‘Social Media Use and Cyberbullying: an international analysis’ – Professor William Pickett

Matt Kempen
Marketing Manager for ACAMH

Posted on

This talk by Professor William Pickett, Queen’s University in Canada, is on ‘Social Media Use and Cyberbullying: an international analysis’. 

This is an excellent presentation from an expert in the field involving a large study involving 50 countries and 400 researchers; results from 2018  include 42 countries and n= 180919 participants,  who self report on use of social media.

  1. The background to social media use is provided including the context of the research which is being presented here ie the Health Behaviour in School Aged Children, survey network (hbsc network)
  2. The opportunities for the study which is presented are set out
  3. The methodology and results are presented in detail
  4. Implications for mental health and clinical practice are highlighted
  5. References and pointers to relevant literature and research are provided.

Slides for this talk are available as a pdf download.

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 William Pickett
Professor William Pickett

Dr. Pickett plays a major role in supporting the Department of Emergency Medicine with its research goals.  His areas of interest include the epidemiology of injury and trauma, pediatric health, agricultural and rural health, and epidemiology as applied to emergency medicine. Dr. Pickett is also Director of the MSc Program in Epidemiology at Queen’s, and acts as a methodological resource person for faculty, residents, and students in the Department of Emergency Medicine. Please see Dr. Pickett’s Google Scholar profile for publications. Bio and photo via Queen’s University, Canada.


Hi everyone, my name’s William Pickett and I’m from Canada and I am presenting a collaborative piece of research called Social Media Use and Cyberbullying and International Analysis. And I’d just like to point out that this is actually part of a broader collaboration, there is a lot of people involved, the two main folks I’ll introduce, so my colleague, Wendy Craig and myself, William Packet. So here is what we’re going to cover during this short seminar. I’ll introduce ourselves and this topic, talk a little bit about our study opportunity, briefly outline our methods, and then we’ll spend some time looking at some data and our findings and then think about the implications for practice afterwards.

So start with introductions. So who am I? I’m a faculty member at a little university in central Canada called Queen’s University. Queen’s is one of Canada’s oldest universities, and I work in the medical school out of the public health sciences department. I’m an epidemiologist and I do research on the health of children and rural health type issues and have done so for 20 to 25 years. And my good colleague Wendy Craig, Wendy is head of psychology at Queen’s and she and I collaborate on a number of things.

Wendy also is in charge of a Canadian non-profit or NGO, I think you would call it a trust in Great Britain, and it focuses its [inaudible 00:01:51]  and it focuses on the prevention of violence and bullying in children and young people. And she’s quite an eminent scholar in that field. Wendy and I belong to a survey network called HBSC, or Health Behaviour in School Aged Children, and this provides the platform for this research. And within that HBSC network or that child survey network, we’re in a focus group called Violence and Injuries, and traditionally we study trauma and injuries related to physical harm. We also study things like bullying and fighting and aggression. But of late, we have been focusing more attention on virtual forms of aggression, including cyberbullying. These are some Canadian data from our survey, and it’s just to show that cyber bullying is sort of on the map. And in our own country of Canada, young people who report being victimized online by the time you hit grade nine or 10, which is age 14, 15, we see 10 to 15% of young people reporting this experience.

And we always see this gender pattern where girls report more than boys, but it’s emerged as an important public health problem. Within this HBSC network we’re also interested in online exposures and particular lately social media use. And there’s a lot of contemporary interest in actually saying; what are these online behaviors doing good and bad for the health of young people? So we wanted to link some of these behaviors with cyberbullying to see whether or not there was a problem or not.

For example, here’s our Canadian data again, this is intensive social media use in Canada. And upwards, by the time you hit age 15, grade 10, 40 to 50% of young people are online virtually all the time with somebody. That’s what an intensive user needs, which is of obvious concern. Their worlds are changing and some of these are what we call problematic in that they’re indicative of more addictive like behaviors. And as much as 10% of young people are expressing problematic social media use by self report. Again, more girls than boys.

So we’re interested in this in Canada and we brought some of our questions to our international network. And here’s our underlying sort of research frame or conceptual framework. We think that some of these expressions of social media use, ultimately end up in mental health problems where they go together. And we’re interested in whether one pathway that might link problematic social media use and intense social media use to mental health problems, is expressions of aggression, cyberbullying.

And this particular analysis is just the first part of that pathway, looking at social media use and cyberbullying. But we have a number of studies planned looking at this full sort of conceptual framework. So a little bit more about our study opportunity you might be interested. Health Behavior and School age Children has been around for about 30 years and it’s a survey in the pre-adolescent – adolescent years, ages 11 to 15 years of age. It is the, I think, the most predominant health promotion survey of its kind in Europe.

And Canada has sort of been along for the ride on it for about 30 years. And its purpose is to gain new insight and increase knowledge and understanding of adolescent health. And we do that with sort of a social developmental lens. We not only sort of monitor behaviours and health experiences; but what are the social and developmental factors that actually lead to those health states? The first objective is research, so we conduct national and international research on a variety of aspects of health and health behavior. Cyberbullying and social media are just one of them.

We have a surveillance function, we monitor and compare young people’s health over time and across countries and across age and demographic groups, and we try to translate the knowledge. So we do webinars like we’re doing here today to actually provide end users, such as clinicians, with new knowledge. As I said, HBSC has been around for about 30 years. It’s affiliated with WHO Europe. It was first conducted in three countries in the early 1980s, and then every four years in cycles since. The most recent data were  2018.

So there’s been 10 cycles so far. There are now 50 participating countries, primarily in Europe and over 400 researchers involved in this network for a variety of disciplines. Wendy is a psychologist, I’m an epidemiologist. We have anesthesiologists, we have psychiatrists, we have all sorts of clinicians, you get the picture. And here’s a little map of our network, so almost complete coverage of Europe, a little bit of the Middle East. It’s expanded into Asia, although not officially within the network and in South America, but there’s sort of satellite projects. And like I said, Canada has been involved since the late 1980s. And the real beauty of this network is the ability to do international comparisons, for example, here is an international comparison, for some example, countries, the most I could fit on a slide, looking at intensive social media use. And this shows that a lot of kids around the world are using social media an awful lot.

And we see gendered patterns consistently across all those countries. The objectives of this analysis were first, we wanted to look at our network and our data and document engagement and social media use, across national. Secondly, we wanted to relate indicators of social media use to the perpetration of cyberbullying, as well as being victimized by cyberbullying to see whether or not there was a problem. And our idea here was to look for universal patterns that crossed countries and cultures.

So what did we do, our methods?

Sampling is done differently across the 50 countries, but the intent is the same. We have an international protocol that insists that we get a representative sample of young people aged 11 to 15 years from across our country, each of our countries. So in Canada, we do a multi-stage cluster stratified sample that goes across our 10 provinces and our three territories, to get a representative picture of young people. The sampling is done within school systems and it’s a school based survey, so we survey our public and private schools, which make up about 95% of our school system. And other countries do like things, but it might look a little bit different in England or Norway or in Russia than it does in Canada, but we all end up with representative samples that we submit.

The survey is done in the classroom. It wouldn’t be done that way now during the pandemic but these were more normal times. Pen and paper survey generally, although online versions exist and are used in maybe one of five schools, and basically the young people sit down in a one hour period, fill out a questionnaire with standard items. Our younger children get a slightly shorter and more age appropriate questionnaire than our older children, but generally done in classes in a one hour time block, and we have a number of items, a range of items that all the countries ask. Here we were interested in social media use, cyberbullying and variables could affect the relationship between the two.

So, for example, we have this scale, it’s called the social media use disorder scale. It comes out of EU kids online and it’s basically getting at how frequently they have online contact with a variety of different types of people. And we were interested in flagging those kids that were on almost all the time to be an intense user.

This is the problematic social media scale, and I don’t expect you to read this, but it gets at things like preoccupation with social media, feeling bad when you’re off it. You’re having trouble managing your time around it, you’re getting into fights with your parents or others who are in conflict about it, and a score of I think five out of nine on this, indicates a problematic social media user. And both of these scales are quite reliable and very well tested.

For the bullying there’s a chap out of – a fellow out of Norway, Daniel [inaudible 00:12:18] who, years ago developed items for the study of face to face bullying. So things like verbal bullying and physical bullying. And we’ve adapted his instrument to actually get at cyberbullying. So the question is, in the past couple of months; how often have you taken part in cyber bullying after being given a definition of that? And you indicate how frequently that is. And here we were looking at any versus no experiences with cyberbullying, either perpetration or victimisation.

Analysis was statistical epidemiological, we described engagement in social media by age and gender and country, we described cyberbullying by age, gender and country, and then we related the two using regression models, I think [inaudible 00:13:11] regression, where we were looking at risks for the negative outcomes associated with the social media use indicators. All of our countries adhere to standard ethical requirements fairly rigorously. We received them from our home institution and our government and other countries do the same. Consent is generally acted. Some jurisdictions require it to be passive, but we adhere to standard ethical requirements required by our scientific community.

So what do we find in this analysis? Well, first of all, our sample, we’ll describe it and it’s big, it’s a really robust sample. Of the 50 countries in the network, 42 had the items or had their data in on time, to participate in analysis. There was about 181,000 participants, this was in 2018, the most recent cycle. Variations in numbers, but between the different countries, so we would have to wait each country, according to their population size. And we have a nice [inaudible 00:14:30] boys and girls from different age groups. And you can see the three sort of age groups that HBSC focuses in on, from preadolescent 11 years, up to full blown adolescents in 15 years.

Prevalence of cyberbullying. So our experiences in Canada are not that different than other places, substantial proportions of young people report being victimized by cyberbullying or engage in slightly less are engaged in the perpetration cyberbullying. We see patterns by age and developmental stage, we see patterns by gender, so here’s the key observations. There are large variations across countries in the reporting of this behavior or these behaviors. We do see gender patterns; boys perpetrate more and girls are victimized more, and we see developmental patterns. Cyber victimization tends to be fairly stable by age, whereas perpetration increases with age. So we do see variations across countries and within our sample and in the different countries, age groups and groups defined by gender.

A similar analysis was done on social media use, we actually had three indicators in this analysis. We had our intense to use measure, we had our problematic use measure, and from the intensity scale, we could find out whether kids were talking with strangers a lot, which is another indicator used in the literature a fair bit. What we saw in these data, again, large variations across the 42 countries. Girls tend to engage in more of this, they’re more intense, more often to do report intensive use and more often report problematic use than do boys. Boys are more often talking to strangers in context; often times that might be playing video games that are online, there could be other explanations for this, but that’s the pattern that we see. And we see developmental patterns, our problematic use increases with age and is higher in girls. Intense use increases with age in both boys and girls, and interaction with strangers increases with age in both boys and girls, and it’s especially prominent in boys.

So then our regression analysis, and if you ever look up our paper, you’re going to see a table like this and I say, don’t look at this output, it’s too much for a presentation. But I want to summarize what we did. We did a series of regression models looking at the three indicators of social media use and how they relate to victimization and perpetration of cyberbullying. And what we found is that all three indicators of social media use are associated with cyberbullying, so you’re talking with strangers a lot, higher risk for both indicators of cyberbullying; perpetration and victimization. If you’re on it a lot, more opportunity, more cyberbullying. If you’re engaged in problematic sort of more addictive type things, more cyberbullying. Associations were strongest for the problematic social media use exposure. It was strongest for perpetration versus victimization, so social media use was most strongly associated with perpetration and again, we saw girls versus boys were most affected by these relationships.

So I give you a little bit of an example graph. These are adjusted relative risks that come out of a [inaudible 00:18:51]  model. And if you have a relative risk of more than one, it indicates that the social media use is negatively associated; well it’s actually positively, but it’s more associated with a negative outcome. So this is – the outcome here is perpetration of cyberbullying. The exposure here is problematic social media use and after adjustment for a number of factors we are finding in boys and girls almost all of the countries, that one seems to be associated with the other. Problematic social media use is associated with perpetration, and what is really important here is the consistency of the findings and the fact that girls seem to have stronger relationships. And I think it’s only one or two countries where you don’t find this association, so very consistent results and that’s something you look for in epidemiology, for something to be potentially causal.

Here’s the same sorts of figures, relative risk estimates, but they’re for the intense measure. We do see statistically significant findings in most countries, but what’s more important here is the pattern that we see, as people reporting intense social media use goes up, perpetration goes up. So our main observations in our regression analysis, all three indicators of social media use were associated with cyberbullying; both types, victimization and perpetration. Associations were strongest for problematic social media use for the addictive like one, for an outcome of perpetration versus victimization. And this was a particularly strong set of relationships in girls versus boys.

Strengths that this analysis, we need to comment really briefly on those strengths. It’s a really large cross national study. We can look for consistent patterns across countries and cultures, and our items are well validated in their survey protocol as well validated. So it makes for a fairly strong, robust analysis. Limitations of this; they are self reports and with all the problems of self reports. And because it’s cross-sectional data, you can’t really infer cause and effect because we don’t have the temporal sequence of events; it’s a snapshot in time.

However, there is other longitudinal evidence in the literature that supports the idea that these relationships exist. So a few implications. I understand the main audience for the seminar, webinar is clinicians, and I think they’re really simple. First of all, I think we all need to recognize that social media use is a normative behavior. We should also recognize it follows gendered and developmental patterns, girls being particularly affected. And social media use   getting more and more as young people age and develop.

So that won’t come as any surprise to anybody but our data strongly supports this sort of normative behavior pattern. I would say, and I think this is especially poignant today; social media use in itself is not a negative behavior. The connections that young people have and the opportunities to communicate and get new knowledge are really good ones but its misuse is a growing adolescent health problem. So it’s sort of a double edged sword. There are lots of good things that come from engagement in social media and it’s normative use, but there are also some flags for some ongoing problems that could lead to aggression and it could lead to mental health problems.

Social media use, indeed provides new opportunities for aggression, often times young people are anonymous. Often times, they’re not sitting face to face so there is less of a hesitancy for people to be aggressive. And it’s certainly a medium that has become a normative part of everyday life where people can experience aggression, and I just simply ask that, what our data suggest that you do as a clinician, to please recognize that some negative effects of social media seem to be very there and very real.

We see these effects consistently and these effects are likely not innocuous. The effects of cyberbullying are well known and have a number of mental health consequences. So please recognize that these negative effects exist, they exist across countries and cultures and our data flag a really, really interesting phenomenon. If you’re interested in reading the article in which this is based, it’s in a special issue of the Journal of Adolescent Health, there’s about 10 different international analyses that are that are present there and the link is provided here.

And that’s really what I had to say. It was entitled Social Media use and Cyberbullying; an international analysis, and on behalf of Wendy and myself, I was really pleased to present this to you today. Thank you.


Leave a reply

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