Thread by @thomasforth

I can't recommend this piece by @sundersays highly enough. theguardian.com/commentisfree/…

Diversity is one of the biggest things we work on @ODILeeds, in part because our sponsors ask us to. It is a challenging topic, and one where the culture and rules of statistical and data analysis can be insulting to people. So we have to tread carefully. odileeds.org/blog/2020-09-1…

I think that data and statistical analysis can add value, help us identify problems, and make improvements. My diversity explorer here is a tool that a lot of people use. I look again every few months to see if/how we need to update it. Hard but important. imactivate.com/diversityexplo…

In the most recent update, we added the option to split all of the data analysis out into much smaller groups of self-described ethnicity. And included a discussion in the blog post (I think still poorly understood, I need to do better) about the drawbacks of this statistically.

Currently the largest thing we talk about and read about is the naming of groupings and the components of groupings. It is a challenge that people like me tend to struggle with. Computers and data do not deal with "it's complicated, varying, nuanced, multiple, and disputed" well.

There are lots of problems with all of BME, BAME, and ethnic minorities. If you're building a website, you probably have to choose one, or exclude some people with extra complexity. If you choose to do no grouping,... you're almost certain not to get statistically valid results.

At the furthest line of thinking about groupings and avoiding putting people in boxes,... you've converged on the French cultural norm. That race (and religion) is not a way of distinguishing people that the state should be involved in. Like we don't collect data on hair type.

That is a position that I find intellectually attractive. But so do loads of very racist people whose reason for supporting it is that they don't want the state to collect the data that would prove the existence of the widespread racism in society that they deny.

It means for example that when Macron said "when you have a skin colour that is not white, you are stopped much more (by police). You are identified as a problem factor. And that cannot be justified" he had no data to prove him right so he was challenged. france24.com/en/live-news/2…

Another good piece on this topic, this time by @samgyimah. The "it's good business" case for making the most of the talent that exists is one of many very good cases. thetimes.co.uk/article/ethnic…

One of the companies that I experience understanding and succeeding because of this is Sky. There's a reason their inclusion page is part of their hiring effort. They do a lot in Leeds on this, because that's how they get better employees. careers.sky.com/inclusion/

Every talented, driven, smart, motivated person that a company didn't hire due to their bias against them, whether it's because of their race, gender, family situation, age, haircut, clothes, accent, or something else, is a person that their competitor can hire and win with.