George E. P. Box Quotes

A man in daily muddy contact with field experiments could not be expected to have much faith in any direct assumption of independently distributed normal errors.

The researcher hoping to break new ground in the theory of experimental design should involve himself in the design of actual experiments. The investigator who hopes to revolutionize decision theory should observe and take part in the making of important decisions.

Remember that all models are wrong; the practical question is how wrong do they have to be to not be useful.

Essentially, all models are wrong, but some are useful.

Since all models are wrong the scientist cannot obtain a "correct" one by excessive elaboration. On the contrary following William of Occam he should seek an economical description of natural phenomena. Just as the ability to devise simple but evocative models is the signature of the great scientist so overelaboration and overparameterization is often the mark of mediocrity.

A mechanistic model has the following advantages: 1. It contributes to our scientific understanding of the phenomenon under study. 2. It usually provides a better basis for extrapolation (at least to conditions worthy of further experimental investigation if not through the entire range of all input variables). 3. It tends to be parsimonious (i.e, frugal) in the use of parameters and to provide better estimates of the response

Statistical criteria should (1) be sensitive to change in the specific factors tested, (2) be insensitive to changes, of a magnitude likely to occur in practice, in extraneous factors.

For the theory-practice iteration to work, the scientist must be, as it were, mentally ambidextrous; fascinated equally on the one hand by possible meanings, theories, and tentative models to be induced from data and the practical reality of the real world, and on the other with the factual implications deducible from tentative theories, models and hypotheses.

I want to tell you how I got to be a statistician. I was, of course, born in England and in 1939... when war broke out in September of that year, although I was close to getting a degree in Chemistry, I abandoned that and joined the Army. They put me in the Engineers (and when I see a bridge I still catch myself calculating where I would put the charges to blow it up). Before I could actually do any of that I was moved to a highly secret experimental station in the south of England. At the time they were bombing London every night and our job was to help to find out what to do if, one night, they used poisonous gas. Some of England's best scientists were there. There were a lot of experiments with small animals, I was a lab assistant making biochemical determinations, my boss was a professor of physiology dressed up as a colonel, and I was dressed up as a staff sergeant. The results I was getting were very variable and I told my colonel that what we really needed was a statistician. He said "we can't get one, what do you know about it?" I said "Nothing, I once tried to read a book about it by someone called R. A. Fisher but I didn't understand it". He said "You've read the book so you better do it", so I said, "Yes sir"

We have a large reservoir of engineers (and scientists) with a vast background of engineering know how. They need to learn statistical methods that can tap into the knowledge. Statistics used as a catalyst to engineering creation will, I believe, always result in the fastest and most economical progress…

All models are wrong; some models are useful.

One important idea is that science is a means whereby learning is achieved, not by mere theoretical speculation on the one hand, nor by the undirected accumulation of practical facts on the other, but rather by a motivated iteration between theory and practice.

The penalty for scientific irrelevance is, of course, that the statistician's work is ignored by the scientific community.

An innovative discussion of building empirical models and the fitting of surfaces to data. Introduces the general philosophy of response surface methodology, and details least squares for response surface work, factorial designs at two levels, fitting second-order models, adequacy of estimation and the use of transformation, occurrence and elucidation of ridge systems, and more. Some results are presented for the first time. Includes real-life exercises, nearly all with solutions.