THE SURVIVORSHIP BIAS!

"Exclusion of an important entity from a specific dataset is the precursor to untrue interpretations." - Mayank Mishra.

The above statement induces a thought inside of me. A thought for the concept of Survivorship Bias. A bias that limits our critical thinking and pushes us towards illogical interpretations. What is survivorship bias you ask?

Survivorship bias or survival bias is the phenomenon in which untrue or false conclusions are drawn because of the exclusion of an entity from a specific list or a group of similar entities because the excluded entity failed to survive in the long run, and hence, was not included in the said list or a group thereby resulting in an illogical interpretation of the dataset.

Let's understand it with an example, we often hear that classical music is way better than today's music. This generalization might be because only the best music from older times is played today while today's music, no matter how good or bad it may be, is readily available. This results in a comparison to be made between the best songs from the past and poor, mediocre, and best songs from today's time. Thus, survivorship bias of better music from an older era results in a false conclusion which causes a generalization that older music is better than today's music as only the best music of the past which has passed the test of time is available for such a comparison.

Another example comes from an actual study in the military during the time of WW-II when American aircrafts were receiving heavy damage on several parts. The military analyzed all the aircrafts which successfully came back from their missions and identified such parts which received the most damage during a mission. They concluded that putting armor around those parts is the logical thing to do as they are the parts that receive the maximum damage. The survivorship bias kicked in here as the military excluded, from their analysis, those aircrafts which did not return from their missions. 

Statistician, Abraham Wald, from the Statistical Research Group at Columbia University undertook this research. He also factored in the crafts which did not survive their missions. After a careful analysis, he concluded that such areas of the aircrafts which took little to no damage were the ones that needed reinforcement (in the form of armor). The logic behind this was that those areas which received the most damage could withstand it and function properly but, when the bombardment happened at those areas which received little or no damage, the crafts did not survive. So, the most important area which needed armored protection was the one that showed little or no sign of damage, as that area was the one that was responsible for the overall functioning of the aircraft.

This shows how survivorship bias can be a root cause for faulty judgment. It's not that ignorance sets up a trap for such a bias but, it's the common thinking pattern that limits our criticality and makes us jump on to illogical conclusions which seem to be logical.

Okay, so we know what survivorship bias is but, why should we care? We need to, we must! Survivorship bias is among such selection biases which manifest themselves many times in our day-to-day life.

Does anybody remember those toothpaste commercials that claimed, "9 out of 10 dentists recommend __________!"

It presents an idea of the brand being superior to its competitors but, is this the logical interpretation of the statement? Think about it, are we told how many brands participated in such a comparative study? Are we told what was the sample size in terms of the dentists who participated in the study? Do we know if there was another brand which was recommended by the dentists as well? Could there be a brand which could have been recommended by 10 out of 10 dentists? Are there some certain circumstances under which the dentists recommend this particular brand?

Without asking any of these questions, our minds instantly jump to the conclusion that establishes the superiority of the brand. This is what marketing campaigns can do. By just telling you enough so as to create a mental outline of an image that they want you to see.

Let's take another example from the restaurant industry. If you look closely, the reviews for restaurants are always skewed in either of the two directions. One being the awful service that you received which deserves 1 or maybe 2 stars. The other one being the best-in-class service that deserves no less than 5 stars. Generally, people give reviews in one of the two situations, when they have an awful time, or when they truly receive a 5-star service. This excludes such customers from the review pool, who had about an average experience with the restaurant service and opted not to share a review. Hence, the exclusion of an entity from a dataset that results in an interpretation that may not represent the data in its true form.

Another example, which kind of makes me nervous nowadays is the success paradox. While thinking about successful people, how many times do we think about such individuals who had the exact same amount of talent, worked nearly as hard and even harder than the others, and devoted their every last second to an idea but still ended up unsuccessful? We see the top one percent of them in interviews where they explain, how working relentlessly for their dream and by never giving up, they finally achieved success. We interpret that maybe this is the formula for success, if we work nonstop, and if we never give up, we can achieve the same success as they have. But, do we think about such people who did the exact same things and still remained unsuccessful? Their exclusion from our analysis compels us to jump on an interpretation that may neither be entirely truthful nor logical.

We need to understand just how easily we are played by such psychological traps. Mere awareness of such things may go some distance but, if we truly plan to go a long way, we need to develop some serious critical thinking abilities. How do we do that? 

Let's not start being overly skeptical about everything in life. That is one sure way to make yourself insane. But, let's train ourselves to ask questions every once in a while rather than blatantly accepting the claims that are thrown our way. Do that, and then maybe we can have our "shields up" to protect our "Starship Enterprise." 

And remember...

One survived but, countless others perished. Without including this in our analysis, we put a brick wall in front of our critical thinking ability. This is what survivorship bias does. And this is exactly why we must care about it.

[Thank you for reading this article. If you liked this article, you will surely like the book "Predictably Irrational" by Dan Ariely. Subscribe to NotionWave on YouTube. Follow me on Instagram]

Support the work by donating.