Select Page

Contrary to what 24-hour news channels will tell you, global quality of life is improving fast.

The late, great statistician, Hans Rossling, co-wrote a must-read book called “Factfulness“, which supports this claim. The book makes an important case that our unparalleled access to information is actively skewing our understanding of the world, with dangerous consequences preventing us from seeing the world as it really is. The overwhelming quantity of bad-news stories leads us all to think that society is much worse than it is.

This idea ties in neatly with another interesting read “The Idiot Brain”. In this book, Dean Burnett makes the interesting case that the process of evolution has rendered the brain inefficient in certain modern-day contexts. For instance, the brain’s incredible pattern recognition capabilities, which protected us from predators for millennia, were trained on speed, not accuracy. The first person to spot the lion survived because that person would outrun anyone who spent too much time analysing the lion.

In short, our minds are not well suited to extrapolating patterns from internet-sized datasets. As a result, our brains look for shortcuts in the data we have consumed, which do not necessarily reflect reality. The art of trying to widen your dataset is one of the biggest growth areas in AI.

Some of the cleverest ideas to emerge out of the 21st century feel thoroughly counter-intuitive. They feel like an affront to common sense. Hungarian statistician, Abraham Wald, proved how counterintuitive complex decisions can be during During World War 2. Wald was a member of the Statistical Research Group, tasked with observing damage to Allied aircraft. He made the remarkable decision to recommend adding armour to aircraft where he saw no evidence of damage on any of the planes he had observed.

This counter-intuitive decision is genius-level lateral thinking. Wald made the clever decision to include ‘unreturned aircraft’ into his data set. Rather than just looking at what was in front of him, he realised that the most important data in his dataset was lying in German fields, having never returned home. Wald’s groundbreaking research led to the development of survivorship bias and selection bias research, where weaknesses in collected data can radically alter the view of the observer, thereby leading to material errors in the interpretation of results.

Wald’s observations are impressive, partly for the demonstration of ingenuity, but also for the anecdotal evidence suggesting that the Defense Department was thoroughly committed to performing the opposite action of adding armour to affected areas. Not only would this have had little impact on protecting aircraft, but it would also have slowed those planes down and therefore, in all probability, increased casualty rates.

This story offers an interesting lesson for modern-day decision-making because we have never been more susceptible to selection biases.

The sheer amount of content available to us forces us to filter our intake of information, meaning that we are all operating under some form of selection bias.

So, how can we channel our inner Abraham Wald and guard against bias?

Here are six questions we ask ourselves when it comes to improving our decision-making processes?

  1. Have I correctly framed the challenges I am facing?
  2. Am I aware of everyone’s point of view?
  3. Do I have the right data for the job?
  4. Am I including all the relevant data?
  5. How am I measuring success?
  6. How set am I in my current hypothesis?