants
Adolescents in this study included 130 ninth grade students (92 females) who participated in both a lab- and school-based data collection. These participants were aged 13 to 15 (M = 14 years 1 month; SD = 5.4 months). They attended five suburban middle schools in the Northeastern United States. Participants’ ethnicities as indicated by self-reports were 98% Caucasian and 2% Hispanic, with two participants not reporting on ethnicity. Participants were enrolled in public schooling within five towns whose average median household income was US$61,800 (range = US$35,087 to US$75,138). According to school records, 2% to 57% of adolescents in the school were eligible for free/reduced lunch.
Half of the participants in this study were selected from a sample of 281 students participating in a larger study examining the social and academic development of youth. Participants in the larger study were recruited via a letter addressed to them and their parents that invited all eighth grade students to participate in a study examining the transition to high school. Within the five schools involved in the study, consent forms were returned by 62% of families (n = 388). Of these, 72% of parents gave consent for their child’s participation (n = 281, 53% of the total population). Only students who obtained parental permission and provided assent participated in the larger study. The sample selected for the current lab-based study was drawn from the larger school study based on participants’ agreement to be contacted regarding participation in an additional lab session. These 165 students were invited to participate in this additional data collection over the course of the summer at a university lab. Of these, 65 students agreed to participate. ANOVAs indicated that those students who agreed to participate were no different from those who did not in terms of indicators of academic, social, or emotional well-being.
Adolescents in the sub-sample were asked to invite their same gender and same grade best friend to the lab. All youth in the current study invited a friend who attended the same school.
Procedure
School-based data were collected over two 45-minute sessions. Adolescents completed self- and peer-rated measures on their own while being monitored by a research assistant. Adolescents were offered a small token of appreciation and entered into a raffle for a larger prize.
Lab-based data were collected with adolescents who entered the university lab with a peer that they identified as a best friend. Each adolescent-best friend dyad engaged in a series of discussions, tasks, and games in addition to completing a series of questionnaires. Adolescents and their peers were compensated US$10 for their participation.
School-Based Measures
Sociometric popularity
Peer nominations were used to measure sociometric popularity. Participating students were provided with an alphabetized list of all grademates in their school and were asked to select an unlimited number of peers who they “like most” and “like least.” The number of like most and like least nominations received was counted for each student and standardized within grade. A continuous score for sociometric popularity was computed by subtracting the standardized like least score from the standardized like most score, and again standardizing the resulting difference score within grade to z scores (M = 0, SD = 1).
Perceived popularity
Peer nominations were also used to measure perceived popularity. Using the same list of grademates, participants selected an unlimited number of peers as “most popular” and “least popular.” Participants were not provided a definition of “popular,” as researchers have shown that early adolescents have a relatively consistent peer-constructed understanding of this status (e.g., LaFontana & Cillessen, 2002). The number of most popular and least popular nominations received was counted for each student and standardized within grade. A continuous score for perceived popularity was computed by subtracting the standardized least popular score from the standardized most popular score, and again standardizing the difference score to z scores within grade (M = 0, SD = 1; Cillessen & Mayeux, 2007). Children were also identified as being perceived popular if their standardized scores were .75 standard deviation above the mean. Twenty-six dyads were identified in this manner. These dyads included 12 dyads (8 female dyads) in which only the target participant was popular, five dyads (four female dyads) in which only the best friend was popular, and nine dyads (four female dyads) in which both the friends were popular.
Participation rate across the schools included in the broader study ranged from 28% to 72%. Although the reliability of sociometric ratings has largely been established within the field due to the numerous raters for each item, the reliability of sociometric ratings even within schools with lower participation rates has recently been demonstrated (Marks, Babcock, Cillessen, & Crick, 2013; Prinstein, 2007). In particular, popularity, which is based on group consensus regarding reputation, seems to be one of the most reliable measures. The use of unlimited nominations, such as in this study, appears to add to the reliability of the measure (Marks et al., 2013). These recent findings lend greater confidence to the reliability of the sociometric measures used in current study.
Lab-Based Measures
Gossip
Gossip was coded within the context of a peer discussion. All dyads were instructed to discuss individual problems that each member of the dyad had indicated willingness to discuss with their friend during a 16-minute period. None of the 65 dyads remained on-topic during the 16-minute discussion—with some dyads engaging in episodes of gossip during their discussions and others lapsing into discussions of other topics (e.g., the weather, what they would do with the money they earned during the session, what they would do after the experimental session was complete). Notably, those adolescents who engaged in gossip during this 16-minute episode were the same adolescents who did so during other tasks. In contrast, those adolescents who engaged in other kinds of talk also did so during other tasks. The consistency with which adolescent dyads were prone to either “fall back” into gossiping or strike up other kinds of conversations at each opportunity led us to believe that spontaneous gossip was indeed a good measure of dyadic gossip. Phenomenologically, we also recognized that this mirrored adolescents’ everyday experiences with one another such that gossip likely occurs in the context of breaks in activity or moments when friends find themselves alone to talk and therefore unlikely to be overhead by the target of the gossip (Rosnow, 2001). Gossip was coded across all 16 minutes of the discussion. On average, 5.36 minutes were spent in gossip or in the discussion of other topics (SD = 3.32). The coding of gossip in tasks not specifically designed to elicit gossip is not unprecedented in the literature as Menzer and colleagues (2012) used a similar approach.
Two female coders who were blind to adolescents’ sociometric status were used for this study. Coders were first trained on a subset of the study transcripts from another task (i.e., another discussion task). When adequate reliability on each of the codes was established, coders proceeded to the coding of the problem discussion described above. Using both the videotape and a typed transcript of the interaction, first the coders identified incidents of gossip separately. Gossip was defined as any talk between the dyad about an absent third party, which could include individuals or groups of people (Foster, 2004). The gossip could be initiated by either partner in the dyad. Distinct incidents of gossip were coded as beginning when a third party was mentioned by either adolescent in the dyad and were coded as ending when the discussion turned to a different topic or person. All transcripts were double coded and inter-rater reliability was high with 89% agreement on incident identification. All discrepancies were resolved by discussion. There were on average 2.97 gossip incidents per transcript (SD = 4.32).
Next, two coders coded these gossip incidents using the Gossip Coding System developed for this project based on the work of McDonald and colleagues (2007). Twenty-five percent of the coding was completed by both coders in order to check for reliability. The target of gossip referred to who the gossip was about and included either individuals such as male peers, female peers, romantic female peers, romantic male peers, family members, experimenters, and celebrities or groups of people including the mixed gender peer group, κ = .93. For the purposes of this article, although all gossip incidents were initially coded, only incidents regarding peers (i.e., female or male peers, romantic female or male peers, and the mixed gender peer group) were included in the analyses. As such, 1.4% of the gossip episodes were about individuals other than peers (i.e., siblings, parents, and celebrities) and, therefore, were excluded from analyses. If one or more female peers were gossiped about then the female peer code was given. Similarly if one or more male peers were the target of gossip then the male peer code was assigned. However, if both male and female peers were the targets of gossip in the incident then the mixed gender peer group code was assigned. Gossip that discussed a peer in a romantic fashion would be coded with either the romantic female or romantic male peer code as fitting. If the gossip was first about one peer and then turned to gossip about another peer this would be coded as two different gossip incidents and each would receive a separate code regarding the target of gossip.
The valence of gossip reflected the valence of the gossip expressed during the incident by both participants and could include positive, negative, neutral, or mixed gossip. Mixed gossip was reflective of incidents in which both positive and negative valence were expressed, κ = .82.
The topic of gossip referred to what was being discussed about the gossip target. During any one gossip incident, multiple topics might be discussed, and therefore, these codes were not mutually exclusive. Topics included the targets’ physical characteristics (e.g., appearance, clothes, fitness; κ = .82), personal characteristics (e.g., individual characteristics, demographic characteristics, or stable characteristics such as intelligence, ethnicity, religion; κ = .80), behavior (e.g., actions that the individual or group engages in including specific actions, morals, or abilities; κ = .79), sexuality (e.g., sexual interests or behaviors; κ = .92), and peer association (e.g., group associations, cliques, peer liking and dating, experiences or actions engaged in with peers; κ = .91).
Next the function of gossip was coded in order to capture the perceived purpose of the gossip. These included furthering intimacy (e.g., gossip that was used to foster closeness within the dyad, “Friend 1: Wow, I know I bet she tells other people about all of that stuff. Friend 2: Sometimes I just don’t feel like you can trust some of those girls. Friend 1: I know, trust is important. Friend 2: Right but they talk about each other and us,” κ = .86); establishing norms (e.g., gossip used to assert what was acceptable/not acceptable within their peer context; “Friend 1: I hated Ann and Betsy because they make out with all these different guys. Friend 2: Betsy at camp was telling us all about how she made out with like nine guys over the weekend at this park. Friend 1: You should have been like, Betsy that is being a slut,” κ = .88); entertainment (e.g., gossiping for the purpose of making the other member of the dyad laugh or as a means for sharing a funny story; “Friend 1: We were at this party and it was so funny. Candy got on the table and she was singing and dancing to the Spice Girls. ‘If you wanna be my lover.’ It was so funny,” κ = .90); share of information (e.g., gossip that shares facts or information with no opinion included; “Friend 1: Every time I turn around someone’s on antidepressants. Friend 2: Yeah, Kim is on them too,” κ = .91); and share of opinion (e.g., gossip that shares opinions or information with personal details included; “Friend 1: Ann and Allen are so mean sometimes. Friend 2: I know. Friend 1: And I hate how they put on a front like, ‘Oh, I’m your best friend!’ Friend 2: Yeah, I know right,” κ = .90).
Friendship quality
The Friendship Quality Questionnaire (FQQ; Parker & Asher, 1993) included 40 items describing youth perceptions of the quality of their best friendship. Each item was rated on a 5-point scale, ranging from not at all true to really true. Subscales represent Validation, Intimacy, Conflict, Conflict Resolution, Help and Guidance, and Companionship Within the Relationship. For the purpose of this study, a mean of all items was used to reflect overall self-reported friendship quality (Cronbach’s α = .90).
Results
Multiple imputation procedures, including relevant auxiliary variables, were used to estimate missing data (Schafer & Graham, 2002). Five multiple imputation datasets were generated using the software program Amelia II (Honaker, King, & Blackwell, 2011). All subsequent analyses were averaged across datasets. Participants who did and did not have missing data did not differ on the variables of interest in the current study. Analyses were also conducted using listwise deletion (i.e., no missing data), and the same patterns of results were obtained (supplemental analyses available upon request from the corresponding author). Means, standard deviations, and ranges for all gossip indicators are included in Table 1.
Descriptive Statistics for Peer Gossip Characteristics.
Preliminary Analyses
Seventy-two percent of the participants were rated by peers as similar in terms of popularity (n = 9) or unpopularity (n = 39). In addition, 12 target participants were rated as popular while their best friend was not, and five best friends were rated as popular while the target participant was not. The t tests were used to examine whether there were differences in time on task, friendship quality, amount of gossip, and the nature of gossip between adolescent dyads in which one or both partners in the dyad were popular (n = 17 and 9, respectively). No differences were evident, and therefore, for all analyses in the article, popular adolescent dyads are considered those dyads in which the target participant, best friend participant, or both friends were rated as popular by their peers.
Preliminary t tests were also run to examine whether there were differences in gossip initiation. Although overall target participants initiated more of the episodes of gossip than did their best friends (i.e., 59% vs. 41%), there were no popularity-based differences in terms of who initiated the gossip in the dyads that were disconcordant for popularity: t(15) = .006, p = .96, M = .59, SD = .19, for not popular adolescents, and M = .59, SD = .25, for popular adolescents.
ANOVA indicated that popular versus non-popular adolescent dyads did not spend significantly different amounts of time on task (popular, M = 9.16, SD = 1.26, and non-popular, M = 8.63, SD = 2.01) suggesting that similar amounts of time were spent discussing other topics.
Gossip and Popularity
In order to examine adolescents’ use of gossip within the peer group, an initial ANCOVA was used to assess whether popular adolescent dyads engaged in more gossip incidents than non-popular peer dyads. In the current sample, adolescents’ sociometric popularity and perceived popularity are correlated at r = .64, p < .001, and therefore, sociometric popularity was entered as a covariate. Consistent with hypotheses, results indicated that popular adolescent dyads engaged in more instances of gossip (Table 2).
ANCOVAs Comparing the Gossip Characteristics for Popular and Nonpopular Adolescents.
Medium effect size.
Large effect size.
p < .05. **p < .01. ***p < .001.
A series of ANCOVAs were next used to examine whether perceived popular adolescents and their peers use gossip in different ways. Again, given the significant correlation between sociometric and perceived popularity, sociometric popularity was used as a covariate in these analyses. As seen in Table 2, dyads with perceived popular adolescents gossiped about male peers, female peers, the mixed gender peer group, and male peers in romantic relationships more frequently than their non-popular peers. Although they were more prone than their non-popular peers to discuss others in a neutral or negative manner, they were also more prone to use a mixture of both positive and negative talk when gossiping about others. Dyads with perceived popular adolescents were also more prone to discuss the behavior of their peers, the kinds of relationships and associations that peers had with one another, their peers’ personal and physical characteristics, sexual behaviors, and characteristics. Finally, the function of popular adolescent dyads’ gossip appeared to be fostering intimacy, establishing norms, and sharing information and their opinion more so than dyads with less popular peers. 1
Exploratory analyses were run to examine gender differences given the small sample size and uneven number of boys and girls in the sample. ANCOVAs were used to test for gender differences in gossip as well as the potential role that gender may play in moderating popularity differences in gossip. Analyses indicated that there were no differences between male and female dyads in the non-popular group; however, a number of differences did emerge between male and female popular dyads. Specifically, popular female dyads gossiped more and were more likely to gossip about female peers and males in a romantic fashion; to use positive, negative, and mixed valenced gossip; to gossip about peers’ physical characteristics, behavior, and peer associations; and to use gossip to foster intimacy, establish group norms, and share information. Effect sizes ranged from .12 to .17.
Discussion
The present study is the first to demonstrate quantitatively the significance of gossip in the lives of popular adolescents. Using both observational lab–based and sociometric school–based methods, findings suggested that not only was the popularity of one or both members of the friendship dyad associated with differences in the quantity of gossip but also the nature of gossip discourse. These findings suggest that as adolescents begin to grapple with a changing social system wherein jockeying for status becomes an important social task, gossip may have a distinct role in adolescents’ group status.
Results indicated that dyads with at least one popular adolescent engaged in higher rates of gossip during their peer interactions. These findings may reflect popular adolescents’ sensitivity to the social demands of the context and willingness to engage in the behaviors necessary to afford them social status; whereas, dyads with less popular peers who do not join in may be penalized accordingly with lower social status. These findings extend previous research that has demonstrated increased social intuition and understanding among popular adolescents (Andreou, 2006) as well as increased willingness to engage in high rates of status gaining behaviors as a means for achieving and maintaining popular status (for a review, see Mayeux, Houser, & Dyches, 2010). Alternatively, it is also plausible that if gossip more frequently characterizes the interactions of dyads with at least one popular adolescent, that non-popular peers may have less experience with these types of interactions and thus may be less prone to engage in gossip in discussions with peers. As such, the failure to engage in gossip may put non-popular adolescents at a further disadvantage in both building alliances and superiority over others.
Notably, no differences in gossip were found between dyads in which both partners were popular or only one partner in the friendship was popular. These findings are striking and perhaps suggestive of the important associations between popularity and the nature and quantity of gossip during adolescence. Popular adolescents in these dyads may be driving the gossip processes in the pursuit of status maintenance or gaining. Alternatively, popular adolescents may have more experience engaging in these types of gossip processes and, as such, may be more likely to engage in these discussions and to direct them in ways that facilitate peer status regardless of their friends’ status. Due to these similar patterns, we will discuss these dyads, in which at least one partner is popular, as popular adolescent dyads.
Although previous research has characterized gossip as a negative or relationally aggressive behavior, it has also highlighted the potential for gossip to create closeness between friendship partners (Banny et al., 2011; Menzer et al., 2012). Consistent with this research, the findings of the current study suggest that friendship dyads with a popular adolescent use gossip at higher rates than non-popular adolescents to establish intimacy. The willingness of popular adolescent dyads to share negative information about peers may relay a sense of trust and valuing of this friendship (Foster, 2004; McDonald et al., 2007) which then may be linked to increased closeness. In addition, popular adolescent dyads’ heightened use of gossip characterized by a mix of positive and negatively valenced gossip may reflect their strategic and astute use of gossip to solidify friendships while simultaneously grappling for status. Consideration of these gossip episodes suggests that partners within these dyads gossip negatively about others while also describing their own positive qualities and the positive qualities of their friend. Thereby, adolescents in these dyads are not only derogating others but concomitantly overtly and covertly building friendship intimacy. These findings are consistent with extant evidence that suggests the important dual use of aggression and social competency in achieving popularity (Closson, 2009; Puckett et al., 2008; Wargo Aikins & Litwack, 2011). As such, the current findings may provide unique evidence that even within one social behavioral domain (i.e., gossip), the balance of aggression and sociability may be at play.
An important future step, when considering the associations between gossip, adolescents’ friendship quality, and peer status, may be to examine the centrality of the gossip target in the peer context. Gossip designed to facilitate status gaining and maintenance may be targeted at those adolescents in lower status groups; whereas, gossip directed at others within their own peer group who are competing for even higher status may create conflict, increase uncertainty regarding the stability of each member’s position within the status group, and perhaps disrupt friendship dyads. These scenarios may have very different consequences for peer status and may contribute to the instability of popularity observed during the adolescent years (Eder, 1985; Merten, 2011).
Similar to early assertions in the literature regarding the use of gossip to establish group norms (Gottman & Mettetal, 1986; Parker & Gottman, 1989), findings of this study indicated that popular adolescent dyads engaged in more norm establishing gossip than less popular friendship pairs. Popular adolescents have been found to be central in the development, socialization, and enforcement of social norms (Brown, 2011); popular dyads may use norm establishing gossip as a means to assert their centrality and superiority in this domain. The findings of the current study also suggest that popular dyads are more likely to use gossip as a means for asserting their opinions than their less popular peers. Consistent with earlier research that popular adolescents have higher self-esteem (de Bruyn & van den Boom, 2005), perhaps these adolescents feel heightened confidence and in turn are more able to use gossip as a means for expressing their points of view. Alternatively, adolescents who successfully express opinions in the peer context may have bolstered self-esteem and develop greater popularity.
Our findings also indicated that dyads with popular adolescents frequently discussed the behaviors, relationships, peer associations, personal and physical characteristics, as well as sexual behaviors of other adolescents. By delineating who is “hanging out with who,” who is in the “in-group” versus who is not, who one’s friends are, and how other adolescents are behaving in their social and growing romantic/sexual interactions, popular adolescent dyads may be further able to define the social hierarchy and use this as a means for both ostracizing others and reaffirming their own place within the group. Discussing what is desirable and objectionable may be one further way to define norms and establish what is acceptable. Moreover, physical characteristics are an easily observable point for discussion that is both relevant to adolescents’ growing conceptions of identity (Pomerantz, 1979) and distinguish popular from unpopular youth (Adler, Kless, & Adler, 1992; Rose, Glick, & Smith, 2011). By setting norms regarding the role of attractiveness, fashion, and appearance in contributing to popularity (Adler et al., 1992; Rose et al., 2011), these adolescents set the tone for group status. Adolescents acknowledge the pressure to conform to social standards as a necessity for ascendance within the social group and recognize that failure to conform likely leads to exclusion (Brown, 2011). As such, this study suggests that gossip may be one social behavior by which social norms are asserted, and examples of nonconformity are ridiculed.
Finally, the findings of the current study indicate that popular adolescent dyads are discussing both their male and female classmates at higher rates than their non-popular peers. In addition, frequent targets of popular adolescent dyads’ gossip is the mixed gender peer group as well as romantic relationships with males. Increasing engagement in mixed gender peer groups and youth’s burgeoning interests in romantic relationships emerge during this time period. Adolescents, rather than adults, appear to provide information and advice regarding these developing relationships, and consequently popular adolescents often seem to take the lead in socializing expectations and setting examples for mixed gender relationships (Furman & Collins, 2009). By discussing the interactions of adolescent boys and girls, popular adolescents may begin to set the standard for what may be expected within the social system in regards to interaction patterns as well as who may be demonstrating the most developmentally normative and non-normative behaviors. The highly female sample included in this study has likely contributed to the prevalence of gossip regarding romantic partnerships with male as opposed to female peers. Future studies with a more even gender distribution of participants will be necessary to examine whether this finding is replicable or rather, partially a byproduct of sample characteristics.
Limitations and Future Directions
Several factors should be considered when interpreting the results of this study. The current study presents concurrent data from early adolescence. As such, we cannot discern the directionality of the association between popularity and gossip. Indeed, we have argued that either directionality is both plausible and interesting. Popular status may not be obtained as a result of gossip but rather may potentiate this social behavior. In this manner, popular adolescents’ broader social knowledge and access to larger social networks may provide them with grist for the “gossip mill.” It is plausible that these social behaviors are used to maintain social status, but without longitudinal data, it is difficult to tease apart how they contribute to the emergence or sustainment of status over time. Future longitudinal research will be necessary to address these questions.
Moreover, although these results appear to be consistent with previous research across other developmental periods and groups, further research examining these processes at other developmental time points is warranted. As adolescents become more at ease with their own identity development, rigid group memberships may begin to feel confining (Brown, 2011). Perhaps gossip that is observed in these early adolescent group contexts wane in frequency as adolescents become less reliant upon gossip as a means for maintaining group boundaries and asserting a strict set of self-definitions. Future longitudinal and quantitative research that measures the frequency, nature, and function of gossip across peer groups and time periods could help further illuminate the associations between gossip and the social hierarchy over time.
The current study is somewhat restricted by the small sample. We were only able to run exploratory analyses to examine potential gender differences in gossip and whether gossip may be more associated with gossip for perceived popular boys or girls. To date, research regarding gender differences in gossip behaviors across age groups has been somewhat equivocal with some evidence suggesting that girls engage in higher rates and focus on different targets (Leaper & Holliday, 1995; Levin & Arluke, 1985; Watson, 2012) though others find few differences (Dunbar, 2004). Our findings demonstrated numerous gender differences in gossip: with popular female dyads gossiping not only more but about different people, about different topics, with more valenced approaches, and seemingly to achieve somewhat different functions. Although these are interesting findings, given the small sample size, the findings should be considered preliminary and further research should examine the influence of gender, popularity, and their concomitant role in adolescent gossip.
In the current study, gossip was coded as a dyadic process; however, future research examining the links between popularity and gossip may benefit from examining each individual’s contribution to the gossip. Such an approach would allow for the use of Actor Partner Interdependence Model (APIM) analyses. APIM has increasingly become a model for examining dyadic relationships in which conceptually the two relationships are interdependent. These models would allow researchers to better understand how each partner’s characteristics predict gossip features, such has been done when examining the influence of other partner characteristics on gossip (e.g., Banny et al., 2011; Menzer et al., 2012).
Overall, sample constraints and the unavailability of longitudinal data were balanced against numerous methodological strengths, including the assessment of gossip through observational methods, the appraisal of popularity via peer reports, and consequently the use of multiple methods. These strengths allowed us to test common conceptions regarding status differences in gossip. Given the importance of peer interactions and status in adolescents’ lives, as well as the frequency of gossip as a means of communication, it is hoped that these findings will shed further light upon salient mechanisms that contribute to adolescents’ social contexts. The current study is believed to function as an important preliminary step and will stimulate additional research regarding the influence of adolescent communications on their socio-emotional well-being.
Footnotes
Acknowledgements
Appreciation is expressed to the Peers and Family Lab staff for their contribution to the data collection and to the participants, their families, and friends who were a fundamental part of this research.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
1.
All analyses were also conducted using the proportion of gossip instances which served different functions instead of the frequency, and we obtained the same pattern of results (all supplemental analyses available from the corresponding author).