During World War II, researchers at the Center for Naval Analyses conducted a study on the damage done to Allied aircraft returning after bombing missions. Based on damage assessments, they recommended adding armor to those areas showing the most extensive damage in order to minimize future losses to enemy fire.

However, Abraham Wald suggested that this kind of analysis was looking at the data inside-out.

Born in 1902 in Hungary, Wald was home-schooled by his parents because, as religious Jews, they would not send him to school on Saturdays; the Hungarian schools did not tolerate such regular absence. But this didn’t stop Wald from going on to earn a PhD in mathematics.

Later, when the Nazis overtook Hungary, he emigrated to the United States and became a member of the Statistical Research Group where he applied his skills to various wartime problems.

From this vantage point, he noticed a simple thing about the damage assessment analysis underway… The planes being analyzed were the ones that returned.

This meant that those particular planes had managed to survive getting hit in places where the damage was subsequently evident on the aircraft.

In turn, this also suggested that, had they gotten hit in other places they would more likely have been among those planes that did not return.

Wald pointed out that conclusions about how best to fortify planes – limited by weight constraints – were being tainted by “survivorship bias” thereby leading to the wrong conclusions.

Analyzing surviving members of a group may be a sound starting point for determining how to survive. But it’s inherently flawed as a way to infer what causes demise.