Introduction
Adolescents’ excessive screen time has been found in some studies to be linked to negative well-being and development (e.g., physical, mental health, behavioral, and neurological problems) (see Fang et al.,
2019; Neophytou et al.,
2021, for reviews). However, these links are much debated, as it has also been suggested in other studies that, for example, the associations between adolescent screen time and mental health issues are inconsistent and small in magnitude, and that these associations have varied in terms of types of screen time as well as mental health problems (see, e.g., Tang et al.,
2021, for a meta-analysis). These small effects or inconsistent findings may be due to the fact that a single assessment of youth screen time has typically been used in previous studies, whereas repeated measures throughout adolescence may provide a more accurate indication of longitudinal screen habits. Using developmental measures of screen time use can provide a more robust examination of their associations with well-being and development, as adolescents may change their screen time patterns (e.g., media types or activities) over time. It is critical to examine screen time patterns using a longitudinal framework. The current study aimed to evaluate the joint longitudinal patterns of time spent on three commonly used media types, i.e., TV/DVDs, videogames, and chatting/surfing on the Internet, over the entire adolescent period, as well as their associations with adulthood outcomes, including mental health (i.e., depression, anxiety, suicidal ideation, and self-injury) and behavioral (i.e., aggression, delinquency, and substance use) issues using data from a representative Swiss adolescent sample.
In adolescence, forming identities and forming/redefining relationships with peers (including romantic relationships) and parents are the “classic” developmental tasks. Digital media is an important place to fulfill these tasks for modern adolescents (Borca et al.,
2015), and adolescence might be a time of changing needs with respect to screen media. Empirical research also supported this; for instance, adolescents’ time spent on media for social purposes (e.g., texting or chatting) increases during mid-adolescence and peaks in late adolescence, whereas their time spent on traditional media devices (e.g., television viewing, videogames) remains stable throughout adolescence (Coyne et al.,
2018). Adolescence has been described as a time of increased arousal and vulnerability in emotional and behavioral regulation, which may pose a challenge for managing the escalating demands of specific media screens.
Empirical evidence has suggested the need to consider variations in adolescent developmental trajectories of screen time. For instance, one study estimated the developmental trajectories of time spent on TV in an Australian cohort (i.e., the Raine study,
n = 2411) from childhood (5 years) to emerging adulthood (20 years) and identified three distinct patterns: consistently high, consistently low, and a sharp increase during the adolescent years (McVeigh et al.,
2016). Considerable heterogeneity was also found in the developmental trajectories of time spent on texting in an American adolescent sample followed from ages 13 to 18 (
n = 425), and four-trajectory groups were detected: perpetuals, decreasers, moderates, and increasers (Coyne et al.,
2018). In addition, a study used the dataset of the Korean Youth Panel Study (KYPS,
n = 3449) to evaluate the developmental patterns of adolescent students’ online game time, and a four-trajectory model was indicated: low, rising, declining, and chronic groups (Hong et al.,
2014). Another study examined trajectories of total screen time (i.e., the sum of time spent on television, videogames, and computers) during a weekday in a Brazilian birth cohort followed from 11 to 18 years old (
n = 3382) and found a three-group trajectory solution fitted the data best: always high, always moderate, and always low (Silva et al.,
2017).
The aforementioned studies have provided evidence suggesting there is considerable heterogeneity in adolescent screen time developmental patterns. To date, most studies examining longitudinal trajectories of screen time in adolescence have evaluated a single media type (e.g., TV) or total screen time. Adolescents may have complex media use patterns, for example, spending a lot of time on several media types during a certain period of time or switching from one media type to another over successive periods of time. Thus, additional research is needed to characterize the longitudinal patterns of time spent on several frequently used media types among adolescents, which will allow us to build a more complete picture of their screen habits over time.
It may be clinically beneficial to identify young adults’ outcomes differentiated by the trajectory subgroups that emerge in the longitudinal pattern analyses (e.g., determined via latent class growth analysis, LCGA), as this could help illuminate the potential distress, costs, and impairment linked to particular trajectories. A small number of previous studies have also evaluated the subsequent correlates of identified screen time trajectory subgroups. For instance, studies have tested the associations of television watching trajectory classes across ages 5 to 20 with percentage body fat and mental health issues (including depression, anxiety, and stress) at age 20 (McVeigh et al.,
2016), the associations between trajectories of texting from ages 13 to 18 and outcomes of depression, anxiety, aggression, relationship with parents and friends, and cell phone problems at age 18 (Coyne et al.,
2018), and the associations between the trajectory subgroups of total screen time (including television, videogames, and computers) from 11 to 18 years and pulmonary function at 18 years (Silva et al.,
2017). The findings of the abovementioned studies provide preliminary corroboration for the notion that adolescents who increasingly or persistently spend a lot of time on a particular type of media screen (or have greater total screen time) are more likely to develop poor outcomes. However, to the best of knowledge, no studies have examined the early adults’ correlates of distinct developmental trajectories jointly defined by multiple types of screen time. The parallel-process LCGA allows the developmental trajectories of several different constructs to be captured simultaneously, and the trajectory subgroups that emerge from this analysis can be used to compare subsequent outcomes. Such findings can potentially inform which screen use trajectories might be markers of mental health and behavioral issues in different domains and inform further research to examine causal directions of influence. For example, knowledge that a person might be a member of a problematic screen usage trajectory in adolescence might be used to inform preventive interventions to reduce the risk of adverse outcomes during early adulthood. A subset of outcomes that have previously been indicated to be related to at least one media type (i.e., TV/DVDs, videogames, and chatting/surfing on the Internet) were included in the present study to examine how screen-time trajectories associated with them. These outcomes include mental health and behavioral issues, i.e., depression, anxiety (e.g., Kandola et al.,
2021; Stiglic & Viner,
2019), suicidal ideation (e.g., Coyne et al.,
2021), self-injury (e.g., Wiguna et al.,
2021), aggression (e.g., Keikha et al.,
2020), substance use (e.g., Boers et al.,
2020), and delinquency (Exelmans et al.,
2015). These outcomes were included also as they are common in adolescence and tend to co-occur (e.g., Murray et al.,
2022; World Health Organization,
2021).
Existing literature provides several theoretical frameworks that imply effects of screen time on these outcomes. Based on the displacement hypothesis (Kraut et al.,
1998), all screen time can have a detrimental effect on health since it might replace time spent engaging in, for example, healthy pursuits (e.g., physical activity) or activities that might be beneficial to youth development (e.g., cognition, impulse control). Review work has suggested that prolonged sensory stimulation from excessive screen time could potentially exert adverse effects on brain development during adolescence, further associated with a wide range of impairments in cognitive function (e.g., deficits in reward and cognitive control), learning, and memory processes (see Marciano et al.,
2021, for a review), which could in turn lead to issues in emotional and behavioral regulation. Another recent review study (Lissak,
2018) has proposed that the associations between excessive screen time and internalizing (e.g., depression and anxiety), externalizing (e.g., aggression), and suicidal behavior may be due to sleep problems, since excessive screen time may affect sleep in multiple ways (e.g., displacing other activities, after-dark use), and the association between sleep problems and mental health and behavioral issues is well-established.
Other theoretical perspectives have suggested that screen time effects may depend on the nature of the media content. For instance, chronic exposure to violent, suicidal-/self-injury-related, and substance-using content in videogames, TV/DVDs, and/or the Internet has been proposed to increase the risk of antisocial behavior (e.g., aggression, delinquency) (Bender et al.,
2018), suicidality/self-injury (Arendt et al.,
2019), and substance use (Davis et al.,
2019), respectively. Besides this, media content (e.g., idealized body image) may drive youth to make upward social comparisons, likely associated with negative self-evaluation or emotional distress (e.g., Hanna et al.,
2017). Taken together, building on the multiple theoretical perspectives that could explain associations between media use trajectories and mental health and behavioral outcomes, focusing on identifying trajectories that may be predictive of these issues may ultimately help inform the identification and/or prevention of them.
Discussion
Evaluation of the joint developmental trajectories of multiple types of media screen time throughout adolescence and their associations with young adulthood outcomes could provide a comprehensive understanding of adolescents’ longitudinal screen habits and early adults’ outcomes in order to inform strategies for improving these outcomes. This study used a parallel-process LCGA in a large longitudinal sample with 9 years of follow-up to estimate joint trajectory groups defined by the amount of time spent watching TV/DVDs, playing videogames, and chatting/surfing the Internet and their age 20 outcomes (including depression, anxiety, aggression, suicidal ideation, self-injury, substance use, and delinquency).
The findings suggested that, overall, from early to late adolescence, youth spent increasing amounts of time chatting/surfing on the Internet, but there was almost no increase in time spent on TV/DVDs and videogames, which is consistent with earlier research (Coyne et al.,
2018). Previous findings also suggested adolescents use the Internet in the completion of their developmental tasks: e.g., exploring identity, developing and practicing autonomy, and formatting relationships outside the family (Borca et al.,
2015). This may account for the increased time spent chatting/surfing on the Internet during this period. However, TV/DVDs and videogames are used for a variety of purposes (e.g., adolescents now increasingly use gaming platforms and online television viewing for social purposes) as technology advances during the longitudinal evaluation period of the current study, future studies would benefit from collecting the purpose of using media screens.
Moreover, the five trajectory subgroups found in this study indicated that the development of adolescents’ screen time was quite heterogeneous. The results showed that more than half of the participants (including low- and moderate-screen use) used each screen type for less than (or almost equal to) 2 h/day, yet their total screen time exceeded the recommended “less than 2 h/day” (M. S. Tremblay et al.,
2016) for the majority of their adolescence. Given that these two groups spend the least time on screens compared to the other three groups, this means almost all adolescents, rather than a small group, may need to be encouraged to reduce the amount of time they spend on media screens, and this should be brought to the attention of parents and educators. Nearly a quarter of participants spent a rapidly increasing amount of time chatting/surfing online (from less than 1 h/day at age 11 to nearly 3 h/day at age 17). This group fits more closely with the overall trends previously discussed regarding the development of screen use in adolescence. Although screen use may link to some positive outcomes in terms of helping adolescents achieve developmental tasks, the concerns about the increasing use of screen time previously discussed suggest there is a need to better understand the pros and cons of different forms of screen use and the correlates of cumulative time spent on devices. About 10% of adolescents spent an increasing amount of time playing videogames (from ~1.5 h/day at age 11 to more than 3 h/day at age 17) and chatting/surfing online (from ~1 h/day at age 11 to ~2.5 h/day at age 17). The increased amount of time spent playing videogames in this group was inconsistent with the average trajectory observed for this type of screen, which more directly suggests that the heterogeneity of screen time trajectories should be taken into account. Additionally, ~10% of youth spent a lot of time playing videogames and watching TV/DVDs (~2.5 h/day for each type of screen) at age 11, and these youth may benefit from receiving guidance for screen use during their early adolescent years. Taken together, these findings indicate the complicated developmental patterns in adolescent screen use, and highlight the importance of monitoring changes in time spent on regularly used media screens in order to provide timely support for overuse behaviors (if their time on screen media exceeds the recommendation, i.e., 2 h/day). The heterogeneity of use also points to a need for tailoring programs based on screen-time trajectory group membership, as opposed to a “one size fits all” approach.
The current study also examined several adulthood outcomes associated with each screen time trajectory group. In terms of adulthood outcomes of depression and anxiety, the group of increasing chatting/surfing had the highest levels of these outcomes (though some comparisons were not statistically significant). These are consistent with some previous findings, for example, suggesting that spending more time on social media may lead to a particularly high risk of internalizing problems (Riehm et al.,
2019). The aforementioned displacement hypothesis (e.g., displacing social interaction in the real world) and upward social comparisons offer two possible explanations for the association between chatting/surfing the Internet and internalizing problems. Some other speculative explanations include that excessive chatting/surfing online may increase dependence on social media/internet (see Lissak,
2018, for a review), that chatting/surfing online may increase the likelihood of negative social media experiences, such as cybervictimization (Craig et al.,
2020), and that night-time use of screen media may lead to sleep disturbances (see Lissak,
2018, for a review), all of which may in turn lead to internalizing problems. On the other hand, this finding may be explained by the fact that youth at risk of internalizing problems are more likely to overuse the Internet. As one example, internalizing problems are associated with sleep difficulties, which may increase youth night-time use. Future studies are needed to investigate this possibility.
Regarding self-injury and suicidal ideation, youth in the trajectory of increasing videogame and chatting/surfing displayed a greater risk of these thoughts and behaviors at age 20. Previous research suggested that the acquired capability for suicide (including pain tolerance and fearlessness about death), i.e., a component of the interpersonal theory of suicide, may explain the link between videogames and suicidal behavior (e.g., Gauthier et al.,
2014; Mitchell et al.,
2015). This may be because overexposure to violent content in videogames and on the Internet may increase habituation to painful and/or frightening stimuli, which might increase the capability for suicide and then lead to self-injury or/and suicidality. This group also spent increased time chatting/surfing on the Internet during adolescence, as found in the increasing chatting/surfing group, which was associated with more internalizing symptoms; similarly, this group had relatively high levels of depressive symptoms, and the depression-suicide link has been well established in the existing literature. Since this group spent increased time both in videogames and chatting/surfing online during adolescence, additional research is necessary to determine whether the high levels of suicidality and self-injury exhibited by this group are due to their involvement in videogames or if there is a potential interaction between videogame use and chatting/surfing online. The aforementioned explanation was merely conjecture, and the association between time spent chatting/surfing on the Internet and videogames and suicidality/self-injury might be reciprocal. Existing research has paid much attention to the pathway from the former to the latter; further research would benefit from examining the directionality of this link.
Regarding aggressive behavior, adolescents in both the early-adolescence screen use group and the increasing videogame and chatting/surfing group exhibited higher levels of this behavior. A common explanation for this link pertains to the presence of violent media content within media activities (e.g., violent movie scenes on TV/DVDs, violent videogames, accessing violent images or videos through the Internet) (e.g., Holtz & Appel,
2011). According to the General Aggression Model (Anderson & Bushman,
2002), for example, exposure to violent content can lead to desensitization among individuals towards both real-life violence and media violence. This, in turn, may increase the probability of aggression due to repeated exposure to violent media and lead to the establishment of stable patterns of aggressive behavior. It should be noted that this explanation is just a speculation, and future research would benefit from delineating specific media content features (e.g., violent content, competition) that may contribute to such associations. Youth in the group of early-adolescence screen use also consumed more tobacco and cannabis at age 20 (the moderate-screen use group also showed greater cannabis use). As previously mentioned, review evidence has indicated that prolonged or/and repeated exposure to screen time during adolescence could lead to a reduction in the efficiency of the cognitive control system, as well as an inclination towards seeking short-term rewards (Marciano et al.,
2021), which may be further associated with risk-taking behaviors such as substance use. In terms of delinquency, both the moderate-screen use and early-adolescence screen use groups had higher levels of it at age 20; however, after controlling for delinquency at age 11 and with Bonferroni correction, only the moderate-screen use group still had statistically significantly higher levels of delinquency compared to low-screen use, but not compared to other groups. This finding is potentially due to the fact that the current sample was sourced from the community and youth reported low levels of delinquency across groups. Again, it is worthwhile to mention the other possibilities regarding the association between screen time and behavioral problems (i.e., aggression, delinquency, and substance abuse). For example, adolescents with more behavioral problems may be more likely to be addicted to screens and pay more attention to violent content, possibly due to, e.g., their lower levels of self-control. This could have the potential to create a vicious circle and should be examined in future research. Besides this, adolescence may be a vulnerable period for developing these problems (i.e., aggression, delinquency, and substance use), since a developmental mismatch may exist where the socio-emotional system underlying sensation seeking matures more rapidly than the cognitive control system underlying self-regulation based on dual systems theories (Murray et al.,
2021; Steinberg et al.,
2008). Longer screen time, aggression, delinquency, and substance use may co-occur in adolescence. Additional research modeling their co-changes would be beneficial to illuminate their links across adolescent development.
Some interesting findings are worth noting, concerning the moderate-screen use group that showed significantly less anxiety but more cannabis use and delinquency than the low-screen use group. The potential explanation needs to be examined in future studies, but these findings might imply that moderately prolonged screen time is not always related to negative outcomes; it varies depending on the specific outcome in question. Future research needs to account for the possibility that there might be both positive and negative aspects to screen use (e.g., social connection versus exposure to cybervictimization, displacement, etc.). It should be noted that the above discussion regarding the potential mechanisms underlying the association between screen time trajectories and mental health and behavioral outcomes was speculative, and this association might be reciprocal (despite the fact that the current findings controlled for previous levels of outcomes), which should be investigated in future research.
Implications of the Findings
The current study detected five trajectory subgroups, and they were associated with different risks for having negative mental health and behavioral issues in adulthood. These findings suggest the value of identifying the heterogeneity in adolescent longitudinal screen time and young adults’ outcomes and that aiming at screen time management to improve youth mental health and behavioral issues could be tailored to these sub-groups. It should be noted that the findings regarding the association between screen time trajectories and mental health and behavioral outcomes are correlational in nature, which provides limited information on the causality of these associations. The practical implications mentioned below would be more convincing if the causal directionality of these associations could be well established in future studies. The findings might suggest that screen time reduction intervention programs targeted at adolescents who spend an increasing amount of time on media screens, particularly chatting/surfing online and/or playing videogames, could help reduce the risk of developing mental health problems including depression, anxiety, suicidal ideation, and self-injury. Interventions aimed at youth who spend much time playing videogames in early adolescence could aid in reducing behavioral issues (e.g., aggression and substance use). However, as mentioned earlier, future research will be required to rule out reverse causality given that youth at risk of mental health issues may be more drawn to spending higher and/or increasing amounts of time engaged in screen use. Existing literature suggests several strategies that may be used to reduce screen time in adolescents, such as self-monitoring, behavioral contracts, screen time budgeting, and parental newsletters focusing on reducing screen time (e.g., Lubans et al.,
2016; Salmon et al.,
2011). Regardless of the direction of causality, the current findings point to the fact that screen use histories may flag the risk of a range of mental health and behavioral issues. Those displaying high or escalating levels may thus benefit from screening for mental health or behavioral issues.
Limitations and Future Directions
The limitations of this study should be noted. First, the assessment of screen time was based on self-reported measures, and the categorization of media variables was limited by the measures available. Future studies using objective (accelerometer-based) measures and with detailed categorization (e.g., separately assessing chatting and surfing the Internet) will provide a more accurate record of screen time and a more precise examination of the link between time spent on specific media activities and mental health and behavioral issues. Second, some important characteristics (e.g., suicide-related content) and purposes (e.g., learning or entertainment) of media screen activities were not assessed in the z-proso study and these characteristics could help to understand the outcomes associated with each type of screen time. Third, the current study only focused on a unidirectional pathway from screen time to mental health and behavioral issues, however, there might be a bidirectional association between these constructs (e.g., Yang et al.,
2022). Fourth, future studies employing designs, e.g., the experience sampling method, to repeatedly measure screen time in briefer intervals or to passively record media use will provide more detailed information on adolescent usage patterns and their association with consequences, whilst overcoming possible recall biases associated with traditional questionnaire methods. Fifth, gender/sex differences in the trajectory of screen time and its association with mental health and behavioral problems are not the focus of current research, future studies should examine possible sex/gender moderating effects using large sample sizes. Sixth, data collection for this study began about 14 years ago, during which time there have been significant changes in media consumption/use patterns, and new media activities need to be evaluated to contribute to the understanding of more recent youth screen usage and its association with mental health and behavioral problems. Finally, the current study was limited by focusing only on negative outcomes and youth time spent on screen media. Future research should evaluate positive outcomes (e.g., subjective well-being) and collect additional information such as youth time spent on other activities (e.g., physical activities) and the availability of other activities that might moderate the link between screen time and mental health and behavioral issues.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.