Predictably overconfident…

As Daniel Kahneman said:

“We are generally overconfident in our opinions and our impressions and judgments”.

What did he mean? Is he not overgeneralizing?

Let’s look at what research on psychology has found on this topic.

Overconfidence has been studied in the context of decision making by asking people certain questions on their chances of answering a quiz, performing successfully a task, guessing rightly an estimate for example. Many studies have shown that we tend to be overconfident and the form our confidence takes is very predictable. So let’s dive a bit deeper.

Bear in mind that the biases explained below are prevalent and no individual characteristics have been shown to mitigate them. That you are a male or female, with such or such personality, with a high or low IQ, we all equally fall for these predictable estimation errors.

Overconfidence takes 3 forms:

  • Overestimation: We tend to overestimate our personal abilities, our performance, our chances of success and our level of control.
  • Overplacement: We also tend to believe that we are better than others. For example 93% of American drivers believe they are better than the average American driver, 75% of university professors believe that they are better than their peers. No matter how educated you are and know about statistic, you will predictably think that you are better than others and certainly better than the average.
  • Overprecision: And finally, we tend to be excessively certain of the accuracy of our private beliefs and estimations.

They are situations where we are overconfident and place ourselves above others, but there are times where we are underconfident and we perceive ourselves performing less than others. Are there some rules, some explanations which could explain why we overshoot in some situations and undershoot in others?

Research has found the following:

  • On easy tasks, we underestimate our performance and overplace ourself. To understand, this phenomena, let’s imagine something you are really good at, let’s say math; in that situation if you have an easy math exams, you will underestimate your performance and believe that you have performed less well than what is your actual performance. Now if you compare yourselves with the people in your class you will have even less information about how they performed compared to how you performed, so in this situation the bias of underestimation is even stronger and you will assume that the other people have performed even less well than they have. In comparison to them, you will overplace yourself, believing that you will do better than other.
  • On hard tasks, it is opposite, you overestimate your performance and underplace yourself. In this situation, you are really not good at math and you have to pass a difficult math exam. In that situation, you will overestimate your performance and the bias will be stronger for other people meaning that you will overestimate their performance even more, underplacing yourself, believing that you have performed less well than they did.

It seems that unless the task matches your skill level, in which cases you will correctly assess your performance, you will fall victim either to overestimation or underestimation depending on the difficulty of the task and the level of your skills.

The principles above apply to other form of estimation biases.

  • Illusion of control: when people have little or no control over a situation, they usually act as if they had. But similarly, when they have a pretty good handle and control on a situation, they tend to underestimate it.
  • The planning fallacy: a common disease of project management. People overestimate their speed of work and delivery when the task is very difficult and underestimate their speed of delivery when it is easy.
  • Pessimism about the future: people overestimates the chance that a rare event (let’s say 9/11, or succumbing to a rare disease) will happen to them but they overestimate that rare event occurring even more for others. Smokers will typically overestimate their chance of getting lung cancer and will also believe that other smokers will probably be at an even higher risk than they are. Pessimism has been prevalent in literature and the media, which overestimate catastrophic and rare scenario.

These findings show that we are predictably inaccurate in predicting future events and our own performance. We are even worse at predicting others. And the less skills and knowledge we have about a situation the stronger the effect,

You might not be able to change your biases, but it might be useful to be aware of them.

 

Subscribe to stay up to date

Loading