Internet-Risk Classes of Adolescents,Dispositional Mindfulness and Health-Related Quality of Life: A Mediational Model

Abstract

This study was primarily aimed at identifying classes of adolescents in relation to their probability of endorsing several risks associated with the Internet (cyberbullying victimization and perpetration, cyberdating abuse victimization, and perpetration, sexting, and grooming). The second objective was to examine a mediational model linking dispositional mindfulness, risk perception, exposure to antisocial content in the media, Internet-risk classes of adolescents, and health-related quality of life (HRQL). The sample comprised 3,076 adolescents (46.2% boys, ages between 12 and 21). Latent class analyses indicated the existence of five classes related to the probability of endorsing Internet risks: No risk (60.75%), only cyberbullying (25.5%), cyberbullying and cyberdating abuse (6.7%), all risks (4.3%), and sexual risk (2.9%). Three mindfulness facets, namely, acting with awareness, nonreacting, and nonjudging, were associated with all the classes of risks. This association was partially explained by the degree of exposure to antisocial content in the media and risk perception. Finally, membership in the Internet-risk classes was associated with a lower HRQL.

Introduction

The Internet contains several risks, such as cyberbullying, cyberdating abuse, grooming, and sexting, which may have a negative effect on adolescents' health-related quality of life (HRQL).^1,2 Cyberbullying is the use of electronic and online communication technology to harass, threaten, and/or socially exclude a targeted individual.^3,4 Cyber aggression may also be aimed at dating partners.⁵ Sexting refers to the creation and exchange, through the digital media, of messages, pictures, and/or videos of sexual content.^6,7 Online grooming is a process by which an adult tricks a minor through the use of new technologies to obtain sexual material of the minor and/or to sexually abuse her/him.^8,9

Unfortunately, many Internet risks do not take place in isolation; instead, there is a tendency for them to occur in conjunction with each other.^4–13 Therefore, it is important to study how several Internet risks combine and negatively affect adolescents' HRQL. Furthermore, protective factors against involvement in Internet risks need to be identified.

Dispositional mindfulness (DM) may be postulated as a relevant protective factor against Internet risks. Mindfulness is a way to attend to present experience, deliberately and without judging, moment by moment.¹⁴ DM has been proposed to comprise five facets: Observing sensations, thoughts, and feelings (OB); describing feelings and thoughts with words (DES); acting with awareness (AA); nonjudging of inner experience (NJ); and nonreacting to inner experience (NR).¹⁵ DM is positively associated with well-being and quality of life.¹⁶ Furthermore, DM may contribute to an understanding of risk-taking behavior on the Internet. Some studies have found an inverse relationship between AA and problematic gaming,¹⁷ problematic smartphone use,¹⁸ and the problematic use of the Internet^19,20 and cyberbullying.^21,22

Several conceptual approaches have proposed the mediational mechanisms through which mindfulness reduces emotional and behavioral problems. According to these proposals, mindfulness enhances self-regulation^23–26 and increases prosocial characteristics through a positive relationship between the self and the other, which transcends self-focused needs.²⁵ Thus, adolescents who are high in some DM facets, such as AA, may pay more attention to scenarios on the Internet, and, accordingly, be less likely to become involved in risky behaviors. Self-regulation could lead to a higher risk perception of some behaviors, such as giving personal data to strangers. Moreover, adolescents are great antisocial media consumers,²⁷ and this exposure is associated with being implicated in cyberbullying behaviors.²⁸ Adolescents with high DM are more motivated toward prosocial behavior and avoiding antisocial media content and, therefore, may be less likely to be implicated in cyber risks.

Research objectives

Research objective 1 (RO1) was to identify classes of adolescents in relation to their probability of endorsing several Internet risks, including cyberbullying victimization and perpetration, cyberdating abuse victimization and perpetration, sexting, and grooming.

RO2 was to examine a mediational model linking DM, risk perception, exposure to antisocial content in the media, Internet-risk classes of adolescents, and HRQL. We hypothesized that (a) DM would explain belonging to classes through the mediation of higher Internet-risk perception and less exposure to antisocial content in the media; (b) DM would be associated with HRQL, whereas Internet-risk classes that accumulate several risks would be associated with lower HRQL.

Materials and Methods

Participants

The sample included 3,076 students (1,422 boys, 46.2%; 1,654 girls, 53.8%; mean age = 13.97 ± 1.45). The participants came from 22 high schools in seven regions of Spain. Data were collected between April and June 2018 by means of convenience sampling.

Instruments

A revised version of the Cyberbullying Questionnaire,²⁹ which comprises nine items for the cybervictimization dimension and another nine for the cyberaggression dimension was employed. The response scale ranged from 0 (never) to 4 (almost every week). Omega: 0.88 (Victimization) and 0.86 (Perpetration).

An adapted version of the Cyberdating Abuse Questionnaire³⁰ was administered. It comprises 22 parallel items, 11 for victimization, and 11 for perpetration, and includes behaviors such as mobile phone control behavior and insults. The response scale ranged from 0 (never) to 3 (almost always). Omega: 0.90 (Victimization) and 0.91 (Perpetration).

The Sexting Questionnaire, which was an adaptation from the original questionnaire³¹ comprised three items related to sending information such as pictures and videos with sexual content to one's partner, an acquaintance, or someone the respondent had met online but did not yet know in person. The items were rated on a five-point response scale ranging from 0 (never) to 4 (7 or more times). Omega: 0.86.

The Online Sexual Solicitation and Interaction of Minors with Adults Questionnaire,³² which contains 10 items that assess the sexual interactions that are part of the initiation, process, and/or result of online grooming, was employed. Items are rated on a four-point response scale ranging from 0 (never) to 3 (6 or more times). Omega: 0.93.

The Five-Facet Mindfulness Questionnaire-Adolescents-Short Form,¹⁵ short version for adolescents (FFMQ-A-SF³³), was used to measure five facets of DM: OB, DES, NJ, AA, and NR. The FFMQ-A-SF comprises 25 items; each facet is assessed by five items answered on a five-point scale ranging from 1 (never or rarely true) to 5 (very often or always true). Omega: 0.80 (OB), 0.78 (DES), 0.85 (AA), 0.84 (NJ), and 0.76 (NR).

Exposure to media with antisocial content was measured by means of the 12 items of the Antisocial Content subscale of the Content-based Media Exposure Scale 2.³⁴ It includes exposure to direct and relational aggressive behaviors, use of drugs and alcohol, and explicit sexual behaviors in the media. Respondents assess the items on a five-point scale (0 = never, 4 = very often). Omega: 0.94.

Risk Perception was assessed with four items:³⁵ Posting personal photos, publishing your real name, your phone number, and the name of your school. Adolescents rated how risky these behaviors were on a 10-point response scale, ranging from 1 (not at all) to 10 (a lot). Omega: 0.83.

We employed the Spanish version of KIDSCREEN-10³⁶ to assess HRQL. The items are assessed by means of a five-point scale, ranging from 1 (never) to 5 (always). Omega: 0.85.

Procedure

The participants completed the measures through Qualtrics©. Participation was voluntary. Parents were given the option to decide whether their children could participate; less than 2% who did not want their children to participate returned the consent form. The project was approved by the (masked for review) Ethics Committee (Ref.231/17).

Statistical approach

We used latent class analyses (LCA) with the maximum likelihood (ML) estimator in MPLUS 8³⁷ to identify classes of adolescents according to their involvement in several Internet risks. We followed several criteria to determine the optimal number of classes.^38,39 We employed the Akaike information criterion, the Bayesian information criterion, the sample-size adjusted Bayesian information criterion, the entropy index, and the Lo/Mendell/Rubin adjusted likelihood ratio test as indicators to evaluate and compare the models.³⁹

We used path analysis with LISREL 10.2 with the robust ML method, which includes the Satorra/Bentler scaled χ² index⁴⁰ (S-Bχ²) to examine associations between DM, antisocial media exposure, risk perception, latent classes of adolescents according to Internet risks, and HRQL. Fit of the model was evaluated by means of the comparative fit index (CFI), the Tucker/Lewis fit index (TLI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). CFI and TLI values of 0.95 or greater and RMSEA and SRMR values of 0.06 or lower indicate that the model adequately fits the data. The significance of mediational paths was tested by means of bootstrapping with 5,000 samples.

Results

Descriptive statistics and correlations between variables

Table 1 presents the descriptive statistics and correlation coefficients between variables. With the exception of OB, all DM facets were positively correlated with HRQL.

Table 1.

Descriptive Statistics and Correlation Coefficients

	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. Cyberbullying perpetration	1
2. Cyberbullying victimization	0.57^**	1
3. Cyber dating abuse perpetration	0.43^**	0.31^**	1
4. Cyber dating abuse victimization	0.40^**	0.35^**	0.66^**	1
5. Sexting	0.35^**	0.37^**	0.29^**	0.23^**	1
6. Grooming	0.38^**	0.44^**	0.32^**	0.35^**	0.60^**	1
7. Antisocial media exposure	0.31^**	0.29^**	0.17^**	0.20^**	0.22^**	0.25^**	1
8. Risk perception	−0.06^**	−0.00	−0.01	0.02	−0.02	−0.01	−0.08^**	1
9. Observing	0.03	0.09^**	0.02	0.04	0.01	0.03	0.06^**	0.08^**	1
10. Describing	−0.08^**	−0.10^**	−0.05	−0.08^**	−0.02	−0.07^**	−0.03	−0.01	0.02	1
11. Acting with awareness	−0.19^**	−0.19^**	−0.12^**	−0.14^**	−0.08^**	−0.13^**	−0.25^**	0.08^**	−0.14^**	0.32^**	1
12. Nonjudging	−0.17^**	−0.24^**	−0.08^*	−0.12^**	−0.10^**	−0.16^**	−0.19^**	−0.04^*	−0.23^**	0.28^**	0.39^**	1
13. Nonreacting	−0.01	−0.03	−0.01	−0.04	0.01	−0.02	0.02	0.06^**	0.37^**	0.11^**	−0.04^*	−0.12^**	1
14. Health-related quality of life	−0.18^**	−0.29^**	−0.09^**	−0.16^**	−0.11^**	−0.20^**	−0.23^**	0.04	−0.00	0.30^**	0.39^**	0.39^**	0.14^**	1
Mean	0.14	0.21	0.07	0.10	0.08	0.08	1.26	5.83	2.89	3.32	3.28	3.54	2.83	4.05
SD	0.31	0.41	0.21	0.26	0.36	0.29	0.96	2.25	0.86	0.74	0.83	0.88	0.73	0.62

p < 0.05, ^**p < 0.001.

SD, standard deviation.

RO1: classes of adolescents in relation to their involvement in internet risks

Table 2 presents the results of the LCA. The indicators improved from the one-class model to the five-class model and were reduced with the six-class model. These results suggested that the five-class latent model was the optimal LCA model for the sample. Posterior probabilities were used to assign each participant to a single class. Figure 1 displays the solution of five classes. The first class (no risk) presents low probabilities in all categories of risks. The second class (only cyberbullying) presents a high probability of reporting cyberbullying victimization and perpetration and low probabilities in the other risks. The third class (cyberbullying and cyberdating abuse) presents moderate-high probabilities of perpetration and victimization of both cyberbullying and cyberdating abuse. Although the fourth class (all risks) was very similar to the third class, the probabilities for cyberbullying and cyberdating abuse were higher. Furthermore, the fourth group had moderate probabilities of grooming and sexting. The fifth class (sexual risk) presents the highest probabilities of reporting grooming and sexting.

FIG. 1.

Latent classes for Internet risks. Conditional risk probability plot. The probability of endorsing each risk was provided by class membership. CB, cyberbullying; DA, dating abuse.

Table 2.

Results of the Latent Class Analyses

Fit indices	Number of classes
Fit statistics	1	2	3	4	5	6
AIC	16,976	14,980	14,500	14,456	14,413	14,411
BIC	17,012	15,058	14,621	14,619	14,618	14,658
SABIC	16,993	15,017	14,557	14,533	14,510	14,528
Entropy		0.708	0.749	0.748	0.801	0.783
Lo/Mendell/Rubin Likelihood ratio test (p)		1,974.8 (p < 0.001)	484.68 (p < 0.001)	57.45 (p < 0.001)	56.16 (p = 0.0014)	16.018 (p = 0.236)
Sample size of each class	C1 = 3,076	C1 = 1,878	C1 = 326	C1 = 447	C1 = 1,863 (60.57%)	C1 = 180
		C2 = 1,198	C2 = 1,878	C2 = 1,039	C2 = 785 (25.52%)	C2 = 422
			C3 = 872	C3 = 1,333	C3 = 132 (4.29%)	C3 = 35
				C4 = 257	C4 = 207 (6.73%)	C4 = 658
					C5 = 89 (2.89%)	C5 = 61
						C6 = 1,720

AIC, Akaike information criterion; BIC, Bayesian information criterion; SABIC, sample-size adjusted Bayesian information criterion.

Table 3 displays the results of the analysis of variance (ANOVA) for the differences between classes in all the study variables. According to multiple comparisons (Bonferroni method, p < 0.05), several statistically significant differences between the classes emerged. The adolescents in the no-risks class scored higher on Internet-risk perception than the other classes, except the sexual-risk class. The lowest scores on antisocial media exposure were presented by those in the no-risk class while those in the all-risks and sexual-risk classes presented the highest scores. The participants in the cyberbullying and cyberdating abuse-risk class scored the lowest on OB. They also scored the lowest on NR together with those in the all-risks class. Those in the no-risk class scored the highest on AA and NJ. Those in the all-risks class, together with the sexual-risk class, scored the lowest on AA and NJ. Finally, the adolescents in the no-risk class scored the highest on HRQL.

Table 3.

Mean Differences Between Classes of Adolescents

	Class1 No risk n = 1,863	Class 2 Only cyberbullying risk n = 785	Class 3 Cyberbullying and cyber dating abuse n = 132	Class 4 All risks n = 207	Class 5 Sexual risk n = 89	F	p
Cyberbullying victimization	0.053 (0.008)_a	0.411 (0.012)_c	0.260 (0.029)_b	0.677 (0.023)_e	0.543 (0.036)_d	296	<0.001
Cyberbullying perpetration	0.020 (0.006)_a	0.316 (0.009)_b	0.068 (0.023)_a	0.53 (0.02)_d	0.42 (0.03)_c	336	<0.001
Cyber dating abuse victimization	0.00 (0.01)_a	0.018 (0.013)_a	0.178 (0.019)_b	0.361 (0.015)_c	0.00 (0.030)_a	108	<0.001
Cyber dating abuse perpetration	0.003 (0.009)_a	0.012 (0.011)_a	0.098 (0.016)_b	0.25 (0.01)_c	0.023 (0.03)_a	67	<0.001
Grooming	0.008 (0.01)_a	0.09 (0.01)_b	0.14 (0.02)_b	0.42 (0.02)_c	0.70 (0.03)_d	305	<0.001
Sexting	0.005 (0.007)_a	0.049 (0.01)_a	0.076 (0.025)_a	0.447 (0.02)_b	0.97 (0.03)_c	339	<0.001
Observing	2.86 (0.02)_a	2.96 (0.03)_a	2.74 (0.07)_b	2.93 (0.06)_a	3.02 (0.09)_a	4	0.006
Describing	3.37 (0.017)_a	3.26 (0.026)_a	3.31 (0.06)_a	3.20 (0.05)_a	3.23 (0.05)_a	6	<0.001
Acting awareness	3.44 (0.02)_a	3.09 (0.03)_b	3.14 (0.07)_b	2.90 (0.06)_c	2.99 (0.09)_b	44	<0.001
Nonjudging	3.71 (0.02)_a	3.34 (0.03)_b	3.47 (0.07)_b	2.98 (0.06)_c	3.04 (0.09)_c	63	<0.001
Nonreacting	2.86 (0.02)_a	2.84 (0.03)_a	2.67 (0.02)_b	2.67 (0.05)_b	2.96 (0.08)_a	6	<0.001
Antisocial media exposure	1.02 (0.03)_a	1.52 (0.03)_b	1.53 (0.08)_b	1.95 (0.06)_c	1.98 (0.10)_c	95	<0.001
Risk perception	6.03 (0.05)_a	5.62 (0.08)_b	5.23 (0.20)_b	5.54 (0.16)_b	5.59 (0.24)_b	9	<0.001
Health-related quality of life	4.18 (0.014)_a	3.91 (0.02)_b	3.80 (0.05)_b	3.64 (0.04)_b	3.78 (0.063)_b	63	<0.001
Age	13.73 (0.03)_a	14.12 (0.05)_b	14.54 (0.13)_c	14.54 (0.10)_c	15.22 (0.10)_d	47	<0.001

_a,b,c,d,eMeans sharing a subscript in a row indicate means that are not significantly different from each other.

Overall, the girls had higher prevalence rates in all the classes except for the sexual-risk class in which the rates for the boys and girls were both 2.9%. The membership in the no-risk class was higher for boys than for girls (64% vs. 57.6%), χ²(4, n = 3,076) = 27, p < 0.001. There were significant differences in age, F(4, 3,071) = 47, p < 0.001). The youngest participants were in the no-risk class, and the oldest adolescents were in the sexual-risk class (Table 3).

RO2: mediational model

We used the no-risk class as the reference class. Dummy variables were used to include membership within the other classes. DES and OB were not associated with any class and were eliminated from the model. Figure 2 displays the significant paths of the model. AA and NJ were negatively associated with antisocial media exposure and this, in turn, was positively associated with all the classes. AA and NR were positively associated with risk perception, whereas NJ was negatively associated. Risk perception was, in turn, negatively associated with only cyberbullying and risk of cyberbullying and cyberdating abuse classes.

FIG. 2.

Mediational model between mindfulness, cyber risks, and health-related quality of life. *p < 0.05, **p < 0.001.

Bootstrapping results displayed that the above mediational paths linking DM with classes were significant, as the confidence interval (CI) did not include zero (Table 4). Moreover, DM facets were positively associated with HRQL, whereas all the classes were negatively associated with HRQL. NJ and NR were directly and positively associated with the all-risks class, and exposure to antisocial media was associated with lower HRQL. Bootstrapping results indicated that the indirect effects from DM facets to HRQL through antisocial media exposure, risk perception, and classes were significant. Indirect effects were more prominent for AA. The fit indices of this model were excellent, S-Bχ²(1, n = 3,076) = 0.47, RMSEA = 0.00 (90% CI [.00 to .044]); non-normed fit index = 0.999, CFI = 0.999, SRMR = 0.0023, S-Bχ²/df = 0.47.

Table 4.

Mediational Paths Between Dispositional Mindfulness and Internet Risk Classes

Mediational path	95% CI^*
Acting with awareness—exposure to antisocial media—sexual risk class	−.00419 to −.00414
Acting with awareness—exposure to antisocial media—cyberbullying and cyber dating abuse class	−.00237 to −.00231
Acting with awareness—exposure to antisocial media—only cyberbullying class	−.01248 to −.01236
Acting with awareness—exposure to antisocial media—all risks class	−.00959 to −.00950
Nonjudging—exposure to antisocial media—all risks class	−.00503 to −.00497
Nonjudging—exposure to antisocial media—sexual risk class	−.00219 to −.00217
Nonjudging—exposure to antisocial media—cyberbullying and cyber dating abuse class	−.00123 to −.00121
Nonjudging—exposure to antisocial media—only cyberbullying class	−.00654 to −.00645
Acting with awareness—risk perception—cyberbullying and cyber dating abuse class	−.00143 to −.00139
Nonreacting—risk perception—only cyberbullying class	−.0014 to −.0013
Nonreacting—risk perception—cyberbullying and cyber dating abuse class	−.00279 to −.00273
Acting with awareness—risk classes—HRQL	.1718 to .1735
Nonjudging—risk classes—HRQL	.0893 to .0904
Nonreacting—risk classes—HRQL	.0055 to .0056

All the mediational paths are statistically significant at p < 0.001, as their 95% CIs do not include zero.

CI, confidence interval; HRQL, health-related quality of life.

Discussion

This study contributes to demonstrating that many of the risks on the Internet for adolescents do not occur in isolation, but cumulatively. Five classes emerged as the best solution to explain the involvement of adolescents in risks on the Internet. Most of the adolescents were classified in the no-risk class. They displayed an almost zero probability of presenting Internet risks. Furthermore, they were the youngest adolescents in the sample and had the highest scores in HRQL, AA, and NJ, and the lowest scores in exposure to antisocial content in the media. There were more boys than girls in this class.

The only cyberbullying class included those with a very high probability of presenting victimization and perpetration of cyberbullying, but very low scores in the other Internet risks. This class concurred with the often reported category of cyberbullying perpetrator/victim.¹⁰ The third class, the all-risks class, displayed high probabilities of most of the Internet risks, especially cyberbullying and cyberdating abuse. Furthermore, those in this class scored high on exposure to antisocial content in the media and low on the DM facets of AA, NJ, and NR. The fourth class, cyberbullying and cyberdating abuse class, presented a high probability of cyberdating abuse victimization, and a moderate probability of cyberdating perpetration and cyberbullying victimization and perpetration.

The final class, the sexual-risk class, displayed the highest probability of performing sexting (100%) and of being a victim of grooming (78%). Furthermore, those in this class had a high probability of being involved in cyberbullying behavior, either as victims or as perpetrators. This group, together with the all-risks class, presented the highest scores in exposure to antisocial content in the media. Furthermore, it included older adolescents and had an equal number of boys and girls. It is noteworthy that those in the sexual-risk class scored high on the perception of Internet risks; the latter may have been a result of the negative experiences of grooming that they had experienced.

We also tested the potential beneficial role of mindfulness facets against Internet risks. AA, NR, and NJ were negatively associated with belonging to these classes. This finding concurs with the results of previous studies, which indicated that DM is associated with less problematic Internet behaviors.^17,18,22 Our study enhances information about the associations between specific DM facets and classes of adolescents that integrate several modalities of Internet risks.

Furthermore, this study contributes to clarifying some mechanisms through which DM is associated with fewer cyber risks. Exposure to antisocial content in the media played a relevant role. AA and NJ, to a lesser extent, were associated with less exposure to antisocial content. This exposure, in turn, was significantly associated with belonging to Internet-risk classes, and, in particular, the all-risks class. Theoretical proposals suggest that mindfulness increases prosocial behaviors through the focus on positive relationships between the self and the others.²⁵ Thus, adolescents who are high in these DM facets could avoid antisocial content in the media, especially content displaying aggressive behaviors. Lower exposure to negative content, in turn, decreases the probability of becoming involved in Internet risks.²⁸

The mediating role of risk perception was modest. On the one hand, AA and NR were associated with higher scores in risk perception. Contrary to our expectations, NJ was negatively associated with risk perception. This facet involves accepting things as they are and could result in young people becoming more confident on the Internet and not avoiding risky behaviors such as sharing their contact details with strangers on the Internet.

Finally, DM was associated with better HRQL, and part of this association was explained by a reduced involvement in Internet-risk classes. AA was the facet that displayed a more prominent indirect effect on HRQL through a lower involvement in Internet risks.

This study had various limitations that provide challenges for future studies. First, the study was cross-sectional and, therefore, future longitudinal studies should examine directional relationships between variables. Second, the exclusive use of self-reports may have increased associations between variables and should have been completed with peer reports. Finally, the sample was not random and therefore, the classes obtained in this study should be replicated in different samples.

Despite the limitations, the findings have implications for interventions. When educators and parents detect that an adolescent is involved in problematic Internet behavior, they should explore the co-occurrence of other risks, as suggested by the classes identified in this study. This approach could better inform the interventions by including specific strategies to cope with the various risks in which the adolescent is involved. Furthermore, the results suggest that risks tend to increase with age. Therefore, preventive interventions should be implemented early. Some mindfulness facets, namely, AA, NJ, and NR, have shown a beneficial role against Internet risks. Therefore, mindfulness-based interventions may promote safe and healthy behaviors on the Internet through the training of awareness and acceptance. These characteristics have been theoretical and empirically associated with self-regulation and prosocial behavior,^25,27 which are important to prevent risk behaviors on the Internet.

In conclusion, findings display five classes of adolescents according to their involvement in Internet risks. Some DM facets reduce the probability of being involved in classes characterized by Internet risks, in part through lower exposure to antisocial media and higher perception of risks. DM is associated with better HRQL, in part through lower involvement in Internet risks.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This research was supported by a grant from the Basque Country (Ref. IT982-16) and the BBVA Foundation (PR[18]_SOC_0096).

References

González-Cabrera

, León-Mejía

, Beranuy

, et al. Relationship between cyberbullying and health-related quality of life in a sample of children and adolescents. Quality of Life Research, 2018; 27:2609–2618.

Wartberg

, Kriston

, Kammerl

. Associations of social support, friends only known through the internet, and health-related quality of life with internet gaming disorder in adolescence. Cyberpsychology, Behavior, and Social Networking, 2017; 20:436–441.

Patchin

, Hinduja

. Bullies move beyond the schoolyard. Youth Violence and Juvenile Justice, 2006; 4:148–169.

Brochado

, Soares

, Fraga

. A scoping review on studies of cyberbullying prevalence among adolescents. Trauma, Violence, & Abuse, 2017; 18:523–531.

Calvete

, Gámez-Guadix

, Borrajo

. (2019) Cyber abuse in romantic relationships: measurement and characteristics from an international perspective. In Giumetti

, Kowalsk

, eds. Cyberbullying in schools, workplaces, and romantic relationships. New York, NY: Routledge, pp. 150–183.

Agustina

, Gómez-durán

. Sexting: research criteria of a globalized social phenomenon. Archives of Sexual Behavior, 2012; 41:1325–1328.

Madigan

, Ly

, Rash

, et al. Prevalence of multiple forms of sexting behavior among youth: a systematic review and meta-analysis. JAMA Pediatrics, 2018; 172:327–335.

Smith

, Steffgen

. (2013) Cyberbullying through the new media: findings from an international network. Florence, KY: Psychology Press.

Bergen

, Davidson

, Schulz

, et al. The effects of using identity deception and suggesting secrecy on the outcomes of adult-adult and adult-child or -adolescent online sexual interactions. Victims & Offenders, 2014; 9:276–298.

10.

Fahy

, Stansfeld

, Smuk

, et al. Longitudinal associations between cyberbullying involvement and adolescent mental health. Journal of Adolescent Health, 2016; 59:502–509.

11.

Quesada

, Fernández-González

, Calvete

. El sexteo (sexting) en la adolescencia: frecuencia y asociación con la victimización de ciberacoso y violencia en el noviazgo. [Sexting in adolescence: frequency and association with the victimization of cyberbullying and dating violence.] Behavioral Psychology, 2018; 26:225–242.

12.

Gámez-Guadix

, Mateos-Pérez

. Longitudinal and reciprocal relationships between sexting, online sexual solicitations, and cyberbullying among minors. Computers in Human Behavior, 2019; 94:70–76.

13.

Evans

, Li

, Whipple

. Cumulative risk and child development. Psychological Bulletin, 2013; 139:1342.

14.

Kabat-Zinn

Mindfulness-based interventions in context: past, present, and future. Clinical Psychology: Science and Practice, 2003; 10:144–156.

15.

Baer

, Smith

, Hopkins

, et al. Using self-report assessment methods to explore facets of mindfulness. Assessment, 2006; 13:27–45.

16.

Siebelink

, Asherson

, Antonova

, et al. Genetic and environmental aetiologies of associations between dispositional mindfulness and ADHD traits: a population-based twin study. European Child and Adolescent Psychiatry, 2019; 28:1241–1251.

17.

Kircaburun

, Griffiths

, Billieux

. Trait emotional intelligence and problematic online behaviors among adolescents: the mediating role of mindfulness, rumination, and depression. Personality and Individual Differences, 2019; 139:208–213.

18.

Elhai

, Levine

, O'Brien

, et al. Distress tolerance and mindfulness mediate relations between depression and anxiety sensitivity with problematic smartphone use. Computers in Human Behavior, 2018; 84:477–484.

19.

Calvete

, Gámez-Guadix

, Cortazar

. Mindfulness facets and problematic internet use: a six-month longitudinal study. Addictive Behaviors, 2017; 72:57–63.

20.

Gámez-Guadix

, Calvete

. Assessing the relationship between mindful awareness and problematic internet use among adolescents. Mindfulness, 2016; 7:1281–1288.

21.

Emirtekin

, Balta

, Kircaburun

, et al. Childhood emotional abuse and cyberbullying perpetration among adolescents: the mediating role of trait mindfulness. International Journal of Mental Health and Addiction, 2019; 1–12, DOI: 10.1007/s11469-019-0055-5.

22.

Royuela-Colomer

, Calvete

, Gámez-Guadix

, et al. The protective role of dispositional mindfulness against the perpetuation of cyberbullying victimization and perpetration among adolescents. Cyberpsychology, Behavior, and Social Networking, 2018; 21:703–710.

23.

Hölzel

, Lazar

, Gard

, et al. How does mindfulness meditation work? Proposing mechanisms of action from a conceptual and neural perspective. Perspectives on Psychological Science, 2011; 6:537–559.

24.

Garofalo

, Gillespie

, Velotti

. Emotion regulation mediates relationships between mindfulness facets and aggression dimensions. Aggressive Behavior, 2020; 46:60–71.

25.

Vago

, David

. Self-awareness, self-regulation, and self-transcendence (S-ART): a framework for understanding the neurobiological mechanisms of mindfulness. Frontiers in Human Neuroscience, 2012; 6:1–30.

26.

Sibinga

, Copeland-Linder

, Webb

, et al. Improving self-regulation in adolescents: current evidence for the role of mindfulness-based cognitive therapy. Adolescent Health, Medicine and Therapeutics, 2016; 7:101–108.

27.

Strasburger

, Jordan

, Donnerstein

. Health effects of media on children and adolescents. Pediatrics, 2010; 125:756–767.

28.

Den Hamer

, Konijn

, Keijer

. Cyberbullying behavior and adolescents' use of media with antisocial ontent: a cyclic process model. Cyberpsychology, Behavior, and Social Networking, 2014; 17:74–81.

29.

Calvete

, Orue

, Estévez

, et al. Cyberbullying in adolescents: modalities and aggressors' profile. Computers in Human Behavior, 2010; 26:1128–1135.

30.

Borrajo

, Gámez-Guadix

, Pereda

, et al. The development and validation of the cyberdating abuse questionnaire among young couples. Computers in Human Behavior, 2015; 48:358–365.

31.

Gámez-Guadix

, de Santisteban

, Resett

. Sexting among Spanish adolescents: prevalence and personality profiles | Sexting entre adolescentes españoles: prevalencia y asociación con variables de personalidad. Psicothema, 2017; 29:29–34.

32.

Gámez-Guadix

, De Santisteban

, Alcazar

MÁ

. The construction and psychometric properties of the questionnaire for online sexual solicitation and interaction of minors with adults. Sexual Abuse: Journal of Research and Treatment, 2018; 30:975–991.

33.

Cortazar

, Calvete

, Fernández-González

, et al. Development of a short form of the Five Facet Mindfulness Questionnaire—adolescents for children and adolescents. Journal of Personality Assessment, 2019:1–12, DOI: 10.1080/00223891.2019.1616206.

34.

Den Hamer

, Konijn

, Bushman

. Measuring exposure to media with antisocial and prosocial content: an extended version of the content-based media exposure scale (C-ME2). Communication Methods and Measures, 2017; 11:289–299.

35.

Paluckaitė

, Žardeckaitė-Matulaitienė

Adolescents' perception of risky behaviour on the internet. In The European proceedings of social & behavioural sciences EpSBS: 3rd icH&Hpsy international conference on health and health psychology, July 5–7, 2017, Porto, vol. 14, pp. 284–292.

36.

The Kidscreen Group Europe. (2006) The KIDSCREEN questionnaires: quality of life questionnaires for children and adolescents. Lengerich: Pabst Science Publishers.

37.

Muthén

L.K.

& Muthén

B.O.

(1998–2015). Mplus user's guide. Seventh edition. Los Angeles, CA: Muthén & Muthén.

38.

Henson

, Reise

, Kim

. Detecting mixtures from structural model differences using latent variable mixture modeling: a comparison of relative model fit statistics. Structural Equation Modeling, 2007; 14:202–226.

39.

Nylund-Gibson

, Choi

. Ten frequently asked questions about latent class analysis. Translational Issues in Psychological Science, 2018; 4:440.

40.

Chou

, Bentler

, Satorra

. Scaled test statistics and robust standard errors for non-normal data in covariance structure analysis: a Monte Carlo study. British Journal of Mathematical and Statistical Psychology, 1991; 44:347–357.