If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Population, Family and Reproductive Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MarylandSoins Primaires et Prévention, CESP Centre for Research in Epidemiology and Population Health, Villejuif, France
Address correspondence to: Chaohua Lou, M.D., Senior Researcher, Department of Epidemiology & Social Science, NHC Key Lab. Of Reproductive Regulation (Shanghai Institute of Planned Parenthood Research), 779 Old Hu Min Road, Shanghai 200237, P. R. China
The purpose of the study is to assess the prevalence of pornography use and its association with a range of perceived gender norms among adolescents aged 10–14 years across five urban poor settings globally.
The study includes 9,250 adolescents aged 10–14 years from Belgium, China, the Democratic Republic of the Congo, Ecuador, and Indonesia, as part of the Global Early Adolescent Study. We examined the percentage of pornography use by sex and site and conducted multivariate logistic regressions to examine the relation between gender norm perceptions and pornography use, adjusting for individual, family, peer, and media exposures.
Ever-use of pornography ranged from 14.5% in Ecuador to 33.0% in Belgium and was more common among boys than girls. Overall, boys who perceived greater permissiveness about romantic relations, adolescents who engaged in such relations, and adolescents who assumed that their friends were sexually active had greater pornography exposure. Pornography use did not systematically correlate with unequal gender norms. Such correlations only exist among boys in two Asian sites, where a supportive school environment, more caregiver awareness, and/or neighborhood cohesion were related to less pornography use.
Pornography use is a gendered experience that begins in early adolescence. Although factors of pornography use vary across the social context, the exposure to pornography has become a normative part of adolescent sexuality development. Young people, especially those from where sexuality remains taboo, need the ability to critically process information and avoid potential risks associated with pornographic gendered and sexual stereotypes, calling for comprehensive sexuality education programs to help them build the knowledge and confidence they need.
Pornography use among young adolescents is more common among boys with unequal gender norm perceptions. More research is needed to better understand this (bi)directional relationship. Comprehensive sexuality education programs are needed to equip young people with the abilities to contest gender and sexual stereotypes and promote healthy sexuality development.
Pornography, defined as videos or pictures intended to sexually arouse the viewer [
]. However, most studies that originate from Europe and North America rarely address intercultural and cross-cultural issues, and even though some have been carried out in Asia, data on adolescents' pornography use in Africa and Islamic countries remain scant [
The predominant lack of perspectives about pornography use among young people outside of the Western world is problematic for several reasons. First, it is based on the assumption that pornography use is mostly a Western issue due to the relative freedom of information and access to the Internet [
]. Second, it fails to understand how pornography consumption relates to sexual and gender attitudes in societies where sexuality may be more taboo and where young people lack access to comprehensive sexuality education—as a means of critically reviewing pornography messages [
]. On the one hand, studies indicate that pornography may be problematic for adolescents' healthy sexuality development. These concerns, generally grounded in moral ideas around the “innocence” and vulnerability of children and adults' responsibility to protect them [
], are particularly salient for adolescents who often lack the media literacy and related critical thinking skills as well as the experiences to properly evaluate the content of pornographic materials [
]. Likewise, the display of unsafe sexual practices (either heterosexual or homosexual) may further promote sexual risk taking, exposing young people to unintended pregnancy and/or sexually transmitted infections [
]. For example, studies among young adults in high-income countries indicate that watching pornography might have a positive effect on body satisfaction, self-exploration, and self-esteem as well as foster more gender-equal attitudes [
]. For example, a longitudinal study among 967 adolescents aged 12–14 years at baseline in the United States found that girls who were exposed to sexually explicit materials at early ages developed more stereotypical gender attitudes over time [
The fact that most data come from Western societies provides little insight into the ways that different social-ecological forces interact to inform pornography consumption and its implications for adolescent sexuality development in different cultural and economic settings [
]. In addition, most research focuses on older youth or young adults. Evidence on early adolescence (ages 10–14 years), which is a particularly critical time of physical, emotional, cognitive, and social development, is largely missing. This lack of perspective is a critical gap as early adolescence is a time when young people start to develop their sexual selfhood and become increasingly aware of socially constructed gender roles and norms associated with being a boy, man, woman, or girl [
]. It is a period when attitudes and beliefs are constantly evolving and influenced by—as well as influencing—young people's emerging (sexual) identity and experiences as they transit through adolescence [
In response to the abovementioned gaps, this study aims to assess the prevalence and correlates of pornography use among young adolescents (aged 10–14 years) across five geographically and culturally diverse urban poor settings. We focus on the gendered patterns of pornography consumption, which we hypothesize to be more common among boys than girls and more common among adolescents who perceive greater gender-unequal norms and those with greater permissiveness toward adolescent romantic relationships [
This is a cross-sectional study using data from the Global Early Adolescent Study (GEAS), a multicountry investigation of how gender socialization affects early adolescent health and well-being. The GEAS operates in 11 urban poor sites across five continents, chosen to represent a diversity of geographical, cultural, economic, and social contexts and because of existing research partnerships. The focus on urban poor adolescents was a deliberate choice to shed light on the emerging but understudied urban poor penalty among adolescents [
For the present study, we used data from five sites which had collected baseline data at the time of this analysis: Ecuador (Cuenca), Belgium (nine cities in Flanders), the Democratic Republic of the Congo, DRC (Kinshasa), China (Shanghai), and Indonesia (Semarang, Denpasar, and Lampung). All sites shared the same study protocol and core survey instruments, although sampling and data collection modes (described in detail elsewhere) were adapted to the local context [
], included random cluster sampling in Cuenca, a combination of exhaustive and convenient sampling in Flanders, multistage random sampling in Kinshasa, stratified cluster sampling in Shanghai, and a combination of clustered sampling and random sampling in three cities of Indonesia. Participants were recruited from schools in all sites, with the exception for Kinshasa, where a random sample of out-of-school adolescents from the same commune was selected to complement the in-school sample because of high school drop-out rates in this setting.
Altogether, 10,841 young adolescents aged 10–14 years were included. After excluding observations that did not reach the preset data quality criteria (<25% missing values for Belgium and <15% for other countries) (n = 490) and observations with missing information on the key dependent and independent variables (n = 1,101), our final analytical sample included 9,250 observations (89.4% of the original sample): N = 532 in Ecuador (85.3% of the original sample), N = 808 in Belgium (79.8% of the original sample), N = 2,720 in the DRC (96.5% of the original sample), N = 1,571 in China (89.7% of the original sample), and N = 3,619 in Indonesia (79.6% of the original sample).
Data collection took place between June 2017 and June 2019 and involved a 1–2 hour survey using audio-computer/computer-assisted self-interview completed on tablets, with the exception of Kinshasa, where the survey was conducted face-to-face because of the low literacy rate in that setting. All participants provided active parental consent as well as their own assent before data collection began. The survey collected information on a range of topics including sociodemographics, family, peer, school and neighborhood circumstances, as well as perceptions of gender norms, knowledge, and practices related to physical, mental, sexual, and reproductive health. The survey instrument (translated, back translated in sites where English is not the primary language) was tested in 14 urban poor cities around the globe, refined, and validated among adolescents in six sites of the GEAS [
Pornography use was assessed via a single question: “Sometimes young people watch pornography, that is, movies or videos that show people's private parts (genitals) during sexual scenes. Have you watched pornography before?”, with a 4-point scale response ranging from “never”, “rarely”, “sometimes” to “often”. Because of the small percentages of adolescents who reported any exposure in some sites, we used a dichotomized variable, opposing any exposure (rarely, sometimes, or often) to no exposure.
Perceptions of gender norms
We used four validated scales to assess young adolescents' perceptions of gender norms in terms of traits, roles, and relationships, grounded in qualitative work conducted in the formative GEAS phase [
The adolescents' romantic relationship scale (ARE), includes five items assessing permissive attitudes toward adolescents' romantic relationships (e.g., “It is normal for a boy/girl your age to want a girlfriend/boyfriend”). The polychoric ordinal Cronbach's alpha ranged from .75 in Cuenca to .90 in Shanghai (for more details, see Table S5);
The sexual double standard (SDS) scale includes six items measuring differential expectations for girls and boys in romantic/sexual relationships (e.g., “Boys have girlfriends to show off to their friends”; “Girls are victims of rumors if they have boyfriends”), with the polychoric ordinal Cronbach's alpha ranging from .81 in Shanghai to .85 in Kinshasa;
The gender stereotypical traits (GST) scale includes seven items tapping into male toughness over female vulnerabilities (e.g., “Boys should be raised tough”; “Girls are expected to be humble”), with polychoric ordinal Cronbach's alpha ranging from .69 in Cuenca to .80 in Flanders;
The gender stereotypical roles (GSR) scale includes four items measuring differential family roles conferring male power in household decisions (e.g., “A woman's role is taking care of her home and family”; “A man should have the final word about decisions in the home”), with polychoric ordinal Cronbach's alpha ranging from .72 in Kinshasa to .86 in Indonesia.
All scale items used five-point Likert response options, and responses were averaged into mean scores with higher scores signaling more unequal gender norm perceptions. Scores were also dichotomized at the median using site-specific thresholds.
We used Blum et al.’s conceptual framework to guide the selection of covariates [
]. At the individual level, we considered age, sex (boy vs. girl in the absence of information about participants' gender identity), and puberty onset. At the interpersonal level, we included the family structure (living with both parents, one parent, no parents) and caregiver awareness (e.g., the caregiver knows their friends by name; the caregiver knows how they were doing in school), peer attitudes toward sexual relations, and participants' romantic experiences (ever vs. never, with either sex). At the community level, we included a measure of school connectedness (perceived care from adults at school) and perceived neighborhood cohesion (a 4-item scale asking if people in the neighborhood look out for, care, and trust each other). Finally, we included a measure of media diet, assessed via the question, “On a typical day, how many hours in total do you spend using social media, chatting with friends online, playing computer games, or using other interactive media?” (0, 1–2, 3 or more).
We first assessed missing data on covariates, which ranged from .08% to 25.37%. Missing data were imputed using K-nearest neighbors based on age, sex, and site [
]. We pursued with descriptive analysis of the prevalence of pornography use by sex and site. Next, we used analysis of variance models to evaluate the associations between pornography use with each gender norm scale as well as social-ecological factors, controlling for sex. This analysis informed the multivariate model, using Firth logistic regressions, to assess the associations between each gender norm scale and pornography use by sex and site, adjusting for social-ecological covariates. We included an interaction between gender norm perceptions and age when significant. Finally, we conducted a sensitivity analysis among the complete case samples, which mostly supported the same conclusions (results not shown). Data were analyzed using Stata 15.0 SE software.
Table 1 shows the characteristics of the final analytical sample in each site. Adolescents were evenly split between boys and girls in each site, with a mean age ranging from 11.93 years in the DRC to 13.05 years in Belgium. Adolescents' family and peer environments also varied by site, with parental closeness ranging from 26.49% in Belgium to 92.43% in China. Conversely, self-reported romantic involvement (ever) and perceptions of peer sexual activity were highest in Belgium (57.28% had ever had a romantic partner and 13.49% thought their friends had had sex), whereas romantic experience was lowest in the DRC (9.41%), and perceptions of peer sexual activity were lowest in Shanghai and Indonesia (3.18% and 4.3%, respectively). Only 16.43% of respondents had access to the Internet in the DRC compared with 98.8% of adolescents in Belgium.
Table 1Characteristics of the study sample by site
Cuenca, Ecuador (N = 532)
Belgium (N = 808)
Kinshasa, the DRC (N = 2,720)
Shanghai, China (N = 1,571)
Indonesia (N = 3,619)
Age (mean ± SD)
With no parent
With one parent
With both parents
Perceptions about whether close friends had sex
Time spent on the Internet
2 hours or less
3 hours or more
DRC = Democratic Republic of the Congo; SD = standard deviation.
The prevalence of lifetime exposure to or use of pornography varied widely across sites from 14.5% in Ecuador to 33% in Belgium (p < .001). Post hoc paired comparison by site indicated similar prevalence in Ecuador, the DRC, and Indonesia (p > .05). Pornography was consistently more common among boys than that among girls in all sites (p < .05), with percentages ranging from 21.9% (Ecuador) to 51.28% (Belgium) among boys and from 6.59% (Ecuador) to 20.64% (China) among girls (Figure 1).
Perceptions of gender norms by site and sex
In general, respondents in Belgium reported more gender-equal perceptions, with relatively high acceptance of early adolescent romantic relationships (ARE) and low endorsement of SDS or GST (Figure 2). Conversely, adolescents in Kinshasa generally perceived more gender-unequal norms than those in other sites (p < .001), with low acceptance of ARE, but high endorsement of SDS, GST, and GSR. Although Shanghai respondents expressed low permissiveness toward ARE, they mostly rejected the SDS, as did respondents in Indonesia. Sex differences were evident in each site: in Cuenca and Shanghai, boys perceived greater permissiveness for ARE as well as more endorsement of GST, GSR, and SDS than their female peers. In contrast, in Belgium and Indonesia, boys were less likely to perceive an SDS than girls, while simultaneously expressing more stereotypical GST and GSR.
Factors associated with pornography use
Table 2 shows the effect size (proportion that could be explained) of gender norm perceptions as well as social-ecological covariates on pornography use (Cohen has provided benchmarks to define small [η2 = .01], medium [η2 = .06], and large [η2 = .14] effects [
]). Overall, pornography use was associated with greater permissiveness toward ARE in almost all sites. Specifically, adolescents who had any romantic experience or who believed their friends had engaged in sexual activities were more likely to have ever watched pornography. None of the other gender norm perceptions were associated with pornography use, except the SDS in Indonesia. Other factors related to pornography use included age, romantic relationship experience, peer sexual norms, and time spent on the Internet for boys and girls alike.
Table 2Effect size of gender norm perceptions and other ecological factors on pornography use
η2 (95% CI)
Adolescents' romantic relationships (ARE)
Sexual double standard (SDS)
Gender stereotypical traits (GST)
Gender stereotypical roles (GSR)
Perceived there are peers having sex
Time spent on the Internet
Note: Effect size η2 was presented for each variable after fitting ANOVA models with the controlling of sex in the total sample and city-specific sample.
ANOVA = analysis of variance; CI = confidence interval; DRC = Democratic Republic of the Congo.
Table 3 presents the findings from the multivariable analysis. The odds of pornography use were elevated among boys with higher perceptions of SDS and GST in Indonesia and boys with higher perceptions of GSR in Shanghai. For girls, the SDS was the only gender norm perception associated with pornography use and only in Indonesia. Among the social-ecological covariates, romantic relationship engagement, peer sexual norms, and time spent on the Internet remained associated with pornography use in several sites. (For additional details on the models, see Table S1-S4).
Table 3Associations between each gender norm scale and pornography use by sex and site: multivariate analysis using Firth logistic regression by sex and by site
OR (95% CI)
High (ref. low)
High (ref. low)
High (ref. low)
High (ref. low)
High (ref. low)
High (ref. low)
High (ref. low)
High (ref. low)
CI = confidence interval; DRC = Democratic Republic of the Congo; OR = odds ratio.
Note:1. ARE: Adolescents' romantic relationships scale; SDS: Sexual double standard scale; GST: Gender stereotypical traits scale; GSR: Gender stereotypical roles scale.2. Covariates adjusted include age, relationship engagement, pubertal onset, family structure, caregiver awareness, if their close friends have had sex, school connectedness, neighborhood cohesion, and time spent on the Internet.3. Interactions between each scale and relationship engagement were tested and were included if significant. 4. ∗: p < .05; ∗∗: p < .01; ∗∗∗: p < .001.
This study adds new knowledge on the prevalence and correlates of pornography use among young adolescents across five countries and four continents. It shows that even in urban poor settings where sexual norms remain restrictive and access may be limited, up to a third of 10- to 14-year-olds report to have seen pornographic videos (intentionally or unintentionally), with substantial variations by site and sex.
Comparing our estimates with previous studies is challenging, given their wide prevalence ranges [
], which likely reflects the younger age of our sample as well as the stricter definition of pornography (revealing of the genital part of the body).
Our findings indicate that young adolescents who report using pornography in several of the sites differ from those without such experiences in terms of their gendered perceptions and romantic attitudes as well as experiences. Some authors call on the sexual script theory to explain these connections, positing that pornography represents a source of sexual scripts that influence behavior [
]. Consistent with this framework, we found that pornography use was more common among adolescents who perceived greater permissiveness about romantic relations, who engaged in such relations, as well as those who assumed that their friends were sexually active. Such associations, which were consistently strong in all sites except Cuenca, may reflect the fact that pornography provides young adolescents with new “sexual scripts” or understandings of boy-girl relationships by portraying these as normative in contexts where premarital sex and adolescent dating or sexuality remain taboo [
]. As such, it may be that watching pornography at early ages leads to more permissive attitudes toward early adolescent boys and girls being in relationships, but it might also be the reverse—that those who already perceive early romantic relations as something normative seek out pornography as a way to navigate these beliefs [
], we also observed gendered patterns in the use and correlates of pornography consumption, providing further evidence that gender differences start to emerge in early stages of adolescent development across diverse cultural settings [
]. Across sites, boys were more likely to report pornography consumption than girls, and in Asia, male consumption of pornography correlated with unequal gender norm perceptions, promoting male dominance and sexual readiness over female vulnerabilities and sexual self-restraint. The correlations between pornography use and male gender unequal attitudes have previously been reported in other settings among older adolescents [
]. However, our results also show that pornography does not systematically correlate with unequal gender norms, suggesting that the meaning and implications of pornography consumption for adolescent sexual development likely vary across the social context.
The directionality of our findings does, however, warrant further longitudinal investigation. Although permissive and stereotypical views about gender and relationships may indeed trigger greater interest in sexual content, previous longitudinal studies suggest a stronger effect of pornography on stereotypical attitudes than the other way around [
]. Both longitudinal and qualitative research is needed to better understand the (bi)directional relationship between pornography use with gendered attitudes and expressions in early and later adolescence, as well as its effects on healthy sexuality development—including any potentially positive impacts of different forms of pornography that challenge traditional gender stereotypes.
Finally, we found that pornography use differs widely depending on adolescents' age (being more common in older age groups) and their social environment. Apart from the peer influences previously discussed, a supportive school environment including connectedness to teachers, as well as more caregiver awareness and/or neighborhood cohesion, was related to less pornography use in the two Asian sites. Pornography use was also more common among those who used the Internet more frequently in the DRC and China. These results underscore the importance of using comprehensive life-course perspectives to understand how young people's social-ecological environments shape how young people make sense of sexually explicit media as part of sexual socialization. Further investigation of the roles of families and school programs as well as media literacy initiatives in equipping young people to contest stereotypical messages, portrayed in mainstream pornographic materials [
], interventions need to help young people understand and navigate the messages that they receive—especially in contexts where parents and adults may find it difficult to communicate about this and whether the exposure is intentional or unintentional, online, or offline. Although appropriate child protection mechanisms and regulations of online pornography is needed, comprehensive sexuality education—when implemented with quality and based on evidence in line with international recommendations [
]. As young people grow up in digital communities (with a majority in our sample using the Internet on a daily basis), carefully implemented web-based sexuality education can also present an avenue to provide young people with fact-based information about their bodies, sexualities, and rights, which promotes consent and nonaggression in relationships.
Strengths and limitations
A number of limitations need to be considered when interpreting our results. First, because of survey limitations (such as using one singular question to assess the outcome), we were unable to assess the intentionality, frequency, and content of pornography consumption, precluding a more nuanced measurement, as well as comparison with previous studies, of the association between sexually explicit video content and adolescents' gender attitudes. Second, although the survey defined what constitutes pornography, the question could have been interpreted differently depending on the respondent's sociodemographic background and context, calling for cautious interpretation of cultural differences as well as qualitative studies to better understand the estimates and associations yielded in our study. In addition, the independent variables included in the analysis do not fully capture the wide range of possible predictors of pornography use. However, we believe that our comparison of a single indicator across different sites using the cross-cultural validated survey instrument, and inclusion of different individual, family, and community covariates, still adds important information to the field, given the lack of studies focusing on 10- to 14-year-olds in low-income contexts. As in any study on adolescent health, questions may have been subject to social desirability bias, especially in the DRC, where data were collected face-to-face, which we addressed by carefully ensuring adolescents' privacy and confidentiality. In addition, although parental consent was used as a way to maximize adolescents' safety, this may have influenced who participated in the study and their willingness to respond. Finally, the focus of the GEAS on urban poor communities, the nonprobability-based sampling method, the absence of data on perceived gender, and the cross-sectional nature of the study prevent any casual interpretation of the findings as well as generalizability beyond these diverse cultural settings and transferability to youth of gender fluidity or atypical gender identities. Some of these concerns can, however, be further explored as the longitudinal follow-up of the GEAS proceeds.
Pornography use and its link with gender inequality in adolescence remains an issue of much global debate. This study shows that lifetime exposure to pornography among 10- to 14-year-olds varies widely across urban poor settings and confirms pornography as a highly gendered experience that begins in early adolescence. The implications of these findings are essential in low-income urban contexts where access to evidence-based sexual education is limited and sexual norms remain restrictive [
], impeding young people's ability to challenge the narratives of gender stereotypes propagated via mainstream pornography. These realities call for investments in comprehensive sexuality education programs (including online) that start early and continue throughout adolescence to equip adolescents with the knowledge and skills needed to contest stereotypes and promote healthy sexuality development.
The Global Early Adolescent Study was supported by Johns Hopkins and World Health Organization as well as other organizations around the world. The authors thank all the researchers and students who participated in the present study and also the administrators and teachers in the target schools or communities who helped facilitating the survey. The authors also thank the GEAS Coordinate Center in Johns Hopkins University for providing the technical support.
Conflicts of interest: None of the authors have potential conflicts of interest to be disclosed.
Disclaimer: Publication of this supplement was supported by the Johns Hopkins Bloomberg School of Public Health Department of Population, Family and Reproductive Health with funding from the Bill and Melinda Gates Foundation.
Specifically, this study was conducted with support from the UNDP-UNFPA-UNICEF-WHO- World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), a cosponsored program executed by the World Health Organization (WHO), the Bill & Melinda Gates Foundation [ OPP1125119 ], the Oak Foundation [ OCAY-17-649 ], and the Packard Foundation [ 2017-66517 ]. Support for each Global Early Adolescent Study site is provided by the Bill & Melinda Gates Foundation [ OPP1197258 ] and the United States Agency for International Development [ AID-OAA-A-15-00042 ] in Kinshasa, the Innovation-oriented Science and Technology Grant from Chinese National Health Commission Key Laboratory of Reproduction Regulation [ CX2017-05 ] in Shanghai, the fund for scientific research Flanders and the Flemish Ministry of Innovation, Public investment, Media and Poverty Reduction in Belgium, the National Secretary of Science and Technology SENESCYT of Ecuador in Cuenca, and the Bill & Melinda Gates Foundation [ OPP1178415 ] in Indonesia.