Media and developing minds header

Digital Media, Mental Health, and Relationships

October 16, 2018

Jennifer Stevens Aubrey, PhD


Jennifer Stevens Aubrey, PhD

Department of Communication, College of Social & Behavioral Sciences, University of Arizona

Linda Charmaraman, PhD

Linda Charmaraman, PhD

Research Scientist, Project Director of the Youth, Media, and Wellbeing Research Lab, Wellesley Centers for Women, Wellesley College

Larry Rosen, PhD

Larry Rosen, PhD

Professor Emeritus and Past Chair of Psychology, California State University, Dominguez Hills; Advisory Board Member, Children and Screens

Emily Weinstein, EdD

Emily Weinstein, EdD

Postdoctoral Fellow at Project Zero, Harvard Graduate School of Education


Dr. Rosen: For the purposes of this conference, the broad topic of social responsibility encompasses three vital questions.  First, what are kids doing with technology?  Second, why are they doing it?  Third, how can we fix the problems that we discover when we answer the first two questions?  The answers that follow reflect five years of data, gathered cross-sectionally from similar samples and evaluated using a constant measurement tool.  

In 2016, 2017, and 2018, roughly 200 older college students (average age about 25) allowed one semester of limited, real-time monitoring of their smartphone use.  Most of the study subjects had jobs and families. Apps installed on their phones measured how often they unlocked their phones and how long those phones stayed unlocked.  An additional study cohort of 36 high school students was added in 2018.

The 2016 cohort of college students turned on their phones an average of 56 times a day (every 17 waking minutes, if they slept for 8 hours).  In 2017, that decreased to 50 times (every 19 waking minutes). It increased dramatically in 2018, to 71 times (every 13.5 waking minutes). The 2018 high school study showed 73 “turn-ons” (every 13.2 waking minutes).  In the college study, total daily minutes of usage and minutes spent on the device per use surged from 2016 to 2017 (roughly coinciding with the increasing popularity of Snapchat and Instagram), but declined in 2018. The 2018 high school cohort spent, on average, ten minutes more on their smartphones per day than the college students, and fractionally more time during each “unlock”.  (Both groups of 2018 study subjects averaged 270-280 minutes of usage per day and roughly 3.7 minutes per use.)

Further self-report data reveal a number of broad behavioral changes during five years of research.  Email use is down. Adolescents have fewer friends on Facebook, and they're spending less time there.  Smartphone and general social media use as well as text messaging are up. So did nomophobia (the anxiety disorder often referred to as “fear of missing out” or “FOMO”).  

In the 2018 study high school students who spent more time on their smartphones had lower GPAs, and experienced more anxiety when separated from their phones.  They also showed more boredom and inattention in class, and were less likely to have study strategies. The more frequently the phone was unlocked, the higher the likelihood of using communication apps (and in particular, social media).

Common Sense Media survey data suggest other changes in teens’ digital media behavior from 2012 through 2018.  In 2012, almost one-half of 13-17-year olds surveyed favored face-to face communications with friends. In 2018, that number had declined to less than one-third, while two-thirds preferred to communicate with their friends electronically.  Over the same period, the percentage in that age range who use social media more than once a day has risen from 34% to 70%.

One way to think about why these behaviors are changing is to look at them in terms of independent variables, mediators, and dependent variables.  Three particular independent variables seem to be technology behavioral drivers: executive function, boredom, and anxiety. Mediators include daily smartphone use, daily social media use, multitasking preference, attention/focus while studying, and digital metacognition (that is, knowing what to do with your phone in different situations).  Current research is investigating the associations among these various independent variables, mediators, and such dependent variables as sleep and academic performance. For example, executive functioning problems and technological anxiety lead directly to poor course performance, but both also lead to poor digital metacognition (which in turn leads to poor course performance).  Boredom also leads to digital metacognition (and thus, to poor course performance) and boredom and technological anxiety leads to more social media use which also leads to poor academic performance.

These findings suggest ways to address the problems that arise from digital media.  First, reduce access to devices during the course of the day. Second, encourage the formulation of metacognition strategies, so that people can manage their digital device use systematically.  Third, reduce technological anxiety and fear of missing out generally and about phone access and social media use in particular. Finally, reduce boredom and increase attention spans.

Dr. Charmaraman:  Despite the federally mandated age limit under COPPA (that is, 13), a 2010 Pew report found that 38% of 12-year-olds in the US used an online social network.  A 2011 survey of 25 European countries found that 38% of 9-12-year-olds already had a social media profile. These data raise questions about these children's safety, what motivates their social media behaviors, and the need for adult oversight.  Compounding these concerns are early adolescents’ limited capacity to self-regulate, their susceptibility to peer pressure, and their tendency to seek autonomy from family networks.

Dr. Charmaraman and Dr. Megan Moreno are investigating these issues.  While the study employs multiple investigational methods and informants (including parents, students, and school staff), this talk concerns only the student survey.  It involves cross-sectional research on 700 children in grades 6 through 8 (average age = 12.5 years), drawn from three schools ranging from a privileged suburban, predominantly white school to a predominantly non-white inner-city school.  (For most of this talk, Dr. Charmaraman did not distinguish among findings for students with different socioeconomic status, school performance, race, gender, or other attributes. She did present social obligation data controlled for age, gender, race, and mother’s education.)

Sixth and seventh graders demonstrated substantially similar online behaviors, but things changed in 8th grade.  In terms of most read online topics, 6th and 7th graders are most interested in animals, science/technology, hobbies, and books/authors.  The top four for 8th graders are TV shows, online games, relationships, and depression.

In contrast to the 2010 Pew finding (that 38% of US 12-year-olds used online social networks), 68% of the 12-year-olds in this study already have social media profiles.  This may be a function of the downward creep in the age at which children are getting their first phones. (Almost three-quarters of the children in this study got their first phone at or before age 11, and 42% signed up for their first social media site at or before that age.)  Comparing first social media use at 11 and under with 12 and over, earlier use of social media experience is associated with greater online social anxiety and having online friends of whom parents would disapprove. Comparing the wider extremes (10 and under versus 13 and over first-time-social media users) shows not only greater online social anxiety, but also fewer hours of sleep and an earlier start to their dating lives.  (The precocious dating may be a function of greater access, but that’s just speculation for now.)

This study also indicates a marked increase in negative online experiences including exposure to rude or mean comments, being the subject of rumors, and being sent unwanted sexual messages or images (Eight percent of 6th graders, 16% of 7th graders, and 27% of 8th graders reported receiving such unwanted sexual content).  On a positive note, students also are giving (60%) and receiving (54%) social and emotional support online, on topics including school, anxieties, dating, and family.

Beginning to use social media earlier is associated with greater likelihood of some psychosocial and physical health consequences, but fewer than this study team expected to find.  For example, there was no association with depressive symptoms, social isolation, difficulty making friends, physical activity, or problematic internet behaviors. In contrast, early adolescents’ social and psychological attributes are associated with the character of their online relationships (including their sense of obligation with respect to those relationships, and their likelihood of receiving unwanted sexual content).  Online social anxiety seems to be the most influential factor in this regard.

Follow-up parent and adolescent studies will further investigate parent-child match, secretive online behaviors, body dissatisfaction, celebrity obsession, and civic engagement.  Planned methodological improvements include merging data gathered from students and parents, and adding online Instagram analysis. In addition, the National Institute of Child Health and Development is now funding a three-year longitudinal study of the effects of early social media in youth on behavioral and socioemotional health.  Recruitment of middle school subjects for that study will begin in Spring 2019.

Dr. Weinstein: A mixed-methods study looked at how adolescents in public school think about the influence of social media on their everyday emotional and social lives.  That research began with surveys of 568 subjects, with follow-up interviews of students constituting a “purposeful sample” spanning the happiest to least happy social media users in the school.  Those interviews elicited comments about the feelings youth experience related to their uses of social apps. The results revealed that all students perceived both positive and negative influences of social media.  

To effectively analyze those stories, it was helpful to differentiate among different dimensions of that use: Self-Expression, Relational Interactions, Exploration, Social Browsing, and Content Browsing.  Each of these dimensions has positive and negative potential. For example, relational interactions can contribute to feelings of closeness and connection and/or to feelings of disconnection and isolation. Self-expression can confer a sense of acceptance and affirmation and/or lead to concerns about judgment and stress over peer feedback. In this study, the vast majority of teens articulated both influential positives and negatives comprising their everyday interactions with social media. This has three important implications.

First, these findings challenge the old question of whether social media are good or bad for kids, in favor of a more nuanced inquiry that prioritizes the combination of ways in which social media are helping and/or hurting adolescents with respect to their social and emotional lives. It is possible to ask, with respect to each principal dimension of any child’s social media narrative, what is contributing to specific positive and/or negative influences? And what effects is that particular child therefore experiencing?

Second, at least in this small sample, the most salient positives stemmed from relational interactions with friends, peers, and physically remote family members.  The most salient negatives related to self-expression, particularly related to concerns about peer judgement and rejection, and about the unknown future consequences of posted content that may persist.

Third, adolescents have varied constellations of social media experiences.  The weight among the dimensions of those experiences shifts over time. So, too, does the balance between positive and negative components within each dimension. This paradigm conjures the image of a see-saw (seen in the work on wellbeing by Rachel Dodge and her colleagues). In thinking about social media experiences as a see-saw, digital media habits ostensibly form the base.

This social media see-saw framework has led to the development of a simple personal assessment, which is now being field-tested in classrooms from middle school to college.  The tool facilitates young social media users’ critical reflection on their experiences and supports their ability to identify and design their own personalized interventions.

According to recent work by Vicky Rideout and Susannah Fox, young people with depressive symptoms reported heightened positive and negative experiences with social media. Ongoing exploratory work by Dr. Weinstein and her collaborators investigates the social media experiences of 30 suicidal adolescent inpatients and also appears to point toward this amplification effect. At-risk youth can find opportunities for (for example) meaningfully supportive online relationships, but they can also encounter heightened challenges in that arena based on the composition of their social networks.

The big picture of all of this is to embrace the complexity of young people’s online digital experiences.  Recognize the positive and negative potentials, even among those who seem to be doing worst or best.  This will make possible more individualized understanding, more targeted supports, and more effective interventions.

Questions and Answers

Audience Question: Research on compliance with emergency alerts is producing surprising data.  Study subjects are notified that they must refrain from mobile phone use in order to keep communication services available for emergency workers.  Compliance is highly, inversely correlated with age. Young adults (18-24) just keep using their phones. Why?

Dr. Rosen: This illustrates young adults’ inability to stay off their phones.  “…[I]t’s oxygen. It’s their air, and they just don’t think about not using a smartphone. It’s part of their life.”

Dr. Weinstein: This also may reflect that age group’s desire for more information, through the portal through which they expect the best real-time updates.  The emergency may actually heighten their desire for constant, instantaneous access.

Dr. Charmaraman: Middle school students aren’t allowed to have their phones on their person, because of their distracting effect.


Audience Question: College students criticize discussions of dating because they consider it an archaic concept.  They hook up, and some have no experience of dating. How is this reflected in Dr. Charmaranan’s research?

Dr. Charmaraman: Taking into account that the study involved children 11-14, the investigators used the term “going out with somebody."


Audience Question: Is it useful to lump all the social media platforms together?  The first premise for this question is that social media interactions are taking place in artificial environments created by third-parties and manipulated to maximize engagement.  The second premise is that engagement is optimized by content that produces negative, rather than positive, effects. So, if social media platforms are designed to make their users unhappy, can there be a net positive social media use strategy, especially for teenagers (who are less in control of their own behavior than adults already)?

Dr. Rosen: These services may not be designed to make users unhappy, but rather, to hold their attention.  The effect on users’ happiness is an independent issue. Regarding the differences among social media sites, our lab’s research looked into the relevance of those differences among the top 11 sites.  A preliminary analysis found that their average student participants had an active presence (checking in at least once a month) on six of the 11. Broadly speaking, young people who check in on one social media site checked in on all of them.  As for user effects, one study found that young adults with more social media friends had less or fewer symptoms of dysthymia, but those who had more social media friends and spent more time online had more major depression symptoms. Because there’s evidence on both sides, it’s hard to lump them together and say they’re all basically harmful.

Dr. Weinstein: In practice, trying to associate interview subjects’ specific positive and negative feelings with specific platforms is difficult.  They tend to toggle back and forth among them, with some exceptions. This highlights the importance of targeting the metacognitive piece so that they can make such distinctions themselves (and modify their behaviors accordingly).


Audience Question: Is Dr. Charmaraman looking at the families and parenting practices in cases of very early social media use?

Dr. Chararaman: Yes.  Conversations with parents of about 25 middle school students revealed an extremely wide range of parental attitudes.  Some felt that they had to constantly monitor their children’s online activities in real time. Others considered social media to be a child’s private space for self-expression.  Conversations with parents and children prior to the onset of social media activity showed how individualized the decision-making can be. Parents recognized that a child’s developmental stage was as relevant, or even more relevant than chronological age (even among siblings).


Audience Question: Is there really a causal factor here, or is the social media platform just a magnifier?  Are there latent attributes, perhaps identifiable at earlier ages, that the social media experience blows up (but does not cause)?

Dr. Chararaman: Cross-sectional research can only identify associations.  Answering the question about bi-directional effects is among the objectives of the upcoming longitudinal study.

Dr. Rosen: It's not necessarily a causal effect, but it interacts with our cognitive variables, our affective variables, and then our boredom variable, which is sort of a combination of both. I don't think we're anywhere close to getting causal effects, I think we're getting predictive effects.  For now, creating predictive models is a priority. It allows students to be warned about the likely consequences of poor metacognition – the practical idea that students who use their phones in class are likely to perform less well than those who don’t.

Dr. Weinstein: Regarding Larry Rosen’s point on the usefulness predictive models - this relates to the broader subject of differential susceptibility.  Warning all students about the consequences of poor metacognition won’t stop all of them from using their phones in class, but it may stop some of them. It may be a partially effective intervention.

Session Materials