Adolescent social media user types and their mental health

Watch this video abstract from Lizzy Winstone on her JCPP Advances paper ‘Adolescent social media user types and their mental health and well-being: Results from a longitudinal survey of 13–14-year-olds in the United Kingdom’.

Authors; Lizzy Winstone, Becky Mars, Claire M. A. Haworth, Jon Heron, Judi Kidger

First published: 10 March 2022

Lizzy Winstone
Lizzy Winstone

My mixed methods research focuses on how young people engage with their peers, family and the wider public through social media, and whether this is determined by, or has an impact on their mental health. I am exploring whether it is ‘healthier’ (in terms of self-harm, depression, anxiety and well-being) to use social media passively compared to using it for communicating with friends or to share content and express one’s identity. I am also focusing on how social media use more generally predicts how connected young people feel to their peers, family and school.

Before starting my PhD I worked for several years as a survey methodologist for the European Social Survey, a cross-sectional general attitudes survey conducted every two years in more than 20 countries. Based at City University London, I specialised in questionnaire design, (cross-national) equivalence, survey non-response and survey project management. I have also previously worked in local authority public health teams as a data analyst, researcher and strategist, specialising in Joint Strategic Needs Assessments of local populations. (Bio and Image from University of Bristol)



SPEAKER 1: Hello. We are the Association for Child and Adolescent Mental Health or ACAMH for short.


LIZZY WINSTONE: Hi. My name is Lizzy Winstone. I’m a senior research associate at the University of Bristol. And part of my PhD research, which was funded by NIHR School for Public Health Research, was recently published in GCPP Advances titled, Adolescent Social Media User Types and Their Mental Health and Well-being. So in this short video abstract, I’m going to give you an overview of the background, methods, results, and interpretation. Social media use is thought to be a novel driver of adolescent mental health. In the UK, almost 90% of teenagers have a social media profile by the age of 13, which is the age group that we’ve been studying. Whereas much of the previous focus of research and policy has focused on screen time, this research examines whether there are mental health implications of how young people use social media. So social media use includes lots of different activities, which we’ve grouped together as passive. So that’s browsing content, direct communication, so messaging, or broadcasting, posting photos, and other content. We used a two wave panel survey across 19 different secondary schools. And the class enumeration part of the model, which I’ll talk about in a minute, included the full available sample at each time point.

So in the first time point in October 2019, there were, with data available for 2,456 young people and a year later 1,820. And of these participants, between 1,311 to 1,321 had complete mental health data at baseline in addition to complete demographic and screen time covariate data. So the survey included 13 different social media activities, past year self-harm. And we included the hospital anxiety and depression scale and the Warwick and Edinburgh Mental Well-being Scale. So we looked at the conditional probabilities of a certain response to frequency of 13 different social media activities, as I said. So we then used plus to compare 2, 3, 4, and 5 plus model solutions both in terms of model fit statistics and also interpretability. So that class enumeration part is represented at the top of this model diagram and the latent class analysis.

And we then went on to test associations with mental health outcomes a year later, so self-harm depression, anxiety, and poor well-being. And then in fully adjusted models, we included demographic and screen time covariates as well as baseline mental health. So in latent class analysis, individuals are assigned a most likely class membership, which gives us an estimate of prevalence of each user type. So among 13-year-olds, we identified four main user types. The largest group we labeled high communicators. These were characterized by frequent direct communication and frequent passive use but moderate broadcasting to close friends or peers.

Around a third of young people were categorized as moderate communicators who engaged in moderate communication and passive use with minimal broadcasting. And then 13% were labeled broadcasters, and they engaged in very frequent activities across the board. And then a smaller group were labeled minimal users, who didn’t really use social media at all or engage very minimally in communication or passive use.

So as I said, we then tested associations between user type and mental health a year later when participants were age 14. And what’s nice about our suggested three step approach that we used in N Plus is the ability to adjust and account for uncertainty in class assignment. So we found differences between user types in the risk of self-harm, anxiety, depression, and poor well-being.

So broadly speaking broadcasters tended to have the poorest mental health outcomes and moderate communicators, sometimes high communicators, generally have the best outcomes a year later. What was interesting as well is that minimal users had the same or worse outcomes as moderate communicators, which suggests that some social media use with minimal broadcasting may represent the optimum level of social media use at this age.

So we situate our findings as new person-centered evidence in support of the digital Goldilocks hypothesis. So whereas previous research has identified optimal social media use in terms of an amount of time spent online, we reconceptualize that optimal use as moderate use but primarily for direct messaging and passive browsing of content. And then we further reconceptualize maladaptive use as excessive messaging, excessive browsing of content but also broadcasting of content or sharing content, well, publicly. We might think of this in terms of an accumulation of risks and benefits.

For example, whereas this minimal user group are not exposed to any of the risks associated with any social media use, they also don’t enjoy any of the benefits of social media use such as being connected with peers, information seeking, entertainment, and so on. And some of those benefits are likely experienced by the moderate communicators group without some of the risks or stresses attached to excessive levels of those activities.

And then in the broadcaster’s group at the bottom there, yes, they may be accumulating the benefits of social media use, but they’re also accumulating risks or stresses in terms of excessive levels of messaging, which may feel smothering or overwhelming to the sender and receiver. They may be experiencing risks attached to excessive levels of passive browsing such as feeling guilty about the amount of time they spend online if it’s not felt to be constructive, for example.

And then they’re also potentially exposed to risks attached to sharing content such as threats to their sense of self and sense of self-worth if an expression of their identity is being rejected by others. In conclusion, our findings emphasize that social media use should not be viewed as this homogeneous activity either in research or policy.

So we find that the majority of young people do not appear to be using social media in a way that’s harmful to their mental health. However, we identify a group of users, so that’s the broadcaster’s group, who may be particularly vulnerable to negative online experiences and perhaps warrant further focus in research or interventions. Thanks very much for listening.


SPEAKER 1: Find out more about becoming an ACAM member and to be part of the advancement of Child and Adolescent Mental Health. Visit

Add a comment

Your email address will not be published. Required fields are marked *