Data002: “Essentially, all models are wrong, but some are useful.”

That’s a quote from statistician George Box. (Any, by the way, by “model” he means a mathematical or statistical formula.) The lesson to carry away from this is that analysis is an exercise in simplification. Simplifications, by their nature, are not identical to the thing they represent, so they are, in that sense, wrong. As such, every analysis is ultimately wrong. Nevertheless, a good analysis will also give you insight into your data and will give you a reasonable basis for making decisions. So you can be both technically wrong and functionally useful at the same time.