Digital technology and media use is integral to adolescents’ lives; adolescents have been labeled digital natives given that they have had exposure to digital technology their entire lives. Previous studies on digital technology and media use have largely focused on assessing behaviors and outcomes within adolescents as a population, which does not allow for an understanding of the heterogeneity of adolescents’ technology use. Few studies have examined specific subgroups to understand the nuances of digital technology and media use across and within adolescent groups [1-3].

Furthermore, many studies on adolescent technology and media use have focused on risk behaviors and negative outcomes. Previous studies have illustrated that digital technology and media use is associated with negative outcomes such as impaired sleep [4-6], decreased physical activity [5,7,8], problematic internet use (PIU) [9-11], and risk of depression [12]. Although several recent review articles have described both the benefits and risks of technology use [13,14], most individual studies take a risk-centered approach [15,16]. Since the COVID-19 pandemic, many teenagers have experienced social isolation as a result of quarantine and remote learning, making digital tools for connection to peers and family even more important.

Critical constructs to consider in the balance of the risks and benefits of technology and media use include factors such as device ownership, frequency of social media use, and the importance of use [17]. Furthermore, other underlying factors in an adolescent’s offline environment may also be critical to consider. Evidence supports that technology use outcomes may also be affected by parental factors such as household rules around media use or the parent-child relationship [18]. To better understand adolescent digital technology and media use, new approaches that can assess distinct profiles within study populations, consider mitigating factors such as technology use and parental involvement, and take a balanced approach to understanding the risks and benefits of digital technology and media use are needed.

Several aspects of adolescent technology use have been studied through previous research. In the area of device ownership, it is understood that US adolescents’ smartphone access is nearly ubiquitous. The 2018 Pew Internet and American Life Project [19] estimated that 95% of US adolescents have their own personal smartphone. These rates increased from 2014-2015, when 73% of adolescents reported personal smartphone access [20]. The age at which teenagers obtain their first phone has also decreased over time, raising concerns that have prompted campaigns to encourage delaying ownership until an adolescent is in the eighth grade [21,22]. Less is known about other device ownership among teenagers. Although approximately 88% of adolescents report access to a desktop or laptop computer at home, it is less clear who owns these devices [19]. Device ownership may be a contributing factor to technology use outcomes. For example, personal ownership of technology may provide more opportunities for frequent use compared with family-owned devices. Personal ownership of devices may also provide access within private spaces such as bedrooms, which are more challenging to regulate by parents [23]. Furthermore, little is known about access to or ownership of newer devices such as virtual reality (VR) headsets and personal assistant devices.

Another area of focus in previous studies is the quantity of adolescents’ digital technology and media use. For adolescents, one way in which quantity of technology use can be measured is via the age at which they first started using technology, such as the age at which a youth acquired a personal smartphone. Earlier initiation to technology use has been associated with maladaptive outcomes such as problematic technology use [24], supporting that quantity of use over time may be an important factor in determining health outcomes. For adolescent populations, quantity of use is commonly measured by self-reporting hours per day spent using technology and media. However, designing research studies to assess quantity of technology use is not without challenges. In particular, self-reporting the quantity of use is subject to recall bias [25,26].

Increasingly, researchers and health care providers are emphasizing that quality of technology use, beyond just quantity of use, may be just as important if not more in understanding links between technology use and health outcomes. Quality of use could include motivations for use, importance of use, or types of web-based interactions. These quality of use concepts are tied to affordances, which describe properties of artifacts that illustrate how they can be used [27,28]. Affordances have been used to tie adolescents’ digital technology use to their developmental milestones [29]. Grounded in an affordance framework, the Adolescents’ Digital Technology Interactions and Importance (ADTI) scale measures the importance of specific technology interactions rather than platform use [17]. Combining quantity and quality measures of technology use may allow for new insights to better understand benefits and risks for adolescents. Thus, a comprehensive approach to understanding adolescent digital technology use patterns could include what devices an adolescent can own and access, the age at which an adolescent obtained their first smartphone, how often they interact with digital technology and media, and how important those technology interactions are to them.

The 2016 American Academy of Pediatrics (AAP) media policy statement, Media Use in School-Aged Children and Adolescents, emphasized the role of parents in adolescents’ media lives [30]. A key role for parents is to serve as mediators of media use for their children, and studies suggest that parental efforts may have varied impacts on adolescents’ media use [31]. Parents often struggle with setting and enforcing media rules such as removing technology from children’s bedrooms [32,33].

Parents’ own technology experiences may also affect their children. Studies have found that parents have varied attitudes toward technology and varied engagement with technology [34]. Parents who struggle with limiting their own technology use may have challenges in moderating their children’s technology use [35]. For parents who prioritize technology use, these behaviors may be modeled for their children. In fact, families with media-centric parents typically have children who report more media use [18].

Parenting style and parent relationship are additional factors in the balance of risk and beneficial outcomes that adolescents experience. Factors associated with positive media use in families have included positive general family functioning, parental involvement, and open communication styles between parents and adolescents [36]. Thus, although it is important to understand the role of technology use in influencing adolescent health and well-being, it is also critical to better understand how the family context affects adolescent technology-related outcomes.

There are several health concerns that have been associated with digital technology and media use influence and affect. First, technology and media use has been shown to negatively affect sleep by delaying bedtime as well as through exposure to light from screens disrupting melatonin levels [4-6]. Second, decreased physical activity has been associated with the sedentary nature of most media use [5,7,8]. Third, PIU is defined as “Internet use that is risky, excessive or impulsive in nature leading to adverse life consequences, specifically physical, emotional, social or functional impairment” [37]. Studies support that components of PIU include compulsive use and anxiety when not able to access the internet [9].

Adolescents’ mental health has also been a common research topic related to potential negative consequences of media use. Studies have found associations between increased social media use and decreased life satisfaction [38,39], increased risk of depression [38,40], worsened body image and decreased self-esteem [40-42], increased fear of missing out (FOMO) [43], and reduced well-being [44]. These studies suggest that media use may have a negative impact on mental health for adolescents. However, other studies have found that social media use does not affect life satisfaction [45,46] nor depression [47]. Some have argued that data support that media use may negatively affect some adolescents but caution that overstating these relationships to apply to adolescents as a whole is not warranted [48].

Studies focusing on ways in which technology influences well-being have found positive associations with increased social support and learning [49,50]. Another study found that adolescents described their affective experiences on social media to include feeling happy and closer to friends, supporting adolescent well-being [50]. Youths may be motivated to adopt digital health technology that includes a social component as it enhances communication skills, enables a sense of belonging and perspective taking and thus increases social support [51,52]. These factors, in turn, may reduce stress or physical illness and improve psychological and physical well-being [53,54].

The purpose of this study was to understand patterns across adolescents’ digital technology and media use, including four evidence-based constructs: (1) technology ownership and use, (2) parental involvement, (3) health outcomes, and (4) well-being indicators. Most studies to date have focused on 1 outcome, such as depression, or on a category of outcomes, such as mental health. This study builds on that literature by using the power of latent class analysis (LCA) to examine critical constructs and understand patterns across and within groups. The emerging understanding that adolescents’ media use is not homogeneous and the critical role of parents in moderating youth media use informed this study’s purpose. The purpose of this study was to use LCA to develop profiles representing benefits and risks as well as parental influence associated with digital technology use.

A national Qualtrics (Qualtrics International, Inc) cross-sectional web-based research panel was engaged during February-March 2019 to collect data for this LCA study.

The Institutional Review Board at the University of Wisconsin approved this study (2018-0781).

Our goal was to achieve a national sample of youths to complete a web-based survey. Compared with traditional survey approaches such as in-person, phone, or mail recruitment, web-based survey panels offer broader reach and lower costs in data collection [55]. We selected the web-based survey platform Qualtrics for several reasons. First, although web-based survey platforms do not use weighting, previous studies have shown that web-based survey approaches using tools such as Qualtrics can achieve demographic attributes that are typically within a 10% range of their corresponding values in the US population [56]. Second, unlike other platforms such as Mechanical Turk, Qualtrics allows for the recruitment of youths via approaching parents for consent as a first step. Third, there is strong and growing literature around the use of Qualtrics to recruit youth samples in the United States, including studies on media [57,58].

Between February 2019 and March 2019 a Qualtrics survey manager recruited adult panel participants who indicated that they had adolescent children aged 13-18 years who spoke English. Parents who met these criteria were provided with information about the survey and an opportunity to complete informed consent forms for themselves. The informed consent process notified potential participants of the study purpose and research team, of the survey length, that the survey had questions for the parents and adolescents to answer independently, that the survey was voluntary, of the Qualtrics incentive points that would be provided upon completion, and of how the study data would be stored and used. The survey information section stated that the researchers would not request any personal information about the participants. Parents also provided consent for their child’s participation if the child was aged <18 years. Once parental consent was obtained, the parents completed early sections of the survey on the Qualtrics platform. After completing the parent portion, the parents were instructed to pass the device to the adolescent participant. The adolescent was provided with study information and an opportunity to provide assent. Adolescents who provided assent were allowed to begin the survey. Adolescents aged 18 years provided consent and were allowed to begin the survey.

The target population for this study was adolescents aged 13-18 years who were US residents and English-speaking. Using Qualtrics panels of adult participants (a closed survey population), we recruited parent-adolescent dyads to allow for parent as well as adolescent input. We established parameters for Qualtrics to recruit a sample consistent with the race and ethnicity representative of the US census population for adults [56]. Qualtrics representatives recruited parents from their panels using emails and SMS text messages. Qualtrics processes ensured that all recruited participants had completed enrollment in Qualtrics panels, and the participants could only complete the survey a single time.

Our sample size estimates were calculated using estimates for LCA [59,60], which supports approximately 1:3-1:4 ratios of number of items to number of participants to achieve a full range of potential number of latent classes with a minimum of 0.8 power. We estimated that approximately 70 items would be included in our LCA process, and we then increased our sample size to allow for investigation of differences in demographic factors such as adolescent age, gender, or race as well as to account for incomplete surveys affecting our final sample size for analysis. Thus, our planned sample size was 4000 parent-adolescent dyads (N=8000 participants).

A total of 4592 adolescent-parent dyads began the survey, of which 3981 (86.7%) completed the survey. Participants were excluded for not completing ≥75% of the survey, responding with single-response selections across multiple survey measures, and 13.3% (611/4592) of adolescent-parent dyads were excluded for these reasons. Regarding the included participants, the adolescents had a mean age of 15.0 (SD 1.43) years, 46.3% (1842/3981) were female, 67.8% (2701/3981) were White, and 75% (2986/3981) lived in a household with an income above the poverty line (Table 1).

Our study findings can first be considered in the context of the few previous studies that have used LCA approaches to study adolescent media use. One previous study focused on quality of life among Swiss adolescents and found 5 distinct classes, with a high social technology use class scoring lowest on moods but highest on social support [90]. Another study focused on technology behaviors, including gaming and internet and smartphone use, among Korean teenagers and assessed psychosocial measures [91]. They found several subtypes, including “dual problem users” who scored highest for addictive technology behaviors and other psychosocial issues. A third study of Australian adolescents found 3 clusters, 1 focused on “instrumental” computer use related to email and general computer use and 2 clusters related to gaming [92]. A fourth study examined physical activity, screen-based media, and self-harm among Chinese adolescents and found that the highest-risk group had low physical activity, high media use, and high self-harm [93]. Our study advances the field by including parents as well as physical health, mental health, and well-being measures. Our study is aligned with previous literature, such as the finding that adolescents who struggle in one domain, such as addictive technology behavior, often have lower health behaviors in other areas, such as psychosocial issues, and concerning health outcomes, such as sleep impairment [93].

Previous studies exploring the potential of media use to negatively affect adolescents have produced mixed results, leading some to argue that these effects are present for some adolescents but not others. Research suggests that the effects of social media on adolescent well-being vary so widely that media use has positive effects on some adolescents and negative effects on others [94]. This finding has prompted a call for research to consider a differential susceptibility model to identify those individual differences that are likely to have susceptibility to negative impacts [95].

Our study builds upon and expands on these previous findings in several ways. First, our study includes both negative and positive impacts, here defined as risk and health outcomes, using previous evidence to define these categories. Aligned with the differential susceptibility model, we found a 2-class model illustrating that most adolescents who use media do well across health behaviors and outcomes. A smaller proportion of adolescents are at higher risk, and this risk extends across their technology behaviors as well as health behaviors. A critical difference between Class 1 and Class 2 was their reported level of sleepiness, with Class 2 participants reporting far greater sleep impairment compared with Class 1 participants. Given the substantial impact that sleep can have on both mental and physical health [96,97], this finding bears further exploration and could be considered a target of future interventions to reduce negative health consequences.

Second, our study included parents as a key construct in understanding the links between technology use and health outcomes. Parents are also the gatekeepers of device acquisition and ownership. We assessed the role of parents as moderators, supporters, and role models in the relationship between adolescent technology use and health and well-being outcomes. Compared with Class 2 participants, we found that Class 1 participants were more likely to have family-owned devices than personally owned devices. Furthermore, adolescents in Class 1 were more likely to have household rules that were often centered on content, coviewing, and communication. Class 2 parents were more likely to report no rules, although Class 2 adolescents were more likely to report strict rules around screen time. These mixed findings may suggest that, for some families, the rules are unclear or are stated but not reinforced. Parents in Class 2 were also more likely to report high levels of their own daily social media use. Finally, Class 1 adolescents were more likely to report positive parent-child communication about technology use and a more positive parent-child relationship in general compared with Class 2 adolescents. The consistency and conceptual connection between these variables is a critical finding of our study and a significant advancement in how we understand the balance of health and risk from adolescents’ technology use. This study’s findings support the positive role that engaged parents can play in promoting health and preventing harm among adolescents related to technology use. Thus, we propose to name the larger Class 1 Family-Engaged Adolescents and the smaller Class 2 At-Risk Adolescents.

When seeking to frame the role of technology in our health and well-being, one often hears terms such as online and offline or online and in-real-life. These terms promote a separation of our daily lives into 2 distinct worlds. News stories often frame technology as something that should be reduced or avoided, which could influence a view of technology as a risk behavior itself. However, today’s teenagers have digital tools woven into the fabric of their everyday existence. For adolescents, technology may be seen as within and not separate from their world and as something that can lead to healthy and less healthy outcomes and experiences. To go beyond this dichotomous view of technology, many researchers, policy makers, and families seek new frameworks to consider and describe the role of technology in how we navigate today’s world. Our study supports this more comprehensive viewpoint as the findings illustrate the strong alignment within each of our 2 classes across physical health behaviors, technology behaviors, parenting engagement, and mental and well-being outcomes.

An emerging framework that aligns with our study findings is the Human Experience framework (HX) [98]. The HX approach seeks to define technology as part of the human experience. As such, it can be associated with both everyday and special interactions and with both positive and negative experiences. Applying HX to our approach to youth technology use may allow us to avoid oversimplified categorizations of technology use as bad or a risk behavior and see technology use as a multifaceted activity that may be used in healthy and unhealthy ways. Thus, the HX approach may be a step toward a more comprehensive approach to understanding the role of technology in young people’s lives and their differential susceptibility to its benefits and risks based on their lived experience. Furthermore, the HX approach may be a useful lens for technology developers to see their products as part of a larger set of life experiences that may affect or influence the adolescent developmental period.

The first limitation of this study is that our results may not generalize beyond the study population of early adolescents recruited via Qualtrics. Recruiting from a web-based panel of participants meant that we could designate the study population size and criteria but limited our ability to assess the external validity of the sample. However, the Qualtrics platform and panels have been used in other studies of early adolescents [58], and the panels have been found to have close approximations to the US population [56]. Second, LCA provides a systematic approach to creating profiles representing critical variables, although our interpretations of those profile groupings may have inaccuracies. Furthermore, other unmeasured variables such as other family or school factors may have influenced our study outcomes. This study used cross-sectional data, which is common within the LCA approach but does not allow for an understanding of long-term predictors or consequences. All measures were self-reported, and future studies including other novel measurement approaches, such as biological or cognitive studies, are needed.

Finally, it may be notable to readers that we did not investigate any specific platforms in this study. Rather, we focused on technology devices and social media use frequency. Furthermore, we incorporated the technology use importance scale (ADTI), which assesses critical interactions and functions of technology that may apply to many different platforms. This was a purposeful approach toward better understanding the mechanisms underlying technology use rather than the role of specific platforms.

This study raises important implications for how we approach the topic of adolescent technology use in health care, policy, and research. Too often, the dialogue around adolescent technology use has been to frame it as a negative risk behavior that all adolescents should cease [99]. Our findings support that, for a smaller group of adolescents in this study who made up At-Risk Adolescents (Class 2), their higher rates of technology ownership and use were associated with higher rates of health risks and lower scores on well-being indicators. However, our findings suggest that two-thirds of the adolescents who made up Family-Engaged Adolescents (Class 1) in our study integrated technology into their lives in ways that were not associated with higher rates of depression, anxiety, or other poor health outcomes. Thus, the study findings indicate that most adolescents using technology do so in ways that do not lead to increased risk of negative health consequences.

We propose that a critical factor that affected Family-Engaged Adolescents’ (Class 1) health and well-being was the role of parents. Family-Engaged Adolescents were more likely to have household media rules at home regarding media and technology use. Participants in this group were more likely to report that the rules they had at home were aligned with the recommendations of the AAP on content, communication, and coviewing. Family-Engaged Adolescents were more likely to communicate with their parents about their technology use and, overall, they reported a positive relationship with their parents compared with the At-Risk Adolescents (Class 2). Furthermore, Class 1 parents used social media less frequently compared with Class 2 parents, highlighting the integral aspect of role modeling that parents have regarding technology use.

There are several concrete recommendations that this study supports. First, consistent with AAP recommendations, we recommend a shift away from rules centered on screen time. Our evidence supports that household rules focused on content, communication, and coviewing were more likely to be associated with lower health risk and improved well-being (see Multimedia Appendix 1 for more information). The findings of this study can direct pediatricians’ and other health care providers’ counseling toward parents and encourage them to leverage these approaches at home. Providers can also partner with parents to ensure that messaging around media is culturally relevant and developmentally appropriate. Health care providers may benefit from using technology within the clinic visit to share these recommendations, such as through electronic health record prompts to ask about technology behaviors and home rules and after-visit summary resources, including recommendations and links to resources.

These resources may include tools that can support parents in creating media rules at home. First, the Family Media Use Plan of the AAP has content, communication, and coviewing as key elements, and this approach is without cost. Other for-profit web-based tools such as Circle and Bark may integrate content restrictions into family rule development, although these programs are also at a cost and often focus predominantly on screen time. Including adolescents in the discussion and selection of rules is a critical tactic to obtaining their buy-in for setting limits and boundaries.

Second, health care providers and researchers should consider the integral role that parents play in their children’s media use. In addition to recommending that parents create and enforce household media rules, pediatricians can support parents in developing positive relationships with their adolescents. These approaches may include coviewing media. Counseling parents about having awareness of their own technology use and their role modeling of technology behaviors may be a critical recommendation to influence health outcomes. Researchers designing interventions must consider the role of parent support, both related to technology and likely more broadly in adolescents’ lives, if they want to truly affect adolescent well-being. Finally, we encourage future research and policy to consider technology as integrated into adolescents’ daily lives. Our study supports the exploration of new frameworks such as the HX approach toward the design of new policies and studies to advance adolescent health (see Multimedia Appendix 2 for more information).