Adolescents’ Risky Online Behaviors and Social Desirability Bias in Vietnam: Privacy Disclosure and Pornography Use

Abstract

This study quantified the prevalence of two risky online behaviors among adolescents in Vietnam: online privacy disclosure and online pornography use. We conducted a field experiment with 1,313 junior high school students aged 13–15 years in Hoa Binh city. In addition to conventional direct questions, we employed list experiments to address social desirability bias in the students’ responses. The results indicated that 49.9% of the adolescents engaged in online privacy disclosure and that 58.5% were involved in online pornography use. This study revealed significant underreporting among adolescents (35.6 and 43.3 percentage points for privacy disclosure and pornography use, respectively). The heterogeneous analyses revealed that recent smartphone ownership and active smartphone and Facebook use were associated with a greater prevalence of these behaviors. Notably, urban adolescents showed greater engagement in pornography use than did their rural counterparts. This study represents a pioneering effort to empirically investigate sensitive online behaviors among adolescents utilizing an experimental approach to address measurement bias. Our findings suggest that list experiments are a robust method for assessing sensitive issues and emphasize the critical need for educational interventions to mitigate online risks faced by young people.

Introduction

The recent global surge in internet access not only provides new opportunities for adolescents, who often lack media skills, but also exposes them to various risks.^1,2 This phenomenon is not restricted to developed countries—it has also become prevalent in developing countries. Adolescents are particularly vulnerable to online interactions, which can impact their physical, behavioral, and mental well-being.^3–5 Among the various online behaviors that pose negative consequences for their future, two are especially concerning: (i) privacy disclosure and (ii) pornography use. This study aimed to identify (i) the prevalence of these behaviors among adolescents and (ii) possible heterogeneity in the prevalence of these behaviors across adolescent characteristics, such as gender and residential location, in Vietnam—a country of fast-paced economic growth.

Online risky behaviors pose critical challenges in developing economies for three reasons. First, information and communication technologies are rapidly advancing even in developing economies; however, these economies often lack robust regulatory frameworks and have less stringent social protection policy enforcement.⁶ Second, digital literacy levels in developing economies are lower than those in developed economies,⁷ making their populations more susceptible to online threats. Third, exposure to harmful online content, such as cyberbullying and graphic imagery, can have detrimental effects on mental health, given the limited access to mental health resources in developing regions.⁸ Thus, understanding risky online behaviors so that online risks can be addressed in developing economies is crucial not only for protecting individuals but also for fostering inclusive and sustainable digital societies.

Privacy disclosure and pornography use can result in various negative consequences for adolescents. First, personal data breaches may precipitate serious risks such as cyberstalking and bullying. Hayes et al.⁹ highlighted a link between online self-disclosure by children and their experiences with cyberbullying. Experiencing this type of victimization increases the risk of experiencing mental health issues and behavioral problems, including depressive symptoms, substance abuse, and physical aggression, over time¹⁰ and may even lead to suicidal tendencies.¹¹ Second, exposure to sexual content can present challenges to healthy sexuality development in adolescents.¹² Specifically, pornography use is connected with premature sexual activity, unintended pregnancies, sexually transmitted infections,¹³ sexting, and sexually aggressive behaviors.^14,15

Notably, most of the past studies^a on these two problematic online activities (privacy disclosure and pornography use) among adolescents have relied largely on traditional survey methods, such as direct questioning and interviews, which often suffer from measurement bias, known as social desirability bias. This bias occurs in self-report surveys when respondents adjust their responses to avoid embarrassment and align with societal expectations, leading to a misrepresentation of their actual behaviors.¹⁶

In this study, social desirability bias was effectively mitigated by applying list experiments. Our study is the first to attempt to measure the prevalence of risky online behaviors among adolescents using an experimental study framework, which is the main contribution to the literature. In addition, this pioneering study estimated adolescents’ behaviors using a dual-method approach that combined direct questioning with indirect questioning. This method was instrumental in eliciting more truthful responses regarding the adolescents’ sensitive online behaviors and in quantifying the social desirability bias in these responses.

Two list experiments—one on privacy disclosure and the other on pornography use—were conducted to evaluate the prevalence of these two risky online behaviors among adolescents. The sample included 1,313 adolescents aged 13 to 15 years from Hoa Binh, Vietnam.^b The findings revealed that 49.9% of participants engaged in privacy disclosure and 58.5% in online pornography use. Adolescents underreported risky online behaviors in direct questioning, likely due to social desirability bias.

Theoretical Framework

This study follows the framework of Bandura’s¹⁷ social cognitive theory (SCT), which highlights the interplay among people, the environment, and behavior in explaining human behavior.^18,c SCT emphasizes the active roles of individuals and environments in shaping behaviors, making it particularly relevant for examining adolescents’ online activities. By situating risky online behaviors within this theoretical framework, this study explores how these factors influence adolescents’ engagement in privacy disclosure and pornography use.

Individual factors, such as self-efficacy, are associated with adolescents’ engagement in risky behaviors.^19,20 Environmental factors, including peer influence, family mediation, and media exposure, play a significant role in shaping adolescents’ behaviors.^20–22 Moreover, social desirability bias can distort adolescents’ responses to sensitive topics,¹⁶ such as pornography use. We discuss these factors and their relevance to online risky behaviors in detail in the Supplementary Data.

Methods

Data and sample

Our study included 1,313 adolescents aged between 13 and 15 years from 18 junior high schools in Hoa Binh city, Vietnam. The city has a total of 29 junior high schools. Seven schools had previously participated in an online safety training program. These seven schools were excluded because we aimed to examine the nuanced prevalence of online behaviors among students. From the remaining 22 schools, four were randomly selected for a pilot experiment. Finally, 18 schools, comprising 110 classes, participated in this study. From these, we randomly selected 40 classes, and all the students in the selected classes were included, except for five students with impairments and nine who chose not to join. Table 1 presents descriptive statistics for our sample.

Table 1.

Descriptive Statistics

	N	Mean	SD	Min	Max
Individual characteristics
Urban residence (dummy)	1,313	0.516	0.500	0	1
Grade (7, 8, 9)	1,313	8.023	0.802	7	9
Female (dummy)	1,313	0.476	0.500	0	1
Kinh ethnicity (dummy)	1,313	0.334	0.472	0	1
Muong ethnicity (dummy)	1,313	0.614	0.487	0	1
GPA	906	6.423	1.036	3.6	8.9
Family background
Paternal age (years)	1,229	43.192	5.729	31	63
Tertiary-educated father (dummy)	1,236	0.163	0.370	0	1
Farmer father (dummy)	1,247	0.760	0.427	0	1
Maternal age (years)	1,298	39.958	4.710	26	63
Tertiary-educated mother (dummy)	1,295	0.192	0.394	0	1
Farmer mother (dummy)	1,298	0.653	0.476	0	1
Both parents at home (dummy)	1,313	0.851	0.356	0	1
Online practices
Home PC user (dummy)	1,313	0.306	0.461	0	1
Home tablet user (dummy)	1,313	0.174	0.380	0	1
Smartphone owner (dummy)	1,313	0.743	0.437	0	1
Duration of smartphone ownership (years)	1,313	1.292	1.358	0	9
Facebook user (dummy)	1,313	0.899	0.301	0	1
Duration of Facebook usage (years)	1,209	1.674	1.002	0	3
Screen time (hours/day)	1,230	4.806	2.603	0	10

N, number of observations; SD, standard deviation.

Experimental design

To evaluate two specific sensitive online behaviors among adolescents, two list experiments were employed in this study. List experiment 1 examined the prevalence of privacy disclosure behavior, whereas list experiment 2 explored the prevalence of pornography use behavior. A total of 1,313 participants were randomly assigned to the control and treatment groups for each experiment. The control group was presented with a list of four nonsensitive items, whereas the treatment group received a list that included the same four items plus an additional sensitive item. The sensitive item for list experiment 1 was “I have shared or posted others’ personal information online without obtaining consent in the last 6 months.” The sensitive item for list experiment 2 was “I have watched pornography online in the last 6 months.” The participants were instructed to indicate the total number of items with which they agreed, without specifying which items they agreed with. All the items employed are provided in Supplementary Data Tables B.2 and B.3.

After answering the list experiment question, respondents in the control group who were not exposed to the sensitive item in the list experiment were asked a direct question (i.e., a yes/no question). In list experiment 1, the respondents were asked, “Have you shared or posted others’ personal information online without obtaining consent in the last 6 months?”. The corresponding question for list experiment 2 was “Have you watched pornography online in the last 6 months?”. This approach allowed us to estimate social desirability bias in the adolescents’ responses by comparing the mean response values between the response to the list experiment question and the direct question.

List experiment assumptions

The validity of the list experiment depended on three fundamental assumptions: (i) randomization of the treatment, (ii) no design effects, and (iii) no ceiling and floor effects. Figure 1, Table 2, and Table 3 present the results of the tests for these three assumptions, respectively. All three assumptions were satisfied. Further details on these assumptions are provided in the Supplementary Data.

FIG. 1.

Balance check (90% CIs). 90% confidence intervals (CIs) used instead of the standard 95% CIs to carefully check the balance of the adolescents’ characteristics between the control and treatment groups. Continuous variables are standardized (denoted as “std”) in the coefficient plots.

Table 2.

Tests for Design Effects

	List experiment 1Privacy disclosure		List experiment 2Pornography use
	Estimated proportion	Robust SE	Estimated proportion	Robust SE
π (y = 0, z = 1)	0.011	0.008	0.011	0.007
π (y = 0, z = 0)	0.017	0.005	0.012	0.004
π (y = 1, z = 1)	0.106	0.025	0.161	0.024
π (y = 1, z = 0)	0.204	0.018	0.165	0.016
π (y = 2, z = 1)	0.192	0.027	0.224	0.027
π (y = 2, z = 0)	0.140	0.027	0.107	0.027
π (y = 3, z = 1)	0.177	0.017	0.165	0.018
π (y = 3, z = 0)	0.127	0.024	0.109	0.024
π (y = 4, z = 1)	0.014	0.004	0.024	0.006
π (y = 4, z = 0)	0.013	0.008	0.022	0.010

Respondent types, $π_{y z}$ , are defined by the sum of yes responses to control items, denoted by $y$ , and truthful answers to the sensitive item. 1 represents an affirmative response, while 0 indicates a negative response. Bonferroni-adjusted p values equal 1.

Table 3.

Distribution of Responses

	Control group		Treatment group
Response value	Frequency	Proportion (%)	Frequency	Proportion (%)
List experiment 1—Privacy disclosure
0	18	2.78	11	1.65
1	201	31.02	143	21.50
2	215	33.18	164	24.66
3	197	30.40	212	31.88
4	17	2.62	126	18.95
5	—	—	9	1.35
Total	648	100	665	100
List experiment 2—Pornography use
0	15	2.30	8	1.21
1	213	32.62	116	17.58
2	216	33.08	177	26.82
3	179	27.41	220	33.33
4	30	4.59	123	18.64
5	—	—	16	2.42
Total	653	100	660	100

Empirical strategy

This study sought to estimate the prevalence of sensitive online behaviors among adolescents. Given the fulfillment of the three previously mentioned assumptions, the difference in the average number of agreed-upon statements between the control and treatment groups could provide an unbiased estimation of the average treatment effect (ATE). To this end, we utilized difference-in-means (DiM), a widely adopted estimator in item count technique analysis,²³ to calculate the ATE. This, in turn, reflected the prevalence of adolescent behaviors of interest: $\hat{τ} = \frac{1}{N_{1}} \sum_{i = 1}^{N} T_{i} Y_{i} - \frac{1}{N_{0}} \sum_{i = 1}^{N} {(1 - T}_{i}) Y_{i},$ (1)where $\hat{τ}$ represents the estimated proportion of respondents with an affirmative response to the sensitive statement. $T_{i}$ is a binary indicator assigned a value of 1 for respondents in the treatment group and 0 for those in the control group. $Y_{i}$ denotes the number of statements with which the respondent $i$ agreed. The total number of respondents in the treatment group is denoted by $N_{1} = \sum_{i = 1}^{N} T_{i}$ , whereas $N_{0} {= N - N}_{1}$ signifies the control group’s size.

We further applied regression analysis as a robustness check of our results by using the following equation: $Y_{i} = α + β T_{i} + X_{i} λ + ε_{i},$ (2)where $β$ is the prevalence of sensitive behaviors, and where $X_{i}$ is a set of covariates.²⁴^,d

Results and Discussion

Privacy disclosure

The first list experiment quantified the extent of privacy disclosure among the adolescents in our study. Table 4 shows that 49.9% of the adolescents engaged in sharing others’ personal data without proper consent (95% confidence interval [CI]: [0.389, 0.610]). In contrast, when asked directly, only 14.4% of the adolescents confessed to such behavior (95% CI: [0.116, 0.171]). A comparison of the results from the list experiment and direct questioning indicated that the adolescents underreported their privacy disclosure behavior by 35.6 percentage points. This difference was statistically significant at the 1% level.

Table 4.

Estimated Prevalence and Underreporting

	Mean number of responses		List experiment	Direct questioning	Underreporting
	Control	Treatment	List experiment	Direct questioning	Underreporting
	(1)	(2)	(3)	(4)	(5) = (3) − (4)
List 1—Privacy disclosure	1.991	2.490	0.499***	0.144	0.356***
	[0.913]	[1.099]	(0.068)
Observations	648	665	1,313	648
List 2—Pornography use	1.994	2.579	0.585***	0.152	0.433***
	[0.915]	[1.071]	(0.067)
Observations	653	660	1,313	653

Standard deviations are shown in brackets. Standard errors are shown in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.

Pornography use

Our second list experiment focused on adolescents’ involvement in online pornography use. Table 4 shows that 58.5% of the adolescents engaged in this activity (95% CI: [0.475, 0.695]). However, direct questioning revealed that only 15.2% of the respondents admitted to engaging in pornography use (95% CI: [0.124, 0.179]). The underreporting of pornography use was even more pronounced than that of privacy disclosure, at 43.3 percentage points, and was statistically significant at the 1% level. Importantly, our ordinary least squares regression analysis results corroborated these findings, indicating similar prevalence rates to the DiM results, even after adjusting for covariates (Table 5).

Table 5.

Estimated Prevalence Using DiM and OLS

	Mean number of responses		Estimated prevalence
	Control	Treatment	DiM	OLS
	(1)	(2)	(3)	(4)	(5)
List 1—Privacy disclosure	1.991	2.490	0.499***	0.499***	0.577***
	[0.913]	[1.099]	(0.068)	(0.068)	(0.062)
Observations	648	665	1,313	1,313	1,055
List 2—Pornography use	1.994	2.579	0.585***	0.585***	0.584***
	[0.915]	[1.071]	(0.067)	(0.067)	(0.065)
Observations	653	660	1,313	1,313	1,055
Covariates				No	Yes

Standard deviations are shown in brackets. Robust standard errors are shown in parentheses. Covariates (19): Individual characteristics (urban residence, grade, female, Kinh ethnicity, and Muong ethnicity); family background (paternal age, tertiary-educated father, farmer father, maternal age, tertiary-educated mother, farmer mother, and both parents at home); and online practices (home PC user, home tablet user, smartphone owner, duration of smartphone ownership, Facebook user, duration of Facebook usage, and screen time). * p < 0.10, ** p < 0.05, *** p < 0.01.

Social desirability bias

In this subsection, we discuss the influence of social desirability bias on self-reported behaviors, particularly in the context of sensitive subjects. This bias often leads to the underreporting of behaviors perceived as socially unacceptable. Considering the hesitance of adolescents in developing countries to discuss taboo topics directly, we compared our findings with those of recent studies in similar contexts. Yu et al.¹² reported that the self-reported prevalence of pornography use among adolescents in developing countries varied from 14.5% in Ecuador to 19.6% in Indonesia, as determined through direct questioning. Notably, these figures align closely with our findings using the direct questioning method, where 15.2% of adolescents reported pornography use in the past 6 months, despite Yu et al.¹² focusing on lifetime exposure.

Our experimental analysis demonstrated the presence of social desirability bias by contrasting prevalence rates derived from list experiments with those obtained from direct questioning. Specifically, list experiment 2 indicated that 58.5% of the adolescents engaged in pornography use, revealing a substantial social desirability bias of 43.3 percentage points. These results imply that the prevalence rates reported in previous studies, especially those using direct questioning methods in developing countries,¹² are likely underestimated because of the presence of social desirability bias. Consequently, the utilization of list experiments is crucial for accurately gauging the extent of sensitive online behaviors among adolescents.

Our study revealed a critical issue: adolescents tend to underreport their risky online behaviors. This highlights a notable gap in our comprehension of adolescents’ online conduct. Overlooking this aspect could lead to ineffective digital protection strategies. Our study provides empirical evidence for policymakers. This scientific evidence is essential for shaping timely and targeted interventions that reflect the actual online behaviors of adolescents.

Heterogeneity analyses

We investigated the disparities in the prevalence of online behaviors across respondent attributes to determine which adolescents were more involved in risky online activities. Supplementary Data Figures B.1, B.2 and Table B.5 show the conditional average treatment effects derived from our two experimental setups. Supplementary Data. The analysis indicated a greater prevalence of risky online behaviors among adolescents who engaged in extensive daily screen use and long-term Facebook usage. Adolescents who spent 5 or more hours per day on electronic devices showed greater engagement in risky online behaviors than those who spent less than 5 hours per day on electronic devices, as illustrated in Figure 2(a). Similarly, adolescents who had used Facebook for at least 2 years were more involved in risky online behaviors than were those who had used Facebook for less than 2 years, as illustrated in Figure 2(b). These observations are concordant with those of prior studies. Livingstone and Helsper¹ identified a positive link between online opportunities and subsequent risks. In a meta-analysis, Vannucci et al.²² confirmed a positive correlation between social media use in adolescence and engagement in risky behaviors. Furthermore, heavy internet usage and prolonged online time are correlated with increased interactions with pornography.^25,26

Our results also demonstrated that adolescents who had recently acquired smartphones (less than 2 years of ownership) presented a greater prevalence of both risky online behaviors than did those who had owned smartphones for more than 2 years (Figure 2(c)). Recent smartphone acquisition may heighten adolescents’ interest in digital exploration, coupled with a potential lack of online privacy awareness. These factors might contribute to adolescents’ increased engagement in risky online activities. This finding is novel in the literature and calls for future research to explore this phenomenon.

FIG. 2.

Heterogeneity (95% CIs).

A notable aspect of our results is the statistically significant difference, at the 1% level, in online pornography use between urban and rural adolescents (Figure 3(a)). The prevalence rate among adolescents living in urban areas was 73.3%, whereas it was 43.2% among those living in rural areas. This pattern aligns with the findings of Yu et al.,¹² who suggested lower pornography use among adolescents in cohesive neighborhood environments. These environments are commonly found in rural Vietnamese settings. Conversely, the relative anonymity and privacy experienced by those in urban areas might allow adolescents more freedom to pursue their interests without community judgment.

Our study also explored gender-related patterns in sensitive online behaviors (Figure 3(b)). Despite observing engagement by both males and females (51.1% vs. 48.7% for privacy disclosure, 65.0% vs. 51.4% for pornography use), the analysis did not reveal statistically significant gender-based differences. The observations regarding gender heterogeneity in the literature vary. Andrie et al.²⁵ and Yu et al.¹² reported a greater tendency for online pornography use among males, whereas Maheux et al.²⁷ reported no significant gender disparity in pornography use. These contrasting findings imply the variable nature of gender influences on online behaviors. Our results align with this variability, suggesting that while there may be differences in online behavior between genders, these differences may not always be significant. It is essential to acknowledge that individual experiences (e.g., online practices) and societal factors (e.g., living environment) may influence online behavior patterns, as shown in our abovementioned analysis.

FIG. 3.

Heterogeneity (95% CIs).

These findings are consistent with the SCT framework.¹⁷ Adolescents who recently acquired smartphones presented a greater prevalence of risky online behaviors, which may stem from individual factors such as a lack of online privacy awareness and limited media literacy. The increased prevalence of risky behaviors among those actively using social media or living in urban areas can be explained by environmental factors. Digital platforms, social networks, and urban environments expose adolescents to engaging in risky behaviors, which are often perceived as normalized or acceptable. Finally, the underreporting of sensitive online behaviors due to social desirability bias highlights the influence of environmental factors. When asked directly, adolescents tended to respond in ways aligned with socially acceptable norms.

Conclusions

This study empirically investigated two aspects of adolescents’ online behaviors: online disclosure of privacy and online pornography use. We applied an experimental approach involving 1,313 students in Vietnam. To address the challenge of social desirability bias in measuring sensitive behaviors, a combination of list experiments and direct questioning methods was employed in this study.

The results revealed the prevalence of risky online behaviors among adolescents and the magnitude of social desirability bias in their responses. Specifically, 49.9% of the adolescents engaged in privacy disclosure, whereas 58.5% were involved in the consumption of online pornography. Further observations indicated that the prevalence of underreporting behaviors among adolescents was 35.6 and 43.3 percentage points in terms of privacy disclosure and pornography use, respectively. The heterogeneous analyses revealed that adolescents who recently acquired smartphones and those who were active users of smartphones and Facebook presented a greater prevalence of engagement in both sensitive behaviors. Urban adolescents demonstrated a greater propensity for online pornography use.

Our findings have three key implications. First, this study advances the field by employing an experimental survey methodology to explore adolescents’ online behaviors. The dual-method approach, blending direct and indirect questioning, effectively captured social desirability bias, enhancing the reliability of the findings. Researchers should adopt this approach to better estimate the true prevalence of sensitive behaviors and reduce bias in self-reports.

Second, the findings emphasize the need for educational programs targeting adolescents, particularly those with excessive screen time or recent smartphone ownership. Policymakers and educators should prioritize digital literacy initiatives that incorporate interactive components such as games, quizzes, and videos²⁸ to help adolescents navigate online risks. Collaboration among governments, NGOs, schools,²⁹ and tech companies can ensure that these programs are accessible, especially in developing countries with limited internet safety awareness. Such initiatives can empower adolescents to make safer online decisions.

Third, given the association between excessive screen time and risky online behaviors, parents must take an active role in managing their children’s screen use. Brauchli et al.³⁰ found that both parenting stress and positive parental attitudes influence children’s screen time, highlighting the critical role that parents play. Parents should engage in open discussions about online safety, use digital tools to manage screen time, and establish boundaries that protect their children from online risks. Educational resources for parents that focus on managing screen time and fostering healthy online habits can strengthen their protective role.

Footnotes

Acknowledgments

The authors express gratitude for the cooperative support provided by ChildFund Australia and Hoa Binh city’s Education and Training Division, as well as the teachers, parents, and adolescents who contributed to this study. Special thanks to Anna Dietl for proofreading this work. Appreciation is extended to Professor Brenda K. Wiederhold, the Editor-in-Chief, and anonymous reviewers for their valuable comments and suggestions.

Authors’ Contributions

T.P.: Conceptualization, methodology, software, validation, formal analysis, investigation, resources, data curation, writing—original draft, writing—review and editing, visualization, project administration. D.G.: Conceptualization, methodology, validation, resources, writing—review and editing, supervision, funding acquisition. M.K.: Conceptualization, methodology, validation, resources, writing—review and editing, supervision.

Ethics Approval

The Hiroshima University Research Ethics Review Board granted ethical approval for the project.

Consent

Informed consent was secured from the adolescents participating in this study and their guardians.

Data Availability

Data and codes will be made available by the authors upon request.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The Japan Society for the Promotion of Science granted partial funding for this study under Grant No. 22K01478.

Supplementary Material

a

We review recent studies in the .

b

We discuss the study site in the .

c

The theory of planned behavior by Ajzen¹⁸ is another relevant theoretical framework in this context.

d

We include 19 covariates in our estimation. Columns (4) and (5) of present the results of OLS models without and with covariate controls, respectively. While unobservable confounders, such as parental supervision and prior exposure to digital literacy programs, may exist, Banerjee and Duflo²⁴ suggest that well-executed randomization ensures balance in both observable and unobservable covariates. Thus, our results are minimally influenced by confounding factors.

References

Livingstone

, Helsper

. Balancing opportunities and risks in teenagers’ use of the internet: The role of online skills and internet self-efficacy. New Media Soc, 2010; 12(2):309–329; doi: 10.1177/1461444809342697

Stoilova

, Nandagiri

, Livingstone

. Children’s understanding of personal data and privacy online—a systematic evidence mapping. Inf Commun Soc, 2021; 24(4):557–575; doi: 10.1080/1369118X.2019.1657164

Jenkins

, Shen

, Dumontheil

, et al. Social networking site use in young adolescents: Association with health-related quality of life and behavioural difficulties. Comput Hum Behav, 2020; 109:106320; doi: 10.1016/j.chb.2020.106320

Reid Chassiakos

, Radesky

, Christakis

, et al.; COUNCIL ON COMMUNICATIONS AND MEDIA. Children and adolescents and digital media. Pediatrics, 2016; 138(5):e20162593; doi: 10.1542/peds.2016-2593

Twenge

, Martin

. Gender differences in associations between digital media use and psychological well-being: Evidence from three large datasets. J Adolesc, 2020; 79:91–102; doi: 10.1016/j.adolescence.2019.12.018

Gatenio Gabel

. Social protection and children in developing countries. Child Youth Serv Rev, 2012; 34(3):537–545; doi: 10.1016/j.childyouth.2011.10.013

Ali

, Qazi

. Countering misinformation on social media through educational interventions: Evidence from a randomized experiment in Pakistan. J Dev Econ, 2023; 163:103108; doi: 10.1016/j.jdeveco.2023.103108

Barbui

, Purgato

, Abdulmalik

, et al. Efficacy of psychosocial interventions for mental health outcomes in low-income and middle-income countries: An umbrella review. Lancet Psychiatry, 2020; 7(2):162–172; doi: 10.1016/S2215-0366(19)30511-5

Hayes

, James

, Barn

, et al. Children’s risk and benefit behaviors on social networking sites. Comput Hum Behav, 2022; 130:107147; doi: 10.1016/j.chb.2021.107147

10.

Copp

, Mumford

, Taylor

. Online sexual harassment and cyberbullying in a nationally representative sample of teens: Prevalence, predictors, and consequences. J Adolesc, 2021; 93(1):202–211; doi: 10.1016/j.adolescence.2021.10.003

11.

Kwan

, Dickson

, Richardson

, et al. Cyberbullying and children and young people’s mental health: A systematic map of systematic reviews. Cyberpsychol Behav Soc Netw, 2020; 23(2):72–82; doi: 10.1089/cyber.2019.0370

12.

, Kågesten

, De Meyer

, et al. Pornography use and perceived gender norms among young adolescents in urban poor environments: A cross-site study. J Adolesc Health, 2021; 69(1S):S31–S38; doi: 10.1016/j.jadohealth.2021.03.008

13.

Collins

, Strasburger

, Brown

, et al. Sexual media and childhood well-being and health. Pediatrics, 2017; 140(Suppl 2):S162–S166; doi: 10.1542/peds.2016-1758X

14.

Koletić

. Longitudinal associations between the use of sexually explicit material and adolescents’ attitudes and behaviors: A narrative review of studies. J Adolesc, 2017; 57(1):119–133; doi: 10.1016/j.adolescence.2017.04.006

15.

Van Ouytsel

, Ponnet

, Walrave

. The associations between adolescents’ consumption of pornography and music videos and their sexting behavior. Cyberpsychol Behav Soc Netw, 2014; 17(12):772–778; doi: 10.1089/cyber.2014.0365

16.

Paulhus

. Two-component models of socially desirable responding. J Pers Soc Psychol, 1984; 46(3):598–609; doi: 10.1037/0022-3514.46.3.598

17.

Bandura

. Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice-Hall, Inc: Englewood Cliffs, NJ, US; 1986.

18.

Ajzen

. From intentions to actions: a theory of planned behavior. In: Action Control: From Cognition to Behavior. ( Kuhl

, Beckmann

. eds) Springer: Berlin, Heidelberg; 1985; pp. 11–39; doi: 10.1007/978-3-642-69746-3_2

19.

Indraswari

, Shaluhiyah

, Widjanarko

. Health risk behaviors and life skills of Indonesian children: A qualitative analysis using social cognitive theory. Southeast Asian J Trop Med Public Health, 2022; 53.

20.

Remondi

, Gerbino

, Zuffianò

, et al. Longitudinal associations between positive parenting and youths’ engagement in sexting behaviors: The mediating role of filial self-efficacy beliefs. Comput Hum Behav, 2024; 158:108320; doi: 10.1016/j.chb.2024.108320

21.

Dolev-Cohen

. The association between sexting motives and behavior as a function of parental and peers’ role. Comput Hum Behav, 2023; 147:107861; doi: 10.1016/j.chb.2023.107861

22.

Vannucci

, Simpson

, Gagnon

, et al. Social media use and risky behaviors in adolescents: A meta‐analysis. J Adolesc, 2020; 79(1):258–274; doi: 10.1016/j.adolescence.2020.01.014

23.

Imai

. Multivariate regression analysis for the item count technique. J Am Stat Assoc, 2011; 106(494):407–416; doi: 10.1198/jasa.2011.ap10415

24.

Banerjee

, Duflo

. The experimental approach to development economics. Annu Rev Econ, 2009; 1(1):151–178; doi: 10.1146/annurev.economics.050708.143235

25.

Andrie

, Sakou

, Tzavela

, et al. Adolescents’ online pornography exposure and its relationship to sociodemographic and psychopathological correlates: A cross-sectional study in Six European countries. Children (Basel), 2021; 8(10):925; doi: 10.3390/children8100925

26.

Nguyen

, Tran

, Nguyen

TTA

, et al. Exposure to sexually explicit Internet material among adolescents: A study1 in Vietnam. Health Psychol Rep, 2021; 9(3):227–239; doi: 10.5114/hpr.2020.99394

27.

Maheux

, Roberts

, Evans

, et al. Associations between adolescents’ pornography consumption and self-objectification, body comparison, and body shame. Body Image, 2021; 37:89–93; doi: 10.1016/j.bodyim.2021.01.014

28.

Andrews

, Walker

, Kees

. Children and online privacy protection: Empowerment from cognitive defense strategies. J Public Policy Mark, 2020; 39(2):205–219; doi: 10.1177/0743915619883638

29.

Pham

, Goto

, Tran

. Child online safety education: A program evaluation combining a randomized controlled trial and list experiments in Vietnam. Comput Hum Behav, 2024; 156:108225; doi: 10.1016/j.chb.2024.108225

30.

Brauchli

, Sticca

, Edelsbrunner

, et al. Are screen media the new pacifiers? The role of parenting stress and parental attitudes for children’s screen time in early childhood. Comput Hum Behav, 2024; 152:108057; doi: 10.1016/j.chb.2023.108057

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.32 MB