Overcoming ‘Confirmation Bias’ for better decision making

There’s an often-quoted story about the legendary mathematician, Abraham Wald. As the narrative goes, in 1942 Wald was one of a few select scientists who were part of the Statistical Research Group (SRG) at Columbia University.[1]

Soon after the U.S. joined the war against Germany, they found their airplanes taking massive hits every time they tried to strike their targets.

The obvious answer was to add a layer of protective armor to the body of the aircraft and the problem would be solved. Except that it created a new problem. The armor had to be heavy to prevent bullets from penetrating the body and added a lot of weight, affecting the aircraft's speed, ability to climb, and maneuverability—in short, the solution didn’t work.

The U.S. air force approached the great minds at SRG for a way out.

There was a lot of statistical data about the hits that the aircraft had sustained. The data showed that most of the hits were around the tail and a few around the motors. The solution was to simply add protective armor just around the tail and the problem would be solved.

Or so it seemed.

Abraham Wald examined the data and came up with an interesting perspective. He argued that the hits near the tail were misleading. The actual damage to aircraft was caused by hits to the motors. The aircraft that were observed to have a lesser number of hits to the motors were actually the ones who survived.

The aircraft that had more hits to the motors were not observed—because they didn’t make it back.

So, Wald suggested that the additional armor should be located around the motors, not the tail, even though all the data pointed otherwise.

The rest, as they say, is history.?

Whether the story is totally accurate or not is debatable. What it does reveal is the importance of being able to look at information and not blindly come to conclusions. One such bias is called the ‘confirmation bias’ and, in the age of AI-driven decision-making, it becomes all the more crucial to be aware of it.

Confirmation bias is the act of listening and giving credence to those facts and arguments that align with or reflect beliefs we already hold.[2]

Why addressing biases is crucial

Look at the impact that biases have in organizations. While it is not feasible to have a more fine-grained view of the contribution of each particular type of bias, this was the result of a survey of 3000 individuals in large organizations.

This literally becomes a matter of life and death in legal judgements such as awarding capital punishment. Imagine a jury that has a confirmation bias and passes sentences without being aware of it!

In HR, while we luckily don’t need to make such grave decisions, there are several important things we do—selecting candidates, giving out rewards and recognitions, deciding on promotions, all of which impact people’s careers and lives in big ways.

Clearly biases are important, and with the proliferation of AI models, are not unlikely to creep into enterprise systems—and get deployed at scale. What a tragedy that would be!

So, every effort to eliminate biases is effort well spent.

Fortunately, when it comes to confirmation bias, there is a clear way out and you can implement it too.

How we avoid confirmation bias at Zuci

In the immensely popular book, Thinking, Fast And Slow , author Daniel Kahneman points out that thinking fast is an automatic shortcut that the brain takes to save precious time, especially in critical moments.

Confirmation bias is a kind of survival mechanism that has in fact helped humankind to quickly gauge threats and take appropriate action—fight or flight. Unfortunately, this has become an unconscious part of our brain, and when applied with a lack of awareness, can lead to disastrous consequences.

The solution is simple: think slow.

·????? Evaluate and re-evaluate

We always look at all angles of any issue and painstakingly deliberate over key decisions as a matter of process.

·????? Get second and third opinions

We involve as diverse a group of people as possible and make sure that no decisions are taken unilaterally.

·????? Take time

It is tempting to give in to pressure and impatience, however that can prove costly. We give every activity the time it needs to be done to perfection, in the best way possible.

·????? Get the facts right

Decisions based on insufficient data is bound to have shortcomings, so we always ensure we have accurate and complete information before making commitments.

Confirmation bias is an unconscious and automatic mechanism of the brain that can inadvertently do a lot of damage. It becomes especially crucial when taking important decisions. The key to overcoming it is to avoid quick fixes and to take the time needed for doing anything well. Thinking slow might seem to cause delays, but they are definitely worth it.

All of these are very simple but incredibly important. Taking time to avoid biases is definitely worth the effort so do try out these steps the next time you have to make an important decision. What biases have you encountered at your workplace? What measures did you take to overcome them?

References:

1.????????? From an article by the American Mathematical Society

2.????????? As defined by Psychology Today

3.????????? As per the Deloitte State of Inclusion survey