January 4, 2019

expert reaction to study on depression and social media

Research published in EClinicalMedicine highlights potential issues with greater social media use in relation to young people’s mental health.

Prof Peter Fonagy, Head of the Division of Psychology and Language Sciences, University College London, said:

“This study, like many others over recent years, highlights a strong association between screen-time and symptoms of depression and the careful analysis of potential mediators offers intriguing potential explanatory hypotheses. However, the study does not investigate activities which are known to be important in promoting wellbeing, but are constrained by absorption in social media: in particular, mutually enjoyable shared activity with parents. Thus, in addition to drawing attention to the impact of social media use on adolescent wellbeing, we should also look at its negative impact on factors that have the potential to promote wellbeing such as family activities.”

Prof Andrew Przybylski, Associate Professor and Director of Research at the Oxford Internet Institute, University of Oxford, said:

“In this study the researchers draw on data from the Millennial Cohort Study to investigate the correlations between self-reported social media time and a range of indicators of psychological experience and behaviour. The study has a number of notable strengths including the use of very high quality self-report data, good use of control variables, and a relatively circumspect tone for what is typically an over-hyped research area. With this understood the paper goes well beyond the data where it makes direct policy recommendations on the basis of correlational data.

“First and foremost, the data are entirely based on self-reporting. With large sample sizes we would expect most variables to modestly correlate with one another (as they do in this paper). The authors should be credited for noting this as a limitation (few do).  That said, results from similar studies such as those conducted by NHS Digital suggest that high levels of social media use might be understood as a consequence, not cause, of low psychological well-being. This much simpler explanation for the correlation, that social media use is a symptom, for the pattern of findings is not seriously considered by the authors.

“Second, a very large number of statistical tests are performed because of analytic decisions taken by the researchers. For example, splitting the sample by gender, looking at a host of outcome variables, and by comparing multiple levels of social media use instead of using a more conservative statistical approach. Because the analyses are exploratory this means a number of the findings might be false positives as the analyses did not correct for multiple tests. This means that many of the comparisons, such correlations between high vs. low social media use might be spurious. Some causal language is used by the authors in how they describe these comparisons – that is not backed up by the data here, and readers could get the wrong impression if they take from this that we know one thing causes the other (we don’t).”

“Finally, and most importantly, the correlations between social media use and indicators of low psychological well-being are indirect. Other factors the authors include such as low levels of sleep, low self-esteem, harassment, and low body image explain most of these relationships. Further, the correlations between psychological well-being and these factors are far stronger than the links between any of these and social media. Because social media is so weakly correlated with either these factors or well-being I’m left to wonder about the wisdom of making social media use the focus of policy. Given social media is by far the least important factor in the central model; why not address sleep, self-esteem, harassment, and body image directly?”

Dr Dennis Ougrin, Clinical Senior Lecturer at the Institute of Psychiatry, Psychology, and Neuroscience (IoPPN), King’s College London, and Consultant Child and Adolescent Psychiatrist at the South London and Maudsley NHS Foundation Trust, said:

“This is not the first paper showing an association between heavy social media use and poor mental health. Although this type of study cannot prove unequivocally that heavy social media use causes poor mental health, it is still of great importance as another piece of much needed evidence pointing in the same direction. Together with other studies, it highlights possible mechanisms of the association. Of these mechanisms, perhaps the most striking is the impact of social media use on sleep, which is incredibly important for good mental health. Another important contribution of this study is demonstrating that the association is stronger in girls. This finding complements recent surveys showing an increase in the prevalence of emotional disorders in girls.”

Professor Stephen Scott, Director of the National Academy for Parenting Research at the Institute of Psychiatry, Psychology and Neuroscience (IoPPN), King’s College London, and Head of the National Conduct Problems & National Adoption and Fostering Services at the South London and Maudsley NHS Foundation Trust, said:

“This cross-sectional study takes data from the Millennium Cohort Study (MCS) of over 10,000 14-year-olds. This study was well carried out with a very large sample, good measures and appropriate analyses, although controlling for parental education would have been helpful. After controlling for family income and other confounders including single-parent families, it still finds an association between increased use of social media and low mood. Some of the effect was due to online harassment, dissatisfaction with body shape, and worse sleep. These are important findings which need to be taken seriously.

“Inevitably there is the chicken and egg question, as to whether more dissatisfied children, who to begin with are less pleased with their body shape and have fewer friends then spend more time on social media. Nonetheless, it is likely that excessive use of social media (and one third of children in lone parent families or in the poorest income group reported using social media for 5 hours or more a day, a huge amount of time) does lead to poorer confidence and mental health.

“To establish the facts, a randomised controlled trial would be almost impossible to carry out, but it would be possible to look at individuals over time and see if there are spontaneous variations between their usage of social media and well-being. It would also be possible to carry out experiments where usage is cut down, but then constructive alternative use of time would need to be strongly supported.”

Professor Sir Simon Wessely, Regius Chair of Psychiatry, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, said:

“It is established that the rates of anxiety and depression have increased in recent years in young people. It is also established that this is particularly true for young women between 16 and 24.  The reasons for this are unclear, but many people assume that this must be due to increased access to social media. However, there have been two problems with this. First, why does it predominantly affect young women? Second, how can one be sure that this is cause and effect – in other words, it could be that depressed or anxious young people are more likely to use social media, rather than the other way round. This study takes us further forward in understanding the gender effect – which they find to be mediated by poor sleep, online harassment, body image issues and self-esteem.  But because it is largely cross sectional, they still cannot definitely say that social media usage causes poor mental health, although the evidence is starting to point in that direction.”

Dr Bob Patton, Lecturer in Clinical Psychology at the University of Surrey, said:

“The paper reports on the findings from a large scale cohort study of 14 years olds in the UK, exploring their use of social media and mental health. Results demonstrate a clear association between time spent on social media and symptoms of depression (it’s important to note that the study reports symptoms rather than a clinical diagnosis of depression), however as this was a cross sectional study it does not provide any evidence that social media use is the cause of the observed symptoms of depression. It was a well-designed study with almost 11000 participants, providing a representative sample of adolescents, and giving confidence on the reliability of the data and the findings.

“The associated press release presents a snapshot of the key findings of the study – that there is an association between social media use and depression, and that girls are more affected by this than boys. The study found that twice as many girls as boys were users of social media (of at least 3 hours per day), and for those using social media for more than 5 hours a day, girls reported twice as many depressive symptoms as boys (in the press release this is presented as the headline finding that “Girls are twice as likely to show signs of depressive symptoms linked to social media use compared to boys at age 14”). While this is correct overall, it does vary considerably depending on the total time engaged with social media – at lower levels of engagement, the difference between girls and boys is markedly lower.

“The press release also suggests that “At home, families may want to reflect on when and where it’s ok to be on social media and agree limits for time spent online. Curfews for use and the overnight removal of mobile devices from bedrooms might also be something to consider.” The evidence presented in the paper does support limiting the time spent on social media (not online as this can encompass a broader range of usage and this was not covered in the data collected). With regard to the “when and where”, this was not something that was reported in the study, and in fact the paucity of data on this is cited as a limitation of the study by the authors themselves in the discussion.

“Overall this is a useful study, adding to the evidence base in this area and supporting the findings of other studies that have linked social media usage and adolescent mental health.”

Prof Naomi Fineberg, Consultant Psychiatrist, Hertfordshire Partnership University NHS Foundation Trust, Visiting Professor, University of Hertfordshire and Chair of the COST Action group into Problematic Internet Usage, said:

“Studies such as this are important as they continue to indicate an association exists between the use of social media and mental wellbeing in young people. These kinds of cross sectional screening studies cannot by virtue of their design definitively tease out the link, such as attributing causation. Indeed, the relationship is likely to be complex and nuanced. The studies do however call out for investment in research designed to look in detail at the way young people use the internet and its consequences on wellbeing.”

* ‘Social media use and adolescent mental health: findings from the UK Millennium Cohort Study’ by Kelly et al. will be published in EClinicalMedicine at 00.01 UK time Friday 4 January 2019, which is also when the embargo will lift. 

All our previous output on this subject can be seen at this weblink: http://www.sciencemediacentre.org/tag/screen-tag/

Declared interests

Dr Andrew Przybylski: “I do not have any interests which might be regarded by a reasonable and objective third party as giving rise to a conflict with my role as an SMC expert in this story.”

Dr Dennis Ougrin: “No COI.”

Professor Sir Simon Wessely: On the board of trustees for the SMC. No other COI to declare.

Professor Stephen Scott: “No COI.”

Dr Bob Patton: “No COI.”

None others received.