Welcome to Science with Shrike! Today we return to one of the more contested ideas in academia and broader society, that of “diversity, equity, and inclusion”. Many have strong feelings on one side or the other for this issue without being able to understand the opposition as anything less than evil. We’re going to dig into some of the priors underlying these beliefs.
Identity Politics
The first thing we need to get out of the way is the political dimension. Identity politics work, and many cynics on both sides realize that “diversity, inclusion and equity” (DIE) all hit on core identities that can be used to persuade people. So from a political aspect, the reason to push it is that you can speak to identities, and as Scott Adams says, identities persuade better than emotions or facts.
But we are going to dig deeper here because the ideas beyond the affinity groups resonate with many people. On one hand, it’s giving opportunities to people who didn’t have them before, and obtaining a larger pool from which qualified applicants can be drawn. On the other hand, the demand for certain affinity groups outweighs the supply, so the quality suffers. This leads to a tension where people are encouraged to redefine “excellence” and lower their standards. There is a tension created here based on an assumption that many jobs cannot be done by just anyone.
But it is this exact assumption— that anyone can do a job— that is one of places where DIE proponents and opponents talk past each other. Many DIE proponents work in a bureaucracy, where chatGPT and/or a pretty pet rock can replace them. Many DIE opponents work in competitive or highly skilled fields where only the top 0.1% or less win. This illustrates a fundamental difference in mindset in jobs.
Threshold jobs vs competitive jobs
Both parties are correct. To illustrate this, we will categorize jobs into two major buckets, what Shrike will term ‘threshold jobs’ vs ‘competitive jobs’. A threshold job is one where anyone who meets a specified set of criteria can perform the job to the same degree as most everyone else who meets those criteria. Think of your fast food cashier. Does a top 0.01% cashier look different than a top 1% cashier than a top 25% cashier? They all take your order and money. In threshold jobs, once past a certain threshold, the performance (output) is the same, regardless of the person’s talent or skill.
But what about cashiers who show up late or mess up orders regularly? These are great questions you might be thinking. We need to clarify that the threshold in a ‘threshold job’ is not ‘the minimum needed to keep you from being fired’, but ‘ability to accomplish a set of tasks’. That means showing up on time is one of those tasks. So the perpetually tardy person might be employed in a position, but failing to meet expectations. Messing up orders is similar, but a little more complicated. You might try to rank cashiers by how many orders they screw up. Do they mess up 10%, 1%, or 0.1%? This comes back to the idea of a threshold. What is the acceptable error rate? That error rate is your threshold. Also, note that the consumer will not see much difference between a 1% error rate and a 0.1% error rate, even though it’s a 10x difference.
Most jobs are threshold jobs. Anyone with a defined skill set can do the job.
In contrast, there are “competitive jobs”. A competitive job is one where your job requires defeating other skilled and talented opponents. Think of the NFL, NBA or other pro-sports. All of the athletes at that level and seeking to be in that level have the required skill set to perform the job. The difference is that there are an estimated 4400 NBA players and ~25,000 NFL players in the entirety of those sports. Relative to the 340 million people living in the US, that means NBA players are top 0.00129%, and NFL players are top 0.00735%. Being in the top 0.01% does not cut it for either sport.
Competitive jobs are winner takes most, and there are huge differences in performance at great and elite levels.
Jobs drive mindsets
Jobs where value accrues to the top require different personalities and approaches for success than jobs where value flatlines above a certain threshold. Notably, once you cross the threshold, extra skill at the job does not translate to extra success in threshold jobs. In these jobs, you will notice people who are ‘adequate’ meet the expectations of the job. People who are ‘good’ at the job stay above the threshold almost all of the time. People who are “excellent” at these jobs take on additional jobs. This is why some retirements require replacement with multiple qualified people. High performers need to do other things with their time, even when they do not like being competitive.
This lends itself to a threshold mindset, where people (especially bureaucrats) believe that meeting a bare minimum set of criteria is all that is necessary to accomplish most jobs. Competitive people are a challenge to these people because to the bureaucrat’s mind, competitive people waste their time trying to be 0.1% when the top 10% will do. This idea that ‘everyone past a minimum criterion is equal’ also forms the basis for “equity” in employment.
In contrast, competitive jobs require all the focus possible because the top 10% are playing in a different league from the top 0.1%, who are different than the top 0.01%, who cannot reach the top 0.001%. There is often no ‘bare minimum’ in these positions because someone else can take your job if they can outcompete you. You must be self-motivated and/or competitive enough to win. You also need high standards because you must outperform other people. This is not the case in a threshold job. Many competitive types do not realize this, and waste their time and energy trying to make a threshold job into a competitive one.
One factor exacerbating the threshold vs elite job mix-up is that it is often intentional. Elite jobs carry more status because they are competitive. Some jobs, like being a physician, have high thresholds that not everyone can make. This is combined with artificial scarcity to create the impression that the job is competitive and thus high status. In some cases, real scarcity of positions makes a job more competitive than it is (ecology, I’m looking at you). Many people will say their job is competitive, when it is a threshold job with either low demand for the job (ecologists), or it requires a complex skillset (most physicians). On the flip side, there are some competitive jobs where the demand is huge, so there are not enough people to fill all the positions available. Shrike would put sales in this last category.
Adding to the confusion, there are some jobs that are treated as threshold positions at some institutions, but are competitive elsewhere. University professors fall into this category. 90% of university faculty have threshold jobs. Do some teaching, maybe play at research, and move on with your life. This is why tenure is hated: everyone is jealous of the job security anyone with the right training can do. However, the top 10% of university professors play in a different arena. They compete for NIH or other agency grants and develop new knowledge no one else can. The Howard Hughes Medical Institute has ~300 investigators. Getting in puts you in the top 0.000088% of people in the US. There are 2000 members of the national academies. Competition for these positions and to be the top is intense. It’s a different world working with HHMI investigators compared to a professor who teaches a few General Biology classes each semester. Many competitive jobs have threshold alternatives (scientist, chef, financier, sales etc) for the lower tier institutions.
In general, there are two ways to separate the type of job. 1) Does the job require being better at it than other people with the same job? 2) If you had a top 0.01% person, a top 0.1% person, a top 1% person and a top 10% person, would you see a meaningful difference in job performance, success, or in some cases wealth?
These two considerations are abhorrent to a threshold mindset because threshold jobs require a minimum level of effort that anyone with the right training can accomplish. They do not understand that there are degrees of competition and success, nor that there are well-trained winners and well-trained losers. They think everyone can be a winner, if only they acquire the needed skills. This also lends itself to ‘skills over talent’ beliefs.
DIE comes from a threshold mindset
To a threshold mindset, DIE is the inevitable conclusion. If everyone with the right training can do the job, all jobs could reflect the demographics of the greater population. If they do not reflect the demographics, there must be something interfering with people of the excluded demographic entering that job. To a threshold mindset, since anyone with the necessary skills can do the job, racism, or exclusivity must be the prime suspects. Perhaps the excluded demographic does not want the job, because they feel excluded. More inclusion to get more of the target demographic to apply. Or maybe the ones in are not going anywhere. Must be racism, so to solve that problem, diversity needs to be valued more.
The nice thing about the problem of “structural racism” is that it has no face, so it is easy to fight and define. You are combatting structural racism in your actions, and your enemies are being racist when they contest your actions. Structural racism will last until the demographics are satisfactory, because the threshold mindset believes success is a matter of receiving a baseline of training, and acting on that training. Amount of skill, and talent, do not matter in threshold jobs, so why worry about it when hiring?
Both diversity and inclusion pair well with equity. Shrike has previously written about the roots of equity in wanting everyone to succeed. Equity is a well-meaning solution that warms the heart of everyone with a threshold mindset: how do we get everyone to the level of skill needed? To a competitive mindset, this does not happen. For every 100 people in a competitive field, the top 1% is one person. That is the winner, and the rest are losers. To a threshold mindset, it could be the top 60% are interchangeable. The problem becomes how to get it to 100%, and/or get 1000 people into the field. These are equity problems.
Advantages of DIE approaches
One strength of DIE approaches is that DIE is one solution to scaling and to teamwork. If you need to hire many threshold positions, approaches that promote diversity and inclusivity will give you a deeper job pool. If you need to get 80-100% of your team to do something, an equity approach will outperform. This works for threshold jobs and tasks. This also works for threshold jobs that pretend to be competitive. In the corporate world, improving compliance with target goals improves efficiency. Equity approaches help you to get everyone to complete the CYA training. There is an advantage to DIE approaches when the job can be done by anyone with a reasonable skillset.
Weaknesses of DIE approaches
There are three key circumstances that kill DIE approaches in time. The first is a competitive job. If you must be in the top 0.1% to realize value, you do not want a deeper pool. That’s more competition. 100% of top teams comply with key job functions because they all want to remain in that top bracket. The motivation comes from the desire to win. DIE is useless in a competitive setting because it takes more than crossing a threshold to win. As a result, the winners get there without DIE.
Artificially inflating status or positions of people who are not competitive via DIE initiatives is not self-sustaining long-term. One example of this is the NIH’s hand-wringing over the lack of Black Principal Investigators. Various explanations exist, though the hypothesis that Black investigators are less competitive than other races due to universities promoting less-qualified Black applicants remains the forbidden idea. Suppose you have a postdoc candidate pool of 1000 people, of whom 5% are Black. Next, suppose that this applicant pool is competing for 100 jobs. In theory, one needs to be in the top 10% of the applicant pool in order to get hired. We will assume all races are equally qualified, so in this scenario, the top 10% consists of 5 Black investigators and 95 other investigators. Now what happens if 20% of those jobs hire a Black person to meet their local diversity/demographic goals? In order to fulfill demand, the universities need to select the top 20 Black applicants out of 50 total Black applicants. That is the selecting the top 40% of applicants. Meanwhile, the remaining 80 people are selected from the 950 other candidates. These candidates are the top 8.4%.
Now in a threshold job, the top 40% could compete with the top 8.4%. But NIH grants are competitive. This means the universities have set their faculty up for failure by hiring weaker applicants. Weaker applicants are going to fare worse in when competing for NIH grants because NIH has eliminated as much bias in grant peer review as possible. To a competitive mindset, the failure in Black NIH grant success comes from a top 40% person facing off against a top 8% person. The top 8% person will win most of the time when competing. To a threshold mindset, racism or something else must be the cause because both people have the needed skills and are therefore equal.
Competitive jobs are one main area where DIE fails, and are often highlighted as the key weakness to all DIE approaches.
The second circumstance that murders DIE approaches is when DIE becomes its own point. When there is pressure to hire a candidate based on a gender, race or some other category, the system breeds resentment. Competitive people do not like losing. Losing to unqualified people is even worse. On the other hand, people who want to be competitive start competing on checking enough DIE boxes. So this changes incentives from ‘good at job’ to ‘checks the needed DIE boxes’. This is how the DIE system becomes self-perpetuating and divorced from any utility it may have given. This is one of the challenges playing out now in the US.
The final death knell for DIE approaches is that laziness undermines equity. Equity places the burden of success on the trainer and/or employer, not on the employee. Many employees realize this, at least in their subconscious, and lower their effort to match. Since many of these practices operate at scale, this starts a cycle of lowering expectations. This provides for the degradation of the skills needed for a threshold job.
The combination of laziness and DIE for its own sake becomes an unmitigated disaster for the company or institution, but not until the skills are degraded below the necessary threshold for the job. As long as the skills stay above the necessary threshold, the system will still work. Humans being short-sighted, it will not be until the skills dip below the necessary threshold that a problem will be observed.
This leads Shrike to two predictions that differ somewhat from the ‘diversity hires in aviation will cause more plane crashes’. The first prediction is that it will not happen until standards are lowered below the threshold necessary for the job. The second prediction is that it will not be restricted to people hired for DIE reasons—it will occur across the board.
The good news is that as long as standards remain above the bare minimum, the impact will not be catastrophic. The bad news is that change is unlikely to occur until a catastrophe happens. When a catastrophe happens, the laziness will make it harder to change.