Let us begin this post by telling you a story, to talk about a bias that relates precisely to storytelling, but also to talk about how the risk of a mechanical view of reality can lead us astray.
During the Second World War, a group of researchers in Britain was faced with the problem of how to redesign aircraft so as to minimise their losses. The starting idea was to analyse aircraft that had returned to base, even though they had been hit by enemy shells. On analysing them, the researchers discovered that the bullets had mainly hit the wings and tail, drawing the conclusion that these parts needed to be reinforced because they were more exposed than others.
Fortunately, Abrham Wald, a mathematician who was involved in the project, had an insight before the team started working on the planes: the sample was missing the crucial part, that of the downed planes. By looking at the unreturned aircraft, one could indeed find an interesting lead for the redesign: it was the engine that was the weak part, not the tail or the wings! The crashed aircraft, crucial to understanding the real reasons for the vulnerability, were not present because they were not returned.
Survivor bias is a type of bias that impacts the selection of the sample to be considered as significant when analysing a phenomenon. It occurs when an individual mistakes a visible success subgroup for the entire group. In other words, we forget to consider all the data on those who did not make it.
Survivor’s bias, as well as a great lesson on how important it is to form consistent samples when we really want to understand a phenomenon, is a good starting point to listen critically to the various storytellers and gurus who tell us success stories: “how I made my first million” “how I founded the start-up that raised 20 million in funds” “how I invented the revolutionary product” etc. But this storytelling does not allow us to also hear the stories of all the other people who failed to launch their start-up, to get huge funding, to get rich, it does not give us elements to refer to the ‘worst practices’ as well.
This is not the only limitation of storytelling, there is also another risk we can run when we take the stories told by and about role models from a mechanistic view of reality.
On a Diversity & Inclusion project on gender, for instance, it might seem like a great idea to elevate a woman and tell her story with the intention of motivating others to follow the path laid out. But this exercise risks making us lose sight of the fact that there are specific elements ONLY of this story that are not found in the other stories: what specific conditions in the context in which the person acted, e.g. what organisational culture, but also what internal conditions, what mental patterns she had to overcome. In role modelling there is therefore a risk of not taking into account the fact that investors lend less to women, poorer people etc.
In short, the exercise runs the risk of being inspiring in the moment but, disconnected from the context and the system in which the person acted, of leaving the people listening with the illusion of being able to uncritically decline the story in their own context. To put it another way, by locating the root causes of the success in the individual, we blind ourselves to the contextual and systemic causes, which are often much more structural for the outcome than personal “heroship”. Moreover, in the case of aircraft the resolution of the problem is relatively simple and mechanical: once the survivor’s bias is discovered in the sample then one can easily intervene on the real reasons and reinforce the engine part.
But if we take the case of personal or professional success, and for example in the story told the person tells us that they had to learn to trust themselves, to negotiate with sceptical investors, etc., can we really think that once we listen to them then immediate change will be triggered? These factors do indeed take a long time to evolve, and do not follow a simple pattern of insight -> resolution.
To the contrary, transforming one’s own disabling psychodynamic patterns requires more than just knowing we’ve got them; it requires inner work, that doesn’t happen at a click of a finger. In the case of aircraft, we are in a system, albeit one with many variables, that is simple: the intuition that the sample needs to be revised is enough for us to solve the problem. When we apply this bias to the case, for example, of a start-up, we are instead acting in a complex and much less mechanistic system, in which it becomes more difficult, even once we have heard the missing story, that of the and the ‘non-survivors’, to really trigger complex change.
Are we therefore to conclude that these practices are useless? Absolutely not! On the contrary, it is important to give visibility to these stories and to continue listening to them and telling them. Just as it is important to also tell the stories of failures, of those who did not succeed, did not get funding, did not get their start-up of the ground.
To tell the stories as whole as possible, providing the contextual elements but also making a rational analysis of them to understand what is applicable to our own, and what it teaches us with respect to our context, our resources, what the story we have listened to highlights with respect to our mental models, without thinking we can reproduce it uncritically. Who knows, it may also point to more structural leverage points that need activating before single individuals can truly blossom, however heroic they aspire to be…