Let’s talk about ethnicity data

This article was originally published on Apolitical.

We’ve all been asked to tick a box defining our ethnicity at some point or another, whether when registering with a GP, enrolling a child in school, or signing up to a local service.

If you’re mixed-race like me, this can precipitate a mild existential crisis in the dentist’s waiting room. Which box should I tick? How many hyphens are appropriate? Is it even any of the dentist’s business? These are questions for another day. What I’d like to ask here is where does this data actually go? And what is it used for?

The answer, unfortunately, is probably nowhere.

Ethnicity data is an underused datapoint in public services. This is surprising, given how frequently this data is asked for. But although many services are asked to collect ethnicity data it is rarely acted upon, often appearing at the bottom of a report that hardly anyone reads.

It is seen as an optional extra, and I have sympathy for the people who collect it. Data entry is a chore, and if no-one seems to be using the data, you’ve painstakingly entered into a spreadsheet what most are likely to skip over.

Then, one day, someone will ask to see a breakdown of service participants by ethnicity, and this suddenly won’t be possible because of ​“poor data quality”. Sometimes partial information is available, other times the question is dropped.

Ethnicity data is underused for three main reasons:

It is difficult to analyse and draw conclusions from ethnicity data

Say you want to assess equity of access to your service, i.e. test whether you are reaching everyone or inadvertently discriminating against people based on their ethnicity. Sounds like a simple question to answer.

However, it is not enough to have accurate information on the people accessing your service. You would also need to know roughly the ethnic make-up of the area you work in so you can tell how many referrals to expect per ethnic group. But you will quickly find that this information is not readily available.

The last UK census was in 2011, meaning that more recent data must be derived from surveys that often do not have a big enough sample size from which to glean meaningful estimates.

Even if you do manage to find an appropriate data source, the categories used to define ethnic groups are not standardised, so you may find your service defines ethnicities in a completely different way to the dataset you are trying to compare against. All this means it is difficult to establish a baseline by which to assess whether your service is reaching the right people.

Ethnicity data is sensitive and must be collected lawfully

The General Data Protection Regulation (GDPR) protects our privacy and safeguards us against the misuse of personal data. This is a yawn-inducing way of saying that laws exist to prevent private companies, the state, and anyone else, from using our personal information for their own ends (nefarious or otherwise).

Ethnicity data is classed as sensitive personal data, largely due to a history of it being used to facilitate systematic discrimination and (in extreme cases) genocide and other crimes against humanity. There are reasons to be cautious about who collects ethnicity data, and what they’re allowed to do with it.

To collect ethnicity data we must be clear from the outset on why we need it. We must ask people for consent to use their data. We must state who it will be shared with. And we should, where possible, use it to benefit the person whose data it is, or the service more broadly.

For these reasons, the introduction of GDPR caused many of us in the social sector to classify ethnicity data as unnecessary: too costly and too risky to keep collecting. We didn’t consider ethnicity data important enough to collect it accurately, ask for consent, and invest time and resources into analysing it, so we lost the right to have this data at all.

We don’t ask questions because we are afraid of the answers

It is no accident that accurate data on ethnicity remains elusive. It is the consequence of a collective unwillingness to ask questions about the inequalities that affect racialised people in our social services.

We lack curiosity because we are afraid of what the data might say, what uncomfortable truths we may unearth, and whether we could handle the conversation around our findings. Better not to ask questions in the first place and to neglect the collection of ethnicity data to the extent that extracting meaningful insights becomes an impossible task.

What’s frustrating is that we don’t even know what we would learn. We could be missing success stories of services that are not just equitable, but actively addressing the unjust inequalities created by our society.

We could turn out to be mistaken in our assumptions that racialised groups fare worse than white people in many cases. But where our services are replicating the racial injustices of society at large, we should not shy away from investigating how and why they are failing.

Black Lives Matter has forced the social sector to begin to engage with the issue of structural racism in our society. But we need to go one step further than ensuring that our processes and procedures are non-discriminatory.

We must also ensure that the services we run are equitable and inclusive. Effective use of ethnicity data is key to this ambition.