Me, trying to sell my students on bias tests, but also warning them: It shattered some of my smug, liberal illusions.

Update: they've complained to my department head that cognitive bias has nothing to do with audio engineering

Show thread

Implicit bias exercise: implicit.harvard.edu/implicit/

These tests are US-centric and use terminology like "european american" instead of "white", - if you feel confused about US terminology and cannot find a good definition, feel free to email for help.

Please pick a test that is related to this peer assessment or you may wish to explore biases that could affect other assessments you may be asked to in future. This could include race, age, gender, disability etc. Some categories such as "weapons" are uniquely american and probably not applicable. These tests try to collect a lot of demographic information at the start, which you can decline to answer.

You will never be asked to share your personal results. Part of the aim of this exercise is to consider how to recognise, assess and mitigate bias - a core skill in engaging with ideas as part of your degree programme. This is a key skill that underpins academic work in much the same way as essay writing skills and referencing.

Show thread

Obviously, the problem is that I didn't sell it very well and they were confused and possibly feeling attacked or defensive.

So there's some text to try to clear things up afterwards

Show thread

What does anti-bias training have to do with audio engineering? 

What does anti-bias training have to do with audio engineering?

One time, a film maker called me up because she wanted to use a hubub of voices as a music statement but it wasn't coming off. I listened to the audio and knew just the thing! A resonant 70s style low pass filter would keep the timbres but jettison the words. I ran her vocal recording through my synthesiser and proudly returned with the improved audio.

She listened. "This does sound a lot better," she said, "but you've completely prioritised the men's voices and removed the women."

Cis normativity aside, she was right. My process had made altos much softer and pretty much nuked sopranos. I had to go back and organise each audio sample by vocal range and run the filter multiple times. It was time consuming, but the result was much nicer.

She noticed this because she knew what it was to be silenced according to vocal range - something that has gradually faded from my mind as its no longer a regular occurrence for me. If I'd been working with a fellow tenor or a bass, we might never have noticed. The audio would have been worse and the while project subtly effected by unconscious bias.

Instead, when the film was released, a reviewer praised the sound in that section.

Unconscious bias isn't just unfair, it makes *your* work worse.

Show thread

What does anti-bias training have to do with audio engineering? 

@celesteh I was expecting a joke punchline and instead got a really informative and wholesome post, A++ would boost again

What does anti-bias training have to do with audio engineering? 

@seachaint @celesteh I was also expecting a joke, something about a bias signal (DC component) that would damage speakers. Nice insight instead 🎉

Follow

What does anti-bias training have to do with audio engineering? 

@caffeine @seachaint

Me to first years: Bias is bad

Me to second years: You must bias your transistors

· · Web · 1 · 1 · 3

What does anti-bias training have to do with audio engineering? 

@celesteh @caffeine
Anti-bias is based, but you gotta bias bases

Sign in to participate in the conversation
LGBT.io

We are a Mastodon instance for LGBT+ and allies!