Edited by: Varun Dutt, Indian Institute of Technology Mandi, India
Reviewed by: Michelle Hood, Griffith University, Australia; David Álvarez-García, Universidad de Oviedo Mieres, Spain
*Correspondence: Eva M. Romera
This article was submitted to Cognitive Science, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Cybergossip is the act of two or more people making evaluative comments via digital devices about somebody who is not present. This cyberbehavior affects the social group in which it occurs and can either promote or hinder peer relationships. Scientific studies that assess the nature of this emerging and interactive behavior in the virtual world are limited. Some research on traditional gossip has identified it as an inherent and defining element of indirect relational aggression. This paper adopts and argues for a wider definition of gossip that includes positive comments and motivations. This work also suggests that cybergossip has to be measured independently from traditional gossip due to key differences when it occurs through ICT. This paper presents the Colombian and Spanish validation of the
Gossip is defined as a type of evaluative conversation about somone else who is not present (Eder and Enke,
Research on the positive side of gossip has mainly focused on its group-serving functions. It has been said that it can be useful to inform members of the group about its norms and sanctions (Grosser et al.,
In addition to providing information, influence, and strengthen group bonds, gossip also has a gratifying individual function: entertainment (Foster,
A recent study with more than 55,000 adolescents from 20 different countries shows the widespread use of ICT in social communication and its importance in social and personal development (Areepattamannil and Khine,
Cybergossip is the equivalent of gossip in the virtual world, but both phenomena must be studied in different ways, as the contexts in which they take place differ and matter. First, face-to-face gossip makes use of verbal language, whilst cybergossip can take the form of written messages, images, and videos. Second, cybergossip messages are recorded and can be reproduced exactly as originally created, which may influence what is shared in the first place (Bertolotti and Magnani,
It is also important to differentiate between cybergossip and cyberbullying. Gossip's negative potential has led numerous studies to identify it as an indirect form of peer group aggression that has the aim to manipulate another's reputation or achieve his/her exclusion from the group (Archer and Coyne,
Despite the differences between the phenomena, work on cybergossip should take research into its off-line version as a starting point. Gossip has been widely studied and with different research methodologies. As anthropology is the dominant field for the study of gossip, observational techniques have been most commonly used (Foster,
The widely known
The
The
Brady et al. (
The scales discussed above have all been designed and validated for (young) adults. For children and adolescents, gossip research exists mainly in the form of narrative and observational studies, and these approach gossip from the same narrow negative angle of indirect agression aimed at damaging others (Xie et al.,
Research into children and adolescents needs to adopt the methodological advances that were made in the study of gossip among adults, and include its positive potential to transmit understanding of group norms and see it as part of the set of socially competent behaviors (McDonald et al.,
To provide in the need for a cybergossip instrument for children the
The first objective of this study was to validate the
Our hypotheses were: (1) the instrument will show a one-factor structure with optimal psychometric properties, homogeneity of measure across gender and country and will have a low association with cyberbullying; (2) Colombian adolescents will be less involved in cybergossip, due to their lower use of ICT (Said-Hung,
The overall sample comprised 3,747 adolescents (48.5% boys) aged 10 to 19 (
To measure cybergossip in adolescents, the GCQ-P scale designed for primary school children was used unchanged (López-Pradas et al.,
This instrument had previously been validated for both Colombia (Herrera-López et al.,
The research had a transversal, restrospective, ex post facto design, with one group and multiple measures (Montero and León,
Mardia's coefficient was calculated to assess the multivariate normality of the dataset using the program “R” (R Development Core Team, 2008) and package “MVN” (Kormaz et al.,
The psychometric properties of the CGQ-A scale were verified through item response theory (IRT) analysis, calculating a three-parameter model (3PL) fitted to polytomous scales (Muraki,
To assess construct validity, confirmatory factorial analyses (CFA) were carried out using the EQS 6.2 program (Bentler and Wu,
The generalization of the model, that is, the degree of robustness of the factorial structure or invariance, was tested through multi-group analysis, with country and gender as analysis criteria. This analysis consists of comparing a set of increasingly restrictive models. In Model 1, configural invariance was tested by imposing the same factorial structure on both subsamples, and checking whether the fit indexes of the combined model indicated good model fit. Subsequently, in Model 2 the factorial loads were restricted and the fit indexes of Models 1 and 2 were compared. Changes (Δ) between the models of NNFI, CFI, RMSEA, and SRMR ≤0.01 were accepted as evidence of measurement invariance (Dimitrov,
The discriminant validity was tested through Spearman's Rho correlations between the CGQ-A scale and the ECIPQ dimensions (cybervictimization and cyberaggression).
Internal consistency was analyzed with McDonald's Omega index (Ω), recommended for categorical variables under the absence of multivariate normality (Elousa-Oliden and Zumbo,
The level of involvement in cybergossip was calculated by taking the average of the nine items. The adopted statistical significance level was 0.05.
Mardia's analysis generated a coefficient of skewness of 63.027;
The 3PL (IRT) analysis provided discrimination values above 1, which are considered good values; the difficulty degree of the items ranged from 0.05 to 1.60, which is acceptable (acceptable values range from −3 to 3); and the probability of failure values were low, indicative of high quality items (Baker,
Mean, standard deviation, skewness, kurtosis, and 3PL analysis (IRT).
CG 1 | I have made comments about other friends or classmates to get into a group on social networks or WhatsApp. | 1.19 | 0.55 | 3.81 | 17.87 | 2.03 | 1.61 | 0.02 |
CG 2 | I talk about others on social networks or WhatsApp because it makes me feel closer to my group of friends | 1.41 | 0.83 | 2.43 | 6.20 | 1.89 | 1.00 | 0.04 |
CG 3 | I have told things about a classmate or friend on social networks or WhatsApp to make the group change their opinion about him/her | 1.45 | 0.83 | 2.10 | 4.50 | 1.92 | 1.00 | 0.08 |
CG 4 | When I'm angry with a classmate or friend, I talk about it on social networks or WhatsApp | 1.39 | 0.78 | 2.36 | 5.80 | 1.73 | 0.97 | 0.00 |
CG 5 | I have said negative things about another person on social networks or WhatsApp without realizing it | 1.40 | 0.74 | 2.24 | 5.72 | 1.99 | 0.76 | 0.00 |
CG 6 | I have shared a classmate's secret with others on social networks or WhatsApp | 1.29 | 0.68 | 2.90 | 9.58 | 1.76 | 1.19 | 0.00 |
CG 7 | I use social networks or WhatsApp to share stories I hear about others with my friends | 1.56 | 0.94 | 1.86 | 3.13 | 2.19 | 0.52 | 0.00 |
CG 8 | When somebody in my group does something bad, I tell the rest of my classmates via social networks or WhatsApp so they know about it | 1.33 | 0.71 | 2.61 | 7.67 | 2.23 | 0.99 | 0.02 |
CG 9 | I talk with my friends on social networks or WhatsApp about what's going on with other classmates for fun | 1.32 | 0.71 | 2.77 | 8.65 | 1.95 | 1.03 | 0.00 |
The inter-item correlation analysis showed values in the upper-middle range. The correlations between the CGQ-A scale and the ECIPQ cyberaggression and cybervictimization dimensions were in the medium range, suggesting a low collinearity and discriminant validity between the instruments (see Tables
Matrix of CGQ-A polychoric correlations.
1 | 1 | ||||||||
2 | 0.61 | 1 | |||||||
3 | 0.54 | 0.52 | 1 | ||||||
4 | 0.43 | 0.43 | 0.43 | 1 | |||||
5 | 0.46 | 0.47 | 0.45 | 0.63 | 1 | ||||
6 | 0.48 | 0.41 | 0.43 | 0.53 | 0.59 | 1 | |||
7 | 0.46 | 0.43 | 0.53 | 0.58 | 0.59 | 0.61 | 1 | ||
8 | 0.48 | 0.49 | 0.50 | 0.52 | 0.56 | 0.54 | 0.59 | 1 | |
9 | 0.49 | 0.49 | 0.43 | 0.49 | 0.56 | 0.53 | 0.58 | 0.59 | 1 |
Spearman's Rho correlations between ECIPQ and CGQ-A.
1. Cyberaggression | 1.12 | 0.26 | – | 5.13 | 41.75 | ||
2. Cybervictimization | 1.22 | 0.35 | 0.49 |
– | 3.85 | 23.06 | |
3. Cybergossip | 1.37 | 0.50 | 0.44 |
0.30 |
– | 2.36 | 7.94 |
The CFA of the one-dimension structure of the CGQ-A scale showed an adequate fit:
CFA of the CGQ-A scale (*
The obtained value of the internal consistency was within the optimal range (ΩCGQ−A = 0.839).
The values of the CR and AVE indexes were also found within the optimal range: CR = 0.904; AVE = 0.513, over 0.70 y 0.50, respectively (Hair et al.,
The results of the multi-group analysis indicated proof of configural and measurement invariance for both the country and gender tests. For both variables, the chi-square differences (
Multi-group analysis of the CGQ-A scale per country and gender.
Country | Mod 1 | 287.072 | 54 | 0.000 | 0.966 | 0.975 | 0.053 | 0.061 | |||||||
Mod 2 | 276.523 | 62 | 0.000 | 0.973 | 0.977 | 0.047 | 0.061 | 10.549 | 0.534 (n.s.) | 8 | 0.007 | 0.002 | 0.006 | 0.000 | |
Gender | Mod 1 | 306.722 | 54 | 0.000 | 0.980 | 0.985 | 0.055 | 0.049 | |||||||
Mod 2 | 317.516 | 62 | 0.000 | 0.982 | 0.985 | 0.051 | 0.050 | 10.794 | 0.793 (n.s.) | 8 | 0.002 | 0.000 | 0.004 | 0.001 |
The adolescent participants in the study had an overall cybergossip involvement level of 1.36 (
The objective of the study was to validate the CGQ-P scale (López-Pradas et al.,
The multi-group analysis proved configural and measurement invariance between Colombia and Spain and between boys and girls, indicating the scale has a robust factorial structure across countries and gender. Put in another way, it means the different items contribute similarly to the overall factor for girls and boys, for Columbians and Spanish. The invariance therefore demonstrates that Colombian and Spanish adolescents conceptualize cybergossip in similar ways when they respond to the CGQ-A scale. This indicates similar dynamics in both countries, despite the cultural, socioeconomic and geographical differences (Romera et al.,
The second objective of the research was to measure how frequenctly adolescents engage in cybergossip, and to investigate differences by country and gender. The descriptive results indicate that they enage in this cyberbehavior quite frequently. Results confirmed our second hypothesis that Colombian adolescents would be less involved in cybergossip than Spanish ones. This could reflect the lower use of ICT in Colombia (Said-Hung,
No differences were found in the frequency with which boys and girls participate in cybergossip, which confirms the third hypothesis. This, together with the results on measurement invariance by gender, supports the idea that boys and girls engage in gossip in similar ways. Previous developmental studies have shown a higher involvement of girls, which they attributed to adolescent girls spending more time on social activities, and having a relationship style characterized by a desire for proximity and an anxiety about rejection (Perry and Pauletti,
The present study has made an important contribution to this field of research by providing it with a valid and reliable instrument with which to measure cybergossip behavior. The CGQ-A scale's optimal psychometric properties and general validity allow its use in comparative studies of descriptive and explanatory nature.
Without downplaying the possible harmful effects of gossip, this paper argues that cybergossip does not have to be understood as a cyberbehavior which must be erradicated or reduced in all situations. This research recognizes and has confirmed the relationship between cybergossip and cyberbullying, but it also supports a focus on positive cybergossip in psychoeducational interventions that promote the learning of new methods of interaction and the development of communication and digital skills to build positive virtual relationships and prevent risky cyberbehaviors.
A number of limitations of this study must be mentioned. No explicit tests of convergent validity have been included. These were omitted as this study's particular aim was to observe its divergent validity, which is why we focussed on the correlation between the concepts. Another potential limitation is that no test-retest measures have been employed to confirm the reliability of the results, although different samples have been used and the optimal psychometric properties of the CGQ-A scale have been established.
It is necessary to continue advancing our understanding of gossip, especially of its effect on group characteristics, following recent evolutionary studies that highlight the influence of social networking on social behavior (Wu et al.,
The study was approved by the Comité Coodinador de Ética de la Investigación Biomédica de Andalucía (Coordinating Ethics Committee of Biomedical Research of Andalusia) and was in accordance with all regulations concerning professional ethics as stated in the International Conference on Harmonization Good Clinical Practice Guideline. The study was approved by the Spanish school boards and the Colombian schools' management, and written parental consent was obtained. In both countries, the students were visited and the anonymous, confidential and voluntary nature and the objective of the study were explained before the survey was taken.
All authors contributed to the interpretation of data, helped to draft, and revise the manuscript and have read and approved the final manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
To the schools and adolescents who participated in this study. To SR and TG who helped to collect the data.
The Supplementary Material for this article can be found online at: