With the new administration, there’s much talk about DEI (Diversity, Equity, and Inclusion) and its systematic elimination from our government. For better or worse, DEI has become a dirty word. We treat it like the new boogeyman. We don’t need to know what it actually does; all we need to know is that it’s bad. Gone are the days of COVID-19, when Trump could blame China (and transitively target Asian Americans with hate crimes). He needs a new target for his blame; this has become DEI.
Naturally, nothing is ever that simple. We often wince at the label ‘socialist’, and it’s weaponized to elicit images of communist countries. However, we’ll object to the elimination of social security. We complain when our roads and bridges deteriorate into disrepair and demand their maintenance. We envision retirement and access to Medicare. All these programs are forms of socialism, yet we don’t object to them. However, despite the name social security, we don’t think of it as socialism.
DEI has been presented as the antithesis of meritocracy. People, disproportionately Christian, Caucasian, and heterosexual men, have argued that employment and compensation should be colorblind and that the most qualified candidate is the best candidate. While this sounds perfect, like a physics problem with point masses and perfectly frictionless surfaces, the truth is much more nuanced than that.
Homogeneous versus heterogeneous populations
The most versatile team is composed of different parts. If the most valuable football position is the quarterback, why not build a team with the eleven best quarterbacks in the league? We understand that building a football team this way is absurd, yet we do not hesitate to build our workforce precisely this way.
Years ago when I established this blog, I wrote a profile page. At the bottom of that page, I list my top five strengths from my assessment: Relator, Maximizer, Learner, Ideation, and Strategic. To maximize the overall strength of our team, I’ll want to team up with someone who possesses different strengths than me. If we can accomplish the task more easily with adaptability and strategic strengths, we may not individually have them, but we’ll collectively have them.
I’ve made the case that having a more diverse workforce, one that represents your target users, will lead to better products. I can list product problems that might’ve been caught by a diverse workforce. In the case of Fitbit, having more dark-skinned employees might’ve identified the problem before launch. In the case of Apple, having more Chinese women employees might’ve addressed the issues with Face ID.
Even if the most versatile tool is a hammer, having a toolbox with only ten hammers (and nothing else) isn’t particularly useful. Heterogeneous is another word for diverse.
Myth #1: a diversity hire is unqualified
Some assert they are uncomfortable boarding a plane with a black pilot. These people imply that because the pilot is black, their qualifications naturally come into question. These assertions are tragically flawed for several reasons:
- That pilot’s race (or gender, sexual orientation, etc.) may not have contributed to their employment. The airline may have indeed objectively picked the best candidate for the job. They just happened to be black.
- Assuming that the first time you see the pilot is as you board the plane, you know nothing about their competency as a pilot. To imply that they may somehow be less qualified because they’re black is, to put it bluntly, racist.
- The standards do not change because they’re a diversity hire. Suppose an airline employs 1000 pilots and this diverse pilot ranks about 500th among that bunch. I understand the desire to have your pilot as one of the top 10% of the pilots in that airline. However, suppose that as you board a plane, that white male pilot ranks 990th among those 1000 pilots. Is that cause for alarm?
While I don’t believe we should lower the standards for diversity hires, intentionally selecting a diverse candidate who meets all qualifications over a non-diverse candidate does not imply they’re unqualified. They are, by definition, qualified for the job and diverse; the two are not mutually exclusive.
Myth #2: by not thinking about diversity, we pick the candidate strictly on merit
Some will argue that the hiring process should be colorblind, excluding any consideration for race, gender, religion, sexual orientation, or gender identity. By excluding all those factors, we can be objective and only look at their qualifications. They will also solve a physics problem with point masses, no friction, and perfectly inelastic (or elastic) collisions, believing this is practical. It’s not.
Allow me to frame it this way. Only 1% of the population is ambidextrous; the remainder of the population is dominant on one hand (only 10% of them are left-handed). This means that 99% of the population is biased. Studies conducted with different names on resumés, demonstrate that racial bias still exists. It’s not as if these companies want to discriminate. However, they read a name like ‘Robert Smith’ and implicitly believe that he’s more qualified than ‘Jamal Warner’ or ‘Stephanie Mills’, based on nothing but their names.
What about gender bias? Some believe we can hear the same response from men and women and evaluate them as equally qualified. Studies show that women can say the same thing and be judged more severely. Keep in mind that we don’t do this intentionally. When someone tosses a ball to us, we instinctively catch it with our dominant hand without thinking about it.
The cold hard truth is the moment we discover anything about the candidate, it biases us. This could be their name, gender, religion, sexual orientation, or gender identity. To pretend that we can know anything about the candidate and evaluate them strictly on their merit without bias is a fanciful bit of fiction.
We generally favor straight white men. Strangely, no one seems distressed about boarding a plane with a Caucasian male pilot who may be less qualified due to implicit bias.
How do we fix it?
Studies have demonstrated that we discriminate. Much like 99% of us have hand dominance bias, let’s stop pretending that bias doesn’t exist. Instead, we can treat it like real physics problems; we account for gravity and friction.
First, we optimize machines by minimizing friction; friction increases wear and tear and generates heat. How do we minimize bias? We limit the information that we know about the candidate. There’s a reason why we don’t ask candidates about religion, marital status, sexual orientation, etc. Unfortunately, gender and race are much easier to determine by simply meeting the candidate (or even reading their name). My friend’s company uses TC (for The Candidate) to conduct all interview discussions; no names nor pronouns. By eliminating unnecessary references to the candidate, we minimize bias.
Second, we account for gravity by adjusting the aim to our target. Understanding that gravity will pull an object down, we aim a little higher. This means that if we know that we favor straight white men implicitly (plenty of studies substantiate that), we’ll adjust to favor the more diverse candidate as a tiebreaker.
Isn’t that discrimination?
Are you distinguishing between candidates? Absolutely! Let’s say you’re creating a product, like Fitbit, and your target users are the entire US population, then I’d expect your employee base to reflect that population. This means your employees should be roughly 50% women, 12.5% black, 5% Asian, etc. If you don’t then you’ll run the risk of:
- Your products (like the Fitbit Ionic) are too large for women’s wrists.
- Heart rate monitors do not work well with dark skin.
- Text notifications do not display Asian characters.
I am suggesting that from a pool of fully qualified candidates, pick the one that brings you closer to reflecting that population. If after months of searching, you only find one qualified candidate (and he’s a white man), offer him the job.
Does this mean that white men will get fewer job offers than they did? First, yes, it does, but studies suggest that companies already disproportionately favor white men among fully qualified candidates; we’re effectively offsetting that bias. Second, conversely reflect upon all those qualified women and people of color who did not get that job offer due to unintentional bias; does that seem fair? Third, once your workforce reflects your target population, then feel free to relax those standards. White men make up about 30% of the US population. If only 30% of your workforce are white men, feel free to disregard it.