Do Notting Hill Carnival and Glastonbury have similar crime rates?

Over the weekend HuffPost reported that new data shows that the arrest rate at Notting Hill Carnival is similar to that of Glastonbury.

There are problems with the calculation which mean that it probably underestimates the crime rate at both events, for different reasons.

As such we can’t be sure which event has the higher crime rate.  

Comparing Glastonbury with Notting Hill Carnival is not a like-for-like comparison

HuffPost took the number of arrests at each event, and divided it by the number of respective attendees. They found that there were 3.10 arrests per 10,000 people at Glastonbury and 3.76 per 10,000 at Notting Hill Carnival.

But this doesn’t account for the fact that Glastonbury lasts for five days, whereas Notting Hill Carnival lasts for two. Imagine that there were the same number of arrests and people at both festivals; you would say that the arrest rate was lower at Glastonbury than Notting Hill because it lasts over twice as long.

The same logic applies when calculating the number of arrests per 10,000 people. If, say, there are 3.1 arrests per 10,000 people at Glastonbury, then on any one of the five days of the festival you’d expect 0.6 in every 10,000 people to be arrested.

If there are 3.76 arrests per 10,000 people at Notting Hill, then on any one of the Carnival’s two days you would expect 1.88 in every 10,000 people to be arrested.

This is a hypothetical calculation, not our estimate of the true arrest rate at either festival (for reasons we’ll get on to). But it helps to illustrate why it’s important to account for the duration of the festival in the calculation.

The crowd size also changes throughout the day

Another problem with HuffPost’s attendance figures are that they don’t account for how the size of each crowd changes over time. Notting Hill Carnival is an event where people are coming and going and are likely not to go for both days, but at Glastonbury the attendees are usually there for the duration of the event.

It’s not only that Glastonbury is a longer event, but the average time a Glastonbury attendee would spend there would be much higher than an attendee at Notting Hill Carnival (who may only stay for a few hours or one of the two days).

So at any given moment you would expect the total crowd at Glastonbury to be close to 200,000 (the average crowd size used in HuffPost’s calculation). But at any given moment at Notting Hill you would expect the total crowd size to be somewhat lower than the one million people estimated by HuffPost.

Unless the calculation accounts for changes in crowd size, it cannot accurately measure what the typical arrest rate is at either event. Given that the crowd size at Glastonbury stays more stable, HuffPost’s calculation could well overestimate the arrest rate at Glastonbury, relative to Notting Hill Carnival.

Arrest rates aren’t the same as crime rates

Another problem with the data is using the number of arrests (converted into an arrest rate) to talk about crime levels. Arrest rates, by definition, can only tell you something about the level of crime recorded by the police (and even then, only crime that would ordinarily lead to an arrest, as opposed to offences that may lead to a caution or other action).

This means the data doesn’t cover the level of unrecorded or less serious crime.

What does that mean from a practical perspective? Well, from HuffPost’s data we know that far more police are deployed to Notting Hill Carnival than Glastonbury festival. And that special measures have been brought in during previous Notting Hill Carnivals to increase the police’s license to stop and search, likely leading to an increased number of detected crimes.

So a higher proportion of crimes which occur at Notting Hill Carnival are likely to be detected (and so more crimes are likely to lead to arrest), compared to at Glastonbury. That means it’s possible that the arrest rate at Glastonbury festival underestimates the true crime rate to a greater extent than at Notting Hill Carnival.  

So in summary HuffPost’s calculations are likely to have used an attendance figure which understates the rate of arrests at Notting Hill Carnival, compared to Glastonbury.

But at the same time it used a metric for measuring crime which understates the level of crime at Glastonbury, compared to Notting Hill Carnival.

It’s possible that these two factors balance each other out. But we can’t be sure.

And so we can’t really use these figures as any reliable indication of the crime level at either event.