How digital technologies affect adolescent psychological well-being and mental health – Dr. Amy Orben

Matt Kempen
Marketing Manager for ACAMH

Posted on

Dr. Amy Orben, College Research Fellow, Emmanuel College, University of Cambridge, delivers a talk on ‘How digital technologies affect adolescent psychological well-being and mental health’.

This lecture was recorded on 6 November 2020 at the Oxford and Reading ACAMH Branch event ‘Enhancing young people’s mental wellbeing through digital technology’.

ACAMH members can now receive a CPD certificate for watching this recorded lecture. Simply email membership@acamh.org with the day and time you watch it, so we can check the analytics, and we’ll email you your certificate.

Dr. Amy Orben
Dr. Amy Orben

Amy’s research uses large-scale data to examine how digital technologies affect adolescent psychological well-being and mental health. She uses innovative and rigorous statistical methodology to shed new light on pressing questions debated in policy, parenting and mental health. She also campaigns for better communication of trends in data and the wider adoption of Open Science. Amy is a College Research Fellow at Emmanuel College, University of Cambridge, and a Research Fellow at the MRC Cognition and Brain Sciences Unit, University of Cambridge. She completed an MA in Natural Sciences at the University of Cambridge before joining the University of Oxford to obtain her DPhil in Experimental Psychology, for which she was award the BPS Award for Outstanding Doctoral Research 2019 and the Understanding Society Young Researcher Prize 2019. During her time at Oxford she was also nominated for the ‘Best Tutor’ award at the Oxford Student Union’s Annual Teaching Awards.

Transcript

Good morning, everyone. Naturally, I would have loved to be there in person. I actually lived in Reading for a while, just a couple of years ago, and I really loved it, so it would have been nice to visit. But now we’re on Zoom, which I think we’re all getting used to as well. So I don’t think I need to introduce myself a lot more. Maybe something to say at the start is that my work is very quantitative, and not focused on clinical practice. So it’s more basic, than some things that we might be hearing later on. But I guess that makes it a nice one to start. So I hope to give you a broad introduction in some of the frontiers that there’s currently debate on around digital technologies, psychological well-being, mental health, especially in adolescence.

So what this talk is today is really much a mixed bag. So I’m going to try to really take you on a little journey of a lot of different parts of my work and other people’s work. So I’ll start with trying to put this debate or concerns about digital technologies and screens and social media that we have into perspective. I think we often forget that this conversation is part of a much larger conversation that’s been had for many, many decades about new technologies and children and teenagers. Then I’ll move on to really… Ooh, I had a typo in here, but I’ll overview the screentime debate, and really try to figure out what do we actually know and what do we not know yet on screen time and social media and mental health. And lastly, I’ll talk a bit about the current work, Covid-related work that has been going on in my group.

So I’ll start with the perspective taking. Nice thing to do with a coffee in the morning. So… and I wanted to start with this paper which first got me interested in taking a larger perspective. And this is Mary Preston’s article in the Journal of Pediatrics, published in 1941. Mary Preston was a pediatrician, child psychologist working out of San Francisco in the University of Stanford. And she was studying, I think, about 112 children to see how they were reacting to both movies and radio dramas. What Mary Preston found was, for her, quite concerning. She found that these terrifying scenes that were consumed, the radio or going to the movies, had an inhibitory effect on the functioning of every organ in the body. She found that the majority of the children seemed to be severely addicted to these forms these new technologies, and that most of them utilised this addiction as an escape from reality, much as an alcoholic does drink.

Naturally, I don’t research the radio; I don’t even research movies. But this struck me that a lot of what was being written about, for example, in the 1940s about the radio, seems to be very similar to the debates we’re having today. Does social media cause major changes both physically and psychologically? Does screen time cause addiction? And why do people consume so much and seem to be so hung to their new technologies, especially our children and young people? I think to understand why these conversations are so similar, we need to actually take a step back and think about what times were the 1940s versus the 2020s or the 2000s?

And in the 1940s, you can see on this graph, that household radio is really experiencing a boom, naturally a bit more protracted boom than the boom of smartphones, but the rate of having a household radio in the US increased from about 40 percent in 1930 to about 80 percent in the 1940s. And so this not only changed the way adults socialized and adults spent their time, it crucially also changed the way that children spend their time and young people. And they found that once the 1940s really were progressing children spent about one to three hours a day listening to the radio. So leisure time was starting to be taken up by this new technology that was reaching a larger proportion of the population by the month and concerns started to emerge.

So I just have a bit of a quiz, and actually, this is a bit harder online, so you don’t have to shout out your answer; I’ll just tell you. But I took out some quotes from parenting magazines or websites, either talking about the radio in the 1940s or talking about smartphones in the last couple of years. And just think about, do you think this is about the radio or about screentime and smartphones? No locks will keep this intruder out, nor can parents shift their children away from it. So if I give this talk in person, most people do get it right. This is actually about the radio; the ‘nor’ might be something that we don’t actually say anymore as much in written, popular language now. But this was from the director of the Child Study Association of America actually, in the 1935.

Have another quote: here’s a device that is everywhere. We may question the quality of its offering for our children; we may approve or deplore its entertainments and enchantments, but we are powerless to shut it out. It comes into our very homes and captures our children before our very eyes. This is normally one of the 50/50 ones. And it’s actually again about the radio, now from a parenting magazine, in 1939, but I feel like you could change this to a lot of the technologies we’re talking about today. And it would really resonate; you wouldn’t really pick it up as something that is very odd in the current day. And we see things like these quotes now: giving your child a smartphone is like giving them a gram of cocaine.

So what this is, naturally, there’s a lot more work in this field trying to figure out what’s different now, and what’s the same from over all these decades. But what I wanted to really start the day with is that these technological… we can call them panic; we can call them concern; we can call them reaction to change… but that our concerns about new technologies seem to be inherently cyclical. And that these cycles of concern really do drive our conversations, and it’s important to try to figure out what is part of the cycle about the conversations we’re having today. And what is something new that we really should focus our attention on. And I think the internet and screen time now is changing; it’s not like the radio or the television. But it’s important for us to really take that into account.

Again, I just wanted to say that the idea of this being cyclical is again not mine to make. In 1935 we have this quote that looking backward, the radio appears as the latest of cultural emergence to invade the putative privacy of the home. Each such invasion finds the parents unprepared, frightened, resentful, and helpless. Within comparatively short time, the movie, the automobile, the telephone, and the cheap paperback book have had similar effects upon the apprehension and solicitude of parents. And we can continue this to the radio, the television show, the video game, the smartphone, social media. So I think this really shows this is an important topic to have a conversation about. We really need to get better at figuring out what around technology should we focus on and be concerned about and target. And what seems to just be concerns that reincarnate that seem to not be that important in the long run.

So, and this is partly the work that I and colleagues have been doing over the last couple of years, is trying to separate what about the screen time that they might actually… should lead to interventions, should lead to policy change, should lead to us being concerned and targeting that, and what might be an overreaction or a spin that is put on things that shouldn’t be that concerning. And I hope that’s… Well, that’s the next part of my talk, I’ll give you an overview about, so we’ll get coming to the next different type of talk.

And what we know about the screentime debate now is that things are really complicated. Screen time is inherently diverse. It can range from Skyping your grandmother to looking at suicidal images on a certain social media platform to looking at a recipe on YouTube and learning how to do a new skill. And it spans a huge diversity of different users. If you work with young people, you know how different they all are, and how oftentimes by averaging and talking about screentime as one thing and talking about their effects on young people or adolescents or children, as one thing, we’re missing that nuance; we’re not appreciating that complexity. And so this part of the talk will try to show you why this appreciation of complexity is so important in these discussions.

So this is where my work started. It was in 2016, 2017, and there was a big wave of concern about smartphones. We had a big article down at the bottom right, written in The Atlantic called Have Smartphones Destroyed a Generation? which actually became the most read science article in The Atlantic in the whole of 2017. And I just pasted a paragraph from that article in the top left, where the author Jean Twenge said that, ‘Psychologically…’ the current generation of teenagers ‘are more vulnerable than Millennials were: Rates of teen depression and suicide have skyrocketed since 2011. It’s not an exaggeration to describe I-Gen…’ and that’s the current generation of teens ‘as being on the brink of the worst mental health crisis in decades. Much of this deterioration can be traced to their phones’. This is a very causal statement that some of the crises we might be still seeing now is being traced to smartphones. Jeremy Hunt, who is the current at that time health secretary, said that social media might pose as great a threat to children as obesity, and there were a lot of parliamentary investigations that started being made into this impact of screen use and social media on young people’s mental health. I think we’re still in part of this debate now.

So naturally, we’re now three years on, what do we know about the link between screen time and mental health and adolescents? Well, firstly, if we just look at correlations, which a lot of this initial work was looking at, it was looking at how does the screen use or the digital technology use of a teenager relate cross-sectionally as a correlation with their well being or their mental health. So this averages across a lot of different types of teenagers with a lot of different types of technology use. And that correlation… and I’m very confident in this result… is we often find that correlation to be negative. So if you give me a large-scale data set of British teens, and we look at how much screens they use and how good or bad they feel, there will probably be a negative relation between the two.

This has been described in narrative reviews, both by myself at the beginning of this year, but crucially also by great researchers, like Candice Odgers, and Michaeline Jensen, who wrote a big annual research review on adolescent mental health, also published in January 2020, which I would really recommend as a starting point. And we all agree that this negative correlation seems to be pretty consistent. And actually, it doesn’t always come up, but averaging across all the different studies, there seems to be that correlation present quite a lot. This correlation… and naturally we’ll get into what we can even interpret out as a correlation later on, but let’s just take the scepticism about correlations aside, and think about that correlation a bit more because another hallmark of the correlation and the negative correlation between screentime and adolescent well being and mental health is that it’s extremely small. In statistical terms, VR is often around 0.1, which is very small if you think about effect sizes.

A way that me and my colleagues have tried to communicate how small this correlation is, is by correlating not just technology use with well being and adolescence, but also other things with well being and adolescence, so this is part of a bigger project. But we correlated four other variables with well being, which we a priori thought should have very little relation with how good or bad an adolescent feels. That is bicycle use, so how often they use their bicycle, their height, whether they’re left or right handed, and if they are wearing glasses. So this is not the best of all graphs, but you see that the correlation between technology use and well-being is negative; it’s here at about negative 0.04. And the correlation between handedness and well-being is very, very, very small. It’s a lot smaller than the one between technologies and well-being, and it’s around zero. Bicycle use and how much you use your bicycle and how tall you are, are actually positively related to well being, a bit more than technology use is negatively related. But wearing glasses is more negatively related to well being than technology use. So naturally, the causal relationships that underlie these correlations are unknown, at least in this study, but it does open up questions around if we’re investing so much time and effort into debating whether we should invest millions of pounds into creating interventions to decrease technology use because of this consistent negative correlation that is very, very small, why aren’t we doing that for glasses, for example?

So it shows us these effect sizes are how my colleague, Andrew Shibilski* often says in teeny, tiny land. And we really need to be careful about how much we interpret into them. And that’s because a correlation is naturally not causal. And it can have a lot of different reasons for existing in datasets, so it could be naturally that increasing your screen time leads to lower well being, which is often the interpretation that we lean to when we see this sort of data. But it could also be the other way around; it could be that adolescents that start feeling worse start using more technologies, and that therefore this correlation starts existing. Or there’s another very real possibility that there’s some sort of third factors that are driving both increases in screen time, for example, and decreases in well being.

One of these that Candice Odgers highlights in her work in the United States is that a lot of lower socioeconomic status families, use more screens. That could be because there’s less afterschool activities available, that the parents have to work longer hours, and then, therefore, that screen time increases in those families. And we also know that families on that have lower socioeconomic status that the adolescents often score worse on well being questionnaires. And so third factors like these could be driving that correlation to exist. So to really understand more than just this correlation, we have to move into longitudinal work and research, which is something that I’ve been focusing on now for the last two and a bit years. The work that’s already published shows quite clearly that we’re probably dealing with a bidirectional relationship. So it’s not just that screen time changes well-being, it’s also the other way around, that well-being might change screen time.

The paper specifically that I worked on looking at this looked specifically at social media, so I’ll just use that to illustrate the point. So this was Working with Understanding Society. So some of you might have already worked with that base that we’re seeing work on it, about 13,000 adolescents between the ages of ten and 15. And crucially, what you can do in longitudinal datasets, which correlational data sets cannot do, is that you can dissociate between person and within person effects. This might look a bit daunting, but it’s actually quite easily explained with a nice example. So let’s just imagine that we’re teachers in a computer IT class in a school. And I remember doing this, that you learn how to type as a child, and you have these typing tests, and the typing test looks at both typing speed and typing accuracy. And so when you look at all the children in your class in this correlational way, and the between person where you’re comparing them, what you probably see is that those who have higher typing speed also have higher typing accuracy. And that’s because they’re probably just better typers. So you would find a positive correlation between speed and accuracy by comparing across all these children in your class. Naturally, that doesn’t tell us anything about the relationship within one child with the question of whether if a child increases their typing speed, how does that affect their accuracy. Probably that relationship isn’t positive. What we would expect is that if you increase your typing speed, if you make a child type faster and faster and faster, their accuracy is probably going to drop. So the within person correlate… the within person relation, sorry… will probably be negative. So this shows that actually the between person comparisons, that are often like the correlations we look at, might show very different results to the within person relations that we can figure out with longitudinal data.

This is what I’ll be presenting now. So using longitudinal models, like a random intercept cross-lagged panel model, we could figure out these within person relations of children in the Understanding Society dataset. And so what we found is that if a child increased their social media use from their own average value across all the years of data we had… so if they increased their use in a specific year, we do predict a small drop in their life satisfaction the year later, so there’s a small negative relation there. An increase from your own average of social media use, predicts one year later, a small drop in your life satisfaction. But the same happens the other way around, so if your life satisfaction drops from your expected value in one year, that does predict an increase in social media use one year later.  And so this shows that these relationships between technology and outcomes, like life satisfaction or well-being or mental health, are probably extremely complicated. There’s definitely no one-way street of the technology affecting us. The way we feel also affects the way we use technologies. And third factors are probably complicating that even much more than what we’re looking at now.

That comes to what I’ve foreshadowed at the beginning of our talk, is that we really need to think about the individuals. So until now, we’ve been averaging across all types of technology use and all types of young people and children and teenagers, but to really understand how social media affects us, we need to actually think about the individual. And the metaphor I often use to explain this is talking about a bar of chocolate. So if we were all researchers, and we had somebody come up to us, and we’re dietician, researchers, and we had a parent come up and say, I want you to tell me how my child will react to eating a chocolate bar, we would probably say we actually don’t have enough information to give you any advice because if your child is a diabetic, then eating that chocolate bar might be extremely harmful; it could even lead on being deadly on certain levels. However, if your child is an athlete, and they’ve just come off the football pitch at halftime, then actually eating that chocolate bar might be really beneficial because it replenishes their energy level. As clinicians, we might also be interested in what the motivations of the child is to eating the chocolate bar. If the child only eats chocolate because they’re lonely or sad, we might react very differently to the parent’s question than if they eat it because they’re hungry. We might want to know the frequency, do they eat chocolate… a chocolate bar every minute? Probably react differently than if they eat it once a week, or on their birthday. So all this information, all this nuance, would be taken into account before we would say we could probably make a prediction.

And this should probably be the same when we talk about screen use, or even more detailed, something like social media use. We would need to think about the type of user. For example, in the chocolate example, are we talking about diabetics? Are we talking about… What sort of age group are we talking about? And often we forget about this in the screentime debate, but preliminary evidence shows that this might be actually really important and this is the area where I’m currently focusing most of my research efforts to extend our knowledge of these different individual differences.

So, for example, this is data from this longitudinal study of life satisfaction predicting social media use one year later. And I’m not going to go through this in detail, but what we…  This is the zero axis, and if the dots are falling on the left, it means that a drop in life satisfaction predicts an increase in social media use. And so we’re seeing this negative relation for girls, so this is just a visualisation for girls. But now if we switch and we start looking at the relation for boys, you see that these longitudinal effects really decrease and become non-significant. So again, this is for girls, we’re finding quite robust directional effects there, but for boys, we see almost nothing. That’s the same when we look at it the other way around, so from social media use predicting changes in life satisfaction. So this is for girls, and then this is for boys. So for boys, again, you see a much smaller link than for girls, who show a much more negative link between social media use, so increase in social media use in one year predicting decreases in life satisfaction.

What I will hint at here is that this is just data from ten to 15 year olds. It seems like maybe teenage boys show different relations at different age groups. So by only focusing on this very young adolescent age of ten to 15, and not maybe 16, 17, 18, 19, 20, we might be getting the wrong picture. But what this work crucially shows is that individual differences really matter. In a really crude way here, we’re just differentiating for gender, but if we actually start looking into these differences in more and more detail, we find that they’ve become quite important in actually predicting what sort of effects we find.

So naturally, the way we… if we were public health experts or if we would need to make decisions about policies, these individual differences really matter. So if we go back to that, across the board, we have a small negative correlation, if every person had this small negative correlation between technology use and well being, and every young person just had this very small negative correlation, we might actually, as a public health expert say, it’s not worth investing this money to address this because maybe glasses show similar correlation, for example. However, it could be that the distribution is much wider, and once we average it, we still get them the same negative correlation, the mean is still the same, but the distribution is a lot wider, so there are now children and young people who are very negatively affected, just like there are children and young people who are very positively affected, and there probably our response would be quite different.

It would be even more different if there were bimodal distribution, so that there are quite a separate group of children or young people who show these very negative links while there’s another group that might show no link or show a positive link. There’s the diabetic model; you might have a way of treating the general population to deal with sugar and maybe sugar overconsumption, but you have a very different way of treating diabetics because they’re a very different group. And so we really do need to understand how different populations vary to actually even understand how we should react to this challenge as a society or as a public health perspective. And actually, what we also need to differentiate is the different types of use. As I said at the beginning, chatting with your grandmother, probably very different to looking at certain types of quite harmful content online.

Now, I’ll just… There’s a lot of work in this area. It’s a bit limited, because we often don’t have access to actual social media platforms, but theoretical work by Verjune* and colleagues, for example, has been thinking about different ways that children and young people could use social media, and they say that, for example, social media could be used very actively; you could chat to people, post things, be inherently connecting with people in a very active way on your social media app, for example, and that can increase social connectedness leading to increases in well being. But social media use can actually also be very passive, the scrolling through a news feeds, the just consumption of other people’s information without any sharing or communication. And so this doesn’t increase connection; it actually increases social comparison. And that might increase envy or low self esteem, and that then actually decreases well-being.

So time spent on social media or screens might be a really bad way to measure their effects because it’s actually how people use them that is really important. But oftentimes, we do still talk about screen time or time spent on social media, and a lot of advice once… There’s a lot of pressure to give time-based advice: children and young people should be online for X amount of hours and minutes every day. But I do have this slide saying that screen time does suck; it’s not really a thing. As we’ve seen, different technology use has very different impacts. We don’t measure it well; we often make people estimate how much time they spent on technologies, and teenagers estimate, and they can’t actually estimate it very well. And I’ve become increasingly sure that actually, always focusing on time spent on technologies is misleading our efforts because it does so much rely on what we’re actually doing, not how much time we’re spending, but that could be a whole separate talk in itself.

So let’s just summarise. What I found… not I… what the field has found over many years now is that the link between screen time… and I’ve sometimes made this into social media use in the last few slides as well… but that the use of modern digital technologies and its link to adolescent well being as a correlation is negative and small. But it’s probably much more complicated; it’s probably bi directional; there might be third factors involved; it’s probably a really highly individual link, depending on both the young person but also on the type of use. And so the last point is crucially that it’s lacking in evidence. We actually don’t know enough yet to give concrete recommendations, to really understand what sort of policy interventions we should design to ensure more positive uses in future.

For example, the chief medical officers, before they were dealing with a pandemic, actually spent quite a lot of time looking into screentime and social media and its impact on child health, and actually found that there was too little evidence to actually give any concrete recommendations. The Royal College of Pediatrics and Child Health has done some really great work also trying to figure out how to give proper guidelines to, for example, parents and clinicians. And they also found that there’s not a way to say that there’s a safe amount of screentime or an unhealthy amount of screen time, it’s really down to what is being used, how the lives of the child are maybe designed around screens or are the screens just helping benefit their lives in certain ways. And if you’re interested in guidelines around screens, this is probably one of the reports to have a look at. It’s very well written and balanced.

Okay, so in the last… So yes, this is now all the evidence for that. This link is actually really complicated. In the last two or three minutes, I just wanted to talk a bit about current work. This might be a bit shorter, but a lot of my work now in Cambridge with researchers like Sarah-Jayne Blakemore and our group had been around lockdown and screen use, and I thought we should just have one or two slides here about this because of the current situation. And actually, for young people and adolescents, lockdown is not social isolation. There is social communication and contact in the home, but there’s less contact often to their peers. And that might actually… adolescence might be a really crucial time where that lack of peer contact might be really impacting how they live their lives or their well being, mental health, and we’re getting more and more understanding about how young people are seen to be struggling more than other age groups in the last couple of months.

This is a pixilated picture from the Peer Research Centre, which actually asked adolescents what they think the positive impacts of social media are and what they think the negative impacts of social media are. And as you can see, 80 percent actually said that it allows them to feel more connected with their peer groups, with their friends, and 69 or 68 percent said that that actually a way of social support and a way of gaining more contact with a more diverse group. Naturally, they’re also negative; 45 percent said that they felt overwhelmed with the drama or the pressure to look good and the pressure to get likes and comments. But as you see there’s a quite big proportion saying it is an important way of socially connecting. So we’ve done a big literature review, and we’re still looking for funding for a study into social connection in lockdown, and how social media might be filling this… might be allowing teenagers to connect in a time where there’s actually very limited ways of connecting with peers. And so lockdown is really changing the way we think about screen time and the way we converse about screen time, and I’m very happy to answer questions you have that are Covid or lockdown related. We might not have the evidence, but we have been thinking about these issues in quite a lot of detail in the last couple of months.

 

Discussion

Awesome webinar.

Add a comment

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

*