Cognitive Bias – Survivorship Bias

Cognitive Bias – Survivorship Bias

Cognitive bias are errors in various thought processes and so-called mental shortcuts that lead to incorrect conclusions about various information. The result is a somewhat distorted view of reality. Find out more about these biases and put your brain to the test in the final quiz.

Biases often arise from a variety of processes that are sometimes difficult to distinguish from each other, such as noise in information processing, limited brain capacity to process information, emotional and moral motivation, etc. The most common cognitive biases include:

Anchoring: describes when the natural human tendency is to rely on a single piece of information or fact in the decision-making process, even though it may not be relevant, complete or true.

Apophenia: is the human tendency to perceive meaningful patterns in random data.

Confirmation bias: is a person’s tendency to favour information and interpretations that support their own beliefs, while ignoring or undervaluing those that contradict their beliefs.

Overconfidence effect: causes people to overestimate their abilities, prospects and chances of success.

Survivorship bias: is a logical fallacy based on the higher visibility of those who have “survived” a process.

The problem of hindsight: there is a tendency to see past events as predictable.

This time the Biostatistics team decided to focus more on the Survivorship bias, with illustrative:

Example 1:

During the Second World War, the Allied Forces decided to arm their bombers to withstand more enemy fire. An analysis was made of where the planes were most often hit (based on an actual sample of hit planes that returned) and these places were to be armoured. However, the statistician Abraham Wald took the Survivorship bias into account and completely reversed the statistics. He concluded that if the planes that returned were hit in the given locations, then the planes that did not return were probably hit elsewhere. He therefore suggested that the unhit areas should be fortified.

Example 2:

At the start of the Second World War, the US introduced a new helmet design designed to significantly reduce head injuries in combat. Almost immediately after its introduction, field hospitals were inundated with new head injury patients. Concerned about the unexpected increase, it was decided to withdraw the new helmets from the market. However, it was later discovered that the sudden increase in head injuries was not a symptom of faulty helmets, but evidence that the helmets were working exactly as advertised. For the first time, field hospitals were seeing patients whose injuries would not have allowed them to survive before. This increase was compelling evidence that lives were being saved.

Example 3:

The 1987 study found that cats falling from higher than the sixth floor were more likely to survive the fall and suffer fewer injuries than cats falling from lower than the sixth floor. Their reasoning was that cats reach terminal velocity after falling from about five storeys and are unharmed after that. In 2008, Straight Dope magazine suggested that another possible explanation could be the Survivorship bias. Cats that die in falls are less likely to reach a vet, so studies do not report cats that die in falls from taller buildings.

Example 4:

A common view is that machines, equipment and goods produced in previous generations are often better designed and last longer than today’s machines. Old machines that have “survived” can still be seen in many plants today, hence the Survivorship bias. This is because all the machines, equipment and products that have broken down over the years are no longer visible to the general public and have been scrapped, recycled or otherwise degraded.

Example 5:

The popular media often feature stories of determined individuals who have managed to overcome obstacles and achieve their goals. But of course, they do not report on the incomparably greater number of people who have also tried, but failed. This bias then tends to create a false positive assumption.


For the following examples, explain what the Survivorship bias is:

  1. The majority of employees are satisfied with the company’s management, according to employee satisfaction statistics.
  2. A preference survey of existing customers showed that they were satisfied with the services offered. As a result, services were tailored to their needs.
  3. Research teams present at the conference and describe how they got their work into a prestigious journal.