Implicit Assumptions About Unconscious Bias
Gaming the Certs
In 1993 I wanted to be a manager. I was living in Boston and in my third year of working for a small software company based in Cambridge. I had already led technical teams in the field, but now it was time to go to HQ and rub shoulders with the bosses. The company hired a psychologist to give us the equivalent of a Hogwarts sorting hat fitment to our future roles in the company. I shied away from all the questions I could interpret as ‘individual contributor’ because basically I… wanted to be a manager. I remember two things about that experience. The first was that the psychologist who administered the examination and I talked briefly about the LA Riots which just occurred the previous year. He was truly shocked about what his psychological profiles of young blacks in LA had told him. In short, LA was just the beginning of the nihilistic potential of the next generation. Secondly, he told me that I would make a very good technical manager.
Maybe he was wrong, maybe he was right about both things, but my test results reflected the profile I wished to put forward, and I grew into it later. I never expected to be the same person my entire life. I did get the management position, but I thought I was Captain Picard. That didn’t work out so well. A couple years later I was clocking big dollars as an independent contractor.
But at the time who was I? I was the guy who never watched TV (I even got sick of Charlie Rose) but bought one and started watching Seinfeld and Letterman to click with my employees. I was the guy who was ready to do what it took to be successful at the goals I set out for myself, which were never easily alignable with any organization I had ever encountered. So, yet again, I was an extraordinary oddment carrying out the necessary tasks at the 65% energy level necessary for slightly upscale bourgeois employment. I still wrote poetry at night.
These days, perhaps you have noticed that HR departments at companies other than those in close proximity to the Charles River have also crept over certain boundaries involving the psychological fitment of their employees to proper profiles. I just recently took such a survey / test / remote-concept-learning thing. Via a disembodied voice and 40 odd panels of well-crafted slides, it walked me through it. Psych evals at scale is a thing, especially for remote workers like me. The subject of this s/t/rcl thing was ‘Unconscious Bias’. Naturally, I aimed to demonstrate my minimal subjectivity to such unbecoming behavior. It wasn’t much of a challenge because I’m Stoic and I don’t pay much attention to personalities.
It also happens that for several years I took various evolutions of Harvard’s Implicit Assumption Test (IAT). Pretty much every time, I demonstrated my racial bias neutrality. Back then (circa 2004 I’m guessing) it gave me some comfort that I passed given that I was something of a raceman. But according to my new friend Stefan, those in the psych sector of the Genius Class, know such tests to be.. erh uhm less than they appear to be. In fact, he turned around in rapid fashion a raft of skeptical reviews that I now present in digest.
IAT Scores, Racial Gaps & Scientific Gaps
1. There is no consensual scientific definition as to what constitutes implicit bias;
2. A definition of implicit bias offered by Greenwald (2017) as describing what it meant to scientists for the prior 20 years is shown to be logically incoherent and empirically unjustified;
3. Exactly what the IAT measures remains unclear, even after 20 years of research.
So it’s this Greenwald cat whose neck is on the line with regard to the applicability of IATs around the globe. Clearly a bunch of HR departments have just run with it. Skeptics like me are just getting their rigorous pants on. Thanks Stefan.
Hacking The IAT
Nevertheless I bothered to go check out the Harvard site and noticed a large proliferation of new IATs, not merely the ones all about race, gender and ethnicity, like the one from HR. I decided to take one and determine my neutrality on the subject of Health and Fitness. In this I basically know that I have a bias towards it. Nevertheless, I have two things working on my behalf that helped me to defeat the test and gain the desired neutral score. The first is my Stoic bearing that really is a proper Type 1 IDGAF. The second is in my newly hatched skills at piano (I’m about 40 days in).
You see the IAT is structured so that you associate a keystroke with a concept. They run you through eight cycles of a mix of associations with two different keys on the keyboard. They flash a word or picture on your screen and you punch the key with the proper association. They time your response with the assumption that if you think long you think wrong and your unconscious bias will revealed in blink think. Since I am a very elementary piano player, I can’t read this music at anything quicker than a slow metronome. So I play those two keys at about adagio speed, roughly 68 BPM. By playing them consistently at the same speed, I defeat the test.
I didn’t set out to do so, it just happened that way. When the associations confused me, and I hit the wrong key, I slowed down to the consistent pace and did the non-blink think. It helped that I said the words out loud. In the end I only made about 6 errors.
Downstream Implications
You and I know that it’s not just corporate HR departments who want you to prove beyond a reasonable civil suit that you are not a latent evil bigot. There are plenty of others who will make various bad faith assumptions about your ability to navigate their born-yesterday mazes of offense. It may very well be the case that IATs are as dubiously rigorous as the entire system of college admissions. That doesn’t change the fact that most of us peasants still leverage our identities based on decisions made about us when we were still in puberty. It’s just another institutionalized essentialism we should all work around, stoically. But now you know some legit geniuses have called them bunk, and you also know that IATs can be hacked. So while the religious devotees of social construction cannot be persuaded by reason, you might choose to send them some flavor of their unreason back at them in the form of a neutral IAT. Remember just like TikTok, you only send the version that makes it look easy.
This is, of course, bad Stoic advice. It is, on the other hand, excellent trickster advice. That makes this Samizdat. What do I believe? I believe everyone has unconscious bias about everything and now peasants are using computer proxy police to collect questionable evidence of thoughtcrime behind a moral figleaf of pseudoscience. Their schemes will be subverted. Conduct unbecoming is as it ever was, fragile dweebs everywhere still hedge their bets against common decency. Maybe that psychologist was right about nihilists.