Sym Roe

Unbiasing

June 28, 2016

I was invited to speak at an event at Newspeak House the other day about the London Civic Tech community’s reaction to the ‘Brexit’ result.

I wanted to talk about something that I’ve been mulling over for a while now called ‘unbiasing’:

Making the unconscious conscious will help you make more objective decisions, facilitate inclusive interactions, and create opportunities. Begin unbiasing with education, accountability, measurement, and more. withgoogle.com

This was partly promoted by this (intentionally provocative, I expect) paragraph in the event description:

Calling angry technologists. We think we know something other people don’t. Let’s do a bit of a post-mortem of the referendum, and talk about how we might be able to apply our skills to make politics better so we can start feeling proud of our country again.

I wanted to challenge the attitude that “We think we know something other people don’t” or that “our skills” automatically give us the ability to “make politics better”.

The ability to write code doesn’t enable you to “fix” society, it enables you to write code.

Code for code’s sake is unlikely to fix anything, so it’s worth starting with an idea, and then validating that idea before writing any code (of course, this applies to non-code based projects too).

When validating ideas, it’s important to recognise that your bias will change the way you look at the world. Because of this, you should try to do anything you can to ‘unbias’ yourself (or, perhaps, learn others’ bias).

There is a high chance that your ability to write code is related to your background, wealth and free time, not your intelligence or social/political outlook.

In the talk I read out a list that I’ve written for myself to try and internalise as much as I can. It’s not perfect and I wont always remember to apply it.

A few people asked me to publish it, so here, in no particular order, is what I wrote.


Make the assumption that you don’t have a representative range of views on any topic at all, especially the ones you’re sure you have a representative range of of views on.

Resist binary options, things are always more complex. Try adding a 3rd option wherever possible.

Questions that can be answered objectively don’t consider how people relate themselves to the question, or the answer. e.g. “The sky is blue” is not as simple a statement as you may think. People’s sense of ‘self’ varies from culture to culture – I could go on.

Try to write down your assumptions as often as you can. If you don’t think you have any, question why and write down ones you think are far too obvious.

When in meetings or public gatherings, note down the mix of gender, age, and other demographic that you can guess at from looking at the people there. You won’t get everything correct, but it will let you reflect on trends over time.

At these same events, try to answer the question “who is unlikely to be here because of the time, location or make up of this event”. If you don’t know, try to find out.

When considering numbers of people, try to mentally frame the number in absolute terms, for example work out the percentage of the population that number reflects.


I would really like to hear feedback on these and particularly improvements or new points that I can try to absorb in to my thinking.

Unbiasing - June 28, 2016 - Sym Roe