4 April 2016

Seeking Mediocrity: Humans are Hardwired to Avoid the Worst, not to Seek the Best Outcome

You probably know about or at least heard of loss aversion: people dislike loses more than they enjoy equivalent gains:

Incurring a loss of 100$ hurts roughly twice as much as gaining 100$ brings pleasure.

We know that all humans are loss averse, to a greater or lesser extent. Why this is the case?

Living on the edge

One very elegant explanation comes from evolutionary psychology. Our evolutionary ancestors lived in environments with relatively scarce resources. Simply put, the resources available in a certain area allowed for survival and successful reproduction, but not for much more.
In this living on the edge setting, it is only natural to develop an adaptation that says: not losing what resources you have is far more important than acquiring new additional resources.

When one has barely enough resources to survive and successfully reproduce, not losing current resources is much more important than acquiring new resources.

Loss aversion is, probably, the most prominent element of an adaptation with much wider effects in judgement and decision-making which I will call a preference for mediocrity.

Economic theory sees people as maximizers – trying to make the best possible decision, trying to get as much benefit (utility) out of their actions and transactions. This is not exactly senseless. After all, who would want to get less when she can get more?

Yet, research in experimental economics and behavioral science found that people are not exactly maximizers; rather they are satisfiers. In plain language:

People don’t necessarily go for The Best, rather, most settle for Good Enough.    

This preference for mediocrity (something that is good enough) makes a lot of sense from an evolutionary perspective.

Imagine our evolutionary ancestors in the African Savanah looking for berries. There are three types of berries: (1) OK taste and OK nutritional value, (2) Delicious with very high nutritional value and (3) Poisonous berries.

When picking berries, from an evolutionary perspective, the most relevant goal is to not pick (and eat) the poisonous ones. Sure, getting more of the delicious ones is very nice, but the difference in benefit between getting OK berries and Delicious ones is smaller than the benefit difference between getting poisonous berries and getting OK berries.

The same broad pattern applies to mating, as well. While both men and women would love to find the best possible partner (in terms of gene quality and parenting quality) to have children with, the reality is that the most important decision goal is to avoid having children with the worst (bottom 10%) potential partner.

Let your imagination take you back to the African Savanah when our evolutionary ancestors lived. Every female would have loved to attract and mate with the healthiest, most handsome, most effective hunter and bravest warrior.  At the same time, settling for the OK-ish guy to have children with would be considerably better than having children with the least reliable male in the tribe.

For our ancestral grandmothers it was more important to avoid getting knocked-up by a deceiving male who would take the fruits and nuts she foraged and leave her alone to take care of their children, who, very likely, inherited some of his defects.

This preference for mediocrity was, most likely, shaped by evolutionary forces throughout millions of years. Not surprisingly, our judgments and decisions are influenced – guided by it even in the XXIst century.

Buying a used car

Nowadays, dealerships use the term of pre-owned car, probably, because it sounds slightly better. Regardless of how they’re called, used cars are a pretty tricky product to buy. They cost a handsome amount of cash and for most consumers cars are difficult to evaluate complex products. Simply put, the car can look great and work just fine for the next two thousand miles (km) and then collapse, leaving the new owner of a used car with huge repair and towing bills.

Our judgment in buying a used car is shaped by the same preference for mediocrity. Most sensible people want to buy a used car that is good enough and, most importantly, want to avoid buying a lemon.

Referring to the used-car-purchase example, in a recent interview, Rory Sutherland mentioned reputation-based heuristics as being effective tools in avoiding disasters. He mentioned that his first used car was bought from someone his parents knew because it was less likely to be cheated by someone who drinks in the same pub as his father.

When my wife and I bought our first used car, one of the sellers asked us to meet him in a supermarket parking lot in a rather dubious area. Comparing that with people inviting us into their home after we took a test drive, I believe we did the sensible thing when we decided to not buy the car we saw in the supermarket parking lot.

I believe that reputation and trust related heuristics are manifestations (second order adaptations?) of our preference for mediocrity.

Another adaptation related to the preference for mediocrity is the compromise effect: our tendency to choose the middle option from a well-balanced choice set. Probably the most used example of the compromise effect is the disproportionate choice share of medium coffee (drinks).

One of the reasons for which people choose the middle option (compromise) is that we want to avoid extremes, since they come with risks associated. The middle option is a safe one since the risk of getting too little quality / quantity is mostly related to the smallest and cheapest option, while the more expensive and higher quantity / quality option comes with the risk of overpaying or waste.

While the preference for mediocrity has a huge explanatory value of human behavior, it also has broad implications.


For applied behavioral science and behavioral design, the preference for avoiding the worst outcome (over getting the best possible option) suggests that removing barriers (anti-nudges) that prevent the desired behavior to occur might (will?) have a greater effect than trying to (actively) encourage the occurrence of the desired behavior.

For improving services and even for our own self-improvement, instead of becoming (even) better at what we already are good at, it might be a good idea to stop sucking at what we’re not doing so well (what we’re worst at).

Take the example of medical practices (offices). Many of the (negative) comments on Yelp regarding doctors concern over-booking, unwelcoming waiting rooms and not-so-pleasant clerical staff (e.g. receptionist). Another big chunk of both positive and negative reviews concern doctors’ bed side manners, or simply put how well they interact with their patients.

Medical competence and the quality of a medical act (e.g. diagnostics accuracy, choosing the right treatment) are very difficult to evaluate, particularly for people without medical training. The upside is that someone who managed to graduate from medical school and completed her / his residency, most likely, isn’t a bad (terrible) doctor.

When it comes to bed side manners and more practical issues such as scheduling, waiting room quality and clerical staff politeness anyone can assess them.

If you would run (own) a medical practice with negative reviews, the first focus should not be to hire even better doctors (from a technical point of view); rather your first priority should be to make sure that scheduling issues do not lead to (very) long waiting times, that the waiting room looks like it is from this century and work a bit on employees’ people skills.  

Such an approach would most likely not make the practice win any prizes or glorifying reviews, but at least it will stop annoying people and reduce the number of bad reviews. Moreover, it should stop the bleeding of clients who just take their business somewhere else.

The same approach works similarly for self-improvement. If someone is a very good specialist in given field, say statistics, but (like someone I know) isn’t all that great at social interactions (i.e. people skills), it might be a good idea to work on the later and not try to become even better statistician.

Let’s give some random, yet illustrative numbers. Mark – the statistician – is very competent from a technical perspective scoring 85 out of 100 on statistics skills. However, on social skills Mark scores only 15 (out of 100) and 25 is the threshold for not sucking.

If Mark is looking for a job, his situation isn’t great. Indeed he is a very good statistician, since most of the jobs on the market require a technical skill level of 70 (remember, Mark scores 85). Despite him being over-qualified for most of the jobs, it might very well be that he can’t find a job even if he is invited for some interviews.

Even in a job such as data analysis (statistician), one has to interact with other people, be able to write reports for non-statisticians and co-exist in an office environment. Mark may very well be rejected for jobs that he’s perfectly qualified because of his poor social skills. From a recruiter’s (hiring manager’s) point of view, very poor social skills are deal-breakers because someone who sucks at interacting with others will probably deteriorate the current team-environment, will most likely require more attention – which translates in putting more on the manager’s plate – and might lead to more serious disturbances in the work environment such as sexual harassment complaints from other employees. Most likely, Mark is neither a horrible person nor a sex-maniac. He just sucks at interacting with others.
Mark can either improve his statistical skills or he can improve his social skills. If he goes for the former, he could try to enter the narrow job market for hyper-specialists (95+ on stat skills) where being a jerk is tolerated (sometimes is part of the job description).

If Mark, however, decides to improve his social skills, then it is possible for him to reach the mediocrity threshold (25) where he’ll be seen as yeah, whatever, he’s not great, but won’t cause any trouble.  Only when Mark will reach this mediocrity level, will his very good statistical skills actually be considered.

I am aware that this is an over-simplification, but I believe I managed to explain the core idea.

Stop Sucking!

P.S. I've redone my website www.naumof.com so that now it doesn't suck (as much). Take a look!

No comments: