This article made me think about how there is an unfortunate side effect of the rise of data science. that is the belief that all these statistical models represent the system under study as opposed to the variance/covariance the model is able unravel. I often have conversations with people who believe their statistical model is essentially the same as a physical model. For example, linear regression and Newton's law of gravity should be held in the same regard.
Interesting the consideration of the difference between 'random controlled trial' articles and 'modelling' articles