This article from the Wharton University of Pennsylvania entitled Confidence Games: Why People Don’t Trust Machines to Be Right highlights the issue of placing trust in our own judgement even though the facts tell us otherwise.
In the article, Wharton professor Cade Massey and doctoral student Berkeley Dietvorst explain that once we use a predictive algorithm that makes a mistake, most of us no longer trust it. We would prefer our own judgement even though the facts point to the algorithm being right more often than we are. An example of an algorithm is one that predicts a winner of a sporting game or the demand vs supply curve for a new product.
Massey even admits to overriding one of his sports prediction models because of his own confidence that a team was going to lose a college football game. He admits it is inherent in our genetic makeup.
Anyone who has spent any time in a corporation would have witnessed this phenomena. Sometimes the boss would be right, sometimes not. What seems to be happening is that decision makers like you and me feel better if we have taken ownership of the decision.
Dietvorst and Massey ran experiments which offered people the option of using an algorithm. They also allowed them to adjust the output up or down a little, depending on their preferences. People were then more likely to use the algorithm rather than disband it and make their own judgement.
So my food for thought is that whenever you are asking someone to make a decision and you have some data or an algorithm which you think provides some telling facts, remember the following: Don’t use absolutes, provide ranges or use words like “indicate” and Do FRAME the decision around the use of the data in order for the most likely best option to be chosen.