Enemy of science and how to circumvent it

It is also called confirmation bias Or affirmation tendency, according to Wikipedia, is the tendency to remember, interpret, or search for information in order to confirm initial beliefs or hypotheses. It is part of our brain dynamics: we seek information that confirms what we actually believe. The process of forming a conclusion is reversed: instead of looking for information that allows us to arrive at a definition, we already have a preconceived conclusion, and therefore, we will be looking for information to support our findings. Meanwhile, in the middle of this process, we tend to ignore or underestimate the information indicating an opposite outcome.

We don’t do it consciously, it’s human behavior. And as scholars, we are also human beings, subject to this, but if we do not take measures to circumvent confirmation bias, there is a real risk of disseminating opinions as facts. An article by Dr. John Lowenides entitled “Why most published search results are wrong“, Strong evidence of confirmation bias among professional scientists. When analyzing 49 well-known research findings in medicine, they claimed to have found 45 effective interventions. Of these 45, in subsequent studies with larger samples, 7. They were contradictory. Not completely defenseless against confirmation bias.”

According to Cassie Kozyrkov, a South African scientist and data statistician and founder of the field of Decision Intelligence at Google, we can adopt behaviors to circumvent this bias.

Do not be bound by your predefined opinions

Easy to talk, hard to do. A good mental exercise is not to take your opinion as an absolute truth, even if it relies on a lot of data. It is important to be able to absorb new information and admit a mistake, if possible.

Emphasize the decision, not the opinion

If an idea does not lead to action, then what is the problem? For example, I can believe in the Easter bunny, and as long as it doesn’t influence my decisions, that’s okay. But when there is a decision to be made, it is important, before opinion, to determine what information will be required to guide me during the introduction.

Focus on what is controlling

If there is a certain action outside of my control (social distance, for example), my opinion on the topic does not matter (I am clearly in favor of science, but let’s continue the mental exercise), we have to keep the distance and that’s it!

Change the order in which you handle the information

“The best antidote to confirmation bias is to plan a decision before seeking information.” By defining the rules for each potential decision, the risk of supporting the stock of information that favors our opinion is avoided.

To work with Analytics, it is imperative to follow these good practices that avoid confirmation bias and allow us to obtain the best results that will guide business decisions. It is imperative to remember that the true value of the data is directly related to the analytical steps to be followed in the process, which includes potential information that can even combat our hypotheses.

Other than that, we do not have the true value of data science, but we are looking for confirmation of personal beliefs or what we are trying to confirm. It is essential to escape bias, because in addition to prompting us to error when making blind decisions, it is in a dark room that distances us from the clarity and light that data can bring.

* Edwarda Espendola It is the pioneer of data science in Where are you, A high-tech startup that helps companies achieve greater operating performance, productivity and quality of life for employees, based on smart data generation.