Can we really compare the numbers?

What did the numbers actually measure?

The figures came from the same international survey of more than 7,600 young people aged 8 to 17 across 25 countries. At first sight, that makes the comparison look straightforward. But the researchers knew that the word cyberbullying might not mean the same thing everywhere. Rather than relying only on that label, they asked young people about particular online experiences, including being treated in a mean or unfriendly way, being teased or made fun of, and being called mean names.

The researchers themselves acknowledged that people experiencing these behaviours might not necessarily describe them as bullying. So the dramatic figures at the top of the page have already become more complicated. They are not simply a league table showing how much bullying exists in each country. They show how young people in particular samples responded to particular questions about experiences online.

That is still valuable information. It is just not quite the simple international comparison the percentages first appeared to offer.

The question can change the answer

Research in Austria shows how dramatically this can happen even within one study. Petra Gradinger, Dagmar Strohmeier and Christiane Spiel asked 1,150 young people about cyberbullying in different ways. When they used a broad question about cyberbullying, 6.3 per cent were identified as occasional cyberbullies. When the same young people were asked instead about particular online behaviours, the figure became 16.2 per cent.

Nothing had happened to the young people between the two measurements. The researchers had changed the way they asked. Even the frequency threshold used to decide when something counted made a difference. A percentage can therefore look wonderfully precise while depending on decisions that most people reading the final figure never see.

So why compare countries at all?

Because good international comparison can show us things that are almost impossible to see from inside one country. When researchers ask broadly similar questions of large groups of young people at roughly the same time, we begin to get an international snapshot.

The World Health Organization's international study of school aged children, for example, gathered responses from nearly 280,000 young people aged 11, 13 and 15 across 44 countries and regions in its 2021 and 2022 survey. Because participating countries followed a common research protocol, the results allow researchers to look across the wider picture and explore whether particular experiences appear across many different societies, whether patterns vary between groups and whether those patterns change over time.

Another major international education study asks 15 year olds in many countries about particular experiences at school. Because some questions are repeated in later survey years, researchers can examine change as well as differences between places. These international snapshots cannot tell us everything, but they allow us to stand far enough back to notice patterns that may be difficult to see from within one country alone.

The purpose is not to discover which country has won or lost. It is to ask:

What might this international picture be showing us?

Even a common survey cannot remove every difference

Using the same research design helps enormously, but it cannot make every context identical. Questions still move between languages. Young people may understand experiences differently, and their willingness to talk about humiliation, aggression or exclusion may be influenced by the family, school and wider society around them.

Age and timing matter as well. A survey asking about the previous two months may produce a different picture from one asking about the previous year, while research involving 15 year olds is not automatically comparable with a study covering children from 8 to 17. This does not make international research unreliable. It helps explain why careful international research spends so much effort trying to make comparisons fair.

Sometimes the research design is the story

The Austrian study gives us a particularly clear example, but the same issue appears throughout research. Before a percentage exists, somebody has had to decide who will take part, what they will be asked, what period of their life they should think about and what threshold will be used before an experience enters the final statistic.

Those decisions may all be reasonable. They become especially important when the final percentage is separated from the research that produced it and turned into a headline. A figure from one country can therefore be perfectly sound within its own study while still being a poor comparison with a figure produced somewhere else.

The number matters, but so does the research that made it.

International snapshots rather than international league tables

This is why AntiBullyingWeek.co.uk will not publish a league table ranking countries by bullying prevalence. That does not mean avoiding international statistics. We want to use them where they help reveal wider patterns, identify questions worth investigating or show change over time.

A surprising difference between countries might reflect a genuine difference in young people's experiences. It might also tell us something about recognition, reporting or the way the research was conducted. Often the useful work begins when we ask why.

Our own visitor study

The same caution applies to the small study we are beginning on this website. Visitors will choose whether to take part, so the people who respond will not represent everybody in their country.

If people from different countries, languages or age groups respond differently to the same fictional situations, that could give us something interesting to explore. It would not justify saying that everybody in those countries thinks the same way. Being clear about what a study cannot tell us is part of taking seriously what it can.

Looking behind the number

When we encounter an international statistic, the useful response is not automatically to trust it or dismiss it, but to understand a little about how it was made. Knowing who took part, what they were asked, what period they were remembering and what threshold researchers used can tell us whether another figure really offers a fair comparison.

Sometimes the comparison survives that scrutiny. Sometimes its meaning changes considerably, and sometimes the difference between the studies becomes more interesting than the difference between the percentages.

The number matters, but so does the story of how the number was made.

What have we learned by looking internationally?

We began with words that do not translate perfectly and ended with numbers that do not always compare perfectly. Along the way we have explored how different places recognise harm, decide when action is needed, understand power, learn about conflict and share ideas across borders.

None of this means that international comparison is impossible.

Translation is not an all or nothing business, and neither is comparison.

We can learn from similarities without pretending that everything is the same. We can learn from differences without turning countries into winners and losers.

Perhaps that is the most useful lesson of International:

We do not have to agree on everything, or experience everything in the same way, in order to learn from one another.