Gender Differences in the Association Between Sexting and Self-Harm Behavior Among Taiwanese Adolescents
Article information
Abstract
Objective
Sexting has been linked to negative mental health outcomes. This Taiwan-based study addresses the links between sexting, cyberbullying, self-harm, and gender differences in the association of sexting and self-harm behavior.
Methods
A two-stage stratified sampling of 5,190 Taiwanese adolescents aged 11–18 completed an online questionnaire measuring sexting, bullying/cyberbullying, and self-harm. Two-way interaction model was used to identify the effect of gender on the relationship between sexting and self-harm behavior. The relationship between sexting and self-harm risk was tested in both the total sample and in different gender groups using univariable and multivariable linear regression models.
Results
2.5% of adolescents engaged in some form of sexting in the past year. 1.7% had sent their own picture or video. 1.0% had experienced non-consensual sexting, and 0.9% had sexted under pressure. Female gender, sexting one’s own information, non-consensual sexting, and pressured sexting were associated with higher self-harm scores.
Conclusion
This is the first study to focus on gender differences in the association between different types of sexting and self-harm in a non-Western society. The impact of sexting on self-harm was found to differ between males and females. Different types of sexting may also be associated with different risks of self-harm.
INTRODUCTION
For the Internet generation, access to the Internet through digital devices is an essential part of their daily lives. Approximately 60% of the adult population in developing countries and 90% in developed countries use the Internet, and 67% and 49% of adults use social media, respectively [1]. Although the widespread use of the Internet improves information equality, it also brings about adverse consequences such as cyberbullying [2] or sexting [3].
Sexting, a contraction of “sex” and “texting,” is one of the emerging issues of the Internet age, especially among adolescents [3]. Currently, there is no universally recognized definition of sexting. Some articles define sexting based on its content (texts, images, or videos), mode of transmission (smartphone or social networking site), consent (consensual or non-consensual), and actions (sending, receiving, or forwarding) [4]. However, most studies agree that sexting is the transmission of sexually explicit content (such as texts, images, or videos) to others through digital technology (such as smartphones or computers) [4]. In our study, we identified three types of sexting: sexting of one’s own information, sexting of non-consensual information, and pressured sexting, based on previous theoretical structures [3,5].
Sexting has been linked to negative mental health outcomes such as depression or anxiety [6], although these internalizing behaviors tend to decrease as people age [7]. A qualitative analysis of posts on a peer support mental health forum by Razi et al. [8] revealed that sexting led to self-harm and suicidal ideation, as well as other negative mental health consequences (such as shame, hopelessness, and rage), although there were still some positive feelings present. Another survey of 6,021 high school students in the U.S. showed that adolescents who experienced sexting, whether consensual or non-consensual, had a higher risk for depression, suicide attempts, and self-harm [9]. Additionally, a survey of Mexican university students demonstrated a relationship between sexting, being a victim of cyberbullying, and depression, all of which may contribute to suicidal ideation [10]. Although the connection between sexting and self-harm is still unclear, sexting should be recognized as a factor that may affect mental health among adolescents.
The prevalence of sexting was estimated to be between 14.8% (sending) and 27.4% (receiving) based on a meta-analysis, with a mean age of 15.16 years [11]. There is no consensus on the gender differences in the prevalence of sexting. Some studies reported female predominance [11,12], while others reported male predominance [13-15]. Gender differences may play an important role in sexting. A qualitative study found that although both male and female adolescents agree that victims who share their intimate images to other people should take some responsibility, there was a gender difference in the consequences of sexting among adolescents in New Zealand [16]. Females are viewed as more sexualized [17], while they are shaped by social norms to protect their purity [18]. The consequence of intimate image being shared without consent is usually harmful for females under most of the social norms [17]. On the other hand, most males were less concerned about privacy and took less responsibility [17]. This difference may lead to different association between sexting and self-harm.
Some researchers view adolescent sexting as problematic behavior, but others viewed it as a social phenomenon which may have cultural differences [19,20]. A multinational survey study revealed that China was ranked last of 11 countries with regard to the prevalence of sexting one’s own information (14.4%) or risky sexting (12.2%) [21]. Our previous work had found that sexting prevalence was 1.3% in the Taiwanese adolescent population with a male predominance (1.5% vs. 1.1%) and associated with self-harm risk [22]. However, the motivation, meaning, and impact of sexting may have gender differences [8,16]. Moreover, different types of sexting may also have different impacts on self-harm. On the basis of our previous work, the aim of our study is to analyze the gender difference between different sexting types, and the effect of gender on the association between sexting and self-harm behavior.
METHODS
Study population
This nationwide survey of a representative sample of Taiwan’s adolescent population aged 11–18 was conducted in 2022. A two-stage stratified sampling, based on age and geographical regions, was used to ensure the representativeness of the average Taiwanese teenage student population. The online questionnaire consisted of demographic data, Internet use behaviors, self-harm behavior, bullying experiences, sexting, and mental health. A total of 5,345 online questionnaires were completed by cluster sampled students from 60 participating elementary or high schools. All answers that were completed in under 60 seconds were excluded due to the high possibility of the respondent answering without fully reading and/or comprehending the questions. After excluding 155 incomplete questionnaires, 5,190 samples were weighted to match demographic data in each geographical region.
The purpose and contents of the online survey were introduced by school teachers. The study was approved by the Institutional Review Board of the National Health Research Institute of Taiwan (EC1100502), in accordance with the Declaration of Helsinki. The guardians’ informed consent was waived by the Institutional Review Board of the National Research Institute of Taiwan due to its anonymous and low-risk nature.
Measures
Sexting
We assessed sexting through three different questions. Firstly, sexting is defined as “receiving or sending messages, photos, or videos with teasing or sexually related content through mobile devices or social network systems.” Participants were asked about their experiences in the past year: 1) “How often have you sent your own photo/video with teasing or sexually related content publicly or privately?” (self-sexting), 2) “How often have you sent someone else’s photo/video with teasing or sexually related content publicly or privately without their permission?” (non-consensual sexting), and 3) “How often have you sent photo/video with teasing or sexually related content publicly or privately under pressure from someone else?” (pressured sexting). Responses were rated on a 5-point scale based on frequency in the past year: 1=never, 2=seldom, 3=2–3 times a month, 4=2–3 times a week, and 5=almost daily. If the participant answered “score ≥2” to one of these three questions, they were considered to have engaged in sexting [5].
Bullying and Cyberbullying Scale for Adolescents
The Bullying and Cyberbullying Scale for Adolescents (BCS-A) scale is a self-reported questionnaire consisting of two parallel 13-item sub-scales that use a 5-point Likert scale (ranging from 0 “this did not happen to me” to 4 “several times a week or more”) to measure one’s experience of bullying perpetration and victimization [23] over the past 3 months. It includes 4 items for physical bullying, 2 for verbal, 2 for relational, and 5 for cyberbullying. Participants who score ≥2 on any cyberbullying-related items are identified as having experienced cyberbullying victimization/perpetration, while those who score ≥2 on any physical, verbal, or relational items are identified as having experienced traditional bullying victimization/perpetration. It showed adequate construct, concurrent, and convergent validity [21], and the content validity of the Chinese version was mentioned in our previous work [22]. The Cronbach’s alpha for the BCS-A sub-scales in this study was 0.871 for traditional bullying victimization, 0.865 for cyberbullying victimization, 0.910 for traditional bullying perpetration, and 0.847 for cyberbullying perpetration.
Risk-Taking and Self-Harm Inventory for Adolescents
The Risk-Taking and Self-Harm Inventory for Adolescents (RTSHIA) is a self-reported questionnaire composed of 18 items with a 4-point Likert scale (ranging from 0 “never” to 3 “many times”) to measure lifetime past self-harm related behavior [24]. Participants with a total score between 1 and 6 are considered as having suspected self-harm, while those with a total score ≥7 are considered as having self-harm related behavior. The convergent, concurrent, and divergent validity had been assured in adolescent population [22], and the content validity of the Chinese version had been mentioned in our previous work [22]. The Cronbach’s alpha for RTSHIA in this study was 0.907.
Statistical analysis
In the statistical analysis, chi-square tests and independent t-tests were used for categorical and continuous demographic variables, respectively. Subgroup analysis by gender was performed. The relationship between different types of bullying experience and sexting was examined using chi-square tests. A two-way interaction model was used to identify the effect of gender on the relationship between sexting and self-harm behavior. The relationship between sexting and self-harm risk was tested in both the total sample and in different gender groups using univariable and multivariable linear regression models. Statistically significant results were defined as two-tailed p<0.05, and the 95% confidence interval (CI) was provided to estimate the effects. All statistical analysis was performed using PASW Statistics, version 18 (SPSS Inc.).
RESULTS
Demographic data and the proportion of having bully/sexting experience
A total of 5,190 adolescents (aged 14.5±2.0 years, 49.0% were female) participated in the survey (Table 1). There were 1,055 (20.3%) students from elementary schools, 2,469 (47.6%) from junior high schools, and 1,666 (32.1%) from senior high schools. Out of the 5,190 students, 117 (2.3%) had experienced traditional bully victimization and 54 (1.0%) had experienced cyberbully victimization in the past 3 months, while 43 (0.8%) had engaged in traditional bully perpetration and 29 (0.6%) had engaged in cyberbully perpetration. There were 130 (2.5%) adolescents who had engaged in some form of sexting in the past year, while 89 (1.7%) had sent their own picture or video, 53 (1.0%) had experienced non-consensual sexting, and 47 (0.9%) had sexted under pressure. The proportion of students who had experienced traditional bully victimization and perpetration was significantly higher among males (2.9% vs. 1.6%, p=0.002, and 1.1% vs. 0.5%, p=0.014) than females, while there was no significant gender difference in the proportion of students who had experienced cyberbully victimization or perpetration. Male students also had a higher proportion of all types of sexting (3.4% vs. 1.6%, p<0.001) and sexting their own information (2.4% vs. 1.0%, p<0.001), non-consensual sexting (1.6% vs. 0.4%, p<0.001), and sexting under pressure (1.3% vs. 0.5%, p=0.003) than females. However, females had a significantly higher self-harm score than males (2.4±5.6 vs. 1.3±4.1, p<0.001).
Participants’ demographic data by different genders
The association between sexting and bullying
Compared to students without a history of bullying, those with a history of any form of bullying (traditional or cyberbully, perpetration, or victimization) had a higher proportion of any kind of sexting (Table 2).
The association between sexting and bullying
The interaction between sexting and self-harm
The effect of gender, bullying, and sexting on self-harm as demonstrated by the two-way interaction model of linear regression is shown in Table 3. As the RTSHIA score is a continuous variable, however, it showed a skew to the right. We then checked the residuals vs. fitted values plot and residuals vs. leverages plot of the linear regression model, and both of them are acceptable. In considering of the aim of this study was to explore the relationship between factors and the sample size was relatively large, we decided to maintain the linear regression model but interpret the result with more caution.
The two-way interaction model between sexting types and self-harm
In the multiple linear regression model with two-way interaction terms, being male had a negative association with self-harm (B=-1.215, 95% CI -1.477 to -0.952, p<0.001). Experience of traditional bullying victimization (B=3.652, 95% CI 2.642 to 4.662, p<0.001) or cyber bullying victimization (B=3.620, 95% CI 2.109 to 5.131, p<0.001) was positively associated with self-harm, but bullying perpetration, either cyber bullying or traditional bullying, was not. Sexting one’s own information (B=3.869, 95% CI 1.918 to 5.819, p<0.001) and pressured sexting (B=7.573, 95% CI 4.839 to 10.307, p<0.001) showed a significantly positive association with self-harm, and this effect was reduced in males.
In our model, gender significantly moderated the associations between certain sexting experiences and self-harm. Specifically, the interaction between sending one’s own sexting information and gender was significant (B=-2.661, 95% CI -5.060 to -0.262, p=0.030), indicating that the positive association between sexting one’s own information and self-harm was weaker among males than among females. Similarly, the interaction between pressured sexting and gender was also significant (B=-4.861, 95% CI -8.288 to -1.433, p=0.005), suggesting that the link between being pressured into sexting and self-harm was notably stronger among females compared to males. However, the interaction between non-consensual sexting and gender approached statistical significance (B=3.433, 95% CI -0.703 to 6.939, p=0.055), indicating a potential trend toward a stronger positive association between non-consensual sexting and self-harm among male adolescents relative to females.
Gender differences in the association between different types of sexting and self-harm scores
The univariable and multivariable linear regression models demonstrating the effect of different types of sexting on self-harm scores are shown in Table 4. Female gender (B=-1.218, 95% CI 0.954 to 1.482 for male, p<0.001), sexting one’s own information (B=2.480, 95% CI 1.332 to 3.628, p<0.001), non-consensual sexting (B=2.724, 95% CI 1.156 to 4.293, p<0.001), and pressured sexting (B=5.748, 95% CI 4.101 to 7.396, p<0.001) were associated with higher self-harm scores.
Effect of sexting type to self-harm (linear regression)
Table 5 demonstrates in both male and female participants, sexting one’s own information (B=1.581, 95% CI 0.412 to 2.751, p=0.008, and B=4.336, 95% CI 2.059 to 6.612, p<0.001, for male and female, respectively) and pressured sexting (B=4.662, 95% CI 2.962 to 6.361, p<0.001, and B=7.772, 95% CI 4.574 to 10.969, p<0.001, respectively) were associated with higher self-harm scores. However, sexting non-consensual information was significantly associated with self-harm in male students (B=3.930, 95% CI 2.370 to 5.489, p<0.001), but this association was not significant in females.
Sexting type to gender and self-harm
DISCUSSION
In our study, we found that Taiwanese adolescents who have experienced bullying (both cyberbullying and traditional bullying, both as perpetrators and victims) are more likely to engage in sexting. Both sexting of one’s own information and pressured sexting, but not non-consensual sexting, were linked to increased self-harm behaviors in adolescents, as well as being victims of bullying (both cyberbullying and traditional bullying, but not perpetration). The impact of sexting on self-harm was found to differ between males and females. In male students, all forms of sexting were associated with a higher risk of self-harm. However, Taiwanese female adolescents who reported sexting their own information or being pressured to sext showed a higher risk of self-harm behavior compared to their male peers in the two-way interaction analysis.
This is the first study to focus on gender differences in the association between different types of sexting and self-harm in a non-Western society. Unlike previous studies [25-27], we found that fewer Taiwanese female adolescents engage in all forms of sexting compared to males. Previous studies have shown that each gender may have different sexting preferences [28]. For example, boys were found to engage in more non-consensual sexting, while girls experienced more pressured sexting [27]. As a behavior that heavily involves sexual and social contexts, the meaning and consequences of sexting may differ between genders. According to Bandura’s social learning theory [29], adolescents can learn behaviors (such as sexting) from peers through social media. In the era of social media, adolescents often judging themselves according to the comparison with their peer’s behavior online [30]. The number of “likes,” friends, and/or followers can be a strong reward in the reinforcement learning of behavior, and the fear of missing out can cause anxiety and push them to imitate their peers and to under-evaluate the possible harmful outcome [30]. For example, girls are more likely to be viewed as sexually desirable and blamable objects in sexual image-sharing, while boys may view sexting as a way to express their masculinity [31]. Moreover, males and females are shaped into different gender roles in different social structures [32]. In the context of sexting, teenage boys usually accumulated “reputation” by exchanging sexual images, while in contrast, girls are often the providers of sexual images and the objects of gaze in sexting [33]. Moreover, girls face more unfriendly social norms than boys in certain culture contexts. For example, slut shaming, the stigmatizing term referring to girls “who engage in, or are perceived as engaging in sexual behavior or sexualized expression,” [34] is a known form of gender-based victimization and has adverse effects on the psychological well-being and distress of adolescent girls [35,36]. This may possibly explain that only female adolescents, but not males, had a relationship between depressive symptoms and sexting [12].
Different types of sexting may also be associated with different risks of self-harm. In our multivariable model, we found that pressured sexting, but not non-consensual sexting was more strongly associated with self-harm, particularly in females. However, in a previous survey among US adolescents, Wachs et al. [27] found that both non-consensual sexting and pressured sexting were positively correlated with depressive symptoms and non-suicidal self-harm. Another study by Frankel et al. [9] also revealed that non-consensual sexting was positively correlated with depression. This difference may stem from the different social context of non-consensual sexting in Taiwan. Non-consensual sexting is a complex sexual behavior that has diverse meanings in adolescent relationships [37], and may have unique implications in the Taiwanese adolescent population. One possible explanation is that, in cases of non-consensual sexting, the power dynamic may be inverted. In Taiwan, more female than male have reported receiving sexting, e.g. unsolicited “dick pics” as a form of sexual harassment [38]. The act of sharing these experiences with peers—often through humor or collective outrage—may serve to mitigate the emotional harm typically associated with such harassment. This coping mechanism could partly explain why, in our findings, non-consensual sexting was not significantly associated with self-harm risk among female adolescents. However, the nature of female non-consensual sexting and the lived experience of female non-consensual sexting needed further exploration.
Our study found that sexting one’s own information, even if consensual, was associated with self-harm, with females at higher risk compared to males. Previous surveys have found that consensual sexting of one’s own information is a common practice among adolescents and is considered a normal part of adolescent sexual development [9,39]. Young people hold differing views about sexting their own information, from positive to negative, across different European countries [40]. Although initially consensual sexting of one’s own information may not cause harm to both parties [39], it can also result in the risk of non-consensual dissemination of sexual images, bullying, and social shaming, particularly for females [41]. This could contribute to the higher risk of self-harm behavior in adolescents who reported sexting their own information, especially for girls who are more susceptible to slut-shaming based on their sexual image.
There were several limitations in our study. Firstly, its cross-sectional design cannot establish causality between bully experience, sexting, and self-harm, as their relationship may be complex and potentially influenced by other factors, such as personality or mental health conditions including anxiety or depression, that were not collected in our study. Secondly, our study only included three types of sexting and did not collect information on other forms of sexting, such as by text, picture, or video, which may have different effects on self-harm. Third, because the RTSHIA score was not normally distributed in this study sample, the results of the linear regression model, especially the hypothesis testing, should be interpreted with caution. Lastly, there is a lack of widely accepted consensus on the definition of sexting and its measurement, which may hinder the comparability of our results with previous literature.
In conclusion, after controlling for possible confounding factors such as demographic data and bullying experience, our study found that sexting may be associated with self-harm behavior among Taiwanese adolescents, with females having significantly higher risks for sexting their own information and being subjected to pressured sexting. This result expands our understanding of the effects of sexting on mental health outcomes among non-Western adolescent populations and highlights the importance of considering gender issues. Further research is needed to better understand the nature of sexting in specific populations, such as females, gender minorities, and in non-Western cultures.
Notes
Availability of Data and Material
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Author Contributions
Conceptualization: Yu-Hsuan Lin, Shan-Mei Chang, Yu-Chuan Chiu. Data curation: Yu-Hsuan Lin, Chun-Hao Liu. Formal analysis: Yu-Hsuan Lin, Shan-Mei Chang. Funding acquisition: Shan-Mei Chang. Investigation: Yu-Hsuan Lin, Chun-Hao Liu. Methodology: Yu-Hsuan Lin, Chun- Hao Liu, Shan-Mei Chang. Project administration: Yu-Hsuan Lin. Resources: Yu-Hsuan Lin. Software: Yu-Hsuan Lin, Chun-Hao Liu. Supervision: Yu-Hsuan Lin. Validation: Yu-Hsuan Lin. Visualization: Chun-Hao Liu, Yu-Chuan Chiu. Writing—original draft: Chun-Hao Liu, Yu-Chuan Chiu. Writing—review & editing: Yu-Chuan Chiu.
Funding Statement
This work was supported by the Ministry of Science and Technology, Taiwan (grant number MOST 112-2410-H-007-099-MY2, MOST 112-2423-H-A49-003-).
Acknowledgments
The study authors thank all of the study participants for their participation.
Data analyzed in the manuscript has been presented in the form of a poster presentation at the 11th Congress of The Asian Society for Child and Adolescent Psychiatry (ASCAPAP 2023).