All statistical modeling has these two frames: the small world of the model itself and the large world we hope to deploy the model in.38 Navigating between these two worlds remains a central challenge of statistical modeling. The challenge is greater when we forget the distinction.
