[ng12]

> you write him off as an idiot and forget him

Not really. I have a lot of faith in our engineering staff so I wonder how he got hired and how he avoids getting fired. The uncomfortable truth is that if he was an employee that "looked good" I would know the answer to those questions.

[praestigiare]

As I said before, my response is not data, I know that. But what you just wrote is exactly how confirmation bias works: If he fit your narrative, he would confirm your narrative. That is why these discussions need to be about data, and not "what everyone sees and no one talks about."

[ng12]

My point is more that the anecdotal stuff matters -- by being ham-fisted in your attempts to address diversity you reduce inclusiveness.

[richmarr]

Your faith in your engineering staff is beside the point. Bias is human nature. Unless you're actively structuring your processes to exclude or minimise it then you're certainly suffering from it, the only question is how.

It's not specifically gender or racial bias either. We ran a study of 700 candidates for a role, running regular hiring against blind assessment and found that the major difference for that role in that organisation was socio-economic... they'd been excluding great candidates who went to less prestigious universities.