survivorship bias : mackenzian

There are so many ways to arrive at faulty conclusions.

When it comes to populations that have undergone trauma, one way to arrive at a faulty conclusion is to learn exclusively from those survive their experiences—taking campus policy advice solely from students who manage to stay in school despite bullying and marginalization, for example.

Another way to arrive at a faulty conclusion is to take cues only from those who don’t survive. A study about contraception recently ended because some study subjects found common side effects intolerable. By ending the study prematurely, researchers have ensured that they can’t track the progress of subjects who were more resilient, and the testing of this medicine might be permanently halted.

As they work to understand their members’ motivations, denominations like mine are commissioning retention research projects about former members as well as current ones. These studies are a tool for mitigating the impact of survivorship bias on church policymakers who would otherwise focus only on people who remain formally affiliated. Of course, the quality of their conclusions from these studies depends on how the studies are designed and what respondents volunteer when asked.

A fascinating example of survivorship bias featuring US Navy researchers went viral on Twitter yesterday:

During World War II, the statistician Abraham Wald took survivorship bias into his calculations when considering how to minimize bomber losses to enemy fire. Researchers from the Center for Naval Analyses had conducted a study of the damage done to aircraft that had returned from missions, and had recommended that armor be added to the areas that showed the most damage. Wald noted that the study only considered the aircraft that had survived their missions—the bombers that had been shot down were not present for the damage assessment. The holes in the returning aircraft, then, represented areas where a bomber could take damage and still return home safely. Wald proposed that the Navy instead reinforce the areas where the returning aircraft were unscathed, since those were the areas that, if hit, would cause the plane to be lost. — Image: Ryan O’Horo (Twitter); source: Wikipedia

It’s a remarkable story: were it not for Abraham Wald and his analyses, the Navy might have protected parts of their planes that didn’t need protecting and left vulnerable the parts that most needed armor.

Consider the range of sources you use to teach you about the populations you work with. Could you be misreading why “survivors” survive and what set them apart from those who don’t thrive?

As far as it lies within your power, learn from all cases, not just those that match your initial assumptions.