I love scientific models as they explain the world and answer the why. Unfortunately, I had to learn, that they are only models. Science is made up of hypothesis and theory, of observation, experience, believe, and assumption. Based on this, models are developed, which might be the truth or not.
In school one of my majors was biology, which I originally wanted to study at University. During school I realized that biology has a lot of assumptions, which might be true or not. I missed the substantial proof. This was one of the reasons, that I decided to study chemistry instead. I believed that the models in chemistry are based on substantial proof and therefore true. I guess I was very ignorant and maybe also a bit naïve, as it took me till my PhD to grasp, that also the models in chemistry are developed based on hypothesis and theory, of observation, experience and believe and might be true or not. I remember, that my professor asked me to develop a model based on the results of all my experiments to be published in my PhD thesis and papers. And I thought, that this will be only an assumption and not the truth. This was the day I finally lost my naivety regarding scientific models.
Daniel Kahneman’s book „Thinking fast and slow” supports that scientific models are only models and not the true reality. In his book he describes how the two ways we make choices: fast, intuitive thinking, and slow, rational thinking work and how and why it leads us to wrong assumptions and decisions. The author refers to a lot of very interesting statistical examples. One of them is the law of small numbers. A study about the incidence rate of kidney cancer in the US had a remarkable result. The counties with the lowest rate of kidney cancer were mainly found in rural sparsely populated federal states. To me it was tempting to conclude, that the cancer rate is lower on the countryside due to less pollution. As it turned out, this conclusion is wrong. The fact is, that less people live on the countryside. And the statistical explanation is, that small samples tend more to extreme results than big samples. During my life, I learned a lot about statistics, and I should have known the right answer. I am not sure if it is reassuring or concerning. I am not the only person, who ignored her statistical knowledge. From the book I learned that this mistake happens to the vast majority of experienced scientists, as well. And these are the people who create scientific models.
This explains why wrong models are developed and persist. A great example is that a lot of people still believe that if you eat a lot of fish your risk for heart disease and stroke is reduced. This is based on a study from the 1970s, in which scientists investigated the data of death of Inuit people. Based on the data the cause of death due to heart attack for Inuits is significant less compared to other peoples. And they eat a lot of fish. So, the scientists assumed, that Inuit people die less of heart attack as they eat a lot of fish. The catch is that at the times of the study the Inuit people had only rarely access to health care. Thus, the cause of death could not be determined medically and the data the researchers used were lacking. In the meantime, Canadian scientists could show, that the heart disease rate of Inuits, Europeans and North Americans is similar, whereas Inuits have a higher death rate due to stroke and on average their life expectancy is about 10 years less.
In spite of this knowledge and oddly enough I still love scientific models to get the world explained and the why answered. Whereas, I got more cautious to trust in them. And maybe, this is my way of faith.
Photo by Andrea Piacquadio