box’s loop

14 february 2017

notes on (Box, 1976).

summary

science and statistics go hand in hand. paper comes from the premise “all models are wrong, but some are useful.”

science is not

• guessing theories, nor
• collecting practical facts, it is iteration between the two: we make a hypothesis, get some data to confirm/falsify, and revise our model.

the theory in statistics is grounded on sound maths. but it still is based on assumptions. if we don’t use it in practice, we can’t know whether it is useful or not. we need to actually work with models and the real life to know.

we can’t have an open loop in which case we are stuck in either of the following:

• cookbookery: doing things without mathematics;
• mathematistry: doing mathematics without regards to reality.

---

references

1. Box, G. E. P. (1976). Science and statistics. Journal of the American Statistical Association, 71(356), 791–799.
   • Bib

   @article{box1976science,
     title = {Science and statistics},
     author = {Box, George EP},
     journal = {Journal of the American Statistical Association},
     volume = {71},
     number = {356},
     pages = {791--799},
     year = {1976},
     publisher = {Taylor \& Francis}
   }
   

[back]