95% of American teenagers use the internet and 81% use social media platforms.

In the time-honoured tradition of being suspicious about young ‘uns and their new fangled gadgetry, concerns have been raised about the potential negative impacts of social media use, with “cyberbullying” one of the more prominent.

I should own up in advance that this is a topic that I think suffers from a fair amount of Greenfield-ian misplaced concern (any excuse to point you here to this wonderful demolition of the Baroness by Dean Burnett), but that’s exactly why this scoping review is useful; what does the evidence actually tell us?

Methods

• The authors aimed to evaluate studies looking at the health-related effects of cyberbullying via social media sites on children and adolescents. They searched 11 databases from 2000 to 2012.
• Studies were included if they reported primary research, looked at social media (defined as collaborative projects, blogs, content communities, social networking sites and virtual worlds, and excluding for example skype or text messaging) and involved children or adolescents.
• 36 studies were included. 58.3% were conducted in the US and 66.7% sampled middle and high school populations.
• Study quality was assessed using the Mixed Methods Appraisal Tool.
• The authors state health outcomes were not defined a priori, in order for them to understand what health outcomes were being measured in the field. It seems predominantly to involve mental health issues or mental distress, though risky behaviours and aggression were also outcomes in some studies.

Results

• There was considerable variation in the quality of the studies. The variation in definitions, populations and measures used, meant the authors could not pool results and makes it difficult to draw conclusions. As a taster of the heterogeneity, reports of just the basic prevalence of exposure to cyberbullying ranged from 4.8% to 73.5%.
• Various negative effects of cyberbullying were reported, including loss of confidence, withdrawal and negative impacts on relationships with friends and family.
• Most studies were cross sectional, which is a big problem in terms of trying to establish whether cyberbullying leads to or causes health problems or distress
• Studies also didn’t tend to check whether the participants had pre-existing mental health problems, another big problem in terms of establishing whether mental health issues are the result of the cyberbullying.

Conclusions

For me, the variability of the measurements and populations and the limitations of the analyses make it too risky to try to draw conclusions. However, the authors say:

Despite some inconsistency in results, social media use is undeniably widespread and can act as a forum for negative behaviours that may adversely affect mental health. Recognition of these issues is key to advancing prevention and management efforts.

I’m not convinced. Riding the bus is undeniably widespread and could be a forum for ‘negative behaviours’, but I haven’t seen any reviews about it. The authors seem here to be stating that because social media could be a forum for problems, then we should ‘recognise’ that and move to prevention and management, which seems a bit of a leap to me. I’d argue that the cross sectional nature of the studies and the heterogeneity of data and definitions make it nearly impossible to state with any confidence what the impact of cyberbullying is.

I’d go further to say that perhaps what the review does tell us is that trying to discern what impact “cyberbullying” has isn’t really possible. Given the diversity in the nature of cyberbullying, it’s frequency, the relationship of the victim and perpetrator, and the type of platform used, is there in fact an argument to be made that “cyberbullying” is too diverse a concept to be understood as a single problem?

We should recognise that there are different and specific types of aggressive interaction which have different impacts and may require different management strategies. For example, we should consider one-off anonymous name calling as a very different type of bullying to the sustained abuse that can take place following the break up of a relationship.

Limitations

• The authors talk quite a lot about subgroups in terms of both victims and perpetrators. However, it’s not clear whether the studies reported were looking for differences between groups and whether they were powered to detect such differences.
• I’ve criticised the studies for not being able to discern causality, but another angle would to be say that maybe asking whether cyberbullying ‘causes’ mental health problems is the wrong question. Knowing that young people suffering mental health problems are more likely to be bullied online would be very important indeed. Unfortunately, the review can’t tell us this either.
• I’m not sure there’s much in the descriptions of cyberbullying to justify considering it as different to other forms of bullying, which might support the idea that social media is an uniquely dangerous platform deserving of special attention. The one exception is perhaps the observation that “The most hurtful or distressing situations involved pictures or videos, in which individuals were asked or coerced into sending pictures of themselves or were covertly filmed or photographed”. This may warrant further attention (considering for example the concern over ‘revenge porn’ as it has been termed in the media.) There’s also little in the findings about whether the ‘social’ nature of social media makes this type of bullying more damaging, as potentially the bullying is public and visible to large audiences.

Link

Primary paper

Hamm MP, Newton AS, Chisholm A, et al. Prevalence and Effect of Cyberbullying on Children and Young People: A Scoping Review of Social Media Studies. JAMA Pediatr. Published online June 22, 2015. doi:10.1001/jamapediatrics.2015.0944. [PubMed abstract]