Young People’s Digital Interactions from a Narrative Identity Perspective: Implications for Mental Health and Wellbeing

Abstract

The current response article began by situating this journal volume in the current COVID-19 pandemic context which has driven young people into digital spaces for far longer periods of time than ever. Many scholars, policy makers, and the public at large are recognizing that these social digital spaces may be the only outlet by which youth and their families can remain safe (at least physically), while also participating in some important social outlets. We discussed the notion of hybrid reality and argued that young people themselves often do not experience their digital and physical contexts as functionally distinct, despite there being good scientific reasons for examining these contexts separately. We then considered whether or not digital activity (at least some) should be considered pathological or inherently unhealthy for young people. Like most of the commentators, we advocated for a nuanced approach and emphasized the importance of a functional, developmental lens. In the second half of our response, we turned to the narrative identity framework itself and clarified misconceptions about what we considered novel about our approach. We found it particularly useful to explain why self-determination theory can be regarded as a constituent of the identity development framework, but that it does not subsume it. We moved on to discuss how the new insights provided by the commentators could be synthesized and suggested directions for new research studies and methodologies. Finally, we ended with an examination of how a more elaborated narrative identity framework offers specific directions for designing new digital interventions for promoting young people’s mental health and wellbeing. It is our hope that the target article with its detailed articulation of the narrative identity framework, together with the thoughtful commentaries and this response, can together inform future research, practice, and policy relevant to digital activity and its impact on young people’s mental health and wellbeing.

Keywords:

• Adolescence
• identity
• digital
• video games
• social media
• mental health
• wellbeing

The COVID-19 pandemic, and the “social distancing” measures that have been put in place as a result, have made this journal issue unexpectedly relevant. Almost immediately after the target article was accepted for publication, the pandemic began and changed all our lives, with digital technology at the fulcrum of this change. At the time we wrote the target article we could not have foreseen the significant increase in the relevancy of the narrative identity framework to understanding young people’s lives today. The COVID-19 pandemic and associated social distancing measures have disrupted young people’s social lives and is impacting normative developmental tasks enormously (Fegert, Vitiello, Plener, & Clemens, 2020; Hamilton, Nesi, & Choukas-Bradley, 2020). Youth are forced to stay home, cut back on their social contacts dramatically, work much more independently for school work, and cope with increasing uncertainty and frightening news (Common Sense Media, n.d.; Fried, 2020; Lee, 2020). In response, many young people have turned to social media and video games which allow them to remain connected with peers and explore their identities while being physically distant (Hamilton et al., 2020).

Recent statistics show spikes in social media use worldwide during the Covid-19 pandemic (Hamilton et al., 2020; Koeze & Popper, 2020). In these times of social distancing, 65% of youth report talking to family or friends via texting or social media more often than they usually do, with 60% of youth reaching out to friends or family one or more times per day (Common Sense Media, n.d.). Yet, this sudden rise in technology use also presents unique challenges to youth (and their parents and teachers): because of the social distancing measures, technology use is oftentimes not voluntary despite being the only way to connect with others. This involuntary use of digital technology might limit young people’s feeling of agency and thus also have negative effects on their wellbeing.

The commentaries were written during what many of us considered the peak (at least the first peak) of the pandemic, during a time of heightened uncertainty and shock, and the impact of this global stressor seems to have seeped into many of the authors’ responses. We were impressed and sincerely grateful for the thoughtful and rigorous scientific inquiry, as well as the big-picture view that characterized the pieces. But there also seemed to be a compelling level of personal reflection and humanism in the responses—an approach we rarely see in these academic dialogues—and this combination seemed uniquely constructive as the world faced the intellectual, emotional, and existential challenges elicited by the pandemic.

Our approach to the current response article is to summarize many of the convergences across the commentaries and to take on some of the most challenging and compelling critiques head on. We do not address each commentary separately in turn but, instead, seek to synthesize core themes in the hope of coming to a more coherent and effective theoretical framework. We trust that, in doing so, the combination of the target article and current response articles will serve as a clearly defined springboard for new directions in theory, methodology, and product design.

Is Hybrid Reality Really “A Thing”?

We argued in our target article that we are currently in the midst of a societal and global transition in which digital “screens”—and the platforms, games, and apps that run behind them—are no longer simply entertainment devices and distractions. Rather, they increasingly serve the same developmental functions as “off-screen” activities, thus constituting an entirely new hybrid digital-physical ecosystem. This proposal was met with conflicting reactions: Singer (this issue) and Nesi and colleagues (this issue) strongly urged us to keep the two contexts separate and to recognize how important it is to study them as distinct, while the other commentaries seemed to accept our characterization with little reservation.

In their commentary, Nesi and colleagues (this issue) take a “contextual approach,” suggesting that there have long been different but interrelated contexts that serve to underpin adolescent development (school, home, peer groups). Although adolescents focus on similar needs and goals across all of these contexts, they need to continue to be studied separately. These authors provided useful examples of spillover studies where online experiences impact on offline relationships and vice versa. They reference the idea of “context collapse” (Marwick & Boyd, 2011)—situations in which offline and online experiences are perceived by youth as interchangeable and mutually instantiated—to explain these spillover effects. However, for us, these effects seem to make the exact opposite point that Nesi and colleagues (this issue) are trying to make. Studies that show “context collapse” highlight the extent to which youth themselves do not distinguish on- and offline contexts, including relationship dynamics and status markers.

Virtually all the examples Nesi and colleagues (this issue) provide make it clear that cause-effect relations between online and offline behavior are bidirectional, and cause and effect are often designated arbitrarily. For example, they cite research that tracks thoughts and behaviors concerned with personal appearance on social media and show how those intrude on offline experiences such as behavior at parties (Choukas-Bradley, Nesi, Widman, & Galla, 2020). According to Choukas-Bradley and colleagues (2020), youth often attend parties and behave in particular ways with the intention of sharing this behavior later (or in the moment) in online spaces. But it is not only that young people’s concern with their “digital selves” affect their offline activities; they are also referencing these digital contexts when they are interacting face-to-face with their friends and then, in turn, changing the way they post or play back in their digital spaces. This is similar to those children who talk to their friends in the playground about the Minecraft projects they collaborated on the previous day and plot who will be invited (and excluded) in the next round of play, then instantiate those plans back in their Minecraft worlds (Lee-Leugner, 2013). So is it the offline context that affects the online context or the opposite? These reciprocal relations create iterative feedback loops in which the elements are tightly coupled, both in time and space. Thus, it seems virtually impossible or at least arbitrary to tease apart their online and offline components. Like many phenomena in science (e.g., cognition-emotion, genes-environment), we agree that there are good reasons that researchers seek to cleave nature at particular joints, even though the elements themselves are tightly interwoven, and important discoveries have been made by separating elements of systems that work as a whole in reality. However, young people’s experience of a divide between the digital and physical worlds may not be as stark as that of older adults; in fact they may not experience such a divide at all. This phenomenological insight may be key to understanding identity negotiations and mental health for the next generations of youth.

In his commentary, Singer (this issue) also provided two arguments for keeping the digital and physical worlds distinct in theory as well as practice. His first concern was about differences in access to digital media, the internet, and mobile devices. Singer (this issue) challenged us by suggesting that before we apply this developmental identity-based theory to our new “hybrid reality,” we should first consider how widespread this reality actually is. Specifically, he points out that during the current pandemic, his vantage point as dean of a large American university has highlighted the digital divide that follows predictable socio-economic and racial fault lines, as well as polarizing students from Western countries in relation to international students coming from non-Western societies (Drane, Vernon, & O’Shea, 2020; Van Lancker & Parolin, 2020). Singer’s (this issue) argument is that, although many young people seem to be living much of their lives online, the Coronavirus crisis has revealed the disparity between these digitally privileged students and those who require financial assistance for remote access to online learning and for whom connecting to the internet is more of a luxury (Drane et al., 2020; Van Lancker & Parolin, 2020). We agree, of course, that these differences do exist and are more stark during this pandemic, when youth need to socialize, get an education, communicate, and generally do so much of their functioning online (Beaunoyer, Dupéré, & Guitton, 2020; Drane et al., 2020; Ramsetty & Adams, 2020; Van Lancker & Parolin, 2020). Accordingly, disparities in access and cost can make a large difference in how youth navigate their developing identity (Kumpulainen, Mikkola, & Rajala, 2018; Metcalf, Blanchard, McCarthy, & Burns, 2008). And in countries and societies that are less steeped in technological advances, we do indeed acknowledge that our model is likely less relevant at this time. However, the time youth spend on the Internet has doubled over the last decade (Ofcom, 2019) and the percentage of people who use the internet and have access to mobile phones is steadily increasing year after year across the globe (Kemp, 2019; Pandita, 2017; Poushter, 2016; Silver, 2019; World Bank Group, 2018). At this pace, it is likely that the vast majority of the world population will have internet access within the next decade (Kemp, 2019; Silver, 2019). At that point, our narrative identity framework will become similarly relevant and may help to define the conditions under which these cultures can build digital contexts that better support young people’s developmental needs and mental health.

Singer’s (this issue) second concern with our discussion of hybrid reality is a more existential one: Especially during this pandemic, where social distancing measures have been enforced (e.g., 1.5 meters apart “rule,” no indoor crowded areas, remote work and school), digital and offline interactions just feel very different to him. Specifically, he is loathe to give up what is uniquely human about our face-to-face interactions—our fundamentally embodied experiences, social or otherwise. He provides a cautionary note about the students he sees losing their capacity to empathize, their skills to navigate difficult discussions and to form meaningful relationships, all because of the encroachment of digital contexts. These contexts usurp youths’ opportunities to train in being uniquely embodied “humans.”

Indeed, we agree that embodied social interactions are fundamental to what makes us human—from the baby’s suckling and cuddling imperatives to the teenager’s hormone-driven needs to flirt, embrace, and have sex, and adults’ needs for physical touch to regulate stress and derive comfort (Slavich & Irwin, 2014). Even cognitive development requires the exquisite synchronization of motor activity with sensory input (e.g., Thelen & Smith, 1994). In no way did we intend to suggest that the digital and physical worlds are the same. Nor do we support the notion that the digital should overtake all or most face-to-face interactions. But there is no doubt that technology has replaced some of those interactions and, we argue, it is only those technologies and digital spaces that speak to young people’s developmental and identity-relevant needs and values that become young people’s most used applications, social media platforms, and games (Davis & Weinstein, 2017). At its core, our approach is not to privilege the digital over the physical but to understand how they are creating unique developmentally-relevant experiences that interweave them (Davis & Weinstein, 2017).

In sum, we agree with Singer (this issue) that humans are fundamentally embodied. We do not wish to privilege digital over physical modes of interaction and learning; we seek to integrate them and inquire about how this integration is promoting or usurping basic identity-relevant needs and goals. In general, but particularly now during the 2020 pandemic, it seems essential to put aside our folk theories of how digital media harms youth and take a critical, open science approach to examining these assumptions.

The Pathologizing of Digital Activity

An important theme that was touched on by all the commentaries was whether or not we should consider young people’s digital participation—whether playing video games or interacting on social media or texting—as something problematic that needs fixing, or whether these activities could confer a host of psychological benefits. All commentaries agreed that it was useful to go beyond conventional measures of screen time—a conspicuous point of convergence and a convenient place from which to move forward. However, they disagreed about the extent to which we should worry about or pathologize certain types of digital activities. Orben and colleagues (this issue) and Ferguson (this issue) broadly argued from the perspective that most of the negative media hype and academic hand-wringing was unwarranted, at least as far as our best data suggest (e.g., Berryman, Ferguson, & Negy, 2018; Elhai et al., 2018; Ferguson & Wang, 2019; Heffer et al., 2019; Hilgard, Engelhardt, Rouder, et al., 2017; Jensen, George, Russell, & Odgers, 2019; Odgers & Jensen, 2020; Orben & Przybylski, 2019a,b). The rest of the commentators took nuanced approaches to the issue. Singer (this issue) and Bushman and Huesmann (this issue) highlighted more negative impacts of digital participation, warning us of the dangers and pathological implications of digital participation. Both Hollenstein & Colasante (this issue) and Nesi and colleagues (this issue) took an equivocal stance, calling for more complex models that incorporate context, timing, and functions. We can appreciate each of these perspectives. However, given that there is far more research focused on the negative aspects of screen time, it is worth highlighting the note of caution regarding the harm that may come from ringing the pathology alarm so persistently (points made by Orben and colleagues’, this issue and Ferguson, this issue). It is these negative messages around young people’s everyday digital use, with so little empirical corroboration, that may themselves stigmatize youth unnecessarily and, in turn, end up causing more harm than the actual digital activity (Odgers & Jensen, 2020; Orben, 2020). Indeed, research in communication studies suggests that people who feel more guilty about their media use (e.g., video games and television), compared to those who feel no guilt, are less likely to experience the benefits of stress recovery usually associated with entertainment media (Reinecke, Hartmann, & Eden, 2014).

Pathologizing screen time, especially in our current pandemic and during social distancing mandates, makes little sense, unless we are prepared to argue that the vast majority of people around the world will be diagnosable at the end of our forced period of digital confinement. Whereas many may be reaching “screen exhaustion” from constant video chatting, online learning, and so on, people clearly need these digital prosthetics for connecting with loved ones and carrying on with their social lives and education (Odgers & Jensen, 2020). This is particularly true for adolescents whose minds, bodies, and brains are structured to learn and thrive through social interaction (e.g., Choudhury, Blakemore, & Charman, 2006; Dunbar, 2003; Siegel, 2020). Their brains are wired to move past the rewards they have learned to value from parents and to seek out novel and intense emotional connections with like-minded peers (Crone & Dahl, 2012). These developmental milestones and mandates cannot be safely met in these current COVID-19 times without the help of social media, social video games, and other types of digital networking spaces (Hamilton et al., 2020). Helping young people understand the different functions of their digital spaces and educating them as to potential benefits as well as risks seems far more useful, and even imperative, than stigmatizing them for “too much” time online.

As we elaborate at length in our target article, digital activity now encompasses not just play and entertainment, but all facets of life, including work, education, learning, and socializing. The web 2.0 does not resemble the conditions, constraints, functions, and interactivity that once characterized digital media and, thus, calls for a much more complex, developmentally-anchored framework that takes multiple levels of analysis into account. Despite the call for more nuanced and complex approaches, we note that the “good” vs “bad” debate about video games (specifically, the violent content) still emerged in the commentaries of Ferguson (this issue) and Bushman and Huesmann’s (this issue) responses. These debates, as Ferguson (this issue) himself clearly states, are exactly what we want to move past at this point, given the robust meta-analyses and reviews that have done the heavy-lifting on this topic. Our aim is to focus on function; that is why we need broader, synthesized theoretical approaches that put each of these piecemeal arguments in a larger developmental context (see Hollenstein and Colasante’s commentary, this issue, for examples).

Singer (this issue) has more complex worries: he stays away from considering any one type of media, game, or online activity as bad or wrong or predictive of pathology. But he voices societal and deeply personal worries: he wonders if we are losing the cohesive, collective sense of a “We” that makes us feel like part of a larger whole, community, nation, or globe, as social media platforms take over and individuals retreat into closed echo chambers. In short, with the advent of these massive social media platforms, Singer (this issue) questions whether we are losing that aspirational identity that can subsume shared master narratives.

From our perspective, this shared aspirational “We” may be an idealization, especially amplified because we are living in a period of history (especially in the United States) when divisions across political factions and race relations are particularly stark. Putting aside whether those shared master narratives of the past actually benefited most young people, we do concede that social media platforms can and sometimes do create echo chambers in which individuals are only exposed to similar expressions of values and beliefs and rarely are confronted with contradictions and their resolutions. Yet it is important to note that the data are actually equivocal about the extent to which digital media itself cause these echo chambers, with some studies suggesting that this claim has been overstated (e.g., Dubois & Blank, 2018). Moreover, there are scientists and digital designers who are recognizing this problem in current social media and attempting to implement digital “interventions” or to design new social media that make people aware of their echo chambers and nudge them out, into more healthy and diverse discourse and actions (e.g., Gillani, Yuan, Saveski, Vosoughi, & Roy, 2018; van der Linden, 2019; van der Linden & Dufresne, 2017).

Importantly, individuals have begun to recognize and analyze the limitations and dangers of currently instantiated social networks. We sometimes forget that we are only at the beginning of the development of the social, interactive, Web 2.0—the Internet as we currently experience it. The internet in its current form is still in its infancy. The mammoth social media (Instagram, Twitter, Facebook), AAA video game enterprises, and multi-use platforms (Google) have no doubt unleashed unprecedented challenges at individual, social, and societal levels (e.g., Livingstone, 2008; Livingstone & Third, 2017). These platforms currently create the homogenous echo chambers through algorithms that work below the perceptual levels of individuals, surfacing only content that users are likely to agree with, and they do so for profit motives. People themselves are not consciously choosing the bubbles in which they find themselves, machine learning algorithms are. But new platforms will and have emerged—some that represent entirely new incentive structures and new ways of acting and interacting in digital spaces (e.g., Kelly, 2016).

Along with technologists and digital sociologists, we propose that we are likely at another transition point in internet culture and preferences, and this transition is happening precisely because many young people themselves are reflecting on the match between their values and goals and the opportunities afforded to them by corporate-run, profit-maximizing media platforms (Tolentino, 2019). Alternative social spaces are being co-created by young people themselves who are interested in finding like-minded peers with shared values and goals. In particular, there is a conspicuous rise of “digital campfires” as alternative modes of online social mini-networks (Wilson, 2020); these are built by young people, for themselves, not by algorithms coded to maximize engagement and dollars. These digital campfires are spaces in which youth choose to interact in smaller groups that feel more intimate and personal than monolithic spaces with 100 s of millions of users (e.g., Facebook, Instagram, and Twitter) and they are organized around mutual interests, values, and goals (Wilson, 2020). The rise of emailed newsletters, the reemergence of blogs, and the development of community platforms instantiated on Substack, Patreon, Discord, MatterMost, and Slack are all examples of digital spaces that are potentiating more intimate communities formed explicitly around shared values and collective feelings of alienation and exhaustion with corporate-run social media. Recent data from Edison Research and Triton Digital’s large-scale surveys (Baer, n.d.) suggest that social media usage overall among Americans 12–34 years old across several platforms has either leveled off or is waning. Research from Global Web Index (Viens, 2019) suggests that the amount of time millennial and Generation Z audiences spend on many social platforms is either flat or declining compared to previous years. They are not spending less time online overall, they are just moving to spaces that resonate better with their shared values and goals (Wilson, 2020).

These new, more intimate spaces may still be thought of as “echo chambers” by Singer and others, and that is the downside of any community (on- or offline) that organize around shared interests. But these new social spaces are different because they are built through young people’s own agency and their need for communion with like-minded peers, not by profit-maximizing algorithms. Many of these online communities are also characterized by a commitment to the very interpersonal storytelling processes we highlighted as critical to the development of a more coherent narrative identity (i.e., opportunities for elaboration, active listening, and the encouragement of grappling with contradictions and paradoxes). Thus, compared to the echo chambers constructed through algorithms, these new emergent digital campfires may have more promising outcomes for young people’s identity development.

From our perspective—perhaps nudged by the measures brought on by the pandemic—we have reached a tipping point from which we cannot retreat. Binary approaches to analyzing youths’ hybrid reality do not seem to fit their lived experience. Thus, we seek to change the conversation altogether: We do not see social media, video games, or any other hybrid technology as constituting any bad or good qualities in and of themselves. We are most inspired by, and would like to work at the edges of, technology that augments developmental processes and can serve as individual and social prostheses for mental health. We see immense potential in harnessing the learning capacities and social impulses of a new generation of youth growing up in a hybrid digital-physical world. There are, of course, reasons to be cautious and to protect children and youth from harm on the internet (e.g., Livingstone, 2008; Livingstone & Third, 2017). But we feel like there is comparatively little attention being paid to the potential of new interactive technologies to inspire new modes of identity exploration and socially augmented self-storytelling (e.g., Anderson & Mack, 2019; Davis & Weinstein, 2017; Kim & Li, 2020). Part of the reason these landscapes of digitally-augmented identity explorations are not being recognized as such is that former generations have had no experience with them. Unlike playgrounds, shopping malls, and telephone conversations, developmental scientists have had little to no experience with TikTok, SnapChat, and Discord and therefore may be particularly challenged to discover how these spaces work (Odgers & Jensen, 2020; Orben et al., this issue; Unicef, 2017). Understanding the promise (and pitfalls) of new social technologies, as they relate to something so complex and dynamic as identity development, may require participation in these spaces. At the very least, being tuned to these new emergent spaces calls for very close collaborations with youth themselves who can articulate how new media augments and supports their narrative identity (e.g., Boyd, 2014; Sclater & Lally, 2014).

With the commentaries in the current issue of this journal, we have been challenged to continue to integrate and acknowledge the theoretical and empirical work that we are reviewing in order to lay the groundwork for theory development regarding narrative identity in the digital age. We are most grateful for those challenges and the help commentators gave with this synthesis. But we are equally inspired and wish to turn our focus to what lies ahead, to anticipate shifts in computing and social media technologies and to guide them into productive, enlivening experiences for youth. The new set of tools, practices, playgrounds, and digital storytelling modes are only barely being glimpsed right now and their legibility will not be apparent for some time. However, our call is for a more updated, nuanced, and a more optimistic view of identity development in the digital age.

Theoretical Threads, Convergent Concepts, and Cultural Considerations

Three main issues have been raised concerning the specifics of our proposed theoretical framework. The first argument, proposed by Bushman and Huesmann (this issue) as well as Orben and colleagues (this issue), is about novelty: these authors argued that the concepts in our model overlap with existing concepts and do not provide novel insight. Indeed, none of the concepts in the narrative identity framework we presented is new. Our citations go as far back as the 1950s, so we by no means wished to suggest we were reinventing any. It was far beyond the scope of the target article to go into the details of each relevant theory but, briefly, our thinking derives from decades of research and theory-development including: Erikson’s (1959, 1968) and Marcia’s (1966) theories of identity development, Dweck’s (2017) unified theory of motivation, personality, and development, Ryan and Deci’s self-determination theory (2000), McAdams’s and McLean’s (2013) models of narrative identity, and Adler’s (2012), Chandler’s (1994), and Pennebaker’s (1997) proposals for using narrative perspectives for clinical interventions. Going into more detail, the commentaries were indeed correct in pointing out that agency overlaps with autonomy and that communion overlaps with the need to belong or relatedness. Dweck’s (2017) review is helpful as she gives an overview of basic psychological needs that have been studied under different names and groups overlapping concepts together. Our intention was to synthesize the bones of these models into one holistic framework. Our contribution comes not from the novelty of any single premise, but from stitching them together into a coherent whole and specifying novel research directions for understanding a relatively new domain: young people’s development in the digital age.

A second criticism was that self-determination theory (SDT) serves as a better framework than our approach for understanding identity development and the impact of digital media on adolescent mental health (Bushman & Huesmann, this issue; Orben et al., this issue). We acknowledge that SDT is grounded in extensive empirical work and can be useful for understanding various human behaviors in terms of needs, motivations, and regulatory styles (Griffin, Adams, & Little, 2017; La Guardia & Ryan, 2002; Ryan, Deci, & Vansteenkiste, 2016; Weinstein, Przybylski, & Ryan, 2012). In fact, we partially drew on SDT, and that is why similarities between SDT and our model are obvious. As pointed out by Bushman and Huesmann (this issue) as well as Orben and colleagues (this issue), the needs that we postulated to underlie identity development are similar to basic psychological needs in SDT. We are grateful for the work done by Orben and colleagues’ (this issue) to present the empirical support for agency and communion needs, as well as research that maps onto person-society integration. These authors also helped us by elaborating on how the autonomy and competence needs articulated in SDT relate to the agency needs in our identity framework. This explicit tracking and linking of concepts with empirical evidence has left us encouraged by the convergence of the narrative identity model with other related studies in the field and should provide clarity for readers confused by the related terms across domains.

Despite its contributions, certain limitations of SDT prevented us from using it for our overarching framework. The main reason why we did not use SDT as our framework is that it is not a developmental theory. Ryan and Deci (2000) note that regulatory styles in SDT (e.g., introjected regulation and identified regulation) are not a developmental continuum. In other words, these regulatory styles can exist at any point in developmental time (Ryan & Deci, 2017). In contrast, we were concerned with different ways in which individuals regulate themselves across adolescent development. Therefore, unlike SDT, our model emphasizes a developmental sequence. Specifically, conformity comes before and is a prerequisite for more autonomous functioning (Erikson, 1968; Loevinger, 1976). People need to learn to socialize and identify with social values before starting to evaluate them for personal congruence. In short, although some of the concepts in SDT were useful for the conceptualization of our model, it could not serve as the overarching framework of our model.

The importance of a developmental lens for understanding identity processes in digital participation was further elaborated on in Hollenstein and Colasante’s commentary (this issue). These authors consider the role of parenting in the digital age: What are the mental health implications for adolescents who are used to vastly different digital ecosystems than their parents? They point out that parents’ supportive role in balancing autonomy-granting and limit-setting will likely make a difference on how and what digital practices are taken up by youth and may lead to a variety of mental health outcomes depending on the success of this balancing (Rodríguez-Meirinhos et al., 2020). Indeed, our target article mostly focused on peers because of their importance in narrative identity during mid-late adolescence, but the larger parental and family context should not be ignored. They remind us that the half century of excellent research on specific parenting practices that could scaffold or impinge upon identity development are probably not only useful for considering bedtime and playground rules, but also limits on Fortnite and Snapchat as well.

Returning to the limitations of SDT it is also limited in its focus on intraindividual processes: The model elaborates on needs, motivations, and regulatory styles, all of which exist within a person. SDT does not deal with how needs and values manifest in interpersonal dialogues which are themselves embedded in cultural contexts. Thus, dialogical processes, which constitute an important part of identity development processes, are not central to SDT. In contrast, we argued throughout our target article that narrative processes are at the core of identity development and, as such, are fundamentally socially constituted through recursive and reciprocal storytelling mechanisms (McAdams & McLean, 2013). Identity development is the gradual process of a person gaining autonomy and equal standing among socially-valued others, in large part through storytelling processes centered around the self. Much to our surprise, this central tenet of the narrative identity framework was largely neglected by all commentaries except Singer’s (this issue).

Finally, a third criticism levied at our approach was about cultural considerations. Singer (this issue) brought up the compelling question of whether our identity framework is universal or whether we are applying a lens that is only relevant to WEIRD (White, Educated, Industrialized, Rich, Democratic) populations (Brady, Fryberg, & Shoda, 2018; Henrich, Heine, & Norenzayan, 2010). For example, Singer (this issue) suggested living up to familial expectations is an important concern of many Chinese youth, and therefore developing a strong personal identity (processes we described at the intraindividual level) may not be a concern for the average Chinese adolescent and may be irrelevant to their mental health. Indeed, the collectivist value that is widely shared in countries like China may be incompatible with the development of personal identity, which entails agency and self-exploration (Schwartz, Zamboanga, Meca, & Ritchie, 2012). It is therefore important to consider in which cultural context the identity development framework is relevant and can be useful.

We agree with Singer (this issue) that cultural considerations are critical to examine and much more work can and should be done to elaborate on master and alternative narratives. These narratives indeed differ across cultural contexts and go on to constrain interpersonal and intrapersonal processes central to identity development. However, we disagree that our model is wholly irrelevant to most non-Western cultures. We argue that although the development of personal identity may currently be a concern primarily of adolescents in the West, it is possible that it will become increasingly more valued by youth outside the western context (Sugimura, 2020). With the growth of information technology infrastructure around the world, more and more people are interacting with those from different cultures with diverse cultural values. In fact it is largely as a result of internet technology that people are becoming less confined to the values and narratives of their local culture, and living increasingly in the global culture and under global narratives (Chareonwongsak, 2002). Indeed, a study conducted on youth in Japan, a country that has thus far considered collectivistic, showed that personal identity was positively associated with well-being for these youth (Sugimura et al., 2016). As individuals are developing, cultures are also evolving. It is possible that the characterization of the East as collectivist may no longer hold in the near future, and the narrative identity development framework we have outlined may become increasingly useful for understanding adolescent mental health globally.

Extensions to the Narrative Identity Approach

Despite our disagreements with some of the criticisms leveled at our theoretical framework, we were generally delighted by the extraordinary breadth of theoretical issues introduced by the commentaries. These varied conceptual viewpoints may constitute the most successful outcome of this journal issue as a whole; taken together, the target article and the commentaries have already brought us so much farther than the conclusions achieved by a majority of “screen time” studies. Indeed, we are encouraged to extend our theoretical framework and further integrate socio-emotional models (Hollenstein & Colasante, this issue), sociocultural considerations (Singer, this issue), social-cognitive models (Bushman & Huesmann, this issue), and motivational models (Ferguson, this issue; Orben et al., this issue), and to situate our approach in the larger domain of developmental psychopathology (Nesi et al., this issue). From our perspective, none of the approaches that were taken by the commentaries directly contradicts ours; instead, they augment, elaborate, extend, and perhaps constrain our model.

The Narrative Core of Identity Development

In the following section, we have endeavored to synthesize some of the most important contributions to our model made by the commentaries, focusing on their specific implications for methodology, future research, and intervention. Before we address these extensions, however, we begin by recalling one of the main premises laid out in the introduction to our target article. That is, with the inception of the internet and mobile digital connections, western (as well as many non-western) societies are in the midst of a massive cultural and societal transformation—one that denotes a paradigm shift reminiscent of the advent of the printing press and the industrial revolution (Granic & Lamey, 2000). Despite this shift, we remain evolutionarily the same humans who initially communicated face to face, without any tools, mainly in small groups, uniquely differentiated from our primate cousins by our storytelling proclivities (Harari, 2015). With these evolutionarily ancient storytelling impulses as our base, we set out to explore the implications that growing up in this digital age confer upon the current and upcoming generations of youth. Curiously, apart from Singer (this issue), no commentator chose to address the fundamental claim that identity is best understood as a continually unfolding narrative, with all the implications that raises. This premise includes the presuppositions that narrative identity processes are (a) inherently interactive and social, constructed through storytelling processes that (b) themselves are constrained by cultural stories shared by majority or minority groups.

We are particularly drawn to this narrative conceptualization of identity formation because it lends itself so well to a transdisciplinary approach to studying the developmental implications of today’s hybrid reality. We see that even beyond psychology, across a wide range of disciplines such as philosophy (e.g., Murdoch, 1997), history (e.g., Harari, 2015), neurology (Sacks, 1985), and across the creative arts (e.g., Emily Dickenson’s poetry; Bethune, on Margaret Atwood, 2017), this narrative structure of identity makes sense as the organizing meta-theoretical lens that elucidates what it is to be human and grow and connect with others in a humane and empathic way. This is why we find it so important to start from a recognition of the narrative basis of meaning-making, and it is why we consider the alternative models proposed by the commentaries complementary and constitutive rather than incompatible with our narrative approach.

We are keen to demonstrate how the narrative identity approach can both integrate and subsume the variety of models that have been put forward by the commentators. In the following section, we summarize these contributions by focusing on suggestions for concrete research directions. We then extend their insights by applying them to the future design of digital spaces that may be aimed at supporting emotional resilience, improving mental health, and constructing digital-based interventions for high-risk youth.

Implications for Future Research and Methodologies

Some of the most useful insights from the commentators were for methodological advances that need to be made in lockstep with the development of our theoretical framework. In the Research Agenda laid out in the original target article (this issue), we described a set of studies that fell into two broad categories of research approaches: (1) studies that test identity processes and mental health outcomes associated with different kinds of digital products and (2) studies that examine identity processes and mental health outcomes associated with individual differences in how young people use the same digital products. Most of the suggestions from commentators were made for how to conduct the latter studies, studies that can broadly be described as mediation and moderation designs. Nesi and colleagues (this issue) provided a range of compelling moderator analyses that could be performed by applying Valkenburg and Peter (2013) Differential Susceptibility Model, a model that outlines several individual difference factors that can predict how the same technology may impact young people differently. We agree that this model is ideal for investigating individual differences based on dispositional and developmental factors. Thus, it will be helpful to synthesize research findings that have come out of the Differential Susceptibility approach. We also agree with Nesi and colleagues (this issue) further advise that studies on identity processes associated with different types of digital products can be framed as investigating “affordances” (Nesi, Choukas-Bradley, & Prinstein, 2018).

Hollenstein and Colasante (this issue) also included some excellent suggestions for research approaches that can examine individual differences, this time by connecting macro-developmental mental health outcomes (e.g., anxiety and depression) with digital experiences in real time (i.e., moment-to-moment). They outlined the benefits of strategies that link real-time, functional mechanisms to proximal process that, in turn, impact on distal outcomes: for example, linking feelings of shame expressed and experienced in real time to processes associated with using digital affordances (proximal process) in particular ways which, in turn, impact on mental health (e.g., De Rubeis & Hollenstein, 2009). One of the biggest advantages to working with video games, for instance, is the opportunity to track real-time behavior and link it to outcomes (Scholten & Granic, 2019; Wols, Lichtwarck-Aschoff, Schoneveld, & Granic, 2018). Real-time recorded gameplay can be assessed to tap behavioral and emotional processes (e.g., avoidance of social interactions, withdrawal after rejection) related to long-term mental health outcomes (e.g., anxiety and depression). For instance, players' responses to game mechanics (e.g., avoiding threats, rejecting bids for cooperative play, ignoring positive cues) can be assessed in lieu of self-report, thus limiting social desirability and related biases that come from retrospective accounts (Wols et al., 2018). Indeed both Orben and colleagues (this issue) and Ferguson (this issue) provide excellent reviews of the serious limitations of self-report and retrospective data in this field and we see both great promise and important challenges ahead in developing sensitive and reliable measures. Hollenstein and Colasante (this issue) also provide suggestions for future iterations of our framework for methodological rigor including (1) consideration of different types of developmental domains (e.g., identity, emotion regulation skills) and how they relate to each other; (2) consideration of developmental processes leading to the adolescent transition to understand different trajectories of identity development thereafter; and (3) approaches to identity development that take a complexity science approach, integrating multiple levels of analysis (this latter suggestion was echoed also by Nesi and colleagues, this issue). We now move to the final section of this response, where we examine the implications for improving young people’s mental health and wellbeing by applying the narrative identity lens to the design and evaluation of diverse digital tools.

Implications for Digital Mental Health Interventions

In our current times, COVID-19 has transformed the avenues for which everyone (including youth) need to seek mental health care (Golberstein, Wen, & Miller, 2020; Pfefferbaum & North, 2020). Digital tools for mental health interventions have proliferated accordingly and become more accepted as potentially effective methods that do not require physical proximity (Shore, Schneck, & Mishikind, 2020). However, it is important to identify the features and conditions under which digital spaces can promote mental health and wellbeing, versus those that are ineffective or can even worsen outcomes. Hollenstein and Colasante (this issue) commented on the potential for digital media to provide the engaging training contexts that build developmental skills such as emotion regulation. They use the example of MindLight, a video game that has been shown to be as effective as cognitive-behavioral therapy for preventing and treating anxiety in children (e.g., Schoneveld, Lichtwarck-Aschoff, & Granic, 2018). MindLight and other digital therapeutic games aim to take evidence-based practices (e.g., exposure techniques, reappraisal training) and embed them into digital contexts that youth find appealing.

Unfortunately, the vast majority of digital mental health interventions for young people do not work, or they work for only a very small proportion of youth who are being coached or who see a therapist at the same time (Clarke, Kuosmanen, & Barry, 2015; Das et al., 2016; Davies, Morriss, & Glazebrook, 2014; Hollis et al., 2017; Pennant et al., 2015; Podina, Mogoase, David, Szentagotai, & Dobrean, 2016; Rooksby, Elouafkaoui, Humphris, Clarkson, & Freeman, 2015; Välimäki, Anttila, Anttila, & Lahti, 2017; Ye et al., 2014). Part of the problem with these digital mental health interventions is that they are poorly designed; virtually all currently tested digital interventions for youth lack engagement and retention, hence high attrition is the norm (for review, see Scholten & Granic, 2019). We suggest that an additional, fundamental problem with these digital mental health approaches is that they exclusively rely on cognitive-behavioral techniques ( CBT; Andersson, 2016; Pennant et al., 2015; Stiles-Shields, Ho, & Mohr, 2016). Face-to-face CBT involves not only the active cognitive-behavioral training, but also interpersonal processes embedded in the therapeutic relationship, processes that tap the interpersonal mechanisms that contribute to healthy changes in narrative identity, as elucidated in the target article. Thus, although CBT is the gold-standard for offline therapy for anxiety and depression (Bennett et al., 2016; Crowe & McKay, 2017; Hofmann, Asnaani, Vonk, Sawyer, & Fang, 2012; Mychailyszyn, Brodman, Read, & Kendall, 2012; Oud et al., 2019; Werner-Seidler, Perry, Calear, Newby, & Christensen, 2017), we propose that CBT approaches are not as effective online because they fail to recruit the interpersonal storytelling processes that are critical for identity development. This hypothesis—that digital interventions aimed at improving young people’s mental health will be more effective when they target narrative identity processes—could be tested in a wide range of studies using therapeutic video games, social media interventions, and virtual or augmented reality treatments.

Before diving into these digital intervention possibilities, we consider Nesi and colleagues’ (this issue) pointed question as to whether digital media contexts can and should be responsible for promoting mental health and supporting developmental processes. Should comprehensive tools be designed for these benevolent purposes or should it be the task of adolescents themselves to construct their narrative identity on their own, piecing together their digital selves without the aid of “therapeutic” digital programs? Singer raises a related concern about whether the face-to-face psychotherapy that he delivers to his clients, with its inherent intimacy and relational potency, can ever be approximated (both in terms of the qualitative experience and its effectiveness) with online interactions. In both cases, there seems to be a contention that we should not even try, either because it invokes excessive adult interference or because that interference will necessarily fail compared to offline guidance and mentorship. Yet from our point of view, commercial products incentivized by profit and promoting youths’ persistent engagement at the expense of all else are already trying to influence these identity processes. Why should psychologists and researchers stay out of this realm of influence, especially given the accumulation of evidence as to what can be most effective? At the very least it seems important for developmental scientists to provide evidence-based principles and practices that can be disseminated to designers in the commercial spaces that are already developing products aimed at improving mental health (Anthes, 2016; Lindhiem & Harris, 2018). We propose that the narrative identity framework has very specific implications for concrete and actionable digital design approaches.

For example, the application of self-affirmation principles—principles that are consistent with our identity framework—may be particularly successful for improving mental health outcomes in youth. Self-affirmations enhance agency, amplify experiences of communion, and help integrate these two core needs through reflection, identification, and affirmations that get expressed in narrative forms that reference the self (often through writing exercises; for a review, see Cohen & Sherman, 2014). Self-affirmation prompts in conventional interventions can be quick and unobtrusive, yet have a holistic impact on the self-system (e.g., Sherman & Cohen, 2006; Steele, 1988). Their effectiveness is especially potent if they are applied during vulnerable or emotionally salient transition periods (e.g., after a romantic breakup or parental divorce) or during developmental transitions (e.g., puberty, beginning of college, entry into the job market; Alwin, Cohen, & Newcomb, 1991; Cohen & Sherman, 2014 ).

Conventionally, when self-affirmation training was delivered in classrooms or therapy sessions, the exercises were intentionally designed to provide participants with a “psychological time-out” (Lyubomirsky & Della Porta, 2010), a short time for self-reflection and grounding. In the digital context, these exercises could be similarly designed to incentivize individuals to pull back from persistently used digital applications and reflect, at different points in their day, on their activity across their hybrid reality, questioning how well that activity promotes positive identity processes. Such exercises could subsequently prompt individuals to seek out and align with social partners who share the values and beliefs they have actively endorsed and integrate those social experiences into a more comprehensive self-narrative. The long-term impact of self-affirmation exercises seems to come from their power to fertilize agency and social connection within identity processes (Cohen & Sherman, 2014). These self-affirmative narratives are then amplified over time and connected with social partners who reinforce them. It is thought that, by these means, they actualize improvements in mental health and wellbeing (Cohen, Garcia, Purdie-Vaughns, Apfel, & Brzustoski, 2009; Cook, Purdie-Vaughns, Garcia, & Cohen, 2012; Sherman et al., 2013; Wilson, 2011).

Another, closely related, potentially fruitful avenue for designing digital interventions are online support forums that are organized around mutual help in coping with mental health complaints. These forums (e.g., emergent spaces on Reddit such as https://www.reddit.com/r/mentalhealth/ or TalkLife: https://web.talklife.co/) are similar in structure to some of the most popular social media platforms, yet their explicit purpose is to provide a context for expressing mental health concerns, recruiting support and advice, and providing social support to other like-minded peers. Past research has estimated that millions of people seek support online every day (Eysenbach, Powell, Englesakis, Rizo, & Stern, 2004), although these estimates are almost impossible to collect accurately. A recent systematic review of studies on young people who seek peer-to-peer online support for mental health problems showed potential benefits, but far better studies need to be designed before reliable conclusions are reached (Ali, Farrer, Gulliver, & Griffiths, 2015). Given that these forums are unmoderated, iatrogenic effects may also result, as when young people with eating disorders convene for “support” but end up amplifying their problem behavior (Arseniev-Koehler, Lee, McCormick, & Moreno, 2016; Bell, 2007; Juarascio, Shoaib, & Timko, 2010; Levine & Chapman, 2011; Yeshua-Katz, 2015).

Studies on digital interventions that apply our narrative identity framework may be particularly informative, because they point to the specific types of peer support that are effective for increasing agency and coherence, improving mood, and ultimately decreasing mental health problems. Machine learning methods, such as topic modeling, and linguistic analysis tools, such as LIWC, in combination with more qualitative coding methods could be used to determine the most effective types of peer support. They can also be used to directly test the general hypothesis that social learning and storytelling mechanisms that have been linked to healthy narrative identities will be causally linked to improved mental health. As just one illustrative example, topic models and LIWC might be used to examine the content of comments related to specific types of posts, with the aim of specifying post and comment classifiers that characterize effective peer support (e.g., Andalibi, Ozturk, & Forte, 2017). Additionally, responses to expressions of contaminated narrative identities in these forums can be tracked through semantic analysis and qualitative methods to investigate the impact of bids for elaboration and whether they lead to expressions of mood improvement. All this information might in the end lead to design improvements in these forums such that agency and communion posts and comments are reinforced through social currency (approval ratings, leadership boards) that matter to young people, potentially boosting integration processes. Then, the influence of these design implementations on young people’s mental health can be evaluated subsequently.

Conclusions

We are genuinely grateful for the careful reading of our target article and the varied insights that were offered by the commentators. There is no doubt that addressing their critiques and considering the implications of their feedback has resulted in a much more robust, synthesized, and useful theoretical framework from which to understand identity development in the digital age. We started this response article by situating it in the current context of the COVID-19 pandemic, a context that has driven young people into digital spaces for far longer periods of time than ever. Most of us are recognizing that allowing for considerable time in social media and gaming contexts may be the only way to keep youth and their families safe (at least physically), while also providing them with some important social outlets. With these peculiar global constraints in place, it is likely that there will be more interest in understanding and capitalizing on the benefits of digital interactions for youth.

In an effort to provide a cohesive response, we focused on some of the most common themes that emerged across commentaries, as well as pinpointing some unique challenges that were brought up. In the first half of our article, we discussed the notion of hybrid reality and argued that young people themselves often do not experience their digital and physical contexts as functionally distinct, despite there being good scientific reasons for examining these contexts separately at times. We then examined a common theme across the commentaries: whether or not we should consider any or some digital activity as pathological or inherently unhealthy for young people. Like many of the commentators, we advocated for a nuanced approach and emphasized the importance of the functional, developmental lens in these discussions. In the second half of our response, we turned to addressing the specific critiques of the narrative identity framework itself. We were thankful for the opportunity to clarify any misconceptions about what we considered novel about our approach and were pleased to be able to map out the many overlapping concepts that appear in disparate models. This problem of having different labels for similar concepts is a notorious one across the psychological sciences and it was important for us to be able to articulate how our approach was similar, as well as distinct, from others. We found it particularly useful to explain why we feel that the SDT model can be regarded as a constituent of the identity development framework, but that it does not subsume it. We moved on to discuss how the new insights provided by the commentators could be synthesized and suggested directions for new research studies and methodologies. Finally, we ended with an examination of how a more elaborated narrative identity framework offers specific directions for designing new digital interventions for young people’s mental health and wellbeing. It is our hope that the target article with its detailed articulation of the narrative identity framework, together with the thoughtful commentaries and this response, can together inform future research, practice, and policy with the aim of supporting young people’s mental health and wellbeing.

Additional information

Funding

This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program [grant agreement No 683262].