Understanding children’s experiences to inform our online safety work

Protecting children is central to Ofcom’s duties under the Online Safety Act. To be an effective regulator we need a rounded understanding of, and evidence base for, children’s experiences online.

Our children’s research programme sets out to do this through an established media literacy research programme, our Online Experience Tracker (13+) and our growing suite of qualitative research. Alongside this existing research we continue to innovate and find new ways to understand children’s online lives.

In summer 2023 we published two reports on research pilots using Avatar and school survey methods. We also commissioned a Children’s Passive Measurement Study to test whether passive methodology can in the future deliver robust metrics of children’s online use.

We want to build on these studies and further develop our knowledge of useful research methods to understand what children do online and their experience of any harmful online content.

In autumn 2023 we organised roundtables in our London and Edinburgh offices to discuss the pros and cons of different research methods, bringing together experts from across academia, charities, government, and other regulators.

Five takeaways from our discussions:

1. We know some children are much harder to research than others, but this doesn’t mean it’s impossible to understand a wide range of children’s experiences. Schools are a good place to conduct research with most children, in particular because they provide a strong safeguarding infrastructure. However, they aren’t always an environment in which all children feel particularly comfortable. Youth groups and sports clubs offer an alternative setting outside of education where children are likely to be more relaxed and able to partake in research. Attendees also shared their experience of conducting research in non-mainstream schools, for example pupil referral units and the additional governance required to set it up.
2. Co-creation methods in which children are an intrinsic part of the development of the research project and materials, can be very valuable. They allow children to be part of the research process so that research is being done with them rather than being something that happens to them. However they can be very time-consuming and can sometimes lead to confusion about goals and who is in charge, and so need to be thought through with particular care. Mosaic methods involve taking a multi-method approach with children which centres their lived experience. This could include activities led by children including photography, giving tours of an area and map-making, alongside more traditional interviews. This is a way of making research age-appropriate and at a level children can engage with, but like co-creation it requires additional time and budget to conduct.
3. It’s important to allow children the opportunity to change their minds about their views during fieldwork, and to be able to modify their previous responses. Children, even more than adults, can take time to become comfortable in an interview setting, and therefore we can expect they might not open up until the interview is almost finished. This also means that long-running ethnographic research, which lets children become comfortable over time, can be particularly beneficial. Children also tend to be more comfortable speaking to other children than to adults, and participants told us about a research approach where children conduct the research themselves and are trained to ask other children questions.
4. Data donation is an innovative method for researchers to access data on a person’s online behaviour with informed consent. This donated data has the potential to provide researchers with a detailed picture of a person’s online activity. GDPR has set up the legal framework for this type of research, but there are not currently examples of it working at scale. Conducting research with donated data from online platforms involves administrative and logistical challenges as there is not a uniform approach across platforms for how users request and receive their data, and different platforms offer different types and levels of information within the data download.
5. Avatar research can be used for a range of research questions, such as testing out how platform algorithms work. This type of research involves the setting up and running of fictional accounts with the aim of understanding the content that is served to children online. Participants discussed the logistical challenge of how closely an avatar can mimic the behaviour of a child, and how the importance of the accuracy varies according to the aim of the research, i.e., whether the avatar aims to understand what a child may see online, or to conduct an AB test of a particular variable.

What we’ll do next

We are already bringing these five takeaways (and more) from the roundtables into our work, including planning for additional ways of researching children. We are continuing discussions started at the events - we are keen to hear from researchers about this work and are currently developing roundtables on other subjects.

Participants at our roundtables included:

Ofcom: Dr Alison Preston, Kate Reynolds, Dr Nicola Sides, Abbie Flewitt, Natalie Mawhinney, Grant McConnell, Dr Uche Onyekpe

External participants: Reid Acton (Young Scot), Kam Atwal (Advertising Standards Authority), Professor Deborah Fry (University of Edinburgh), Will Gardner OBE (Childnet), Dr Richard Graham (Online Harms Lead, Child & Adolescent Faculty, Royal College of Psychiatrists), Dr Lizzie Jones (Office of the Children’s Commissioner), Dr Anna Lavis (University of Birmingham), Fraser Macdonald (Data for Children), Professor Andrew Manches (University of Edinburgh), Izzy Millward (Office for National Statistics), Jill Morton (Scottish Government), Dr Carrie Myers (City University), Alison O’Connor (ICO), Dr Amy Orben (University of Cambridge), Hillary Phillips (YouthLink Scotland), Professor Andrew K. Przybylski (Oxford Internet Institute), Professor Noel Purdy (Stranmillis University College), Dr Eleni Romanou (NSPCC), Tom Sweetland-Michie (Education Scotland), Willow Warder (ICO), Chris Willman (ICO), Stuart Wood (Internet Matters), Dr David Zendle (University of York).