‘Adolescent peer relationships and mental health: an epidemiological perspective’ Professor Lucy Bowes 

Matt Kempen
Marketing Manager for ACAMH

Posted on

Professor Lucy Bowes lecture on ‘Adolescent peer relationships and mental health: an epidemiological perspective’

This was recorded on 16 March 2018 at the Emanuel Miller Memorial Lecture and National Conference 2018 ‘Focusing on Adolescent Mental Health’.

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.

Professor Lucy Bowes
Professor Lucy Bowes

Lucy’s research focuses on the impact of early life stress on psychological and behavioural development. In particular, Lucy has focused on the effects of victimization on young people’s adjustment and wellbeing. Her research integrates methods from social epidemiology, developmental psychology and behavioural genetics in order to understand the complex genetic and environmental influences the promote resilience to victimization and early life stress. The aim of her work is to guide intervention work by identifying proactive factors that promote positive outcomes among vulnerable children.


Thank you so much. It’s an absolute pleasure to be here today. Thank you so much for inviting me to talk and share with you some of my research and research from my colleagues as well. Today, I want to talk a little bit about adolescence as a very interesting period in terms of mental health. I know we all feel that everyone in the room has much more experience than I in working with adolescents in this field. But I also want to talk about it as a potential window for opportunity because I think a lot of the time we focus on the risk factors in adolescence and there are many; it is a period of risk. But because of that and because of the fact that some of these risk factors are modifiable, we can intervene. And because there are a lot of also protective factors around this period of time, it is a massive period of opportunity, which I don’t think we should lose sight of.

I’m going to talk particularly about peer relationships and a lot of my research, like Chris’s, has been on bullying, and I’m going to talk more in terms of the epidemiology of bullying and the links between bullying and peer mental health. But then because, again, we shouldn’t always think about peer relationships as being doom and gloom, despite a large focus on the damaging impact of peer relationships in the research literature as well as the media, I want to focus a little bit on the bright side of peer relationships as a protective factor. And adolescence is a complicated, wonderful period of time when we think all the way through from early to mid to late adolescence and of course adolescence, it’s part of development. Adolescents are functioning and doing as they are because of pre-existing factors, because of genetic innate characteristics, because of the early environment that individuals receive. We heard earlier from Ian’s talk about the impact of maltreatment. But it’s also a period of great transition. We have social transitions going on.

There’s a lot of research about the increase in the importance of peer relationships during this time. We have school transitions. If we think in terms both the secondary school transition and then, of course, we mustn’t forget the transition to university in late adolescence. I always talk to my students and I say if I wanted to have a good natural experiment to induce depression or psychological disorder, I would get a bunch of high achieving students who are big fish in a small pond, I would remove them from their environment, remove them from their social relationships, their families. I would put them in different surroundings that they’re not used to, a highly pressurised environment, I would disrupt their sleep, change their dietary patterns. I might expose them to some substances they don’t normally take. Guess what? That’s what we do in university. And then we’re surprised when we see this elevation… we see this elevation in rates of depression, psychological disorders. Should we be surprised? I think not.

So there are important social transitions going on in this time, which I think are crucial, really, in understanding some of the epidemiology of mental health. But we also… that’s sort of interplayed with biological transitions, with this emergence of puberty, and although the genome doesn’t change, it remains static, we’re learning more and more about the epigenetic changes that can go on across the life course. But I’m going to present to you at least some behavioural genetic evidence that suggests that there are important changes going on in genes that influence even our social relationships during this time. And the reason I present my jigsaw puzzle is we cannot… We try to isolate this and research this in our little silos, but of course, they’re intrinsically connected and we never really can.

This is just a study, and I wanted to exemplify some of the patterns we see across adolescents in terms of mental health. So on this graph, we’re looking at depressive symptoms that are self-reported from the Short Mood and Feelings questionnaire. This isn’t my data. I was just interested. But this is the fairly typical finding that we see, and that is we see very low rates of depressive symptoms in mid childhood. And we often see this interesting finding, no sex differences going on. I’m talking about biological sex differences here. If anything, males slightly higher than females, but not by very much. But then as you go across puberty, you see this massive rise in increase, at least in self-reported depressive symptoms, and you start seeing this difference by sex here with girls. This two to one ratio typically going on, girls reporting more symptoms of depression than boys. And really by the time… So all of this is occurring around that puberty time, and by 18, you’re looking at prevalence rates that match that in the general adult period. So there’s a lot going on. And I could… This is depressive symptoms; I’m particularly interested in depression. But actually most of the psychological disorders that we deal with, we would see a similar pattern in large-scale epidemiological studies.

But I thought it was also important to present findings such as this, again, another colleague’s paper, in our very own JCPP. And this is just the data for females, for girls, but what I wanted to point out, much like the previous lecture, although you can have this general trend towards an explosion in psychological disorders and poor mental health in adolescence, there’s a huge amount of individual variation. So you have the lucky individuals that are low and depressive symptoms. They remain low or in effective problems here. And you will see that the sort of classic group of individuals who suddenly have this exponential growth in terms of these types of problems, of effective problems. But you also have more slow, chronic increasing trajectories as well. So there’s a lot of noise behind the data, a lot of individual variation, and we still need to do a lot more to understand why that is. Okay, so there’s a lot going on in adolescence. It’s an important period, and particularly when we think in terms of mental health this is a very key period of time.

One of, I believe, the most important risk factors for depression at least, during this period, is bullying. I genuinely believe it is, but I’m going to present lots of research from different types of epidemiology to say that we need to be cautious when interpreting our research, and we say how much of an impact bullying may actually have. So you already heard a lot about bullying today, and we know bullying can take many different forms. It can be the more physical violence, more typically by boys, but not always, name calling, exclusionary types of bullying, being left out of the crowd. And, of course, a more recent focus on cyber bullying, bullying using an online media. And again, we often segment this and some researchers try to look at the independent effects of different types of bullying. I’ll even present my own research where we tried to do this. But more and more, I think what we tend to see is these types often co-occur. And it is much like when we tried to desegregate different forms of violence, particularly in terms of maltreatment; it’s more common than not that you will see multiple different types of victimisation.

And victimisation and bullying in particular is often in the news, and every year there are more and more stories of incidences of tragic cases where a young person has committed suicide and bullying has been said to play an important role. And in fact, actually, it was high profile suicides in Norway that led to the explosion of research on bullying to start with, by Dan Olweus, but the picture is slightly more complicated and that’s what I wanted to talk through.

But firstly, what’s the scale of the problem? Is bullying a big issue? Well, this is data that we published last year, and it was from the What About Youth study, which is a huge study of 15 year olds in England. And they were asked a bunch of questions, bullying being one of them. And in fact, there were over 110,000 fifteen year olds in England that reported on their experience of bullying, so a huge study. And what was really concerning is that 30 percent of our young people are reporting some form of regular bullying. And I’m using the classic research cut off of what’s regular bullying? How long’s a piece of string? Well, what we tend to use is two to three times a month or more. That’s classified in the research literature as regular bullying. So this isn’t occasional, this isn’t lifetime, this is regular bullying in the last couple of months, 30 percent of our young people are self-reporting this. That is a problem.

You may be asking, what about cyber bullying? Cyber bullying is very interesting and there’s been a lot of good reason to suspect that this could be even more worrying than more traditional face to face forms of bullying. And some of the very good reasons for this is that there’s a potential for a much larger audience, whereas face to face bullying at the very worst it could be whole school, but it remains confined to the whole school, and even that’s somewhat unlikely. When we’re talking about social media, the audiences can rise exponentially. The individual can be targeted anywhere. Again, I mean, okay, this is a bit of a fuzzy distinction, but at least if bullying, which typically occurs in the school environment, is occurring there, a young person can go home and potentially, hopefully, feel safe at home. At the very worst, they may choose to change school. There may be certain areas of the school that are more risky for them that they may choose to avoid. But when bullying happens on social media, it becomes very difficult to avoid that. And this becomes a generational thing. There may be some of us in the audience thinking, switch off your phone, don’t log in, surely you can avoid it. And actually, this distinction between an offline and online world doesn’t really exist for most of our young people. This is just another way of talking. And that would… telling someone to switch off their phone and not engage would be like saying, that’s fine, just don’t talk and don’t interact with any people. That will solve the problem.

So an individual can be targeted anywhere. Removing Internet access could be potentially isolating and damaging in and of itself. There is difficulty in removing online content. When something is posted it can be very difficult to remove. Perpetrators of bullying can be anonymous, so it’s not always easy to find out who’s doing the bullying and to take action. There’s a lot of discussion about whose responsibility cyber bullying is, so although the perpetrators are often anonymous, oftentimes actually it is the kids involved in face-to-face bullying anyhow. So it is often the kids at school. However, it’s not necessarily occurring at school, so does it still remain the school’s priority to deal with that? Should teachers be dealing with that? Should parents? Should internet providers be dealing with it? Should it be due to particular sites? And there’s a lot of discussion around this particular problem. And then there’s concern as well because we know that social media use is rising hugely, so could this lead to more opportunities for bullying to occur?

So firstly, is cyber bullying on the rise? I think there’s lots of reason to suspect that it should be, but actually not many people have really looked at whether it’s on the rise. Now this is really down and dirty data. So this is from the Northern Ireland Kids’ Life and Time survey. And we’ll do some fancy analysis on this, and we’re looking at growth curve models. But actually, you don’t need to; you don’t need to do anything complicated to just have a look at this. So this is cross-sectional survey sweeps. It’s not the best measure at all of cyber bullying. Kids were simply asked, have you ever been bullied by someone sending nasty texts or putting up bad things about you on the internet? It’s a yes, no; it’s a binary. It is very, very basic. However, it gives us a little bit of an insight into what’s happening here. Is it increasing? And so I’m just sharing with you some data from 2008 all the way through to 2014, and actually, no, there’s not good evidence to suggest. Certainly there’s not good evidence to suggest… to be anywhere near supporting the kind of media concern on cyber bullying at the moment, that there’s this massive rise and all of our young people are exposed. That doesn’t seem to be the case. That’s not to trivialise cyber bullying at all. But that is to say, we should be careful. I think there’s a lot of focus, socially driven focus on the negative impact of the internet and what it might be doing for our young people, and I think some of that is fear rather than data driven.

Well, what about this idea that cyber bullying might be more harmful than face-to-face bullying for all those reasons that I mentioned earlier? And they are good reasons. They are sensible reasons. Well, again, this is from the same paper that we studied. It is… it’s just cross-sectional; it is just self-report. There are always limitations with that kind of data. And we’re just looking at the relationship between what we’ve called traditional bullying, so face to face bullying, and cyber bullying, here with mental wellbeing, using the Warwick-Edinburgh Mental Wellbeing Scale. And what we find is that traditional bullying is associated when we look at… If we use the cut off of bullying as two to three times a month or more, so I can look at odds ratios because they’re more meaningful to me… we find that actually, if anything, there’s a slightly larger relationship between traditional bullying and mental wellbeing, so more bullying, poorer wellbeing. There is a relationship between cyber bullying.

And actually the strongest relationship is in the combination of both cyber bullying and traditional bullying. And actually, that’s what the smaller studies as well tend to find, that oftentimes the kids who are reporting severe and chronic cyber bullying are the same kids who are also reporting face-to-face bullying. So our findings, like a lot of the findings, is that this isn’t some new scary form of bullying. Actually, this is just another avenue. It’s another resource. It’s another string on the bow for bullying to take place in. I think bullying is a massive problem, but I don’t think we should start panicking that cyber bullying is something new. That said, of course, interventions targeting cyber bullying, there may need to be some differences, but the fundamental concepts, I think, are probably very similar.

So that was just a little bit about bullying and cyber bullying in terms of mental well-being, but you may be asking yourself, well, what about psychological disorders and particularly does the impact of bullying last into adulthood? And one thing I didn’t warn you about is that I would actually be talking about hamsters in this talk [laughter], but I will, and I tried to find a cute picture, but I couldn’t find one. So, as a… epidemiology, I love some of the old school epidemiology.  In Bradford Hill we talk about biological plausibility. And I think there’s a lot of social and cognitive explanations for why bullying might lead to things like depression. But what about is there any good evidence for biological plausibility? And I love thinking about these studies.

We’ve got a very good… We… They have got a very good paradigm in animal research. And here’s one. They have Syrian hamsters. They’re naturally very aggressive. I don’t know if anyone knows what Syrian hamsters are. I just got my six year old one and I can testify they bite and they hurt. They are aggressive little creatures and they only come out at night. You have to spend a lot of time cleaning the cage for no gain at all [laughter] but kind of cute. So the loser… this conditioned defeat paradigm is one in which they have Syrian hamsters, and normally if you placed a new Syrian hamster into the cage, into the territory of your test hamster, it will attack it and it may even fight to the death. They are very aggressive things. So they have a paradigm where they have a specialised cage and a Plexiglas and then they can place other hamsters opposite the test hamsters to gauge the reaction. And in this particular paradigm, they put a great big, butch looking Syrian hamster. This is how I like to imagine a big, bully Syrian hamster in the cage, but separated by a Plexiglas to your test hamster. And they do that until your test hamster displays signs of social defeat, tail covering, teeth chattering, submissive posture. I know. I feel sad for the Syrian hamsters, but not that sad; they hurt.

But what’s really interesting is when you then remove this big bully hamster and you replace it with a little puny hamster. Now, normally you would expect your test hamster to regain their aggressive and dominant approach. And that’s not what’s seen. After conditioned defeat what they’ve seen that upon being paired with a small and nonaggressive opponent, the defeated hamster displays no normal territorial aggression, but shows defensive and submissive behaviours, such as flee, tail lift, teeth chatter and defensive postures. They maintain this in their ongoing social relationships, even in contexts where they would be the dominant hamster. These are just hamsters, but this similar what’s been dubbed the loser effect has been reported in a wide variety of species, and I think it’s a very interesting model for what might be going on in terms of behavioural changes at the very least when you have social defeat in bullying.

You heard a bit about cortisol in the last lecture, so I also wanted to put this slide in again, talking about some of the potential impact of bullying. And this was a study that I helped collect data on as a PhD student, which was great fun and probably disturbing when I talk through it. We looked at identical twins and later in this talk I’ll talk to you about how neat identical twins are and why we love them for research. We brought these twins in because they were discordant on their experiences of bullying. Now, most twins, these are identical twins, 100 percent genetically identical, is more often the case than not that if one twin is bullied the co-twin is bullied, and I will talk about what that means later in the lecture. But for this particular sample, we had discordant twins. One had been bullied, one hadn’t. But obviously they grew up in the same home environment. They share the same genome. We bring them into the lab and we do an experimental paradigm to look at their cortisol reactivity. So this isn’t baseline morning cortisol. This is cortisol reactivity to a stressor. And the stressor was the complex mathematical task designed to be too tricky, and we tell them it’s a competition with their twin. We don’t give them any feedback. And then we make the poor kids stand up and talk about their worst experience at school while we do the still face paradigm, and it was really difficult for me not to smile at all during this process. And all the way through they’re spitting. We took about eight saliva samples, so that we could test for their cortisol reactivity.

And these are the results. So we’ve got the bullied twin shown in the solid line and the non-bullied twin in the dashed line. PST is the psychosocial stressor, and cortisol reactivity takes some time to kick in, at least in saliva, to be able to measure it in saliva, so what you see is this is the psychosocial stress and then you see a change in cortisol reactivity after. But check this out. This is what’s interesting. The non-bullied identical twin in each pair displays the normal elevation in cortisol that we would expect to see, but the bullied twin does not. They do not show the expected rise in cortisol, and interestingly this kind of pattern of down regulated, a flattened cortisol response is often seen in kids who are maltreated. It’s not always seen. It is a contentious area. Replications can be difficult.

But some of the suggestion is that when you have chronic stress, chronic high levels of adversity, it may cause a dysregulation in our cortisol systems. Potentially, it could be the body’s way of conserving energy, so that you’re not always being flooded each time with a new cortisol response in terms of stress. In another paper, Isobel looked at whether this could actually be a helpful and adaptive response. You might think it might be better not to display a peaked kind of elevation in cortisol response, and actually that wasn’t found to be the case. If anything, it was associated with more emotional and behavioural problems. So this is an experimental study, but with a particularly nice sample because they provide a very good control for each other, being that they’re identical twins.

So we have some biological plausibility. If we know that bullying seems to be having a role in terms of cortisol regulation, we see some animal experiments, and there’s a lot of suggestion about the cognitive changes that were going on, it makes sense that bullying might be an important risk factor for mental health. However, it’s also a tricky risk factor because who are the kids that are most likely to be bullied? They’re the kids with the lowest self-esteem. They’re the kids that already have high symptoms of depression or depressive symptoms, I should say, as a child. They are much more likely to be bullied. On the flip side, you also have kids that are more in your face. They have some impulse control. They may have symptoms of ADHD, autism. They are also more likely to be bullied. You tend to get two different types. This is particularly kids in early adolescence I’m talking about these risk factors. Children who are maltreated are twice as likely to be bullied by their peers. So they have that awful experience at home and then they’re going to school. It should be safe, but they’re more likely to be targeted. Socioeconomic deprivation, there’s some suggestion that bullying rates are higher. So there’s a lot of potential confounds in this relationship because a lot of obviously these risk factors for bullying are also, in and of themselves, risk factors for depression.

So I was interested in seeing whether bullying might still be associated with depression, at least at age 18, when we’re talking about rates of depression that match those in adulthood and whether the associations remained when we used both a clinical diagnosis of depression and whether we adjust for all of these other pre-existing risk factors that may confound the relationship. So for this study used data from ALSPAC, Children of the 90s study. It’s a self-report of bullying, which has some limitations because it’s also the same individuals who are more likely to be depressed, and we know all about depression and negative memory recall. There’s lots of reasons why self-report, particularly for bullying, could be tricky, but actually there’s no gold standard. Who do we ask in adolescence? Do you ask the teacher? Teacher doesn’t always know what’s going on. Teenagers themselves will tell you they don’t tell their parents if they’re being bullied. Only about 20 percent say that they tell their parents. So who do you ask? If you do peer nomination that works really well for a school study, but it doesn’t work well for a large cohort study because you’d have to go to every school that each of your participant kids live in. So you’re kind of stuck with self-report.

And this is what we find, so we… Just categorising a scale for bullying from no, occasional to frequent, frequent is happening weekly, and we see a typical dose response pattern according to the extent of self- reported bullying and depression, and this remained even when we adjusted for all of those other risk factors that I spoke about for depression. And because the rates of bullying are quite high, if you stick this into a population attributable fraction, it could suggest that up to 26 percent of depression in that sample could be attributed to bullying, and this is factoring out all of these other risk factors. This is correlational analysis. However, that’s a very large population attributable risk. And I think this does tell us that this is at least an area that we should be paying more attention to than I think we are.

However, yes, that was lovely and I got to publish in BMJ and it’s all exciting, and you can’t really let go of the fact that are we actually adjusting for all of our confounding variables in those types of longitudinal cohort studies? Is the relationship as large as we think? And some of my earlier work was looking at twin studies, but this time comparing identical and non-identical twins. If you compare them because you know that their shared environment is the same, you know that their additive genetics is either one if they’re identical twins or 0.5 if they’re dizygotic because they share 50 percent of the segregating genes on average, you can bung it into a model and you can get some estimations of heritability. And we know that when we do this, and other studies have found the same, that being bullied is not purely environmental; it is partly heritable. So, for example, I was interested in the transition between primary and secondary school, and some of the risk factors for this, and when I use a twin model to look at this, I find that genes do seem to play an important role. This type of analysis can’t tell me which genes at all. It’s just looking at variance in a population that may be explained by additive genetic factors.

But what we found is that there were genetic factors that seem to influence one’s risk of being involved in bullying in primary school. Some of them had an impact on your risk of being involved in bullying in secondary school. But there were new genes or different sets of genes that seem to have a risk in secondary school victimisation. So genes seemed to matter. The shared environment also played a role. So this is… I mean it is a little bit of a statistical artefactual. Anything about twins growing up in the same family that makes them more similar is one way you can think about this. We can’t say again what the shared environment is. People tend to think of it in terms of things like socioeconomic status, but actually that can impact on different kids in different ways, so it’s not necessarily a shared environment. But that said, we find evidence that the same shared environmental characteristics that seem to predict bullying in primary school also predict bullying in secondary school. And that there are non- shared environmental experiences that seem to increase individual’s risk, so perhaps we shouldn’t be surprised.

So what does that mean, genes play a role in bullying? Well, genes play a role in everything. Everything is partly heritable. Your likelihood of being divorced is partly heritable. Heritability we should take with a pinch of salt. Any human complex behaviour is going to have some element of it that the variance in which genes play an important role, so we shouldn’t be surprised. However, it does mean that bullying isn’t completely environmental, and it does mean that my earlier ALSPAC study isn’t adjusting for one very important variable. And in a lovely paper from Jean Baptiste Pingault that came out last year… I like to show this. I like science. I like being shown that we have to be careful. I had this worry when I published the paper on ALSPAC data, that are we overestimating risk. This paper suggests that, yes, indeed, we may be. However, bullying does appear to have a relationship with a range of different mental health disorders. In this paper they looked at depression, anxiety, psychosis. But what we see is if we just look at the phenotypic relationship… That’s what I did in that earlier paper where I showed you the pie chart… if we just look at this, this is where we get the strongest relationship. We find that it does… It’s strongest concurrently, as you would expect, and it decreases a little over time when you get up to five years from exposure, as you would expect, but it remains strong and true.

You could look at the difference between dizygotic twins; that’s a little bit more stringent because they grow up in the same family or you could be more stringent still and you could look at identical twins like we did in that study when we looked at the cortisol reactivity. So for identical twins you’re effectively adjusting for the role of genes, genetic confounding, as well as all of those other shared environmental characteristics, and what you find here is a much greater attenuation of the relationship between bullying and different psychological disorders. It remains; it is still there, but it does attenuate. This kind of information is important, not just because it’s fancy statistics and we get to use twin models, but it also tells us the extent to which we might expect bullying to have an impact on different outcomes. So when we’re thinking about bullying interventions as well, how much should we expect them to reduce things like depressive symptoms or psychosis? We need to get a better handle on the strength of this relationship in order to know whether it’s our intervention that needs tweaking or whether, in fact, that relationship is not as strong as we think.

Okay, so the dark side of peer relationships with the focus on bullying, we know that bullying is highly prevalent. It is highly prevalent, but we don’t really have great evidence that cyber bullying is on the rise. That’s a positive. We know that youth with pre-existing symptoms of poor mental health are more likely to be bullied and that some of the relationship between bullying and poor mental health may be confounded by these factors, by genetic factors and by other factors. However, even when we adjust for these factors, there does remain a relationship and it does remain over time. So peers can be risky, but peers can also be positive, and I promised that I would talk a little bit about the positive impact.

So my lab… So we know a lot about social support. There’s the social support buffering hypothesis. And that typically goes along the lines of it’s good to have social support and it may be even better to have social support in the context of adversity. And for adolescents, of course, peers are a primary source of social support. Now, I’ve been… I am interested in the impact of maltreatment on development, but I also recognise the fact that maltreatment, at least if we’re talking about maltreatment recorded at the level of who gets referred to social services, is the tip of an iceberg. And actually, when we think about parenting, there is a continuum of parenting from wonder parents who I see in the schoolyard who are always on time and their kids have their hair brushed all the way through to the extreme end.

So in our analysis here, we’re looking at poor parenting. In particular, harsh and neglectful parenting with low levels of emotional warmth. So it’s not maltreatment, but it is more physical discipline and it is low warmth. And this is using data from the TED study. This is unpublished data, so I’m not putting effect sizes or anything yet, but what we find is typical in that poor parenting at the age of four is associated with outcomes even at the age of 16 in adolescents. So when we look, for example, at the child behaviour checklist, we see a strong relationship with emotional difficulties at 16 and also with behavioural difficulties at 16. And this relationship remains even when we adjust for the kids’ emotional and behavioural problems before our measure of poor parenting or concurrent with our measure of poor parenting. Why do we do that? Well, it isn’t just about the parent’s relationship with the child and the parents’ impact on the child’s behaviour. The child’s behaviour does also impact on our parenting, so we do need to adjust for that. But even when we adjust for this, we adjust for gender, social class, maternal depression, there is still a strong and consistent relationship. This is not news. Everybody knows this.  It’s found in every study. From this, you can… I’m very interested in resilience and how we might promote resilience.

So we know that poor parenting is a risk factor for emotional and behavioural problems. Okay, what can we do about it? Given that you’ve had poor parenting, what are the kinds of protective factors that we may be able to strengthen or support in order to promote better outcomes? This is just a little diagram to show how I indirectly measure resilience or how I think about resilience, and I think about resilience as being doing better than expected. So if I think about someone’s adjustment difficulties and I have the extent to which they experienced poor parenting… This isn’t actually the scale; this is more of a representation for you, so you know what I’m doing. If I were to plot my regression line, it would look something like this. And I would have individuals that are doing about as I would expect in terms of their adjustment difficulties, given their exposure to poor parenting. They’re doing about as well as the norm. There would also be individuals who are doing worse than the norm, would be doing… They would have more adjustment difficulties than we would expect given their experience of poor parenting. And we would have individuals doing better, fewer adjustment difficulties than you would expect given their exposure to poor parenting. And I like this because if you were to plot the residual scores, it gives you a better idea of the full range of functioning, rather than just arbitrarily dividing kids up as being resilient or non-resilient on some threshold that we’ve decided is important. It gives us a better concept of the range from vulnerable all the way through to resilience. And we can use this as our dependent measure to look at factors that are associated with this type of resilience.

And that’s exactly what we did, and one of the key factors I was interested in is peer support. So here the young people themselves are talking about their peer relationships, the emotional warmth they obtain from their peer relationships in particular. And what we find is that peer support is associated with behavioural resilience at 16. And this is behavioural resilience in the context of negative parenting in the early years. So peers are a buffer, a positive buffer in terms of behaviour. We often talk about adolescents as being risky and promoting bad behaviour, and although that’s true, we shouldn’t forget that they can also promote better behaviour and better functioning, particularly amongst kids exposed to early adversity. And these effects remain, even adjusting for early emotional and behavioural difficulties, prosocial behaviour, the kinds of characteristics that might increase your likelihood of finding a friend and also showing resilience, and all of these other factors. Peers seem to be important.

I’ve talked to you a little bit about genes and the importance of genes on relationships, so you might be thinking are some kids just better able, innately better able to make friendships and maintain friendships? Some kids are born more sociable than others. And again, we can use this monozygotic twin difference design to try to unpick whether this social relationship effect that might be genuinely protective or whether it might just be another confound. So, again, these are identical twins, they share the same shared environment, same genotype. They’re a very good case control design. They are the perfect match for each other. We love twins. And here we have twins who have both been exposed to poor parenting, so they’ve grown up in families, they’ve been exposed to poorer parenting, they’re identical twins, but they differ on their levels of self- reported peer support. And what we find is when we look at the… we relate the differences in peer support in this at risk group of twins who’ve experienced poor parenting and see whether they correlate with differences in behaviour problems, it’s still correlational, but you’re adjusting for many other confounding variables than you are able to in a standard cohort analysis. And what we find is that among twins who experience poor parenting, the twin who experienced higher levels of peer support had the fewest behavioural problems at 16 years of age.

And again, it suggests an association between peer support and behavioural functioning among children exposed to poor parenting, and it suggests that this association isn’t simply confounded by kids’ genes or environmental factors. And in these analyses, we control for the child’s own behaviour that may relate to their likelihood of attracting a friend. We can’t look at the particular characteristics of their friends because we don’t have that data and there’s some lovely studies that do that and look at the bidirectional relationship. We’re not able to, but they are beautiful, and I can send papers to anyone that’s interested.

So in conclusion then, adolescence is an important developmental period in terms of mental health. It’s also a period of wonderful opportunity in promoting better mental health. Teenage bullying is an important risk factor that might increase the risk of poor mental health and that risk may be sustained into adulthood, but positive peer relationships may be a crucial source of support for adolescents, particularly for at risk teens. Thank you very much. [Background noise] Shall I take my own… Thanks Cathy.

Speaker 2

Sorry, hello, thank you. I just want to ask about the difference in the… so, with the adolescent girls having higher rates of depression than adolescent boys, do you think that could be something to do with the nature of adolescent girl peer relationships and that actually they’re not, even though they might superficially appear supportive, they often can be quite competitive? Or, I mean, I was wondering your thoughts on that? Or may the male peer relationships in adolescence are actually… I don’t want to use the word simpler but may actually be more supportive.

Dr Lucy Bowes

Thank you. I don’t think it’s anything to do with bullying or bad peer relationships because actually, in all of our analysis, when we compare the association strength between bullying and any given outcome between girls and boys, it’s usually the same. So I don’t actually think it’s that. I think the reason why we see this two to one ratio of depression in girls to boys is very complex. There’s a lot of other reasons. There’ll be a lot of social reasons as well. Girls are more exposed to a broad range of different impacts for mental health, not victimisation interestingly. Boys are more likely to be bullied than girls and to bully others than girls, at least in early childhood. There’s also hormonal attributes, there’s… Some of it is also the likelihood of a girl talking about mental health and disclosing poor mental health. So there’s a lot of really interesting work trying to unpick what this gender difference is. There is some suggestion, particularly for girls, for early menarche, for example, can be related to depression, at least in the short term. The results don’t seem to persist over time. But for girls who reach puberty before their peers, they’re more at risk of depression than their peers. So I think this is a really complex question, and I don’t think, unfortunately, it boils down to just peer relationships. Can peer relationships be more supportive for girls than boys? I don’t know. I’d love to look at that, the social support buffering hypothesis. I don’t think we…, actually, so I can’t answer that, but I would love to have a look at it. Thank you.

Mike Davies

Mike Davies, neuropsychologist. Very interested in what you were saying about poor parenting. Obviously it’s very important, but it’s a term that’s used very loosely and I’m very concerned about that. So I’m just wondering how you defined it and how you measured it?

Dr Lucy Bowes

Yes, thank you very much. I completely agree. This is in TEDS. In terms of parental warmth, they have expressed emotion, which gives the five minute speech sample. We’ve used that in E-Risk as well. When they talk about… You talk about each of your twins in the family for five minutes and then you code it for the warmth expressed. So it’s not just the positive words used, it’s also how you say it. And I can testify because we have talks where they say, oh, he’s a little bugger but said in a way that’s clearly warm, very high inter rater reliability. That expressed emotion is actually… often tends to be the biggest risk or protective factor in all of the measures we have of parenting. It seems to work better than observer reports. So it’s also the home observation report. So independent assessors going in the home and coding on their parenting as well. So that’s the harsh and it’s maternal self-report of harsh parenting and use of physical discipline. It’s only mum because in those days still, but particularly in those days, only mums came in and did the appointment. We had such a minority of dads that we can’t really do anything with the analysis. But of course, there are problems with that as well, even as physical discipline, because it’s completely socially unacceptable nowadays to say that you use physical discipline, whether or not you do, so that’s going to play an impact on the likelihood that they would report this.

Mike Davies

The reason for my question was how you responded it seems to be about the quality of parenting and my interest is in we hear a lot about attachment disorder. And I think what you’ve described for me would not cause attachment disorder in the sense of how I was thinking you were using poor parenting, but that’s fine.

Dr Lucy Bowes

Yes. We don’t have, for example, it’s not a good measure to look at changing, so inconsistent parenting, which for attachment disorder, I guess, would also like harsh and inconsistent. And we don’t have a good capture of that in this study. It’s one of the things with large cohort studies. This is a very beautiful study and they’ve got very good measures of parenting for the size of the study. But it’s fairly typical that in the very large scale studies, you get this kind of very broad brush measure of risks like parenting. And it’s only in the smaller psychological studies that they’re able to disentangle these different types of parenting that may be more risky, but I’d be interested to hear from you. What do you think is the riskiest type of parenting in terms of attachment disorder?

Mike Davies

I think that’s one of the problems I’m sort of talking about. I don’t think we really have good measures of that. And I think one of the earlier speakers also said it has… in my view anyway… that the parenting has to be really very severe to have a long term effect on the child of the sort of thing that we all of the professionals today should be worried about. That’s why I really asked the question because I didn’t know how, as you say, how I said, how you measured that particular term. Thank you.

Dr Lucy Bowes

You’re welcome.


I just wanted to ask, I’m really intrigued about the monozygotic, the dizygotic and controls, and of course the environmental impact of having a sibling that’s the same age as you. What are the links and concordances and discordant aspects of the capacity to make good peer group relationships and have… you know, supportive peer group relationships with looking at those sort of variables?

Dr Lucy Bowes

Those dynamics. I guess I could rebrand your question and say do twins generalise to other children because they have their own sibling who could be supportive? And we have looked a bit at that, and we know that the sibling relationship itself is supportive and is a buffer particularly against harsh parenting, not maltreatment. So there is an element of support there. There’s also a flip side. So we’ve looked at sibling bullying. Whether we agree with the term sibling bullying is almost mute, but certainly where you have behaviours that if they occurred in school would be classified as bullying when they occur weekly in the home environment that seems to be detrimental and is associated with self-harm, suicidal ideation and depression at 18 years. So I think it depends on the strength of the sibling relationship. Whether it’s closer in twins compared to singletons, you hear people say it is, but I don’t think we have good evidence to suggest that that’s the case. And certainly when we look at things like the prevalence rates of bullying, prevalence rates of poor mental health, prevalence rates of lots of other variables, twins, broadly speaking, are comparable to non-twin children. So they seem to be a good kind of group to study. But for bullying, we have to be careful. And you’re right, that’s one of the assumptions of twin modelling, do our results generalise?

Add a comment

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