White and other defaults – learnings and resources from our webinar with Edafe Onerhime

 

Edafe Onerhime, Data Person - DEI Lead - Cat Mum

Last month we had the privilege of hosting Edafe Onerhime to give us her talk about White and other defaults, understanding assumptions – how they affect our data and our world. The view expressed by Edafe in this event are her own and don’t reflect that of her employer or organisations she’s associated with.

The talk cultivated a safe space to explore the assumptions we hold and how these may affect our data practice. Edafe shared with us how she became aware of her own assumption of “white by default”, how we may be able to recognise our own assumptions, and why it is important to do so. She shared some context around how the British empire and colonialism impacted data collection and use (and how data collection and use contributed to reshaping society). We learned how the effects of colonisation are still present in our use of data today – with the particular example of the UK census that makes it difficult to choose ethnicity options that are not white.

 
 

We also looked into the brain processes that cause us to form assumptions – using a theory from the book Thinking Fast and Slow by Daniel Kanheman. The theory states that we have two brain systems– system 1 that is automatic, biased thinking, and system 2 that is more logical slower thinking. Edafe challenged us to engage our system 2 thinking to try and resist an initial bias.

Key takeaways

For me, the main take-aways from Edafe’s talk were:

  • Data is not neutral

  • Think about what biases are baked into your data

  • It is not your fault that you have biases, but it is your responsibility to think about what assumptions you are making

After the talk, we had some time for the audience to pose questions to Edafe. The answers Edafe gave were thoughtful, nuanced and practical – I have done my best to summarise and package up some of that wisdom below.

How to influence decision-makers

One question asked for advice on how to better push for more inclusive data collection when not in a position of power to be able to implement these sorts of changes, particularly where you are the only black/non gender-conforming person in your company. Edafe prefaced the advice by stating how important self-care is, and to remember to put your health first before embarking on any form of activism. She then gave four methods that can be used to change a company’s mind:

  1. Evidence what they are already doing, eg. if you show how they are already putting lots of time and money into data collection, then you can say they may as well do it properly.

  2. Show them how it will save them money.

  3. Pull out the FOMO card – ask them if they want to be left behind.

  4. Talk to other stakeholders – sometimes getting external people to ask for the same thing can be effective.

How to (gently) point out biases

Another question asked for tips on how to highlight assumptions that other people may be making. This can be a difficult thing to do, depending on your relationship with that person. The first tip Edafe gave was to give the person ‘plausible deniability’ – assume the mistake is due to a lack of attention, not an inherent bias and lead them towards the realisation that something needs to change. For example, you may ask “who do you expect to be filling in this survey” if you think the questions do not serve the demographic. Finally, she reminded us to point out everything that is great about the project/proposal you are discussing, and frame everything as ‘we can improve this’.

What good looks (and doesn’t look) like

Someone asked if Edafe had any good examples of personal demographic data capture and she pointed us to look at the Canadian census. Part of her presentation questioned why ethnicity questions weren’t listed in alphabetical order – and how hard you have to work to ‘find yourself’ if you’re not white and top of the list.

Where we can improve

Finally, Sian asked Edafe to look 10 years ahead, and describe the changes she hoped to see. Edafe responded, both in terms of the census (that we had been discussing) and wider technological and data advances. For the census, she hopes to see the data collected more accurately and quickly. Since it holds such importance, and we now have the internet to aid data collection, this seems possible. In terms of wider development, Edafe spoke about the surge in AI and how this technology makes very quick decision with bad or biased data, so we need to be aware of this and be careful how we embed AI into society.

Further learning

Throughout the talk and following discussion, many useful resources were shared – both by Edafe and also from the attendees. We have listed them below.

Resources from Edafe:

Resources shared in the chat channel:          

What people said about the event:

“I hugely enjoyed this talk. I’m new to thinking about data and had always chalked it off as cold and faceless. Edafe’s stories brought the field to life in a way I could engage with and has really unlocked some of my thinking about how to connect with this domain. Thank you so much. I also really loved how everyone started sharing resources and compliments.”
— Attendee
“Edafe gave great examples and explanations for the ways that data is subjective and encodes bias and some questions to think through when shaping a new data initiatives, which will be really useful to bring up in discussions with peers and clients”
— Attendee
“I have straight away challenged my organisation to review the questions they are going to ask when looking at the next state of the sector review - and I will continue to do so, inspiring!”
— Attendee

Data Orchard runs a regular programme of data events for people in nonprofit organisations. Do keep in touch. You can:

 
 
Libby HarkinsData4good