Survivorship Bias: The Statistical Importance of Failures

“XYZ appointed as the CEO of a huge multi-national company!” read the newspaper headline, almost as if to mock me. This CEO gets so much attention. He managed to make his way to the position after dropping out of college. What about me though? The long hours when I sat behind not just one, but five different books just to understand one concept. Acing college classes and getting perfect scores. Looking at all these college dropouts, was there really a point in me getting my degree? Maybe I should have dropped out too. I could have been so much more successful. It makes sense, doesn’t it? Look at Steve Jobs and Bill Gates. If they could achieve such huge feats by not wasting their time in universities, what did I gain from it? Imagine the luxuries I would have been able to enjoy if I had made it big. Only if I had dropped out, right?

In life, we often end up victim to such judgements. We end up not seeing the facts, or even bother researching about them. And why would we? After all, it is easy to give into survivorship bias.

Definition

Survivorship bias is a type of selection bias where the results, or ‘survivors’, of a particular positive outcome are disproportionately evaluated. Those who fail, or don’t survive, might not even be considered. Focusing on said ‘survivors’ can result in a false, or incorrect, estimate of probability. In a nutshell, it could skew results one way when the facts actually point to another. 

Survivorship bias is how a sleazy car salesman can throw facts and figures at you to make you buy a dirt box. It’s how a barmy idea can be sold based on data that doesn’t assess the entire scenario. It is the reason behind your elevated hopes of winning a lottery because someone you know just did.

What can it do?

 

Companies that had to be shut down weren’t taken into financial assessing — a reason why no one could see the 2008 Financial Crisis coming. Survivorship bias can mislead studies and statistics. It can make the financial system seem like a booming greenhouse when in reality, a recession might be right around the corner.  

In a survey, if our sample happened to only consist of eight-year-olds, we can pretty much prove the existence of the tooth fairy. This is what survivorship bias is — it is a sample selection bias that takes data sets that have ‘survived’ and not ones that have ceased to exist or failed.  

In today’s world, there are various examples of survivorship bias which we often do not pay attention to. The main reason for this is that we end up focusing and remembering successes more than failures. We can see this not only on big levels such as businesses but also in our own lives.

It’s a common notion that music just isn’t that great anymore. It’d be easy to agree to that when listening to a decade’s hits playlist. However, this doesn’t exclude the fact that throughout the decade there were more than just the limited songs that you see in the playlist. Of course, they might have not made it big or not been properly publicised. Yet we believe that a particular era could’ve been better than the current one for music.

Sitting in your house on New Year’s Eve, you often end up scrolling through social media only to see photos and videos of people who have gone out. The brain, on seeing this, makes itself believe that everyone except you is enjoying themselves. However, you fail to think about all the other people who are staying back at home, seeing the same kind of photos and videos as you. Only sixteen per cent of people actually do go out on New Year’s eve, whereas a majority stay home.

 

Ever wondered why there are no stories you can read up about people who failed on their way to success? On the internet, you will find people who might have initially failed but eventually succeeded. Steven Spielberg, J.K. Rowling and Walt Disney are such examples. That, however, is not what we are looking for here. After all, these people did gain success at some point. Do you remember any stories off the top of your head about people who actually didn’t make it —the vast majority?

These are some of the everyday cases that fall under survivorship bias. The tendency of humankind to overestimate the ease of success might seem discouraging by now. In materialistic terms, achieving success is extremely difficult. With the knowledge of survivorship bias, you are now aware that it might be even tougher than you thought. However, that in no way shows that perseverance does not pay off. At the end of the day, if you love what you do, success is just a byproduct.

Most stats out there usually undergo multiple testings and observations and are omitted, to overcome survivorship bias. When a statistic is released, it is checked and cross-checked by other statisticians by passing it through their own test scenarios and accounting for any unidentified bias. In that regard, survivorship bias is a problem long solved, but even if it does creep into important life-shifting areas, human effort can’t be pegged by any stat and that’s all in our hands. 

So go back to sipping your coffee in peace and just being you. And take more calculated risks while you are at it.

 

Written by Kaavya Azad and Noel Pereira for MTTN

Artwork by Chirag Bansal

Edited by Chintan Gandhi

Proudly powered by WordPress | Theme: Baskerville 2 by Anders Noren.

Up ↑

%d bloggers like this: